About the "only the 1N4007 should exist" thing: As far as I can tell, only the 1N4004 and 1N4007 actually exist. Diodes Inc gives their typical capacitance characteristics as two graphs: One for "1N4001..1N4004" and one for "1N4005..1N4007", suggesting that these bunched-together parts actually contain the same die. The lower-specced ones are just down-labeled (and, funnily enough, more expensive than the higher-specced ones... Probably because there's so little demand). I highly doubt that any of the diodes that get labeled as a "1N4001" actually failed the tests for a 1N4004. Could be a fun thing to try - buy a few 1N4001 and 1N4004 from the same manufacturer and check their actual breakdown voltage (and/or forward voltage) to see if they're in fact the same part labeled differently.
During manufacture of my PDVS2mini's I used to have about 1 in 25 that would randomly switch themselves off during soak testing. Turns out it was a leaky reversed biased diode in the soft power switch circuit. I tested brand new diodes out the bag and could never measure a leaky one......till I reflowed a bunch in my oven. The pitfalls of 250degC lead free paste! Changed to a different brand diode with similar spec and it never occured again.
I had problems with Bournes 0805 smd inductors. The inductance would change 20% after soldering. The temperature specs on the datasheet were followed. These inductors had good specs for Q and current but were unusable.
I would certainly be suspecting the caps before suspecting the diode, they often fail from physical stresses or corrosion, with a known flex issue in that area I would be 99% sure one of the caps has a fracture which results in them shorting out.
i had one fail on me just from hand-soldering (or may have been faulty from the factory). I used it as a dc blocking cap in an audio amp, and had some dc at the output because of the leaky capacitor. I've seen cracked ones go short a fair number of times (cracked because pcb bends; i drop my multimeters fairly often and i bet others do too).
No! Corrosion is only applicable / known for Thin Film resistors, where the protection layer has pinholes. Have never heard of corrosion failures on MLCCs in Automotive.
I recalled repairing a TV decades ago. The complaint was the channels keep drifting. The owner had a new tuner unit replaced but that did not solved the problem. I was asked to have a look at it. I was at my wit's end and after a couple of days of meddling, I decided to check the diode 1N4148 for short but turned out OK. Anyway, I had it replaced and bingo the drifting stopped. Another time, a Luxman L85V amplifier having cracking noises on one channel that was caused by a noisy 1N4148 diode. PS During that time, my repair tools were and still are : a SANWA analog multimeter and my brain. No schematics. Now life is much easier with the internet.
I do find it quite disheartening that the manufacturer will happily promote repairs by creating a DMM, yet won't release schematics to repair their own devices 🙄 great video as always Dave, look forward to these videos
@@sylviam6535 yeah but if they had half a brain they'd know schematics have been around for macboooks for years and no nock off macboooks are on the market lol
@@TheCod3r - Macs use lots of custom hardware which their suppliers are not allowed to sell to anyone else. See the Louis Rossmann channel for the loops he has to go through to find parts for repairs.
OH HAI. That was my forum post! (I am 10 months late to this video!) Looks like I'm going to need to dig out my old cheap multimeter to finally fix this thing! (For these past 4 years Ive been in the habit of removing the batteries every time I used it)
A possible failure mode for the diode: if someone did install the batteries backwards, you would dump a lot of current through that diode. That could leave it with permanently degraded leakage.
I tried to post a more detailed analysis, but my post was blocked--maybe because I tried to link the relevant data sheets? Anyway, depending on some assumptions, a pair of fresh Energizers would end up dumping anywhere from 2 to 10 amps through the 4007--well in excess of its 1A rating. It would not surprise me to see permanently elevated leakage after that abuse.
@@danmenes3143 Interesting. That deserves further investigation. If it is the case then this kind of battery draining could be significant on a global scale?
@@danmenes3143 My same thinking. It's also so easy to do and not notice for a few moments or even minutes. It could even create some magic smoke if left for a little while to sit. Frankly the reverse diode protection only works well if there is a fuse in series with the power source. Otherwise the best alternatives are a series diode or similar solution (N or P MOSFET circuits)
@@nlhans1990 That's what I'm thinking. Suppose you screwed the cover back on, then flipped it over and noticed it wouldn't turn on. "Magic smoke smell" might well be your first hint about what was wrong. A fuse or a PTC thermistor could be used to protect the crowbar diode.
@@danmenes3143 I was thinking the same thing during the video. Snap some primary lithium cells or rechargeable NiMh cells in there and it will easily exceed the maximum forward current of the diode. There would be at least a few watts of dissipation in the diode if high-capacity cells were used. If I were to use something like that, I would want to mold some additional features into the battery compartment to prevent the cells from making contact if they are put in reverse. The diode would then be a belt and suspenders in case the 200 lb. gorilla crammed them in hard enough to make momentary contact.
I had a digital tyre inflator /pressure gauge that used a coin cell, right out of the box it drained the battery over night and did several new batteries the same so I contacted PCL who made or sold them and they just sent me a new one. Having been told I could bin or do whatever I wanted with the first one I took it apart and had a look at the board , first thing that caught my eye was a big fat ceramic capacitor right across the battery so I remove it and not having a replacement I put the board back in and the gauge worked perfectly and the coin cell is still working the unit two years later. Not sure why the designers thought the cell needed a bypass cap. when the microchip had several of its own. Excessive battery drain is often down to leaky by pass capacitors and always worth a look if you have that type of trouble.
In Automotive applications, there might be MLCCs on the battery input, to suppress spikes from the cars supply system. It's forbidden though, to just put a single MLCC there, only two in series, as these components might develop leakage currents due to vibration, bending, even arising from the manufacturing process. This might even cause fire in extreme cases. Therefore, these two MLCCs are good candidates for excess leakage currents.
I work as a PCB design engineer in the auto industry. Not only are there two capacitors in series, but they're also rated for much higher voltages than required for the application and are physically rotated 90 degrees to each other. That way they won't physically both be on the same flex plane.
@@Aerospace02 I think it's no secret, that small case sizes are forbidden in this case, e.g. 0805 preferred. Edge values, i.e. max. CV product of given technology should be avoided. As well, technologies like Open Mode or Flex Termination may be used, but unfortunately not available any more in big quantities. Root cause of failures in MLCCs are mechanical cracks, as well as bad spots /contamination in the ceramic layers, which gradually may cause leakage currents to shorts.
Ford specifies any capacitor across un-switched Vbatt has to use fail safe capacitors such as Muratas GCE series, they do not allow two separate capacitors anymore, have not for a number of years now. it is in the Elcomp.
@@KasperPilsted Ford allows us to use two separate caps on unswitched Vbatt. In fact, they specify it. This is for instrument clusters and HUDs. Does it perhaps depend on the product?
There are Experiments I have done once with no success, but there are times when I want, semiconductors are wonderful, I dreamed that I was doing experiments on electronics, and I also dreamed of videos Your sharing is very successful, I want to share with the community good knowledge! thank You Share!
For a jellybean diode making all the parts for 1kV, and likely also having most of them likely to survive 2kV in reality is cheap, so no reason to make the lower ranges, unless you are having a bad process month, and they are failing at 1kV, so you are looking at the voltage they fail, and binning the bad ones as 1N4001, so you can get them out, or simply making the entire run as 1N4001, even if only 1% are not passing at 1kV. Most likes will only give you the lower voltage if you special order them, and in any case will give the 1kV part as standard, just marked down. Generally for things like military or aerospace parts, where they originally put in 1N4001, so changing part number is a pain, but the supplier can provide the part, marked as 1N4001, and it will pass the entire test spec for the 1N4001 part. Originally IIRC the 1N4007 used to be a special process part, with a mesa etch to have the corners able to withstand 1kV, instead of the regular 1N400x, which was simply a scribed die. But turns out easier to not skip the process step, to do the mesa etch, so they made all of them as that instead, and got the process dialled in right. Slight difference in surge capability, but if you are doing something that runs the diode at that ragged edge you are doing it wrong, because there will be some that do not meet that "typical, not tested in production" spec off the line. If you want an avalanche diode there are better diodes, with actual process controls for the voltage. I have a large number of 12V and 24V zener diodes, which I use more often than not as regular diodes, as they do work well there, replacing the 1N4148 in most cases. Have accidentally made a few of them turn into LED's as well, at least for a second or three.
@@SeanBZA The 1N400x series were first made back in 1963 when binning would certainly have been required to get a decent yield. These days I'd be surprised if they aren't all 1000v parts which are just remarked.
@@ferrumignis That is why you also got a 1N4000 diode, which was guaranteed to be a diode, though not above 25V, and it was very common in the surplus market in bags of assorted components. IIRC Sinclair used to use the reject transistors and diodes from Ferranti to make his first amplifier modules, taking all those production line packaged rejects, for scrap metal value, and then sorting them into categories of dead, not too dead, very leaky, not too leaky, and will work under 15V, and used them in his amplifiers, at least those with some gain. Ferranti tested at 30V, so those that failed there could still be useful at 12V or less. Was worth it when at the time a transistor was not something you bought for a penny, but instead was up to a pound, and a pound was worth something.
I just pulled my BM235 apart as I was watching this video and noticed the lack of the diode across the battery terminals in mine as compared to the one Dave was showing, I have had mine for many years now so is an older revision. My BM235 just started doing a strange fault on resistance or continuity mode will go to 0 when the leads are shorted then stay at zero when they have been opened again.
Putting the reverse battery protection diode after the switch doesn't sound like a terrific plan, the entire short-circuit current of the batteries will flow through the range switch, which is most likely way above its rated current.
13:40 unless there's a poly fuse or something (and even if there is), I could see not wanting the crowbar diode to act through the switch. It would subject the switch to the short circuit current, which, If they used lithium primary cells, could be quite substantial. It also wouldn't protect anything upstream of the switch.
Hi Dave, I kinda stopped using mine because I was tired of replacing the battery every time I needed to use it. Even then it was the least used of my meters. I've had so many other things going on I never got round to finding out if anyone else had the problem.
There's another option for excessive power consumption when switched off; during shipping the NSA intercepted it and installed a wiretap in it. Also if that tantalum is in their unit I would be rather suspicious of it, they tend to develop very severe leakage currents when operated outside their ratings like in a momentary reversed battery scenario.
No clue how these are set up internally, but is there a chance that when turning that rotary switch you'd potentially create some EMI connecting/disconnecting power to all the circuits that they're trying to filter out in all cases? (with the cap/inductor)
Fascinating, I had somebody mention that they had an issue with her guitar pick up draining batteries as in put new battery in and by the next day be completely dead. My first thought was capacitor or some sort of component sitting between the switch and the battery or the switch was not working right which is usually a quarter inch and shorting the ring shield. Lo and behold this video pops up in my feed and confirms my suspicion. If I do eventually work on it I will have to make a video of it because apparently this is a common issue for that pick up.
Be very aware that the X7R ceramic MLCC's (which are used in the higher value capacitors) are EXTREMELY piezoelectric and can function as microphones causing feedback in amplifiers. Much like many old Triode Vacuum tubes could because the grid would vibrate.
This is a timely video as I am chasing another issue with my BM235. A few days ago I turned it on and it initially showed the panel test (all segments lit up) and then the panel went blank. Believing it may have been flat batteries I changed them but the same result. Tried a few more changes of batteries having checked the voltage beforehand, but the display refused to come on after the initial panel test. Both fuses are intact also. I then hooked up my bench power supply directly and fed the meter 3.2V. Still didn't turn on. Increased to 3.3V and the meter powered up and stayed on. I slowly reduced the voltage from 3.3V until I got to 2.4V when the low battery indicator on the meter came on, yet the meter was still operating as expected. I shall be buying some AAA-size lithium batteries in the next few days, and hopefully, the inherent higher voltage of those will get me out of trouble, but still puzzled as to why normal fresh alkaline batteries have now stopped working. I'm suspecting the initial test for voltage has failed or has drifted. Any ideas?
just because the batteries are fresh doesnt mean they are good. ive had numerous ones fresh out of the package volt test ok but under the slightest load go flat.
It makes sense to put your primary decoupling caps at the base of the star grounding junction. That’s your true ground reference, don’t want to add track, lead, esr and connection resistance back to the battery on top of what your reference plane is
I believe you mentioned moisture-conductive dirt, especially under the SMT components. In my years of occasional conductive problems troubleshooting I recall two cases when there was a buried copper (or tin?) whisker between traces. Pretty hard to locate, but eventually chasing micro-ohms or even nano-ohms along the traces, I found the location. Looked perfectly OK on the top, even after removing the solder mask. Hmmm??? But digging into the epoxy, there was a hairline of copper (in my case). Looking with a good baclight and a good microscope it became evident that the whisker was just long enough to bridge the tho traces. A little more scraping and the whisker was gone. I don't recall exact resistance, but it was some ohms. Just a hair width or even less of copper. My PCB manufacturer told that he had seen some similar cases of buried whiskers in raw laminate. So my two cases were not unique.
Those capacitors will filter out the voltage surges due to the sliding ('bouncing') contact as the switch is rotated. I don't know whether this their primary purpose, but such surges on memory cells thoroughly confuses state machines.
@@inse001 yeah, the only thing I can see is when you have the switch on and you change the batteries...but it seems to me a little too much to worry about from a design point of view....
Those two ceramic capacitors would also be my second item to check after the protection diode. The worst thing about the MLCC caps is that they tend to break short circuit. (Although there are some more expensive variants that are guaranteed to break open circuit, those did not appear to be that variety.) Near the end of the video you mentioned PCB design concessions. I know what it's like. My most recent board is a very simple MCU (ATtiny1614) controlled power switch with 5000KV isolation, and a pair of LED's for operational status. Even on that simple board I had to partially rewrite the firmware because I needed to change pin mapping, and my MCU is only 14 pins! I can only imagine what happens in the back and forth at a company that is designing a board with hundreds, or even thousands, of connections, and the software, schematic, and layout guys trying to get a schematic that can be routed and still functions correctly.
I have a Fluke 87 which I bought in 1994 and it has a similar issue. It's also a known problem with this model (and a few others) and it seems that there is a residual draw on the battery even when the switch is in the OFF position - so off isn't really off at all. This meter also drains batteries when it goes into sleep mode. One thing I found was that when the switch was in the OFF position, sometimes the display would flicker a bit and the buzzer would emit a weird noise. Tapping the rotary switch would affect these behaviors. The meter was then dismantled and the contacts on the rotary switch cleaned and lubricated. This stopped the above reactions, and slowed down the battery drain. I don't know why it's so hard for manufacturers to actually have the unit completely OFF when the switch says OFF. There should be ZERO current draw.
If you want to compare the boards, do another scan then open both in Photoshop or Photopea and overlay one over the other in one of the odd modes like divide or difference mode. Use "distort" to drag the corners to line up and the board will black-out as it lines up... except where the differences are. I've got to pull my 89 IV apart at some point and probably the 189 at the same time, I'll see if I can do the same.
I am very interested in learning more about the failure modes of electronic components since a large part of my job is troubleshooting a complicated board. I have no answers to questions about why they failed in particular ways. Where can I learn more?
Capacitors so that, even if it is switched off, the large area of the battery does not act like an antenna, and couple RF energy into or out of the circuit, but instead connect it to the common terminal. Yes the capacitors can be leaky, especially with the board flex issue, which likely has cracked them, and they started to grow whiskers, that are acting like a partial short or have reacted with moisture to make a conductive channel internally to the capacitor. Probably a good idea on the next revision of the board to move those 2 capacitors, likely 2 different values, like 1000p and 100n, so as to get different self resonant frequencies, and move them to another orientation, so the flexing forces do not crack them. Probably assembly force applied direct to them is causing this.
at 5:44 it says rated peak reverse voltage in percent, so what you are calling 10V would be 10% of the rated peak voltage right? which for the 4007 with a peak voltage of 1000V would be 100V then? or am I missing something? I mean doesn't change your point just makes it even more clear I guess.
I don't know the construction of the range switch, but by the looks of it the positive strip isn't always connected to the same sub-circuitry. So maybe they wanted/"needed" these capacitors here in whatever mode is selected. And by designing this way they covered them all at once.
My eevblog BM325 died an early death due to battery leakage. I suspected something was draining it. It was just out of warranty and I wasnt impressed - my much older and much cheaper meter is still going strong
I have a cheap studio light with excess battery drain. Luckily i have two so i got to do some measurements and compare them. The affected unit has a very high (12mA!) standby current, and oddly enough, when switched on at lowest brightness it actually draws less than that. My only conclusion so far is that it has a faulty mcu or firmware causing it to draw excess current in standby. Current draw when switched on was completely normal and matched the other unit.
Mhm they probably have the same firmware for years and never change it. It would be a weird fault for the MCU as well to not enter standby properly, but then, weirder things have happened, so it is a possibility. It might be worth digging further for educational purposes. It could be powering something external to the MCU down and that thing is malfunctioning for some reason. Difficult to guess without seeing the device.
If the 4007 was the culprit then it would be easy enough to find a replacement, but if it’s the mlcc caps then how do you determine their value since they are not marked. That’s something that annoys me about smd caps.
@@bertblankenstein3738 Yes in general if you are using a MLCC you are not too worried about value, as most are used for decoupling, DC isolation or something else in that realm, and are not too critical value wise, which is a good thing as they also are very much going to change value depending on bias and time. Only when you get into RF circuits do they get critical as part of tuned circuits.
13:00 Isn’t this the only way they could have wires those capacitors? If the caps were after the switch, then the shunt would also be after the switch and you’d have to take switch resistance into account
I've seen this issue on UT181a battery drain when off I can't understand why not simplify on-off switch just the battery and especially on an alkaline powered multimeter
The 1N4007 is a 1A diode. What if someone puts in fresh new AA batteries in the wrong way.. how much current is going to flow? AA batteries have a reasonably low ESR, so it's easy to get >1A. I wouldn't be surprised if it was a multifold of 1A.. (like 3A or more) If the meter would sit there with these batteries for a while.. I'm sure that diode isn't so happy anymore. The meter may have survived, but perhaps that diode could start leaking and be the culprit. I think it's also quite easy to test. Just put in 2xAA in backwards, leave it for 5min, and then see if anything has changed(or smoked).
leadfree ist more deadly for component and less for all people. 🤗 Keep the reflowtemperature as less as possible. Test if 5sec longer on a lower temperature fixes this problems. Make test PCB with ~ 100 parts and measure each one after reflow
Agreed. The bureaucrats that came up with that lead free mandate need to be drawn and quartered (euphemistically of course). It is probably the worst decision forced on the electronics industry ever.
at the 6 to 7 min mark that graph is in percent of max volts not volts so that 25C leakage current graph ends around 100v not 10v. extrapolating I'd expect less than f-all current (sorry for the imperial units).
I’d put my money on routing - especially if there were _any other revisions before the one pictured - which then likely wouldn’t feature those caps. Still just a friendly wager though. All three components are fairly easy to test (or just replace, at a cost of about 10 cents).
Is there a video where you explain your deal with brymen, im interested did you design it yourself and they just manufacture it or is it more complex Also wouldn't mind link to the video about start grounding, the youtube search is utter garbage and until recently kept adding videos from recommend past second video and was basically useless for two weeks and then magically fixed itself
Could it maybe be they wanted EMI protection at each switch position and routing all the nets to the caps after the switch would develop some crosstalk or power unintended areas?
I am so confused. I thought the switch on the DMM physically opened and closed the circuit to the MCU and the other components. So does the OFF position on the switch actually do?
It cuts the circuit to everything except that one diode and those two caps. Which are on the battery side of the switch. Diode makes sense to be there, caps are a mystery.
I had the some issue with a fluke 176 old version i don't have the MSP microcontroller i replaced with new when i found a bizarre behavior when I was measuring 480vac the multimeter shut down and restart but it work fine with 110vac
I don't have a Brymen branded meter, but I do have a Greenlee DM-210A (similar and made by Brymen but not true RMS) which hasn't had any of these issues in ~5 years. I'll compare the board
Design flaw. This would be a complete non-issue if the off position was a physical clicky-clacky switch that mechanically and physically broke the circuit between B+ and the first part of the meter proper. Put a proper on/off switch on that thing. this 'soft off' bullshit is why this is even a thing.
@@henryD9363 Just like Nearly Every Device on market, The only expection is If it's Very Cheap and don't even have any Protection or tremendously well-Engineered or Expensive so they used better measures. But I'll say 90% of them uses it. It's so that the Fuse or the Diode Blows up and You replace/repair them instead of Repairing or throwing away the entire device
the red is actually a stand off and it USED to be soldered to a daughter board that had the actual banana jacks (for scale of course!) on it. Dave had desoldered those two connections in order to see the bottom third of the main board.
I was googling this today lol. Is this question relevant? "Why connect capacitors to ground?" e.g. from a post: A high-frequency signal will see the capacitor connected to ground, and travel through it, since it is a low impedance path, but a low frequency signal will not be affected by it. The capacitors to ground form a low-pass filter for the lines they're connected to, as they remove high-frequency signals from the line by giving those signals a low-impedance path to GND. Edit: I have no idea hoping someone can figure it out in comments and explain
And you were wondering, Dave...Why are Flukes so expensive? I remember your older videos where you were dropping a Fluke meter from a bridge onto rocks below. Then, you took the Fluke swimming with you, to see if that would kill it. Do you truly believe that this meter could hang with that Fluke?
The caps could be to smooth the voltage if there are some oxidation in the battery terminals and instability due to meter trepidation. But I also agree they could be after the rotary switch. The reverse protection diode after the switch with some long traces could mean if the battery is left reversed some time the traces could fuse with the current no?? Regards!
About the "only the 1N4007 should exist" thing: As far as I can tell, only the 1N4004 and 1N4007 actually exist. Diodes Inc gives their typical capacitance characteristics as two graphs: One for "1N4001..1N4004" and one for "1N4005..1N4007", suggesting that these bunched-together parts actually contain the same die. The lower-specced ones are just down-labeled (and, funnily enough, more expensive than the higher-specced ones... Probably because there's so little demand). I highly doubt that any of the diodes that get labeled as a "1N4001" actually failed the tests for a 1N4004.
Could be a fun thing to try - buy a few 1N4001 and 1N4004 from the same manufacturer and check their actual breakdown voltage (and/or forward voltage) to see if they're in fact the same part labeled differently.
During manufacture of my PDVS2mini's I used to have about 1 in 25 that would randomly switch themselves off during soak testing. Turns out it was a leaky reversed biased diode in the soft power switch circuit. I tested brand new diodes out the bag and could never measure a leaky one......till I reflowed a bunch in my oven. The pitfalls of 250degC lead free paste!
Changed to a different brand diode with similar spec and it never occured again.
all the more reason why banning lead in solder is stupid.
Can you provide the company of the diodes so I may be able to avoid them?
I had problems with Bournes 0805 smd inductors.
The inductance would change 20% after soldering. The temperature specs on the datasheet were followed.
These inductors had good specs for Q and current but were unusable.
I would certainly be suspecting the caps before suspecting the diode, they often fail from physical stresses or corrosion, with a known flex issue in that area I would be 99% sure one of the caps has a fracture which results in them shorting out.
Absolutely agree with
I think you've nailed it
i had one fail on me just from hand-soldering (or may have been faulty from the factory). I used it as a dc blocking cap in an audio amp, and had some dc at the output because of the leaky capacitor. I've seen cracked ones go short a fair number of times (cracked because pcb bends; i drop my multimeters fairly often and i bet others do too).
No! Corrosion is only applicable / known for Thin Film resistors, where the protection layer has pinholes. Have never heard of corrosion failures on MLCCs in Automotive.
just repeated thermal stress can crack them, especially X7R curve capacitors because the ceramic is very piezoelectric
I recalled repairing a TV decades ago. The complaint was the channels keep drifting. The owner had a new tuner unit replaced but that did not solved the problem. I was asked to have a look at it. I was at my wit's end and after a couple of days of meddling, I decided to check the diode 1N4148 for short but turned out OK. Anyway, I had it replaced and bingo the drifting stopped. Another time, a Luxman L85V amplifier having cracking noises on one channel that was caused by a noisy 1N4148 diode.
PS
During that time, my repair tools were and still are : a SANWA analog multimeter and my brain. No schematics. Now life is much easier with the internet.
I do find it quite disheartening that the manufacturer will happily promote repairs by creating a DMM, yet won't release schematics to repair their own devices 🙄 great video as always Dave, look forward to these videos
Specially cause anyone can just open the meter and reverse engineer it anyway, it’s not like they are stopping anyone from copying it 🤦🏼♂️
They are probably concerned about Chinese knockoffs. It can be reverse engineered anyway, but having to do it is still somewhat of a barrier to it.
@@sylviam6535 yeah but if they had half a brain they'd know schematics have been around for macboooks for years and no nock off macboooks are on the market lol
@@Kc12v140 yeah it wouldn't be difficult for someone like BigClive to reverse engineer within a few hours, a day at most lol
@@TheCod3r - Macs use lots of custom hardware which their suppliers are not allowed to sell to anyone else. See the Louis Rossmann channel for the loops he has to go through to find parts for repairs.
OH HAI. That was my forum post! (I am 10 months late to this video!)
Looks like I'm going to need to dig out my old cheap multimeter to finally fix this thing! (For these past 4 years Ive been in the habit of removing the batteries every time I used it)
A possible failure mode for the diode: if someone did install the batteries backwards, you would dump a lot of current through that diode. That could leave it with permanently degraded leakage.
I tried to post a more detailed analysis, but my post was blocked--maybe because I tried to link the relevant data sheets? Anyway, depending on some assumptions, a pair of fresh Energizers would end up dumping anywhere from 2 to 10 amps through the 4007--well in excess of its 1A rating. It would not surprise me to see permanently elevated leakage after that abuse.
@@danmenes3143 Interesting. That deserves further investigation. If it is the case then this kind of battery draining could be significant on a global scale?
@@danmenes3143 My same thinking. It's also so easy to do and not notice for a few moments or even minutes. It could even create some magic smoke if left for a little while to sit.
Frankly the reverse diode protection only works well if there is a fuse in series with the power source. Otherwise the best alternatives are a series diode or similar solution (N or P MOSFET circuits)
@@nlhans1990 That's what I'm thinking. Suppose you screwed the cover back on, then flipped it over and noticed it wouldn't turn on. "Magic smoke smell" might well be your first hint about what was wrong. A fuse or a PTC thermistor could be used to protect the crowbar diode.
@@danmenes3143 I was thinking the same thing during the video. Snap some primary lithium cells or rechargeable NiMh cells in there and it will easily exceed the maximum forward current of the diode. There would be at least a few watts of dissipation in the diode if high-capacity cells were used. If I were to use something like that, I would want to mold some additional features into the battery compartment to prevent the cells from making contact if they are put in reverse. The diode would then be a belt and suspenders in case the 200 lb. gorilla crammed them in hard enough to make momentary contact.
I had a digital tyre inflator /pressure gauge that used a coin cell, right out of the box it drained the battery over night and did several new batteries the same so I contacted PCL who made or sold them and they just sent me a new one. Having been told I could bin or do whatever I wanted with the first one I took it apart and had a look at the board , first thing that caught my eye was a big fat ceramic capacitor right across the battery so I remove it and not having a replacement I put the board back in and the gauge worked perfectly and the coin cell is still working the unit two years later. Not sure why the designers thought the cell needed a bypass cap. when the microchip had several of its own. Excessive battery drain is often down to leaky by pass capacitors and always worth a look if you have that type of trouble.
In Automotive applications, there might be MLCCs on the battery input, to suppress spikes from the cars supply system. It's forbidden though, to just put a single MLCC there, only two in series, as these components might develop leakage currents due to vibration, bending, even arising from the manufacturing process. This might even cause fire in extreme cases.
Therefore, these two MLCCs are good candidates for excess leakage currents.
I work as a PCB design engineer in the auto industry. Not only are there two capacitors in series, but they're also rated for much higher voltages than required for the application and are physically rotated 90 degrees to each other. That way they won't physically both be on the same flex plane.
@@Aerospace02 I think it's no secret, that small case sizes are forbidden in this case, e.g. 0805 preferred. Edge values, i.e. max. CV product of given technology should be avoided.
As well, technologies like Open Mode or Flex Termination may be used, but unfortunately not available any more in big quantities.
Root cause of failures in MLCCs are mechanical cracks, as well as bad spots /contamination in the ceramic layers, which gradually may cause leakage currents to shorts.
But there you have other spike sources on the battery, just the battery alone like here won't produce spikes
Ford specifies any capacitor across un-switched Vbatt has to use fail safe capacitors such as Muratas GCE series, they do not allow two separate capacitors anymore, have not for a number of years now. it is in the Elcomp.
@@KasperPilsted Ford allows us to use two separate caps on unswitched Vbatt. In fact, they specify it. This is for instrument clusters and HUDs. Does it perhaps depend on the product?
There are Experiments I have done once with no success, but there are times when I want, semiconductors are wonderful, I dreamed that I was doing experiments on electronics, and I also dreamed of videos Your sharing is very successful, I want to share with the community good knowledge! thank You Share!
Dave, the abscissa shows percentage of rated reverse voltage. It's not 10V but 10% of 1000V for the 4007. ...
When you say "it's 10V" for a 1N4007 that's actually the chart of 10% or 100V
For a jellybean diode making all the parts for 1kV, and likely also having most of them likely to survive 2kV in reality is cheap, so no reason to make the lower ranges, unless you are having a bad process month, and they are failing at 1kV, so you are looking at the voltage they fail, and binning the bad ones as 1N4001, so you can get them out, or simply making the entire run as 1N4001, even if only 1% are not passing at 1kV.
Most likes will only give you the lower voltage if you special order them, and in any case will give the 1kV part as standard, just marked down. Generally for things like military or aerospace parts, where they originally put in 1N4001, so changing part number is a pain, but the supplier can provide the part, marked as 1N4001, and it will pass the entire test spec for the 1N4001 part.
Originally IIRC the 1N4007 used to be a special process part, with a mesa etch to have the corners able to withstand 1kV, instead of the regular 1N400x, which was simply a scribed die. But turns out easier to not skip the process step, to do the mesa etch, so they made all of them as that instead, and got the process dialled in right. Slight difference in surge capability, but if you are doing something that runs the diode at that ragged edge you are doing it wrong, because there will be some that do not meet that "typical, not tested in production" spec off the line. If you want an avalanche diode there are better diodes, with actual process controls for the voltage.
I have a large number of 12V and 24V zener diodes, which I use more often than not as regular diodes, as they do work well there, replacing the 1N4148 in most cases. Have accidentally made a few of them turn into LED's as well, at least for a second or three.
D'oh, yeah, totally missed the %
@@SeanBZA The 1N400x series were first made back in 1963 when binning would certainly have been required to get a decent yield. These days I'd be surprised if they aren't all 1000v parts which are just remarked.
@@ferrumignis That is why you also got a 1N4000 diode, which was guaranteed to be a diode, though not above 25V, and it was very common in the surplus market in bags of assorted components.
IIRC Sinclair used to use the reject transistors and diodes from Ferranti to make his first amplifier modules, taking all those production line packaged rejects, for scrap metal value, and then sorting them into categories of dead, not too dead, very leaky, not too leaky, and will work under 15V, and used them in his amplifiers, at least those with some gain.
Ferranti tested at 30V, so those that failed there could still be useful at 12V or less. Was worth it when at the time a transistor was not something you bought for a penny, but instead was up to a pound, and a pound was worth something.
@@SeanBZA Some of the old Sinclair amplifier kits were rather notorious for blowing up, so that doesn't surprise me at all.
I just pulled my BM235 apart as I was watching this video and noticed the lack of the diode across the battery terminals in mine as compared to the one Dave was showing, I have had mine for many years now so is an older revision. My BM235 just started doing a strange fault on resistance or continuity mode will go to 0 when the leads are shorted then stay at zero when they have been opened again.
Putting the reverse battery protection diode after the switch doesn't sound like a terrific plan, the entire short-circuit current of the batteries will flow through the range switch, which is most likely way above its rated current.
Dave is too good. Always impressed.
13:40 unless there's a poly fuse or something (and even if there is), I could see not wanting the crowbar diode to act through the switch. It would subject the switch to the short circuit current, which, If they used lithium primary cells, could be quite substantial.
It also wouldn't protect anything upstream of the switch.
Follow the MONEY!
Big battery at it again!
Hi Dave, I kinda stopped using mine because I was tired of replacing the battery every time I needed to use it. Even then it was the least used of my meters. I've had so many other things going on I never got round to finding out if anyone else had the problem.
Dave don't care he already sucked you into buying the POS and spent your money long ago.
There's another option for excessive power consumption when switched off; during shipping the NSA intercepted it and installed a wiretap in it.
Also if that tantalum is in their unit I would be rather suspicious of it, they tend to develop very severe leakage currents when operated outside their ratings like in a momentary reversed battery scenario.
The tantalum is not across the battery
No clue how these are set up internally, but is there a chance that when turning that rotary switch you'd potentially create some EMI connecting/disconnecting power to all the circuits that they're trying to filter out in all cases? (with the cap/inductor)
Good thinking, it's possible I suppose. The switch could rapidly ramp certain voltage levels up and down, inducing EMI interference.
"I had to go through a couple dozen datasheets" sounds frightening to me but sounds perfect for this channel
My Fluke 89 IV and 189 both munch through AAs when off. My old 79-III only has a 9V battery and it's absolutely fine.
Fascinating, I had somebody mention that they had an issue with her guitar pick up draining batteries as in put new battery in and by the next day be completely dead. My first thought was capacitor or some sort of component sitting between the switch and the battery or the switch was not working right which is usually a quarter inch and shorting the ring shield. Lo and behold this video pops up in my feed and confirms my suspicion. If I do eventually work on it I will have to make a video of it because apparently this is a common issue for that pick up.
Be very aware that the X7R ceramic MLCC's (which are used in the higher value capacitors) are EXTREMELY piezoelectric and can function as microphones causing feedback in amplifiers. Much like many old Triode Vacuum tubes could because the grid would vibrate.
And i thought 007 drained only martinis
This is a timely video as I am chasing another issue with my BM235. A few days ago I turned it on and it initially showed the panel test (all segments lit up) and then the panel went blank. Believing it may have been flat batteries I changed them but the same result. Tried a few more changes of batteries having checked the voltage beforehand, but the display refused to come on after the initial panel test. Both fuses are intact also.
I then hooked up my bench power supply directly and fed the meter 3.2V. Still didn't turn on. Increased to 3.3V and the meter powered up and stayed on. I slowly reduced the voltage from 3.3V until I got to 2.4V when the low battery indicator on the meter came on, yet the meter was still operating as expected.
I shall be buying some AAA-size lithium batteries in the next few days, and hopefully, the inherent higher voltage of those will get me out of trouble, but still puzzled as to why normal fresh alkaline batteries have now stopped working. I'm suspecting the initial test for voltage has failed or has drifted.
Any ideas?
just because the batteries are fresh doesnt mean they are good. ive had numerous ones fresh out of the package volt test ok but under the slightest load go flat.
My theory is you were fooled into buying a P.O.S. meter
my 2cents. I'd connect it on the bench and have a look with an IR camera to see what is heating up while off.
i understand none of this but I still enjoy watching
It makes sense to put your primary decoupling caps at the base of the star grounding junction. That’s your true ground reference, don’t want to add track, lead, esr and connection resistance back to the battery on top of what your reference plane is
I believe you mentioned moisture-conductive dirt, especially under the SMT components. In my years of occasional conductive problems troubleshooting I recall two cases when there was a buried copper (or tin?) whisker between traces. Pretty hard to locate, but eventually chasing micro-ohms or even nano-ohms along the traces, I found the location. Looked perfectly OK on the top, even after removing the solder mask. Hmmm??? But digging into the epoxy, there was a hairline of copper (in my case). Looking with a good baclight and a good microscope it became evident that the whisker was just long enough to bridge the tho traces. A little more scraping and the whisker was gone. I don't recall exact resistance, but it was some ohms. Just a hair width or even less of copper. My PCB manufacturer told that he had seen some similar cases of buried whiskers in raw laminate. So my two cases were not unique.
Those capacitors will filter out the voltage surges due to the sliding ('bouncing') contact as the switch is rotated. I don't know whether this their primary purpose, but such surges on memory cells thoroughly confuses state machines.
You would have to place them on the other side of the switch for debouncing.
@@inse001 yeah, the only thing I can see is when you have the switch on and you change the batteries...but it seems to me a little too much to worry about from a design point of view....
@@inse001 Then the springed contacts to the board and battery terminal contacts.
Those two ceramic capacitors would also be my second item to check after the protection diode. The worst thing about the MLCC caps is that they tend to break short circuit. (Although there are some more expensive variants that are guaranteed to break open circuit, those did not appear to be that variety.)
Near the end of the video you mentioned PCB design concessions. I know what it's like. My most recent board is a very simple MCU (ATtiny1614) controlled power switch with 5000KV isolation, and a pair of LED's for operational status. Even on that simple board I had to partially rewrite the firmware because I needed to change pin mapping, and my MCU is only 14 pins! I can only imagine what happens in the back and forth at a company that is designing a board with hundreds, or even thousands, of connections, and the software, schematic, and layout guys trying to get a schematic that can be routed and still functions correctly.
Had a design where some reverse bias diodes suffered mechanical damage which lowered the resistance.
It's great the meter is saved with that reverse diode, but do the batteries get hot, leak and explode this way?
Normally the user would detect that in seconds as it does not work. You can short alkalines briefly without harm. NiMH is another matter.
Very interesting and informative video, as always Dave! 🥰
I have a Fluke 87 which I bought in 1994 and it has a similar issue. It's also a known problem with this model (and a few others) and it seems that there is a residual draw on the battery even when the switch is in the OFF position - so off isn't really off at all. This meter also drains batteries when it goes into sleep mode.
One thing I found was that when the switch was in the OFF position, sometimes the display would flicker a bit and the buzzer would emit a weird noise.
Tapping the rotary switch would affect these behaviors. The meter was then dismantled and the contacts on the rotary switch cleaned and lubricated. This stopped the above reactions, and slowed down the battery drain. I don't know why it's so hard for manufacturers to actually have the unit completely OFF when the switch says OFF. There should be ZERO current draw.
I've seen wiper noise affect readings, so maybe those components filter for that.
If you want to compare the boards, do another scan then open both in Photoshop or Photopea and overlay one over the other in one of the odd modes like divide or difference mode. Use "distort" to drag the corners to line up and the board will black-out as it lines up... except where the differences are. I've got to pull my 89 IV apart at some point and probably the 189 at the same time, I'll see if I can do the same.
My Swissmicros DM42 calculator was eating CR2032s for breakfast. Had it fixed under warranty. The culprit was an MLCC. Now it is perfectly fine.
Because they forgot to change the washer in the tap or faucet that's why it is draining as shown in Thumbnail.
I am very interested in learning more about the failure modes of electronic components since a large part of my job is troubleshooting a complicated board. I have no answers to questions about why they failed in particular ways. Where can I learn more?
IPC-7711 may be off interest if you're serious about it.
So that 10nA at the beginning were the caps charging...
Some old battery might have leaked (a tiny bit) onto the PCB..
LC series resonance frequency might give a clue about the purpose of those?
Capacitors so that, even if it is switched off, the large area of the battery does not act like an antenna, and couple RF energy into or out of the circuit, but instead connect it to the common terminal.
Yes the capacitors can be leaky, especially with the board flex issue, which likely has cracked them, and they started to grow whiskers, that are acting like a partial short or have reacted with moisture to make a conductive channel internally to the capacitor.
Probably a good idea on the next revision of the board to move those 2 capacitors, likely 2 different values, like 1000p and 100n, so as to get different self resonant frequencies, and move them to another orientation, so the flexing forces do not crack them. Probably assembly force applied direct to them is causing this.
So, they put a point of failure to avoid rf energy to do what exactly when the switch is off?!
at 5:44 it says rated peak reverse voltage in percent, so what you are calling 10V would be 10% of the rated peak voltage right?
which for the 4007 with a peak voltage of 1000V would be 100V then? or am I missing something?
I mean doesn't change your point just makes it even more clear I guess.
I don't know the construction of the range switch, but by the looks of it the positive strip isn't always connected to the same sub-circuitry. So maybe they wanted/"needed" these capacitors here in whatever mode is selected. And by designing this way they covered them all at once.
It can also be a failing tantalum capacitor if one is connected during off-state.
My eevblog BM325 died an early death due to battery leakage. I suspected something was draining it. It was just out of warranty and I wasnt impressed - my much older and much cheaper meter is still going strong
I have a cheap studio light with excess battery drain. Luckily i have two so i got to do some measurements and compare them.
The affected unit has a very high (12mA!) standby current, and oddly enough, when switched on at lowest brightness it actually draws less than that. My only conclusion so far is that it has a faulty mcu or firmware causing it to draw excess current in standby. Current draw when switched on was completely normal and matched the other unit.
Mhm they probably have the same firmware for years and never change it. It would be a weird fault for the MCU as well to not enter standby properly, but then, weirder things have happened, so it is a possibility. It might be worth digging further for educational purposes. It could be powering something external to the MCU down and that thing is malfunctioning for some reason. Difficult to guess without seeing the device.
Dave, I would suspect that large capacitor C1 right below the diode that you ignored.
If the 4007 was the culprit then it would be easy enough to find a replacement, but if it’s the mlcc caps then how do you determine their value since they are not marked. That’s something that annoys me about smd caps.
You have to guess based on the functionality, or maybe it's still readable just with higher leakage.
Probably safe to whack any cap on there as long as it fits.
@@bertblankenstein3738 Yes in general if you are using a MLCC you are not too worried about value, as most are used for decoupling, DC isolation or something else in that realm, and are not too critical value wise, which is a good thing as they also are very much going to change value depending on bias and time. Only when you get into RF circuits do they get critical as part of tuned circuits.
Really interesting - would like to see an equivalent tour of a Fluke 87.
Maybe the diode is considered not fast enough to shunt reversed polarity. Maybe charging capacitors delay the reversed voltage rise
Don't rule out the tantalum capacitor. I've seen plenty of them go leaky.
Seen plenty of them go bang in my younger days working on TTL designs.
I always have a small strip of thin double sided PCB for battery current tests. small thick wires soldered to it.
13:00 Isn’t this the only way they could have wires those capacitors? If the caps were after the switch, then the shunt would also be after the switch and you’d have to take switch resistance into account
battery spring contacts are bent and touching nearby traces or vias. they are very near the landing pads!
At 6:00 the x axis is not voltage but % of peak reverse voltage.
So it's not 10V but 10% of 1000V ie 100V !!!!
I've seen this issue on UT181a battery drain when off I can't understand why not simplify on-off switch just the battery and especially on an alkaline powered multimeter
Why use the star ground/net instead of a polygon?
The 1N4007 is a 1A diode. What if someone puts in fresh new AA batteries in the wrong way.. how much current is going to flow? AA batteries have a reasonably low ESR, so it's easy to get >1A. I wouldn't be surprised if it was a multifold of 1A.. (like 3A or more)
If the meter would sit there with these batteries for a while.. I'm sure that diode isn't so happy anymore. The meter may have survived, but perhaps that diode could start leaking and be the culprit. I think it's also quite easy to test. Just put in 2xAA in backwards, leave it for 5min, and then see if anything has changed(or smoked).
I'm thinking it has something to do with getting the relative voltage of the battery at level with the probes, or just some self probing protection?
Would the caps across the switch be for debouncing?
leadfree ist more deadly for component and less for all people. 🤗
Keep the reflowtemperature as less as possible. Test if 5sec longer on a lower temperature fixes this problems.
Make test PCB with ~ 100 parts and measure each one after reflow
Agreed. The bureaucrats that came up with that lead free mandate need to be drawn and quartered (euphemistically of course). It is probably the worst decision forced on the electronics industry ever.
at the 6 to 7 min mark that graph is in percent of max volts not volts so that 25C leakage current graph ends around 100v not 10v. extrapolating I'd expect less than f-all current (sorry for the imperial units).
I’d put my money on routing - especially if there were _any other revisions before the one pictured - which then likely wouldn’t feature those caps. Still just a friendly wager though. All three components are fairly easy to test (or just replace, at a cost of about 10 cents).
Wow, naked multimeters look very vulnerable!
Is there a video where you explain your deal with brymen, im interested did you design it yourself and they just manufacture it or is it more complex
Also wouldn't mind link to the video about start grounding, the youtube search is utter garbage and until recently kept adding videos from recommend past second video and was basically useless for two weeks and then magically fixed itself
Could it maybe be they wanted EMI protection at each switch position and routing all the nets to the caps after the switch would develop some crosstalk or power unintended areas?
What are the values of those Caps? How would you measure them?
You buy 1N4004 because it is the only thing in stock, might be a reason this past year.
I am so confused. I thought the switch on the DMM physically opened and closed the circuit to the MCU and the other components.
So does the OFF position on the switch actually do?
It cuts the circuit to everything except that one diode and those two caps. Which are on the battery side of the switch. Diode makes sense to be there, caps are a mystery.
I had the some issue with a fluke 176 old version i don't have the MSP microcontroller i replaced with new when i found a bizarre behavior when I was measuring 480vac the multimeter shut down and restart but it work fine with 110vac
16:57 ... you'll have to link it in?
I don't have a Brymen branded meter, but I do have a Greenlee DM-210A (similar and made by Brymen but not true RMS) which hasn't had any of these issues in ~5 years. I'll compare the board
since it probably uses the same Analog Devices chipset as the Fluke 7X series, you may be able to convert it to RMS like Dave did the other day.
Bah. I was just about to buy a bm235. So hold off on getting it?
Design flaw. This would be a complete non-issue if the off position was a physical clicky-clacky switch that mechanically and physically broke the circuit between B+ and the first part of the meter proper.
Put a proper on/off switch on that thing. this 'soft off' bullshit is why this is even a thing.
What is the point of keeping a schematic from you...?
Why didn't you remove those caps you suspect of leakage for testing ?
It's not the Customer's Board, so he don't have to same issue to replicate or test
So the diode just shorts out the battery if they are in the wrong way?
Yep!
Amazing! How often is this design used in consumer appliances?
@@henryD9363 Just like Nearly Every Device on market, The only expection is If it's Very Cheap and don't even have any Protection or tremendously well-Engineered or Expensive so they used better measures. But I'll say 90% of them uses it. It's so that the Fuse or the Diode Blows up and You replace/repair them instead of Repairing or throwing away the entire device
Does the other eevblog multimeter have the 10 nano amps resolution
what are the two red cut off wires that are coming out from the multimeter PCB?
the red is actually a stand off and it USED to be soldered to a daughter board that had the actual banana jacks (for scale of course!) on it. Dave had desoldered those two connections in order to see the bottom third of the main board.
Nintendo Wii remotes surely have a similar fault, always drain in weeks of not being used 😟
While AA batteries are great low cost option, they seem to leak almost every time as compared to the 9v battery
I was googling this today lol. Is this question relevant? "Why connect capacitors to ground?" e.g. from a post: A high-frequency signal will see the capacitor connected to ground, and travel through it, since it is a low impedance path, but a low frequency signal will not be affected by it. The capacitors to ground form a low-pass filter for the lines they're connected to, as they remove high-frequency signals from the line by giving those signals a low-impedance path to GND.
Edit: I have no idea hoping someone can figure it out in comments and explain
Possibly to filter switch noise when turning the unit on to prevent the MCU from having issues arising from a bad reset pulse?
It would be nice if he sent it to you, so we can see you solve it.
Were the dudes cells NiMH?
And you were wondering, Dave...Why are Flukes so expensive? I remember your older videos where you were dropping a Fluke meter from a bridge onto rocks below. Then, you took the Fluke swimming with you, to see if that would kill it. Do you truly believe that this meter could hang with that Fluke?
Diminishing returns. These are still made very well
Hi!
im guessing them caps are there to stop the measurements bouncing around by providing a clean rail..... but.. i could be just talkin out my arse! :P
The caps could be to smooth the voltage if there are some oxidation in the battery terminals and instability due to meter trepidation. But I also agree they could be after the rotary switch. The reverse protection diode after the switch with some long traces could mean if the battery is left reversed some time the traces could fuse with the current no?? Regards!
How can I get a bm786
No schematics, not even for a trusted partner.... No worries for Dave.
It's the caps, it's always the caps 😁
Since you're having these meters made for you can you the company to provide schematics?
I wish I had watched this I had an expensive camera that would drain BATTERIES very fast.
One word, batterizers
Poor diodes, they are simple d(i)udes.
Let's have "A Squizz of Sheila shall we?" Beuller? Beuller?
he he
My money was on capacitors.