In-depth Look at the HP AC-4G Decade Counter
HTML-код
- Опубликовано: 29 июл 2024
- I asked, y’all delivered! This was by far and away the most requested of all the projects I’m currently working on, so without further ado, enjoy the beautiful Hewlett Packard engineering!
If you’re not sure what on earth I’m talking about, check out this video: • Let Me Know Which Proj...
Check out Ray Fantini’s fantastic video on the HP Frequency Counter: • FR38B/U Vintage Freque...
Here’s the excellent reference data on the HP materials:
hparchive.com/Catalogs/HP-Cata...
www.nixiebunny.com/hpac4/hpac4...
Also, we now have some epic shirts for sale!
my-store-11554688.creator-spr...
Come join us on Discord and Twitter!
Discord: / discord
Twitter: / usagielectric
Thanks for watching! 
Наука
What a nice video. The filming is exquisite. The HP folks were analog circuit wizards from the start. Their tube-based instruments achieved a level of perfection which you’d think would be only achievable with far more modern solid state components.
Thank you so much! It means a lot to hear that the filming looks good!
Those HP guys were definitely on another level. To build a binary counter that so perfectly blends flip flops with analog wizardry is just inspired. The lateral thinking required just to kind of understand what they created had my brain feeling like it was melting, I couldn't imagine the level of skill it took to design this stuff from the ground up!
This makes me realize, how grateful I am for stuff like TTL logic.
On the other hand it is amazing how many brilliant ideas came before it. I guess at the time of vacuum tubes, you would do everything to keep the tube count down, so it would be perfectly fine to have weird voltages and waveforms.
The way of outputting different analog levels in a kind of grid configuration so that only the voltage across one neon bulb is enough to trigger it, blew my mind.
I did a similar thing when I was at school (propably around the fourth grade), where I took the voltage drop across different amounts of 1N4148 diodes to build a decoder from an analog voltage to a seven segment clock display made from christmas light bulbs and transistors.
The input was driven by a single transistor current sorce that charged a capacitor to generate a linear ramp.
It worked but it also was horribly inaccurate.
A small side note, respect for taking the results of your project ideas video so serious and making a list with all the responses.
Thats a whole new level of connecting to your viewers.
Greetings from Germany,
Michael
Oh absolutely, TTL makes life so much easier! But, to make a 4-bit 1-2-2-4 weighted binary counter out of just four dual triodes is absolutely mental! The engineers of the day were so amazingly good at using every tool in their kit to it's most potential. I never would have thought of using analog signals to create different voltage potentials to convert a binary code into decimal.
That's awesome that you were already building wild stuff like that in 4th grade! I think in 4th grade the only thing I cared about was being the fastest person to run to the swings, haha. Whether it was horribly inaccurate or not, it worked, which is absolutely awesome!
Really, you guys are awesome! Y'all take the time to not only watch my videos but leave comments/suggestions/ideas that are absolutely invaluable. It's really humbling and awesome to have such an active community of like-minded crazy individuals trading ideas and inspirations!
Side note to the side note, I have a good friend who lives in Germany now! We worked together in Japan for years, but he recently moved back home to Germany. I'm terrible at remembering German city names, but I think he's just outside of Frankfurt.
That is ingenious. Hewlett Packard commanded a lot of respect in their day, and with good reason. Thanks for demonstrating it. 🙂👍
Just brilliant, thanks for making this video! Please please consider connecting this to a rotary phone dial to count the pulses...
Thank you so much!
I had never though about that, but that's a really interesting idea! It seems rotary phones use a Unary coding system (1 pulse for 1, 2 pulses for 2, etc.), so this would be perfect displaying the appropriate number, a kind of Unary to BCD to Decimal converter. Hmm, I wanted to build a display for the unit, but it would be properly awesome to also add in a rotary phone dial to the display unit...
I'll do some thinking on this one, but I really, really like this idea!
I love the past future merge of the phrase "you scroll down" and literally turns the page
this is such a great video! you explained everything so well, it's awesome that you went so deep into learning the logic. And that shot of the HP tubes was awesome
Thank you so much! It's really a level of engineering that's quite a few steps beyond my capabilities, but it's such an exquisite example of pure engineering skill, I'm glad I dove into it!
And those HP 5963 tubes are just great! Whoever said that tubes were unreliable? Oh, I did make a mistake in the video though, I said the module was 50 years old, but in fact, it's now 60+ years old! Talk about a brilliant chunk of engineering that truly stands the test of time!
When i saw this notification, i jumped out of bed immediatly. What an amazing device!
Thank you so much for doing what you do and be careful!
Now to find a way to replace the 4017 chip in my sequencer
Thank you so much for checking the video out!
It would be interesting to see what it would take to modify the circuit to become a straight 4-bit binary counter or a straight ring counter. I think all the ingredients are there, it would just take a little more analog design skill than I have, haha!
If you ever get your hand on an HP3000 Series III Mini computer system, you'll see the build quality ! All think gold plating everywhere and pure beauty engineering
I don't remember seeing a tour of the room a couple of episodes back in this playlist. A couple of episodes back was the vacuum tube toggle flip flop episode.
Thanks for analysing these cute little modules !
Thanks for suggesting them!
It's hard to describe the feeling of joy I felt once the fundamental operation finally clicked in my brain!
That was a fascinating video! I did chuckle whenever you called that decade counter "small", though. ;-) Also, the required voltages made me feel like the blood had drained from my face, LOL! I'm glad you were able to demonstrate it. :-)
Nice job on finding the instructions for the counter, and powering it up.
Thank you!
I'm glad I dove into it, it's truly a beautiful example of the type of engineering skill HP was capable of!
Awesome video ! Love your work ❤️
Thank you so much!
Fascinating technology, ingenious circuits and absolutely lovely documentation! These timing diagrams are so clear that I'd be hard pressed to find anything of that quality in modern IC datasheets.
Oh, and a mandatory wascally wabbit :)
Thank you Keri!
HP's early documentation was amazing! Beautiful schematics, timing diagrams, illustrations of the types of waveforms to expect at different parts, breakdown explanations, etc. I love old documentation like this because not only is it great for troubleshooting, but it's also like a perfect learning tool!
And the little bunny is happy and hoppy as ever!
awesome in depth explanaition! a very impressive piece of engineering, that unit would nmake aglorious output display for some decorative project like a clock..or sub counter.
Thank you so much!
And, building these into a sub counter is a really brilliant idea! Although, they're a little pricey, so I'll have to see if I can build up a lower voltage version of these to maybe drive some 7-segment VFDs I've got laying around.
Fun fact: The nixie variant is the same thing with a bunch of photoresistors stuck to the front which are then used to light the nixie cathodes.
Amazing that it is still working perfectly after all that time since it was first manufactured, and they say that Tube technology is fragile and unreliable, I've seen devices made with the latest solid-state electronic technology available today fail after only a month or so.
This thing is so well designed, it truly is an amazing bit of history. Of all the tube projects I've worked on, I've come across a handful of bad tubes, but more often than not it's a blown resistor or capacitor that was the root cause. Tubes themselves (aside from being fragile due to the fact they're glass) have proven themselves to be extremely hardy, and a majority of the stuff I have with 60 to 70 year old tubes in them still work like a charm!
amazing stuff....cheers.
Thank you so much!
I actually really like the design of this counter and may have to try to modify it a bit for some projects in the future!
I actually just picked up one of those 522B counter at an estate sale recently. I have no idea it it works because there is a fuse missing from it. Thanks for the video on the counters tho!
That's awesome! The 522B is an awesome piece of kit!
I also recently got my hands on a 521A counter, which is really similar. It needs some work though, so hopefully I can get it going and there should be some videos to accompany the restoration sometime soon!
Fantastic!
I've got a series of books from the 1960's explaining valve technology so wasn't surprised at the -80V pulse, but it is great to see such a circuit in action and so well explained.
These books use the term "cumulative action" to describe the instantaneous response when a pulse is applied (covering your whole explanation on how the grid capacitors work). I love this phrase, so close to "magic happens"!
You must find an excuse to interface this unit into your one bit valve computer.
can you post the title of those books? I'm interested, I'd like to find a pdf for them and have a good read
@@willjackson1955 Most of the links are for ebay listings, so rather than post links that will die here's a Google search that will find them
common core "basic electronic circuits"
This is a two part follow on to the "Basic Electronics" series that you should really read first.
Some of them are available on archive.org if you've got an account to borrow books from them (free).
Thank you!
Yeah, the voltage levels aren't all that surprising given that it's tubes. The majority of tube logic happens at ridiculous voltages (but tiny current). The response from the pulse is wild though. I like the term "cumulative action" though, because there's a whole lot of things that get set in motion as soon as that pulse is detected! But, at the end, it is indeed a bit magic, haha.
@@willjackson1955 Also check out bitsavers.org, there's a ton of great pdfs on old tube systems saved up there. In particular, check out the IBM 604, it's been my primary go to resource!
www.bitsavers.org/pdf/ibm/604/
I have a Hp 522B picked it up recently in a lot haul, i have not turned it on or tested it, but I would conceder selling it, however, I'm sure shipping will be high as its a boat anchor lol
Just as interesting is that the original documentation is still available. Lucky we have internet sites that provide all manner of instruction manuals, etc.
These kind of counters and the decimal modules were common back then. There were very similar models made by Beckman and others. I bought one of those from a thrift store around 1990 for $7. My memory of fuzzy but I think it worked perfectly. Unfortunately I cannibalized it for its power transformer and crystal oven.
I always find it super fascinating how many different design methods were being employed back then! The engineers of the day really were just figuring these things out and using every trick they had at their disposal to get the product working correctly!
i'm really confused by two +ve voltages to neon. will the same method work for LEDs? without Gnd connection, how will electrons flow? should this setup use batteries as source or SMPS or bench power supply?
Very cool. I have a similar beast, but it was made by Beckman. Different build style. I didn't use a PCB as I remember.
Thank you!
I just looked up "Beckman frequency counter" and that's crazy that they have some models with a design that's almost identical to the HP with the vertical rows of neon illuminated decimal digits. Very cool equipment!
I remember seeing one of those frequency counters actually working, back in the seventies, on a military comunications workbench! It was already obselete then, but quite a thing to behold! it was sitting next to a "newer" nixie display version, both from HP. Was very interesting to see that hibrid way to decode. Makes you think there is always a different way to get from A to B!
The HP Nixie tube version actually uses the same basic design, only the neon bulbs are used as sort of optocouplers to select the appropriate nixie element. Check out the schematic for the AC-4J, AC-4K, or AC-4L in this manual (pdf page 32, 33, 34): www.nixiebunny.com/hpac4/hpac4.pdf
You can see it's pretty much identical to the shcematic for the AC-4G that I use in the video, except there's a separate circuit with LDRs that sit right next to the neons. Each LDR creates a voltage divider with the 270k resistor in series with it, and that's what drives the nixie tube element selection! It's such a cool design and a really great way to implement a better readout without redesigning the entire thing!
@@UsagiElectric thank you for the circuit insight! I will check it
@@UsagiElectric sorry, but the link is not working (for us, at the southern hemisphere, at least!)
@@AlejandroLopez-qd3xm Hmm, weird that link wasn't working. Here's the same manual hosted at a different place: www.steampoweredradio.com/pdf/hewlett%20packard/Hewlett%20Packard%20AC-4%20Decade%20Counter%20Operating%20and%20Service%20Manual.pdf
@@UsagiElectric got it! Just a thing with my phone. Thank you
I managed to the much higher speed HP AC-4E version of this. The resistors are mostly ceramic power resistors
So Cool!!
Thank you!
Here is an idea to try. Connect your vacuum tube 555 circuit to the input of it, and see if it counts on its own.
I wonder if these kinds of circuits were used in other applications back in the day. Adders, Counters, Early Computers, Military equipment, Elevator Floor Indicators, Scoreboards, Ticker Tape Marquees, Teletypes, Industrial Machinery Sequencing, and others.
One other thing to note, the 5963 tube is a special version of the 12AU7 with special cathode coating and a bigger cathode to better handle not wearing out as fast being in one state for extended lengths of time, and handles a lower plate voltage that was developed for digital circuits such as computers and counters. Early computers had issues with tubes going bad because of cathode poisoning from what i read, and this was caused from being in only a few states for long lengths of time (audio and analog circuits on the other hand use a much greater voltage variation and keep the valve moving across its range).
Now that is a cool idea! Unfortunately, the vacuum tube 555 is such low voltage, I don't think it'll provide enough of a voltage change to actually get it to trigger. The AC-4G is looking for a -80V pulse, but the 555 timer can only provide about an 18V to 20V pulse at max. But, that actually leads nicely into something I definitely want to try, which is rebuilding this design as a 4-bit 1248 style binary counter that runs at just 24V. I think the fundamentals are there and it can be done, though it may require some pentodes, which will double the tube count. Still, I think it would be awesome to have a binary counter triggered by the 555 timer!
The Multivibrator design here (with the cross coupled grids and plates) was the fundamental building block for nearly every tube computing circuit I've seen. IBM used multivibrators in just about every configuration they could to perform all sorts of wild computing trickery, it's a really versatile design! IBM had a pretty similar design for a single input flip flop as well: i.postimg.cc/8PVjKhwH/FlipFlop.png
Although, HP using capacitors as state memory and interrupting the plate voltage is the first time I've seen that switching method, which is really cool.
The 5863 is a great tube! And, they do seem quite popular among the audio crowd as well, despite not being originally intended for that purpose. IBM really loved to use the 5965 in their designs, which I believe is a specialized version of the 12AT7 built to similar specifications as the 5963. Though, the mainly used it where they needed more power on tap. I think they preferred to use the 6J6 type 7-pin dual triode where possible because it was a more physically compact tube.
@@UsagiElectric Maybe you could use an amplifier tube to boost the -20V to -80V. If the gain was 4 you would boost the pulse to -80V
@@stphinkle I've done a bit of experimentation with boosting voltage levels up to power some neon bulbs using just 24V input and a tube oscillator. Ideally, I need to wind my own specific transformers. But, I never had great luck in getting any reasonable current out of it, tubes are just a bit too high impedance to do some really tricky things like that with. Though, I have recently done some experiments with a 2D21 Thyratron to generate negative voltage from positive voltage by building an oscillator and negative charge pump, and the Thyratron was actually able to provide about 10mA of current at -10V, which isn't bad at all! With some extra work on that, the design could maybe be refined a bit to get to some nice HV for power neons and other things! Hmm, I'll have to play around a bit more with that...
That was brilliant. Thank you so much for that. It would be fun to replicated a low voltage version of that flip-flop.
Add: if you were to make something like this as a kit I would totally buy this.
Thank you so much!
That's actually something I really want to do! Though I think I might change it a bit to count in a more traditional 4-bit 1248 weighting style, but I'm very interested in seeing how this design works scaled down to much lower voltages!
@@UsagiElectric I now got one from eBay as well but have yet to light it up. You didn't happen to find a socket compatible with the footprint? I'm also thinking I'll need a opto-coupled interface so I may drive safely it from modern electronic.
@@tommythorn That's awesome that you picked one up on eBay!
Depending on which model you got, the footprint might a little different. If it's an 8-pin, a standard vacuum tube Octal socket should work. If it's a 12-pin one like the AC-4G one I have, I think something like the socket I linked below might work, but I'm not sure. When I get around to building up a proper display for this piece, I'm planning on giving that 12-pin a socket a go.
An optoisolator is definitely a good call if planning to control it with something modern! Even though the currents are very low, the super high voltages would make quick work of any modern silicon!
Socket: www.ebay.com/itm/194014637116?hash=item2d2c2c6c3c:g:1NoAAOSw0EBgaMEf
@@UsagiElectric Yep AC-4G like yours (12 pin I think, I’m not near it). If I get around to it, I’d get some more and make it into a clock (or variations there of).
I worked on countless numbers of these things during my active years in the United States Air Force PMEL (Precision Measurement Equipment Laboratory. They were well engineered, and well built. But they were a drag when they broke. (My time was 1965 - 1972).
That's awesome that you got to see these things in the wild!
Given how insane the analog digital mixture that's going on here, I can imagine they need a really good scope and a serious familiarity to be able to diagnose them when they're misbehaving!
@@UsagiElectric The Air Force used HP 524 series, almost exclusively.They were large, heavy, hot (Lots of tubes...) and usually quite reliable. The reference in them was a crystal, in a heated oven, and in the calibration environment, required extremely long stabilization periods. Even then they often tended to wander a bit. I sometimes wonder how much better the performance might have been if current technology components had been available for construction. Even so, they were quite the machine.
@@harryshector That's awesome, thank you for the extra insight into these! It's interesting you mention long stabilization times, because even this simple little module needed about 5 minutes to properly stabilize and be reliable. Not a problem when filming because it always takes time to get the camera setup, but was infuriating in initial testing!
That's crazy that they kept the crystal in a heated oven! HP truly liked to overengineer everything!
I wonder what it would take to make a clock out of these displays. I’ve always wanted a Nixie tube clock (preferably as a kit I could build), but now I want a clock with these displays (which would also be an awesome kit!) too! … Yep! I definitely need to download your brain into mine! LOL! Question: Do you yet have out a video that talks about how you learned this stuff? I have always been fascinated by hearing about other people’s backgrounds in what they’re good at; not just electronics, but anything and everything, from music to engine wrenching. Some are university trained, and I’ve found that many more are self-taught, but either way, I’m a firm believer that what one person can do, another can learn, if the same steps are followed, and that’s certainly what I tell my music students. So do you have a video talking about your background and training? If not, I think that would make an AWESOME video! 😊😊😊
Oooh, making a clock out of these would be an interesting challenge. Since clocks count relatively weird compared to decimal or binary, it makes it tough to cascade them. I'm sure there are more efficient ways, but the way I would tackle it would be to breakout each digit of time and then have four of these display that. So, for example, if the time is 15:37, unit 0 would need to display 7, unit 1 display 3 and so on. Then every minute, have a controller that resets all to zero, then pulses the input pulse the appropriate amount of times. The fourth unit would only display 0, 1 or 2, unit three would go through the full range, unit two would display from 0 to 6, and unit one would go through the full range.
Now, that could be relatively easily done with something like an Arduino or other microcontroller, but the real challenge would be to have a completely hollow state way to do it. But that's a bit beyond my skill level and would require a massive amount of tube logic, but it sure would be cool, haha.
Thank you so much for the praise! I haven't actually had any formal training on this stuff at all. I'm pretty much just making things up as I go along. I've had an interest in electronics stuff in general for a long time, building my first 386 computer when I was about 10 out of scrap parts. Then I learned a lot more about individual components and circuit design by building little BEAM robots. I think that's probably the best entry way into learning electronics design as they have some really innovative designs and excellent information out there. This is an excellent site with some very cool examples of BEAM robots: www.smfr.org/robots/
This is also an excellent resource for some good BEAM circuits: solarbotics.net/library/circuits/default.htm
I only really recently got into vacuum tubes (like last year recently). My primary source of information for learning about tubes and tube circuits has been the IBM 604 Customer Engineering manual (www.bitsavers.org/pdf/ibm/604/). That manual is full to the brim of amazing explanations of how tubes work and how they're being used to build more complex computing circuits. Without that manual, I'd have been mostly lost in trying to figure out this stuff.
So, generally, if someone is looking to get into electronics or tubes/basic computing circuit design, my two recommendations are BEAM robotics and the IBM 604 manual!
Hey, with 2 modules, do you know what you can do ?
A 2digits voltmeter !
The staircase output is a DAC so you only need a pair of tubes as comparators, then an oscillator that will be inhibited by the comparator.
Thank me later ;)
And this is one of the reasons why this output exists, IMHO (the other for making nice patterns on TV screens, probably)
That's actually really smart!
Also, some cool patterns on an old CRT sounds really epic too, haha.
@@UsagiElectric That's not smart, that's how it was done back in the days ;-) (I've seen a video where it was done with relays...)
I wonder if the staircase output is an early form of a digital to analog converter.
I think you may be on to something there, that sounds about right!
I have no clue what it was actually used for in the full machine, but that certainly makes sense!
@@UsagiElectric I am wondering if it was some kind of feedback loop, controlled the intensity of a light, drove an analog meter, changed the volume or tone of a speaker sound, triggered the carry when it got over a certain voltage, controlled a magic eye tube, or something. I am not sure.
@@stphinkle I did some more digging to see if I could find out what it was being used for in the original frequency counter, but the only manual I could find, used modules that didn't have the staircase voltage output. SO, it wasn't used in earlier versions of the frequency counter, but I still have no clue what HP's original intended use for it was. There are lots of interesting situations where DAC could be useful though, but given how masterful the HP engineers were at seamlessly mixing digital an analog, I'm sure the genuine use is pretty impressive!
Any chance of getting the links to the power supply converters?
Sure thing!
This is the one I use to step down to 6v: www.amazon.com/gp/product/B018JPGKQK/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
And this is the one I use to step up to HV: www.amazon.com/gp/product/B07T6L61D9/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1
Impressive! You might consider to convert it to a permanent working exhibit. It is much too nice, to put it back in the cabinet in a non working state.
Thank you!
I was actually thinking about building a display unit for it in this video, but the video would have been much longer and I was too impatient to wait for parts to arrive, haha. Though, I'm glad I waited because C000kie had a brilliant idea to add a rotary phone dial to it so the counter decodes the pulses, which I think would make an awesome addition to the display piece!
But, either way, I'm going to start collecting parts (notably a socket and some good push button switches) and there'll definitely be a follow up with the display piece one of these days!
@@UsagiElectric take some more tubes to build an oscillator and a comparator, and you get yourself a vintage 2 digits voltmeter.
Wow, those things weren't cheap!
No kidding! The AC-4G I have here was listed at $35 in 1959, which when adjusted for inflation is $320! And it was the cheap one, the AC-4E was listed at $80, which is $730 today!
13:00: I'm not gonna lie, "MEG" confused me.
HP is always just so damn confusing. That's why I don't like them. How on earth am I supposed to program an MCU to read that? It's interesting how the entire track is filled up with solder. Probably to protect the copped and give it more current handling capacity? I wonder if they had a PCB company etch that board or if they just did it themself. No solder mask, nothing. It doesn't even look like FR4. It looks like one of my prototypes that I make before I send it to a PCB manufacturer for production.
Yes, KiCad warns me of any potential mistakes that I have made. Just connect the dots pretty much when plotting the board, but I still prefer to etch my own prototypes first and test it. It has saved my arse on a few occasions.
Speaker micro phone audio in audo out
Mayan