In an effort to be completely transparent, we decided to make a video with regards to grey-market chips. It's better than the Super Bowl halftime show, so queue it up! ruclips.net/video/eIInfDmITPU/видео.html
Thanks a lot for the shoutout! The first one you tried to put acetone on are of the "hardened and cooked and laser etched" variety, just like me only a little something comes off using acetone. However your YM2151 I _knew_ would come off clean haha.
You can check that your YM3812 are remarked by comparing the top and bottom texture (they should match if they are non-tampered with), and the ejector holes should be shiny and not covered
I appreciate the tip! Out of morbid curiosity, I've gone ahead and ordered fifteen OPL2 chips from various vendors to test myself. This is a fascinating rabbit hole.
@@LGRBlerbs Please tell us you will record your unboxing? haha. Oh here is my full checklist : 1)Texture must be the same on bottom and top 2)Pins should be straight with no solder or flux traces. 3)Acetone should not alter the chip's surface. 4)Datecode should be in the 80s or 90s (for most DIP YMxxxx chips) 5)Ejector pin marks should be shiny for most casing types (and not covered or sanded) 6)No laser etching on those 80s or 90s YM chips Anything else is proof that the chip has been tampered with!
@@flamestoyershadowkill More like feaudsters. But genuine new old stock would look like that set of 5 and might be sold as such. For those unaware, "new old stock" means original things that were never used for anything and have been sitting in a warehouse since they were made back in the day. MOS/Commodore is no longer around as a company, but Yamaha is and might still be doimg a production run every 6 years to stock up the spare parts supply to musicians with Yamaha keyboards.
Here's that video that goes more in-depth: ruclips.net/video/k72SFBOZ_lw/видео.html To be clear, I'm cool using reproduction or resurfaced chips so long as they function the same or better. I just wish that the chip makers would _make it clear that they're clones_ if that's the case, since I like to know if I'm dealing with brand new or vintage stuff when it comes to hardware. Buying old sound chips is enough of a gamble already, no need to obscure things further.
LGR Blerbs I’d prefer having properly licensed chips if the cost was reasonable. I can‘t imagine reasonable licensing cost for something like OPL2. Building a input/response tester should be doable. That way chips could have verified behaviour. I have a hard time imagining it being profitable to manufacture new clones. Resurfacing I can imagine. Silly, though.
When I was in school the instructor passed around a few Cisco interface cards... I wanted to know how much they would cost, and one of the first things that came up in a Google search was a page all about how you can tell if a card is a fake or not. One of the example cards the school had turned out to be a bootleg. Found a number of fakes at my last job in the piles of cards they had, and the effort put into making them look real varied considerably, everything from faking the hologram stickers to not even remotely the same color on the components.
When the patent ran out on "FM" synthesis (more on that in a moment) lots of skeezy third party compatible chips started to appear. The thing is, the patent sort of lied because Yamaha FM sound chips actually use phase modulation not frequency modulation. Early on Yamaha learned that FM produced nasty harmonics and they could get the same synthesized sounds without the unpleasant artifacts if they used phase modulation in the operators rather than frequency modulation, and then kept that information a closely guarded secret within the company. Remember that FM chips are really a form of digital synthesis, it's a bunch of digital adders working with 12 bit digital waves so how it works isn't immediately obvious if you decap the chip and look at the circuit and it was a very long time before anyone thought to check to see what exactly was going on inside a Yamaha FM chip. People just assumed it was doing what the patents said it was doing when it fact it was actually slightly different than what was described in the patents.
I assume that's why they never sued Casio over their phase distortion synths - because they didn't want to have to admit that their own synths weren't actually FM? But then again I guess the difference is modulation vs distortion. I personally prefer the Casio/PD sound but you can't get the kinds of metallic sounds FM/PM gives you unless you have a VZ-1 - AFAIK the only synth to have fully implemented the idea of PD synthesis. I had one for a while but it was an absolute pig to program so I went back to the more limited architecture of the CZ-3000 and more recently the CZ-1. Only thing from the VZ-1 that I miss is that you could set the pitch wheel to have a six-octave bend range - three in each direction!
@@theidealcopy_ No phase distortion synthesis and frequency modulation/phase modulation synthesis actually work differently. Professional FM synthesizers like the DX7 used 6 operators and were much better sounding than 2 operator and 4 operator FM chips in consumer hardware. There was also the TG-77 and it's relatives that used a 6 operator FM synthesizer which could use sampled sounds as operator waveforms instead of just sine waves like the DX-7. It had a 6 operator FM synthesizer (basically a DX-7 but with fully programmable operators) and a PCM section that could be used as a ROM sampler (it had a set of built in sounds and card slots for additional sound libraries) or the ROM sounds could also be used as waveforms in the FM section.
Also look up the Yamaha FS-1R which was the last FM synthesizer Yamaha made. Ir was quite powerful but somewhat flawed. A bad hard to understand interface, slow to download parameters and proper editing tools never materialized during it's production (there were editors created for it long after it was discontinued though). It still fetches $1000-$1500 on ebay for a working unit in good condition.
Where's the money in spending about half a million give or take to spin up a slightly shoddy pin compatible replacement of OPL2 or OPN? I wonder if Yamaha simply sold a number of genuine chips with different markings, or if they ever sold bare dies to be packaged elsewhere. As to late 90s flood of OPL3 software-compatible logic made by CODEC manufacturers, those really made sense since the CODEC was to be spun anyway, so it would be a cost saving to put the synthesizer right on there.
Clint, don't worry; those of us who've been following you for some time know that you would never try to mislead people. I for one can't imagine a more honest and down-to-earth guy than you.
Replica would be a great word if sellers were to disclose the info. I would also love to see someone playing around with the functions of the chips on an oscillioscope to see how much difference there is if any.
Often, modern counterfeits of old logic chips are software reimplementations on a commonplace microcontroller (a small self-contained computer). I don't know if there is "firmware" openly available to make your own. I can see someone making a bsnes/higan style "cicle perfect" replacement.
As someone who repairs all types of vintage electronics, I'm REALLY sick of all the fake semiconductors and IC's out there. We need more places that salvage authentic parts from scrapped equipment. I was pretty happy to save some 6532 and 6802 IC's from some scrap boards the other day. Great for fixing arcade stuff :-)
It won’t work the same. They sound like crap. I had the misfortune to only hear OPL clones through my life (first a shitty OPL3 clone made by Oak technologies/OPTi found on a cheap “Pro Multimedia PMM-240” Card, another Creative’s infamous CQM OPL3 clone that was found on the AWE64) and they all suck! Grand Pianos sound like toy pianos! Guitars sounds like they had plastic strings! Drums lack depth and reverb and sound like tiny little toys! And in AWE’s case, the instruments were quite soft in some games and needed the volume to be turned up ridiculously high!
I once asked Yamaha (office in Germany) directly if the Chinese YM3812 were "legal" to import to Germany. And they replied: (original text at bottom of long comment) ------------------- The chip is not supplied by Yamaha anymore. Whether the offered chip works, we can not guarantee and Yamaha can also not guarantee that it is not a security risk. and asking if i get trouble from them if i import it from China: Whether you buy this chip from China is up to you. (And they repeat that they take no guarantees or know if the chip works or is safe to use) ------------------ Now their reply was strange to me. They seemed to know about them and the only thing they cared about is that they were not liable. If you compare that behaviour to Apple, well, they are different. I assume they know that the chips are still made in China, but have no connection to the factory. Having some experience with the chinese market in that area, the parts might either be old RMA returns (where a defect is maybe only a wrong font used for the printing) OR some chinese guys got hold of the old factory equipment. I worked for a korean company in the past, arranging QC inspections at various locations and sending parts for customers back to China via our RMA process. And the parts almost ALWAYS showed up on Alibaba or Aliexpress some months later. ------------------- Original text of the emails: Email 1) Der Chip ist von Yamaha direkt nicht mehr lieferbar. Ob der angebotene Chip funktioniert können wir nicht sicherstellen und Yamaha gibt dafür auch keine Garantie, weil es gegebenfalls ein Sicherheitsrisiko darstellt. and asking if i get trouble from them if i import it from China: Email 2) Ob Sie diesen Chip YM3812 in China bestellen ist Ihnen frei überlassen.
Jedoch geben wir keine Garantie, dass dieser Chip funktioniert und wir weisen nochmals daraufhin, dass es gegebenfalls ein Sicherheitsrisiko darstellt, diesen einzubauen bzw. zu verwenden.
@Luis Wouldn't that be copyright? The FM patent expired in like 1995 (and immediately there were legal OPL3 clones such as Emu's CQM used in later Sb16/Awe64's)
Does Apple go after people making parts or clones of Apple II’s or similar old hardware? I’m pretty sure Apple only targets people repairing products that keep them from selling new products. If they were making clones of current Yamaha products, I’m sure they’d be more concerned.
@@Dan-TechAndMusic That's why I don't raise any qualms about using "abandonware" sites when I can't find authentic copies of old games/software for personal use. It's still technically copyright, but nobody's enforcing it, because the original authors are no longer interested in selling it, which means that you aren't competing with them. That said, if I CAN get an authentic copy, I most certainly will, if for nothing than for it's value as a collectible (especially complete box sets).
laser marking tends to be a dead giveaway. I have seen plenty of AY-3-8910's on ebay with 2015+ date codes, and the modern microchip logo, and laser marking! While the '8910 was made by microchip, it will be printed and never laser marked, and have a date code no later than probably 1993 or so. I bought a bunch of SID chips a long time ago that were remarked. They ended up using white printing on them, instead of the silver printing real chips use. The chips were indeed SID chips, but they all had bad filters on them and were obviously culls from testing.
My understanding was that the real/original factories use laser printing for the markings. The fake ones use some sort of paint. That's why acetone took the paint off the fake chip.
I bought 2 OPL2LPT (directly from the source seller as a DIY kit) right after you released the video on it in 2017. I got different YM chips in the packet. OPL2LPT: YM3812 9115 EA1B OPL2LPT: YM3812 9116 EADB The look like new old stock. Even the Pins are a little bit bent outwards and not straight like on used chips. They also work perfectly fine. The markings on mine are white and not the brownish white some chips have.
EDIT: This guy ran into some problems with fake chips, and his findings may answer some of your questions: ruclips.net/video/2OMMHwY5V5k/видео.html AFAIK they're often reverse engineered clones, and your mileage varies a lot... some clone chips work fine, some are intermittent, some don't work at all, and some are *dangerously* fake. I don't think anyone has the motivation to re-etch the labels on real chips for no reason, so if it seems off, it probably is. There are plenty that are *completely* fake chips with BS labels: me and a few other people have run into supposed SP0256 text to speech chips that were actually dangerously fake. Considering you *can't* get most of these kinds of chips from reputable sources anymore, this really makes things tough. I'm sure the guys making those boards have had to do a good amount of testing, and have thrown out boxes of chips lol
I'm making C64 SFX Sound Expander cards (OPL2). Aliexpress chips in atrocious packing, but the 10 chips were all different, just the legs were re-dipped. (+they are OBVIOUSLY thoroughly tested before I send them out)
OPL2 (YM3812) and OPL3 (YMF262) are actually fully digital so internally they have such a thing as perfect output that doesn't age (as long as the chips are still working). The conversion to analog was done using a separate DAC chip (Adlib boards used Y3014B and OPL3 boards used YAC512). However, nowadays it would be fairly easy to replace the old DACs with brand new modern ones for "perfect" sound or just take the output digitally, although one might have to put an FPGA in between for compatibility conversion. Never the less, the idea of an Adlib card with optical digital output kind of appeals to me.
I saw a recent video on Adrian Black's channel where he got some "resurfaced" RAM chips from China and he had problems with them. They worked on one model of machine but not another, while the Real chips worked on both machines.
Clint, I love that you started this channel. I’ve been following you for I think... almost a decade? I understand you upping your professionalism means less videos on your main channel. Quality takes time. But if you’re posting largely unedited and ad-hoc videos on a regular basis here, I’m totally down.
Unless you're planning on investing in an xray, we'll probably never know for sure. But like you said, repros can be just as good (if not better) than original hardware. But it's the manufacturers responsibility to let you know if that's what they are.
Kevtris's website has an analysis of eBay SID chips from over 15 years ago where he found small issues with each one that almost no one would have noticed. Yes, they "worked," but they appear to be factory rejects.
Hah, I saw a photo of my ebay listing of Sound Blaster reproductions flash by there. I would not at all be surprised that the YM3812 on them are fakes, especially since they have such a recent date code on them (13th week of 2013) but they do work exactly as they should. Yamaha in particular has been known to change the way they mark their chips over the years, with different fonts, spacing, sizing, and even different date code codes, which makes it pretty hard to determine which ones are real and which ones are not.. You know, assuming that it isnt just painted on like the one you had there. I had a different project based on the YM3802-X, and out of the 25 that I ordered from 4 different vendors, only 2 of the chips actually worked and both of them had significantly different looks to them. I did indicate the origin of the parts on my listing just in case the question ever came up.
Hi Clint, I absolutely love your new side channel. It’s absolutely LGR content how it’s meant to be: Just rambling about some sorta stuff, you know! Please keep going. Would you consider to do a a live thrift or any streaming on this channel?
When you were testing each of the chips, I immediately thought of that scene in "The Thing" where MacReady does a blood test on everyone. I need to watch that movie again.
Interesting, I needed a Yamaha YM2149F chip for my Atari ST after a botched project so I bought a couple from a China-based seller on ebay. The one in the ST has been working perfectly fine for a couple of years. From what I remember the new chip was identical but might crack it open again and have a closer look!
I can understand the issue you are trying to get across. It's like with retro game cartridge's label art. They can be tattered, scuffed, or even losing color on them yet someone could come around and replace the label to make it look brand new. It's still the same game, it's just it looks brand new even though it was a used cartridge.
(7:45) When you bring that SAAYM card up _really_ close to the camera, I can see a slight striping pattern on the Yamaha chip, almost parallel with the chip, but the striping is slightly lower on the keyed side than the unkeyed side. I don't know if I'd be able to see it in person without a magnifying desk lamp, but in this RUclips video, it is very obviously resurfaced before applying the acetone swab.
So a while back, I was liquidating large boxes that were from a computer store in the early 2000s. And I remember that within a few days, all of the old stock of sound cards were purchased up. I was thinking to myself, "Why would anyone want these old ISA/PCI sound cards??".
I remembered that 8 bit AdLib midi card. It was great for making noises on games like King's Quest and Leisure Suit Larry. Even made the scratchy sounds for the Battletech game (Crescent Hawk's revenge.)
Oh shit, I didn't know this was something to look out for O.o There's a couple projects I've been wanting to build with YM2612's, so I'm glad I saw this before just finding some random ones on the internet and buying them.
Not clones, just tatty old chips freshened up cosmetically and made to appear like "new". There really isn't the money to actually spin up a clone in an IC package.
@@LGRBlerbs Given that new-old stock is more highly regarded than recycled pulls, then in Shenzhen they are going to paint pulls to make them look like new-old stock. They could well claim that they ARE new-old stock. Whether new-old stock or recycled pulls, the key is testing the chips from the source. Tex-Elec is upfront the the chips are not necessarily "factory fresh".
@@SianaGearz That depends on whether or not you think you could get your money back. There *are* boutique factories that will make chips for you, but only if you order at least $20,000-$30,000 at once, because of the cost of retooling a production line. Having said that, though, I doubt any of them are new (aside from guaranteed new-old stock).
@@SpearM3064 The problem is that just looking at a die, you can't exactly make a copy of it. Either you need to get a copy of genuine masks and run those, or you need to fully reverse engineer the logic. Some structures like look-up tables and mask ROMs are easy to read from the die photograph, and structures giving a clue to design approaches used become apparent, but in general decoding every single gate is not entirely possible, there is a fair bit of educated guesswork and validation involved. Die shots have been invaluable in resolving reverse engineering ambiguity in hobbyist communities and getting FPGA-based clones from kinda alright to likely 100% correct, but the evaluation of die shots performed has been selective and based on prior knowledge. I had heard of someone running a very minimal new mixed signal chip for a smaller company and this ending up costing them almost half a million in setup costs. So replicating the weird YAC512 or YM3014B companion DACs from scratch is kinda out of question. If you were to run your own digital portion FM clone, you should redesign for an I2S standard audio interface. As far as digital circuits go, there are a number of very affordable processes that are available IF you have a synthesizable representation, i.e. you have fully reverse engineered and re-engineered it. But then you look at major CODEC manufacturers' attempts at designing their own OPL3 compatible circuits, like by OPTI, AD, Crystal and Creative, and all of them are very far from exact clones, super easy to distinguish, only ESS displays a commendable amount of effort at getting it almost right. So it would be a big mystery where an exact clone design would come from and how much effort someone would be willing to spend to get it right.
Awesome video, thanks for sharing! Feel free to send us the components, we'd love to x-ray them for you. There's a lot we can tell by looking inside these chips.
Fake chips are a big problem in electronics manufacturing. I worked for a manufacturing service back in the mid 2000's we had to issue a recall because we were shipped a few batches of a fake Motorola SRAM chip on an assembled controller board. We sent some failed chips to the OE and the date code was the correct format but not one they had on record! Sneaky sneaky. The issue was the chips worked in the factory but they had a high failure rate in the field. The package was not sealed (SMT) properly on the fakes so the flux and water from the assembly process got inside and corroded the leads/die. No such issue with the legitimate chips. If you are purchasing for hobby project be sure to buy your parts from a reputable vendor like Mouser or Digikey.
Everyone would be fine with these chips if the manufacturer explicitly marked them as compatibles/replacements instead of going down the counterfeit route, but I guess the margin is higher if you're selling a "vintage chip" instead of a modern reproduction. I mean it's actually kind of cool that they still send you a working product. I've heard about the fake electronics components situation before and a common tactic is to just straight up sell duds, or for large orders mix a batch of duds into some of the real ones to pad them making it look like a high failure rate instead of an outright scam. Sparkfun got hit by this during an ATMega shortage that drove them to buy from unfamiliar suppliers; they delidded the broken chips and later found out they were engineering samples of some motherboard chip that someone in China pulled out of a dumpster and relabeled.
It is common in the IC business for compatible parts to carry the same part number, just a different manufacturer mark. You can search Digikey for a part number like LM358 and virtually every major chip manufacturer has that part number, despite that the LM prefix usually means a product of Texas Instruments (via National Semiconductor.)
@@hikaru-live No one would complain about using the same part number and clearly marking that the chip is compatible or a reproduction, as long as the chips actually worked. It's people relabeling chips that don't work or dishonestly labeling non-compatible chips that causes problems.
The chip on that white card looks dodgy as even before you got the acetone out, I look forward to this series....and you are the men to do it with your stash of oldschool legit cards! Loving this new content yo!
0824 --> 24th week of 2008 ... (it could mean something else...) were they even making this chip in 2008? I saw the original video, thanks to the Hack-A-Day Blog. Very nice in-depth review. Cheers,
Interesting topic. I have a batch of YM2612's that I got off ebay for next to nothing (about 10 of them) Can't say if they're real or fake, but their appearance does suggest the ebay listing that claimed they were 'new' is pretty dubious...
The main thing from my perspective, as a user, is that it doesn't matter as long as the chips work correctly. If I was a collector and I wanted everything to be as original as possible, I could see myself having a problem there. The real issue is when you order something like some memory chips that simply will not work as stated.
I've started having to worry about fakes since I've started getting into collecting old video game stuff. Honestly I don't see a problem with 'fakes' as long as they work right and are advertised as such. Since some of that stuff is just too expensive to pay for now a days but it'd still be nice to have something close enough to add to your collection/to use.
This is a big issue in the audio world. I do audio repair and restorations and there are transistors and ICs (opamps) out there that are what work best in certain circuits. There are modern equivalents that can work, but can send the design into oscillation if you’re not careful. Anyways, there are a lot of fake transistors and ICs out there and if I can I try to avoid them. Avoid eBay and other direct from China options. People have broken these devices down and the findings are that basically they’ll work, but usually they use smaller dies and the device won’t be anywhere as robust or meet the same specs as the originals (voltage, distortion, etc). Right now I’m repairing a Yamaha integrated amplifier and it uses a Sanken power module for the output. My only real option for fixing it is to buy a fake from China, or wait for years until a used one shows up. I guess you can redesign the circuit, but at that point it all isn’t worth it. It would be incredible if a company took pride in their product and started manufacturing all of these devices that are sought after but no longer made to original factory standards.
Thanks for the video. I have an OPL2LPT too, but it doesn't work. I am fairly sure that I soldered everything correctly. It could be that the chip is defect. Even if it wasn't a forgery, now I know to watch out if I go looking for a replacement. (I wonder how I can test if the chip is working)
With regard to fake chips, the SID chips don’t really work quite right for a number of reasons, but also there’s the issues with AY3-8912’s. I have experienced fake AY’s that whilst they never specified a significantly asymmetric clock input would work, the originals _do_ but the fakes *don’t*.
I heard somewhere that some new old stock chips are surfacing but in reality they are old faulty chips that fell out during initial production, but they weren't destroyed and today they are selling them. I don't mind if they are fake chips as long as they work. The Yamaha chip is not a big problem because they are digital, but the SID chip is much more complex, parts of it analog, parts of it digital, parts of it using 9VAC and stuff... The original is no longer manufactured and the exact technology of making them is not out there or something...
I forgot the Raiden PCB was such a bootleg-looking board with bodge-wire solutions and the daughterboard. Shaped up immensely for Raiden II that followed.
That's why you always buy things through Paypal, document the testing / validation process, and always file a dispute and get your money back if the parts are fakes or scams.
So for sound chips this is pretty innocuous I think... but I wonder if this is also happening for chips with batteries like the Dallas Realtime Clock chips. I have at least a few dead RTCs that were seemingly 'new' :(
Yeaa, I think absolutely for reproduction ICs. I kind of appreciate having something like Yamaha branding on a chip that WAS after all a Yamaha design. Like even if they were 100% fabricated as a new IC, and granting it worked out of the factory, like??? Hey??? An otherwise dead card just got a second wind.
As I understand, actual fake chips are fairly rare. Usually what happens is that these places in China remove chips from broken equipment and resurface them in an attempt to pass them off as new/unused. Most of the time, they will work fine. Occasionally, you may get one that is damaged in some way. Sometimes, if multiple versions of a chip were available, they may remove the markings of a less desirable version and replace them with the markings of a more desirable version. In those cases, the chip may work, but perhaps not perform as well as its markings would indicate.
LOL. But let me tell something. RS232 to USB chips made a SUPER BAD reputations because of fakes. This is extensive to PROLIFIC chipsets PL2303. In the end the are ruining their image. Getting originals is complicated. after having issue over issue with the PL2303 I moved back to old laptops with XP and REAL serial port to avoid conerters. But this FT232 are ALL OVER in developement kits In the end I like WCH, they are reasonably good and cheap enough to defeat counterfeits. But WCH driver install process is not direct, their download page is in chinesse so people find it a bit frustrating. I have used a few WCH products and no probs.
Here is a better question. Why not hook it up to a dynamic signal analyzer or a spectrum analyzer and see if the probably counterfit chips are performing on spec. If they aren't well that's an issue. I think your qtip is probably just taking the ink from the label on the chip off. Are they laser engraved or just stamped with ink?
I got some YM3812-F's off EBay a while ago, and I think the ones I got are legit, at least at the die level... haven't tried the acetone thing to check for resurfacing, but I only bought them after someone I trust on twitter bought some from the same seller and de-capped the chip and got the die under a microscope.
If the chips work, it's fine for me. If I could find a decent quality USB/MIDI module based on a OPL2 chip on the market, to make my synth music with it, it would be an instant buy (fake or not)
Yeah I had to take my time when sourcing YM3438 chips for sega modifications. I finally found a place that was from China of course, but they provided me with an actual shipping/packing slip of the order and the chips were in an anti-static tube etc. Having said that...they still all looked the same and had date codes from 2004? right....... But they all worked without issue as I made a testing rig just for that purpose to make sure before installing them into client systems.
Being a watch collector, I understand the fake stuff is a real problem. Fake Rolex watches are bad. Not so sure that fake chips are a big deal. If they function as they should, does it really matter? I guess for the hardcore collector/hobby person it matters. But if one is looking for function only, I think to myself "no big deal". But that's just me.
The people who can answer this is Yamaha. They should make a statement whether they did indeed make a batch of chips in mid-2008. It may have been a new run, or it may simply been wafers that have been in storage that they diced up and packaged. The demand for these things are tiny in the scheme of things. A single run of thousands of chips, less than a week, could easily satisfy the demand.
@@LGRBlerbs I saw that but it didn't answer the question. They said that they couldn't guarantee the authenticity of chips sold on the second-hand market. All companies will say that if you're not getting the chips direct from an authorized distributor. If they are indeed fake, Yamaha should have no problem saying that all chips with a datecode of 0824 are fake. If they can't, then it must be because some are real.
Is there a difference in sound quality between legit and fake chips? How about a video using genuine and fake chips in the same board and comparing the audio?
As mentioned in the video, these particular chips sound good to my ears. It's just a question of where they originally came from, whether they're new old stock or resurfaced used chips or what. No one really seems to know and that's odd.
Early OPL clones often had bad sound because the cloners hadn't figured out that Yamaha's patents on FM synthesis contained subtly wrong information. Specifically Yamaha was actually using phase modulation not frequency modulation because frequency modulation caused unpleasant harmonics with certain frequency combinations. FM is a purely digital synthesis unlike e.g. a SID chip so it's easy to reproduce once you know exactly how the chip works but Yamaha had omitted one key piece of information from their patents and never told anyone (TBF they figured this out later and then neglected to update the patents)
If someone is actually making a 1:1 copy, the chip would sound the same. It's actually fully digital! Analog audio only comes out of the external DAC (YM3014)
@@kveemusic There is a filter on the output but other than that yes it's completely digital inside. Early copies were apparently just based on a block diagram of the chip and what the patents said about FM synthesis and so had sound issues because they didn't know Yamaha was actually using phase modulation. Even if you decap the chip it would not be immediately obvious that it's using PM instead of FM. They are almost identical except for one small difference and even that is pretty slight (I know because I built a virtual phase modulation operator inside Synthedit years ago to replace the flawed FM operator) Years ago I built a virtual DX7 using Synthedit and found it didn't sound quite right and, after some digging online I found out about this and changed the operators to PM (had to actually build a PM operator inside Synthedit because it didn't have a premade one) and then it sounded like a real DX-7. The difference comes down to one small change in how the operator modifies the carrier other than that they are identical. The thing is I really didn't even know exactly what the difference was I had read that Yamaha used phase modulation rather than frequency modulation in their "FM" chips (there wasn't a lot of information on this) and I thought "hmm well maybe if I just have the modulator modulate the phase of the carrier rather than it's frequency" tried it and suddenly it sounded like a proper DX-7.
Found this animated GIF that visually represents what's going on: en.wikipedia.org/wiki/File:Phase-modulation.gif This is what the synthesis inside a Yamaha "FM" chip is actually doing. The reason they use PM instead of FM is because FM causes nasty sounding harmonics at certain operator / carrier frequency combinations. Other than the unpleasant artifacts the same operator/carrier parameters with FM and PM synthesis sound basically identical to the human ear even if mathematically the output waveform is a slightly different.
Don't get your hopes up, especially not with some of them like the 6581 SID chip used in the Commodore 64. It's a mix of analog and digital circuits, using a manufacturing process nobody's used in over 20 years. Even the "boutique factories" that will make production runs of other computer chips won't touch them. (The 8580 used in the 64C is a different story. You can still get a boutique factory to make those, *if* you order about $30,000 worth.)
I have no problem with people selling replica chips that are functionally identical to the original, just communicate that they are replicas; it's not like MOS or Yamaha are making any more (to my knowledge), and the originals are aging, and won't last forever.
The RUclipsr ElectronUpdate might be interested in this. He specializes in de-encapsulating chips and breaking down the structures. It would be interesting for him to compare a real and suspect chip on the inside. That would tell you for sure.
@@cassandra-show But then someone has to send him a surely original Yamaha IC for comparison. A known original BUT dead Yamaha chip, because decapping a known working original is a sin. But if these are clones, I very much doubt they are re-designed, pin compatible dies. It would be a huge work and money to make. Maybe someone got the original films.
yeah I don't like the fact that they aren't stating that they resurfaced the chips, I would also be okay if they were reproductions as long as they say they are, at the end of the day if it sounds and performs the same thats all that matters.
I agree, companies should be more open about there products. Especially in markets where cloning is a big problem. Refurbished and rebranded, that's all I'd like to see on a chip if it's an up-cycle.
Could it also be that these chips aren't made in huge demand, so when a batch is made, they're made in batches of tens or hundreds of thousands all at once. So the low volume would mean a batch would sell out slowly, thus a lot of them would all have the same date codes.
The only problem I see is they don't advertise them as repo and try to pass them off as counter fit. However the chips you have could be a refurbish too.
Pretty sure those things were just made in a huge batch towards the end of their life or maybe even after their life. From that point, tons of them just sit around in the warehouse until they're sold off in bulk as new old stock, where I think most of the "new" ones come from - all the same batch.
On the flipside LGR...I imagine that getting older, proper chips are not all that easy to come by these days for things like this. Older chips winding their way into e-waste because the PC no longer turns on and everything else just gets tossed/recycled, just plain burning out, that sort of thing. And by now, you can obviously find the schematics for making these chips online, 8 Bit Guy's used them before. Another idea to propose: Could the chips in question...be reverse engineered new stock chips created to answer for a shortage in what's rapidly becoming a newfound field of demand? More and more people are wanting to get these old machines up and running, without being able to find old stock parts now, after all. There are bound to be companies out there who've sourced out a Chinese factory to manufacture these for sale on the cheap for use cases like this. And in the case of the SAAYM card you mentioned, it wouldn't shock me if cards like that were using said chips. In cases like this, if they put the time in to make them sound accurate...is there really anything dubious about meeting said demand in the best, easiest way possible? I know that's what's going on with some SID chips, would make sense that some of that is happening with these too.
Actually, that's *not* what's going on with some SID chips. No legitimate manufacturer will even try to reproduce them. What you're seeing is hardware emulation of the SID chip using an FPGA. Which is great, but it's only as accurate as the emulation it's based on.
'Fake' in the sense of compatible doesn't have to be bad however when they sell it as the real thing (with company name and such) there is a problem. The only way to find out it is different is to decap the thing (with some toxic goodies) and compare the layout of the die (you need a (electron) microscope for this) to the original (few people did this with with branded 'clones' of originals). However, these are old chips and it could be possible the manufacturer update/improve the layout to make it cheaper to produce. Over time, it can be very hard to tell what is original or not, the painted one is definitely not an original (could be a bitbanged microcontroller, a die clone, an EPROM or even a small FPGA with fancypancy painting on top). Do you know the date code is really a date code, maybe it is a batch code and a batch can be huge or a version code, could be also thousands or millions of the same. Hard to tell when hardware gets older. A good example of (more recent) counterfeit is FTDI chips, google it, very interesting.
If there's a dime to be made then China will make repros of an item. These are definitely repros. It's a common problem with old cartridge games too. Ebay does nothing about the countless auctions from China selling repro games for Nintendo 64 and Snes.
It's a well known practice. Take 75 working ICs and take 25 random ICs that have the same number of pins/form factor. Relabel all 100 ICs and sell them. You just increased your profit with minimal/acceptable impact on your customer feedback.
In an effort to be completely transparent, we decided to make a video with regards to grey-market chips. It's better than the Super Bowl halftime show, so queue it up! ruclips.net/video/eIInfDmITPU/видео.html
Ahh, most excellent. Thank you for putting this together, this should prove helpful in the future 👍
Thanks a lot for the shoutout! The first one you tried to put acetone on are of the "hardened and cooked and laser etched" variety, just like me only a little something comes off using acetone. However your YM2151 I _knew_ would come off clean haha.
You can check that your YM3812 are remarked by comparing the top and bottom texture (they should match if they are non-tampered with), and the ejector holes should be shiny and not covered
I appreciate the tip! Out of morbid curiosity, I've gone ahead and ordered fifteen OPL2 chips from various vendors to test myself. This is a fascinating rabbit hole.
@@LGRBlerbs Please tell us you will record your unboxing? haha. Oh here is my full checklist :
1)Texture must be the same on bottom and top
2)Pins should be straight with no solder or flux traces.
3)Acetone should not alter the chip's surface.
4)Datecode should be in the 80s or 90s (for most DIP YMxxxx chips)
5)Ejector pin marks should be shiny for most casing types
(and not covered or sanded)
6)No laser etching on those 80s or 90s YM chips
Anything else is proof that the chip has been tampered with!
I pictured Clint trying to prove a point and reaching for a wood chisel. "See!? It totally comes off!"
@@PJBonoVox "See, hydrochloric acid removes it!" XD
It'a frustrating that they don't label them as repo, refurb, or old stock.
lazy people
@@flamestoyershadowkill More like feaudsters. But genuine new old stock would look like that set of 5 and might be sold as such.
For those unaware, "new old stock" means original things that were never used for anything and have been sitting in a warehouse since they were made back in the day. MOS/Commodore is no longer around as a company, but Yamaha is and might still be doimg a production run every 6 years to stock up the spare parts supply to musicians with Yamaha keyboards.
@@johndododoe1411 ok
Here's that video that goes more in-depth: ruclips.net/video/k72SFBOZ_lw/видео.html
To be clear, I'm cool using reproduction or resurfaced chips so long as they function the same or better. I just wish that the chip makers would _make it clear that they're clones_ if that's the case, since I like to know if I'm dealing with brand new or vintage stuff when it comes to hardware. Buying old sound chips is enough of a gamble already, no need to obscure things further.
LGR Blerbs I’d prefer having properly licensed chips if the cost was reasonable. I can‘t imagine reasonable licensing cost for something like OPL2.
Building a input/response tester should be doable. That way chips could have verified behaviour. I have a hard time imagining it being profitable to manufacture new clones. Resurfacing I can imagine. Silly, though.
When I was in school the instructor passed around a few Cisco interface cards... I wanted to know how much they would cost, and one of the first things that came up in a Google search was a page all about how you can tell if a card is a fake or not. One of the example cards the school had turned out to be a bootleg. Found a number of fakes at my last job in the piles of cards they had, and the effort put into making them look real varied considerably, everything from faking the hologram stickers to not even remotely the same color on the components.
"People should know things." Clint, 2020
When the patent ran out on "FM" synthesis (more on that in a moment) lots of skeezy third party compatible chips started to appear. The thing is, the patent sort of lied because Yamaha FM sound chips actually use phase modulation not frequency modulation. Early on Yamaha learned that FM produced nasty harmonics and they could get the same synthesized sounds without the unpleasant artifacts if they used phase modulation in the operators rather than frequency modulation, and then kept that information a closely guarded secret within the company. Remember that FM chips are really a form of digital synthesis, it's a bunch of digital adders working with 12 bit digital waves so how it works isn't immediately obvious if you decap the chip and look at the circuit and it was a very long time before anyone thought to check to see what exactly was going on inside a Yamaha FM chip. People just assumed it was doing what the patents said it was doing when it fact it was actually slightly different than what was described in the patents.
I assume that's why they never sued Casio over their phase distortion synths - because they didn't want to have to admit that their own synths weren't actually FM? But then again I guess the difference is modulation vs distortion. I personally prefer the Casio/PD sound but you can't get the kinds of metallic sounds FM/PM gives you unless you have a VZ-1 - AFAIK the only synth to have fully implemented the idea of PD synthesis. I had one for a while but it was an absolute pig to program so I went back to the more limited architecture of the CZ-3000 and more recently the CZ-1. Only thing from the VZ-1 that I miss is that you could set the pitch wheel to have a six-octave bend range - three in each direction!
That is super interesting. Thanks for sharing
@@theidealcopy_ No phase distortion synthesis and frequency modulation/phase modulation synthesis actually work differently. Professional FM synthesizers like the DX7 used 6 operators and were much better sounding than 2 operator and 4 operator FM chips in consumer hardware. There was also the TG-77 and it's relatives that used a 6 operator FM synthesizer which could use sampled sounds as operator waveforms instead of just sine waves like the DX-7. It had a 6 operator FM synthesizer (basically a DX-7 but with fully programmable operators) and a PCM section that could be used as a ROM sampler (it had a set of built in sounds and card slots for additional sound libraries) or the ROM sounds could also be used as waveforms in the FM section.
Also look up the Yamaha FS-1R which was the last FM synthesizer Yamaha made. Ir was quite powerful but somewhat flawed. A bad hard to understand interface, slow to download parameters and proper editing tools never materialized during it's production (there were editors created for it long after it was discontinued though). It still fetches $1000-$1500 on ebay for a working unit in good condition.
Where's the money in spending about half a million give or take to spin up a slightly shoddy pin compatible replacement of OPL2 or OPN? I wonder if Yamaha simply sold a number of genuine chips with different markings, or if they ever sold bare dies to be packaged elsewhere.
As to late 90s flood of OPL3 software-compatible logic made by CODEC manufacturers, those really made sense since the CODEC was to be spun anyway, so it would be a cost saving to put the synthesizer right on there.
blerbposting just keeps getting better
Indeed
Clint: "I could do this all day"
Me: *I could watch this all day*
Yeah, imagine him going through all the chips like this he owns on video.
Me too
Clint, don't worry; those of us who've been following you for some time know that you would never try to mislead people. I for one can't imagine a more honest and down-to-earth guy than you.
Yeah
Replica would be a great word if sellers were to disclose the info.
I would also love to see someone playing around with the functions of the chips on an oscillioscope to see how much difference there is if any.
The OPL chips require an external DAC, so you don’t even need an oscilloscope - just a logic analyzer! :-D
Often, modern counterfeits of old logic chips are software reimplementations on a commonplace microcontroller (a small self-contained computer).
I don't know if there is "firmware"
openly available to make your own.
I can see someone making a bsnes/higan style "cicle perfect" replacement.
replica might infringe on copyright layouts
As someone who repairs all types of vintage electronics, I'm REALLY sick of all the fake semiconductors and IC's out there. We need more places that salvage authentic parts from scrapped equipment. I was pretty happy to save some 6532 and 6802 IC's from some scrap boards the other day. Great for fixing arcade stuff :-)
I'm not particularly worried as long as they work fine and they aren't priced as if they're genuine.
Unfortunately, some of them are indeed labeled "genuine" and sold at a markup. I've ordered some myself to try and show this in the future.
If they are not manufactured by Yamaha then it is not ok. It is counterfeit. Even if they work perfectly the same, and have same layout internally.
The problem is that they're always priced marked as genuine. China doesn't make repros for any other reason than to make as much money as possible.
It won’t work the same. They sound like crap. I had the misfortune to only hear OPL clones through my life (first a shitty OPL3 clone made by Oak technologies/OPTi found on a cheap “Pro Multimedia PMM-240” Card, another Creative’s infamous CQM OPL3 clone that was found on the AWE64) and they all suck! Grand Pianos sound like toy pianos! Guitars sounds like they had plastic strings! Drums lack depth and reverb and sound like tiny little toys! And in AWE’s case, the instruments were quite soft in some games and needed the volume to be turned up ridiculously high!
RAMChYLD Modern clones can be made from die scans if they wanted to, Yamaha won't care anymore.
I once asked Yamaha (office in Germany) directly if the Chinese YM3812 were "legal" to import to Germany. And they replied: (original text at bottom of long comment)
-------------------
The chip is not supplied by Yamaha anymore.
Whether the offered chip works, we can not guarantee and Yamaha can also not guarantee that it is not a security risk.
and asking if i get trouble from them if i import it from China:
Whether you buy this chip from China is up to you.
(And they repeat that they take no guarantees or know if the chip works or is safe to use)
------------------
Now their reply was strange to me. They seemed to know about them and the only thing they cared about is that they were not liable.
If you compare that behaviour to Apple, well, they are different.
I assume they know that the chips are still made in China, but have no connection to the factory. Having some experience with the chinese market in that area, the parts might either be old RMA returns (where a defect is maybe only a wrong font used for the printing) OR some chinese guys got hold of the old factory equipment.
I worked for a korean company in the past, arranging QC inspections at various locations and sending parts for customers back to China via our RMA process. And the parts almost ALWAYS showed up on Alibaba or Aliexpress some months later.
-------------------
Original text of the emails:
Email 1)
Der Chip ist von Yamaha direkt nicht mehr lieferbar.
Ob der angebotene Chip funktioniert können wir nicht sicherstellen und Yamaha gibt dafür auch keine Garantie, weil es gegebenfalls ein Sicherheitsrisiko darstellt.
and asking if i get trouble from them if i import it from China:
Email 2)
Ob Sie diesen Chip YM3812 in China bestellen ist Ihnen frei überlassen.
Jedoch geben wir keine Garantie, dass dieser Chip funktioniert und wir weisen nochmals daraufhin, dass es gegebenfalls ein Sicherheitsrisiko darstellt, diesen einzubauen bzw. zu verwenden.
Interesting, thank you for sharing!
Those FM chips were made in the 80's in, like, NMOS... Do Chinese clone chips really make sense?
@Luis Wouldn't that be copyright? The FM patent expired in like 1995 (and immediately there were legal OPL3 clones such as Emu's CQM used in later Sb16/Awe64's)
Does Apple go after people making parts or clones of Apple II’s or similar old hardware? I’m pretty sure Apple only targets people repairing products that keep them from selling new products. If they were making clones of current Yamaha products, I’m sure they’d be more concerned.
@@Dan-TechAndMusic That's why I don't raise any qualms about using "abandonware" sites when I can't find authentic copies of old games/software for personal use. It's still technically copyright, but nobody's enforcing it, because the original authors are no longer interested in selling it, which means that you aren't competing with them.
That said, if I CAN get an authentic copy, I most certainly will, if for nothing than for it's value as a collectible (especially complete box sets).
laser marking tends to be a dead giveaway. I have seen plenty of AY-3-8910's on ebay with 2015+ date codes, and the modern microchip logo, and laser marking! While the '8910 was made by microchip, it will be printed and never laser marked, and have a date code no later than probably 1993 or so. I bought a bunch of SID chips a long time ago that were remarked. They ended up using white printing on them, instead of the silver printing real chips use. The chips were indeed SID chips, but they all had bad filters on them and were obviously culls from testing.
My understanding was that the real/original factories use laser printing for the markings. The fake ones use some sort of paint. That's why acetone took the paint off the fake chip.
Other way around Ian. The acetone takes the marking off because it also is taking off the surface below.
I've the OPL2LPT too, it works, and has the SAME datecode: "0824 EAYB"
I've a pack of other OPL2 Chips (new-old stock) with the datecode: "0938 GAGD"
"0938 GAGD" is a remarked code. No OPL2 chips were made in week 38 of 2009, and no OPL2 chips are die revision D.
I'm only a minute and a half in... But this looks dangerously educational. Or at least informative. Not the blerbs we've become accustomed to!
😁
I bought 2 OPL2LPT (directly from the source seller as a DIY kit) right after you released the video on it in 2017. I got different YM chips in the packet.
OPL2LPT: YM3812 9115 EA1B
OPL2LPT: YM3812 9116 EADB
The look like new old stock. Even the Pins are a little bit bent outwards and not straight like on used chips. They also work perfectly fine. The markings on mine are white and not the brownish white some chips have.
EDIT: This guy ran into some problems with fake chips, and his findings may answer some of your questions: ruclips.net/video/2OMMHwY5V5k/видео.html
AFAIK they're often reverse engineered clones, and your mileage varies a lot... some clone chips work fine, some are intermittent, some don't work at all, and some are *dangerously* fake. I don't think anyone has the motivation to re-etch the labels on real chips for no reason, so if it seems off, it probably is.
There are plenty that are *completely* fake chips with BS labels: me and a few other people have run into supposed SP0256 text to speech chips that were actually dangerously fake.
Considering you *can't* get most of these kinds of chips from reputable sources anymore, this really makes things tough. I'm sure the guys making those boards have had to do a good amount of testing, and have thrown out boxes of chips lol
I've seen that, definitely a good video on the topic!
If they're functioning the same, I dunno if I'd really care, unless I'm paying a premium for a real vintage part
I'm making C64 SFX Sound Expander cards (OPL2). Aliexpress chips in atrocious packing, but the 10 chips were all different, just the legs were re-dipped. (+they are OBVIOUSLY thoroughly tested before I send them out)
Currently watching this on a 10" Panasonic black and white tv from 1969
OPL2 (YM3812) and OPL3 (YMF262) are actually fully digital so internally they have such a thing as perfect output that doesn't age (as long as the chips are still working). The conversion to analog was done using a separate DAC chip (Adlib boards used Y3014B and OPL3 boards used YAC512). However, nowadays it would be fairly easy to replace the old DACs with brand new modern ones for "perfect" sound or just take the output digitally, although one might have to put an FPGA in between for compatibility conversion. Never the less, the idea of an Adlib card with optical digital output kind of appeals to me.
"people should know things" ~ Clint 2020
I knew you had another channel, but I didn't know about this one...that's pretty cool
I saw a recent video on Adrian Black's channel where he got some "resurfaced" RAM chips from China and he had problems with them. They worked on one model of machine but not another, while the Real chips worked on both machines.
Clint, I love that you started this channel. I’ve been following you for I think... almost a decade? I understand you upping your professionalism means less videos on your main channel. Quality takes time. But if you’re posting largely unedited and ad-hoc videos on a regular basis here, I’m totally down.
Unless you're planning on investing in an xray, we'll probably never know for sure. But like you said, repros can be just as good (if not better) than original hardware. But it's the manufacturers responsibility to let you know if that's what they are.
Kevtris's website has an analysis of eBay SID chips from over 15 years ago where he found small issues with each one that almost no one would have noticed. Yes, they "worked," but they appear to be factory rejects.
Hah, I saw a photo of my ebay listing of Sound Blaster reproductions flash by there.
I would not at all be surprised that the YM3812 on them are fakes, especially since they have such a recent date code on them (13th week of 2013) but they do work exactly as they should.
Yamaha in particular has been known to change the way they mark their chips over the years, with different fonts, spacing, sizing, and even different date code codes, which makes it pretty hard to determine which ones are real and which ones are not.. You know, assuming that it isnt just painted on like the one you had there.
I had a different project based on the YM3802-X, and out of the 25 that I ordered from 4 different vendors, only 2 of the chips actually worked and both of them had significantly different looks to them.
I did indicate the origin of the parts on my listing just in case the question ever came up.
Hi Clint, I absolutely love your new side channel. It’s absolutely LGR content how it’s meant to be: Just rambling about some sorta stuff, you know!
Please keep going. Would you consider to do a a live thrift or any streaming on this channel?
When you were testing each of the chips, I immediately thought of that scene in "The Thing" where MacReady does a blood test on everyone. I need to watch that movie again.
According to Mr. H (another RUclips channel), they're doing *yet another* remake of The Thing. Here's hoping this one is actually good. ;^)
Interesting, I needed a Yamaha YM2149F chip for my Atari ST after a botched project so I bought a couple from a China-based seller on ebay. The one in the ST has been working perfectly fine for a couple of years. From what I remember the new chip was identical but might crack it open again and have a closer look!
I can understand the issue you are trying to get across. It's like with retro game cartridge's label art. They can be tattered, scuffed, or even losing color on them yet someone could come around and replace the label to make it look brand new. It's still the same game, it's just it looks brand new even though it was a used cartridge.
A kind of disturbing amount of people in these comments that are apathetic to being mismarketed to.
(7:45) When you bring that SAAYM card up _really_ close to the camera, I can see a slight striping pattern on the Yamaha chip, almost parallel with the chip, but the striping is slightly lower on the keyed side than the unkeyed side. I don't know if I'd be able to see it in person without a magnifying desk lamp, but in this RUclips video, it is very obviously resurfaced before applying the acetone swab.
So a while back, I was liquidating large boxes that were from a computer store in the early 2000s. And I remember that within a few days, all of the old stock of sound cards were purchased up. I was thinking to myself, "Why would anyone want these old ISA/PCI sound cards??".
Thanks for the info! I had heard about fake chips but it was always in regards to CPUs or GPUs. Glad to find out any chip is fair game. :(
I remembered that 8 bit AdLib midi card. It was great for making noises on games like King's Quest and Leisure Suit Larry. Even made the scratchy sounds for the Battletech game (Crescent Hawk's revenge.)
Oh shit, I didn't know this was something to look out for O.o There's a couple projects I've been wanting to build with YM2612's, so I'm glad I saw this before just finding some random ones on the internet and buying them.
I always assumed the OPLs you buy from Texelec or Serda are good quality clones at least. I'm sure they would get complaints if they weren't.
As mentioned in the video, the audio quality is quite good IMHO. It's just origin of the chips that is suspect.
Not clones, just tatty old chips freshened up cosmetically and made to appear like "new".
There really isn't the money to actually spin up a clone in an IC package.
@@LGRBlerbs Given that new-old stock is more highly regarded than recycled pulls, then in Shenzhen they are going to paint pulls to make them look like new-old stock. They could well claim that they ARE new-old stock. Whether new-old stock or recycled pulls, the key is testing the chips from the source. Tex-Elec is upfront the the chips are not necessarily "factory fresh".
@@SianaGearz That depends on whether or not you think you could get your money back. There *are* boutique factories that will make chips for you, but only if you order at least $20,000-$30,000 at once, because of the cost of retooling a production line. Having said that, though, I doubt any of them are new (aside from guaranteed new-old stock).
@@SpearM3064 The problem is that just looking at a die, you can't exactly make a copy of it. Either you need to get a copy of genuine masks and run those, or you need to fully reverse engineer the logic. Some structures like look-up tables and mask ROMs are easy to read from the die photograph, and structures giving a clue to design approaches used become apparent, but in general decoding every single gate is not entirely possible, there is a fair bit of educated guesswork and validation involved. Die shots have been invaluable in resolving reverse engineering ambiguity in hobbyist communities and getting FPGA-based clones from kinda alright to likely 100% correct, but the evaluation of die shots performed has been selective and based on prior knowledge.
I had heard of someone running a very minimal new mixed signal chip for a smaller company and this ending up costing them almost half a million in setup costs. So replicating the weird YAC512 or YM3014B companion DACs from scratch is kinda out of question. If you were to run your own digital portion FM clone, you should redesign for an I2S standard audio interface.
As far as digital circuits go, there are a number of very affordable processes that are available IF you have a synthesizable representation, i.e. you have fully reverse engineered and re-engineered it. But then you look at major CODEC manufacturers' attempts at designing their own OPL3 compatible circuits, like by OPTI, AD, Crystal and Creative, and all of them are very far from exact clones, super easy to distinguish, only ESS displays a commendable amount of effort at getting it almost right. So it would be a big mystery where an exact clone design would come from and how much effort someone would be willing to spend to get it right.
Awesome video, thanks for sharing! Feel free to send us the components, we'd love to x-ray them for you. There's a lot we can tell by looking inside these chips.
Fake chips are a big problem in electronics manufacturing. I worked for a manufacturing service back in the mid 2000's we had to issue a recall because we were shipped a few batches of a fake Motorola SRAM chip on an assembled controller board. We sent some failed chips to the OE and the date code was the correct format but not one they had on record! Sneaky sneaky. The issue was the chips worked in the factory but they had a high failure rate in the field. The package was not sealed (SMT) properly on the fakes so the flux and water from the assembly process got inside and corroded the leads/die. No such issue with the legitimate chips.
If you are purchasing for hobby project be sure to buy your parts from a reputable vendor like Mouser or Digikey.
Everyone would be fine with these chips if the manufacturer explicitly marked them as compatibles/replacements instead of going down the counterfeit route, but I guess the margin is higher if you're selling a "vintage chip" instead of a modern reproduction.
I mean it's actually kind of cool that they still send you a working product. I've heard about the fake electronics components situation before and a common tactic is to just straight up sell duds, or for large orders mix a batch of duds into some of the real ones to pad them making it look like a high failure rate instead of an outright scam. Sparkfun got hit by this during an ATMega shortage that drove them to buy from unfamiliar suppliers; they delidded the broken chips and later found out they were engineering samples of some motherboard chip that someone in China pulled out of a dumpster and relabeled.
It is common in the IC business for compatible parts to carry the same part number, just a different manufacturer mark. You can search Digikey for a part number like LM358 and virtually every major chip manufacturer has that part number, despite that the LM prefix usually means a product of Texas Instruments (via National Semiconductor.)
@@hikaru-live No one would complain about using the same part number and clearly marking that the chip is compatible or a reproduction, as long as the chips actually worked. It's people relabeling chips that don't work or dishonestly labeling non-compatible chips that causes problems.
That reminds me... I still need to put an ad-lib in my DEC Venturis...
I don't care as much about if the chip is authentic, as long as it is accurate and compatible.
The chip on that white card looks dodgy as even before you got the acetone out, I look forward to this series....and you are the men to do it with your stash of oldschool legit cards! Loving this new content yo!
0824 --> 24th week of 2008 ... (it could mean something else...)
were they even making this chip in 2008?
I saw the original video, thanks to the Hack-A-Day Blog. Very nice in-depth review.
Cheers,
Interesting topic.
I have a batch of YM2612's that I got off ebay for next to nothing (about 10 of them)
Can't say if they're real or fake, but their appearance does suggest the ebay listing that claimed they were 'new' is pretty dubious...
The main thing from my perspective, as a user, is that it doesn't matter as long as the chips work correctly. If I was a collector and I wanted everything to be as original as possible, I could see myself having a problem there. The real issue is when you order something like some memory chips that simply will not work as stated.
Perhaps a video showing off the difference in sound, if any, between a real and fake chip is in order?
I've started having to worry about fakes since I've started getting into collecting old video game stuff. Honestly I don't see a problem with 'fakes' as long as they work right and are advertised as such. Since some of that stuff is just too expensive to pay for now a days but it'd still be nice to have something close enough to add to your collection/to use.
This is a big issue in the audio world. I do audio repair and restorations and there are transistors and ICs (opamps) out there that are what work best in certain circuits. There are modern equivalents that can work, but can send the design into oscillation if you’re not careful.
Anyways, there are a lot of fake transistors and ICs out there and if I can I try to avoid them. Avoid eBay and other direct from China options. People have broken these devices down and the findings are that basically they’ll work, but usually they use smaller dies and the device won’t be anywhere as robust or meet the same specs as the originals (voltage, distortion, etc).
Right now I’m repairing a Yamaha integrated amplifier and it uses a Sanken power module for the output. My only real option for fixing it is to buy a fake from China, or wait for years until a used one shows up. I guess you can redesign the circuit, but at that point it all isn’t worth it.
It would be incredible if a company took pride in their product and started manufacturing all of these devices that are sought after but no longer made to original factory standards.
Thanks for the video. I have an OPL2LPT too, but it doesn't work. I am fairly sure that I soldered everything correctly. It could be that the chip is defect. Even if it wasn't a forgery, now I know to watch out if I go looking for a replacement.
(I wonder how I can test if the chip is working)
With regard to fake chips, the SID chips don’t really work quite right for a number of reasons, but also there’s the issues with AY3-8912’s. I have experienced fake AY’s that whilst they never specified a significantly asymmetric clock input would work, the originals _do_ but the fakes *don’t*.
I heard somewhere that some new old stock chips are surfacing but in reality they are old faulty chips that fell out during initial production, but they weren't destroyed and today they are selling them. I don't mind if they are fake chips as long as they work. The Yamaha chip is not a big problem because they are digital, but the SID chip is much more complex, parts of it analog, parts of it digital, parts of it using 9VAC and stuff... The original is no longer manufactured and the exact technology of making them is not out there or something...
If it's a 1:1 copy, personally I don't see a problem...
I didn't even know this channel existed. Thanks RUclips recommendations
I forgot the Raiden PCB was such a bootleg-looking board with bodge-wire solutions and the daughterboard. Shaped up immensely for Raiden II that followed.
I am having this issue trying to find specific cache ram for a DEC 486. I order chips and the speed is not what is printed on the chip.
That's why you always buy things through Paypal, document the testing / validation process, and always file a dispute and get your money back if the parts are fakes or scams.
So for sound chips this is pretty innocuous I think... but I wonder if this is also happening for chips with batteries like the Dallas Realtime Clock chips. I have at least a few dead RTCs that were seemingly 'new' :(
Even if the are new chips the important thing is their internal cct is the same as original so thanks for building old chips
This feels like OG LGR 😂😂 keep em coming!
Yeaa, I think absolutely for reproduction ICs. I kind of appreciate having something like Yamaha branding on a chip that WAS after all a Yamaha design. Like even if they were 100% fabricated as a new IC, and granting it worked out of the factory, like??? Hey??? An otherwise dead card just got a second wind.
Already love this "new" channel
As I understand, actual fake chips are fairly rare. Usually what happens is that these places in China remove chips from broken equipment and resurface them in an attempt to pass them off as new/unused. Most of the time, they will work fine. Occasionally, you may get one that is damaged in some way.
Sometimes, if multiple versions of a chip were available, they may remove the markings of a less desirable version and replace them with the markings of a more desirable version. In those cases, the chip may work, but perhaps not perform as well as its markings would indicate.
I dont why, but I love this format
It'd be interesting to use a logic analyzer to compare the output of some remarked or possibly fake chips to one that is known to be real.
Hmmm, not sure if I'm as skeptical regarding the origin of these chips, but what I am pretty certain about is; I have the same doorbell as you
"Made in Anger"
Reminds me of a few years ago when FTDI was bricking fakes with a driver update.
LOL. But let me tell something. RS232 to USB chips made a SUPER BAD reputations because of fakes. This is extensive to PROLIFIC chipsets PL2303. In the end the are ruining their image. Getting originals is complicated. after having issue over issue with the PL2303 I moved back to old laptops with XP and REAL serial port to avoid conerters. But this FT232 are ALL OVER in developement kits In the end I like WCH, they are reasonably good and cheap enough to defeat counterfeits. But WCH driver install process is not direct, their download page is in chinesse so people find it a bit frustrating. I have used a few WCH products and no probs.
I'm loving these off-the-cuff videos.
As long as it performs the same and it's not advertised as (new) old stock, shouldn't really matter.
You were pressing down on those SID chips with the q-tip. You expected it to come off.
Here is a better question. Why not hook it up to a dynamic signal analyzer or a spectrum analyzer and see if the probably counterfit chips are performing on spec. If they aren't well that's an issue. I think your qtip is probably just taking the ink from the label on the chip off. Are they laser engraved or just stamped with ink?
That SAAYM board chip was obviously painted. I am refering to the others where something came up but not the whole thing.
I got some YM3812-F's off EBay a while ago, and I think the ones I got are legit, at least at the die level... haven't tried the acetone thing to check for resurfacing, but I only bought them after someone I trust on twitter bought some from the same seller and de-capped the chip and got the die under a microscope.
If the chips work, it's fine for me. If I could find a decent quality USB/MIDI module based on a OPL2 chip on the market, to make my synth music with it, it would be an instant buy (fake or not)
Yeah I had to take my time when sourcing YM3438 chips for sega modifications. I finally found a place that was from China of course, but they provided me with an actual shipping/packing slip of the order and the chips were in an anti-static tube etc. Having said that...they still all looked the same and had date codes from 2004? right.......
But they all worked without issue as I made a testing rig just for that purpose to make sure before installing them into client systems.
Being a watch collector, I understand the fake stuff is a real problem. Fake Rolex watches are bad. Not so sure that fake chips are a big deal. If they function as they should, does it really matter? I guess for the hardcore collector/hobby person it matters. But if one is looking for function only, I think to myself "no big deal". But that's just me.
this test probably depends on whether they took the time to laser etch the chips again
Could they be coming from the same new old stock supplier? If so, they could have the same date code and printing
The people who can answer this is Yamaha. They should make a statement whether they did indeed make a batch of chips in mid-2008. It may have been a new run, or it may simply been wafers that have been in storage that they diced up and packaged. The demand for these things are tiny in the scheme of things. A single run of thousands of chips, less than a week, could easily satisfy the demand.
There's another commenter earlier that got in touch from Yamaha. They couldn't vouch for them.
@@LGRBlerbs I saw that but it didn't answer the question. They said that they couldn't guarantee the authenticity of chips sold on the second-hand market. All companies will say that if you're not getting the chips direct from an authorized distributor. If they are indeed fake, Yamaha should have no problem saying that all chips with a datecode of 0824 are fake. If they can't, then it must be because some are real.
Is there a difference in sound quality between legit and fake chips? How about a video using genuine and fake chips in the same board and comparing the audio?
As mentioned in the video, these particular chips sound good to my ears. It's just a question of where they originally came from, whether they're new old stock or resurfaced used chips or what. No one really seems to know and that's odd.
Early OPL clones often had bad sound because the cloners hadn't figured out that Yamaha's patents on FM synthesis contained subtly wrong information. Specifically Yamaha was actually using phase modulation not frequency modulation because frequency modulation caused unpleasant harmonics with certain frequency combinations. FM is a purely digital synthesis unlike e.g. a SID chip so it's easy to reproduce once you know exactly how the chip works but Yamaha had omitted one key piece of information from their patents and never told anyone (TBF they figured this out later and then neglected to update the patents)
If someone is actually making a 1:1 copy, the chip would sound the same. It's actually fully digital! Analog audio only comes out of the external DAC (YM3014)
@@kveemusic There is a filter on the output but other than that yes it's completely digital inside. Early copies were apparently just based on a block diagram of the chip and what the patents said about FM synthesis and so had sound issues because they didn't know Yamaha was actually using phase modulation. Even if you decap the chip it would not be immediately obvious that it's using PM instead of FM. They are almost identical except for one small difference and even that is pretty slight (I know because I built a virtual phase modulation operator inside Synthedit years ago to replace the flawed FM operator)
Years ago I built a virtual DX7 using Synthedit and found it didn't sound quite right and, after some digging online I found out about this and changed the operators to PM (had to actually build a PM operator inside Synthedit because it didn't have a premade one) and then it sounded like a real DX-7. The difference comes down to one small change in how the operator modifies the carrier other than that they are identical. The thing is I really didn't even know exactly what the difference was I had read that Yamaha used phase modulation rather than frequency modulation in their "FM" chips (there wasn't a lot of information on this) and I thought "hmm well maybe if I just have the modulator modulate the phase of the carrier rather than it's frequency" tried it and suddenly it sounded like a proper DX-7.
Found this animated GIF that visually represents what's going on:
en.wikipedia.org/wiki/File:Phase-modulation.gif
This is what the synthesis inside a Yamaha "FM" chip is actually doing. The reason they use PM instead of FM is because FM causes nasty sounding harmonics at certain operator / carrier frequency combinations. Other than the unpleasant artifacts the same operator/carrier parameters with FM and PM synthesis sound basically identical to the human ear even if mathematically the output waveform is a slightly different.
I just hope all vintage computer chips get cloned 😀
Don't get your hopes up, especially not with some of them like the 6581 SID chip used in the Commodore 64. It's a mix of analog and digital circuits, using a manufacturing process nobody's used in over 20 years. Even the "boutique factories" that will make production runs of other computer chips won't touch them.
(The 8580 used in the 64C is a different story. You can still get a boutique factory to make those, *if* you order about $30,000 worth.)
GOOD LUCK WITH THE VIC II
I have no problem with people selling replica chips that are functionally identical to the original, just communicate that they are replicas; it's not like MOS or Yamaha are making any more (to my knowledge), and the originals are aging, and won't last forever.
Even if I couldn't use it I'd love to have a LRG badged Adlib.
PS
Love the doorbell tone
I wish I could have them manufactured! Haven't come across a decent way to do so yet.
Who was at the door? :P
@@LGRBlerbs if you ever do I'll definitely find the money to buy one lol
Seriously, I'd hang it on my wall.
@@antdude mail carrier.
The RUclipsr ElectronUpdate might be interested in this. He specializes in de-encapsulating chips and breaking down the structures. It would be interesting for him to compare a real and suspect chip on the inside. That would tell you for sure.
Someone really needs to make a candy microchip. Xylitol, anise, mint, minerals, sugar...
Send one to electronupdate, decap and check with a microscrope
Mateus Pinesi de Souza this seriously would be the way to determine if they were pin-compatible clones or die copies
@@cassandra-show But then someone has to send him a surely original Yamaha IC for comparison. A known original BUT dead Yamaha chip, because decapping a known working original is a sin. But if these are clones, I very much doubt they are re-designed, pin compatible dies. It would be a huge work and money to make. Maybe someone got the original films.
yeah I don't like the fact that they aren't stating that they resurfaced the chips, I would also be okay if they were reproductions as long as they say they are, at the end of the day if it sounds and performs the same thats all that matters.
I agree, companies should be more open about there products. Especially in markets where cloning is a big problem. Refurbished and rebranded, that's all I'd like to see on a chip if it's an up-cycle.
Maybe some sound cards were quite rare and expensive collector's item. (Since you can find on eBay recently...)
Could it also be that these chips aren't made in huge demand, so when a batch is made, they're made in batches of tens or hundreds of thousands all at once. So the low volume would mean a batch would sell out slowly, thus a lot of them would all have the same date codes.
i really love when the sunshines trhough your window, gives it a warmer more cozy touch. Dont close it pls :3
The only problem I see is they don't advertise them as repo and try to pass them off as counter fit. However the chips you have could be a refurbish too.
Pretty sure those things were just made in a huge batch towards the end of their life or maybe even after their life. From that point, tons of them just sit around in the warehouse until they're sold off in bulk as new old stock, where I think most of the "new" ones come from - all the same batch.
On the flipside LGR...I imagine that getting older, proper chips are not all that easy to come by these days for things like this. Older chips winding their way into e-waste because the PC no longer turns on and everything else just gets tossed/recycled, just plain burning out, that sort of thing. And by now, you can obviously find the schematics for making these chips online, 8 Bit Guy's used them before. Another idea to propose: Could the chips in question...be reverse engineered new stock chips created to answer for a shortage in what's rapidly becoming a newfound field of demand? More and more people are wanting to get these old machines up and running, without being able to find old stock parts now, after all. There are bound to be companies out there who've sourced out a Chinese factory to manufacture these for sale on the cheap for use cases like this. And in the case of the SAAYM card you mentioned, it wouldn't shock me if cards like that were using said chips. In cases like this, if they put the time in to make them sound accurate...is there really anything dubious about meeting said demand in the best, easiest way possible? I know that's what's going on with some SID chips, would make sense that some of that is happening with these too.
Actually, that's *not* what's going on with some SID chips. No legitimate manufacturer will even try to reproduce them. What you're seeing is hardware emulation of the SID chip using an FPGA. Which is great, but it's only as accurate as the emulation it's based on.
'Fake' in the sense of compatible doesn't have to be bad however when they sell it as the real thing (with company name and such) there is a problem. The only way to find out it is different is to decap the thing (with some toxic goodies) and compare the layout of the die (you need a (electron) microscope for this) to the original (few people did this with with branded 'clones' of originals). However, these are old chips and it could be possible the manufacturer update/improve the layout to make it cheaper to produce. Over time, it can be very hard to tell what is original or not, the painted one is definitely not an original (could be a bitbanged microcontroller, a die clone, an EPROM or even a small FPGA with fancypancy painting on top). Do you know the date code is really a date code, maybe it is a batch code and a batch can be huge or a version code, could be also thousands or millions of the same. Hard to tell when hardware gets older.
A good example of (more recent) counterfeit is FTDI chips, google it, very interesting.
If there's a dime to be made then China will make repros of an item. These are definitely repros. It's a common problem with old cartridge games too. Ebay does nothing about the countless auctions from China selling repro games for Nintendo 64 and Snes.
It's a well known practice. Take 75 working ICs and take 25 random ICs that have the same number of pins/form factor. Relabel all 100 ICs and sell them. You just increased your profit with minimal/acceptable impact on your customer feedback.