Capacitor ESR testing

Repairs & upgrades.
User avatar
exxos
Site Admin
Site Admin
Posts: 23488
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Capacitor ESR testing

Post by exxos »

I actually did these tests way back in early 2018 intention of writing a more in-depth article. Though I thought I would throw this information out there meanwhile. This page is purely information to illustrate how capacitors can vary tolerances and values on the test. It is not a how-to guide on how to test capacitors!

Test were done with 3 capacitors removed from STF/E power supplies. There were actually a lot more capacitors tested during these experiments. But I basically cherry picked the ones below to illustrate what happens. One of the problems with testing capacitors in such a way is we do not actually know the tolerance or ESR rating to start with. You can probably assume 20% tolerance. But I have seen capacitors up to 50% in the past.

I was doing some other tests under load originally but tests were not exactly portraying a fault condition. Because the capacitors which were measuring higher value were actually regulating better because of the inherent higher value. But of course the capacitor being so far out of tolerance indicates it is failing. So I did not include tests or take it any further because I really needed a better is test setup and never got around to designing one.

The first set of images (taken on carpet) were the "cold" tests. The images taken on my workbench were after they were gently warmed in my ultrasonic cleaner heater tank. This was to simulate them running for a long period of time.


CASE 1 - 1000uF JAMICON

1,000uF capacitor @ 20% tolerance , 800uF - 1200uF
COLD - 1106uF 0.06R
HOT - 1174uF 0.02R

This one seems to be within 20% tolerance but the actual ESR value varies somewhat.

IMG_2824.JPG
IMG_2824.JPG (359.92 KiB) Viewed 2889 times
IMG_2827.JPG
IMG_2827.JPG (339.03 KiB) Viewed 2889 times

CASE 2 - 2200uF WANGS

2,200uF capacitor @ 20% tolerance , 1760- 2640uF
COLD - 2860uF 0.02R
HOT - 3121uF 0.02R

ESR remains consistent but the value changes a fair amount.

IMG_2819.JPG
IMG_2819.JPG (356.99 KiB) Viewed 2889 times
IMG_2829.JPG
IMG_2829.JPG (316.69 KiB) Viewed 2889 times


CASE 3 - 4700uF WANGS

4,700uF capacitor @ 20% tolerance , 3760 - 5640uF
COLD - 5052uF 0.00R
HOT - 5408uF 0.02R

ESR actually gets lower as it heats up and capacitance increases.

IMG_2822.JPG
IMG_2822.JPG (432.09 KiB) Viewed 2889 times
IMG_2828.JPG
IMG_2828.JPG (294.29 KiB) Viewed 2889 times

CASE 4 - 1500uF PANASONIC FR - 20% 0.02R

1,500uF capacitor @ 20% tolerance , 1200 - 1800uF
COLD - 1664uF 0.02R
HOT - 1742uF 0.00R

As expected Panasonic is well within tolerance hot or cold. ESR does drop down to below what the meter can detect.

IMG_2820.JPG
IMG_2820.JPG (381.64 KiB) Viewed 2889 times
IMG_2826.JPG
IMG_2826.JPG (345.42 KiB) Viewed 2889 times


In most cases you could assume the capacitors are all intolerance and ESR is "good" but unfortunately things are not as black and white. the Panasonic as a base reference is what we are likely to expect from a good capacitor. It varies just only 78uF from cold to hot.

With case 2, value is clearly out of tolerance when cold or hot which is approximately 42%!

Coming from a company where we service endless thousands of power supplies, I can confirm that when the value starts getting higher the capacitor has reached its end of life. The power supplies used to power fruit machines where they would be generally pushing several amps or higher. We could always visually tell when the capacitors were failing because the lamps started flickering. generally the capacitors would measure a lot higher value than would ordinarily be expected. unfortunately this is not a surefire method of capacitor failure is one of the main ones. Capacitors could still measure perfectly well and still cause lamps flickering.

One thing to watch out for is the apparent "good" ESR value. Generally when the capacitor starts reading higher value than it should be, the internal dielectric is starting to break down and fail which is bringing the internal foil layers closer together increasing the value. Because of this failure the ESR value will generally get lower because it is approaching a short circuit type of failure.

In some cases we just used to measure capacitors on basic multimeter resistance test. Sometimes the resistance would keep on climbing and then hit a point and simply stop. For example a faulty capacitor as it charged could reach 500ohms and go no higher indicating the capacitor has failed. Again this is not a surefire method because this failure may only happen when the capacitor has been in operation for some time. Measuring "cold" clearly can give a different results than when it is warmed up. In general using a oldskool AVO analogue meter is generally better for testing capacitors because it has a much lower impedance than digital mean the meters which have very high impedances so do not "load" the capacitor the same as oldskool meters do.

So while these ESR meters can give you a indication, and obviously if the capacitor is measuring way higher than it is supposed to be, then sure, that capacitor is bad. BUT, you basically cannot trust them to give you the real picture. You are personally introducing a discrepancy if you are just testing the capacitors "cold". Also if you are not measuring the resistance with a analogue meter, then you are skipping a series of tests completely.

So again while these meters can give you a indication, you really need to actually know how to use them properly, and understand their limitations and limitations of the tests you are actually doing.


EDIT:

Going back to the ESR value again. These old capacitors seem to be like the ESR value of modern day capacitors. I would assume this is unlikely that 30-year-old technology would equal today's technology. But without finding the datasheets for the original capacitors this is very hard to prove. For example the original capacitors could may well have been 0.10R or even 0.20R etc. as a dielectric fails the resistance progressively goes down until it reaches 0.02R. So you could mistakenly think that this capacitor is really good as it is on par with modern capacitors.

The reverse problem is that a capacitor could actually read 0.10R and you could then think is faulty. But again, without the datasheet, this may or may not be the proper value for that particular capacitor. I think even the modern high-voltage caps have around 0.5R resistance in general for example. Of course if the ESR is like 0.5R or higher for values like 4,700uF then you can pretty much assume that capacitor is faulty. But overall the ESR value as tested is somewhat meaningless unless you have some sort of baseline comparison to compare against.

The only indication you could actually go on is the measured capacitance value. But again because of tolerances, it is very difficult to go on that value alone. Of course if the capacitor is higher than 20% value, then the likelihood is that capacitor is failing. The Panasonic capacitor varies about 5% in capacitance value from cold to hot. Obviously this has to be a "good" figure for a modern new capacitor. While the capacitors tested are generally varying 10% from hot to cold. You could use this as a indication, but without the datasheet for any sort of method of comparison then there is not really any reliable way to test them.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
Steve
Posts: 2570
Joined: Fri Sep 15, 2017 11:49 am

Re: Capacitor ESR testing

Post by Steve »

Very informative, thanks for posting :dualthumbup:
Maximilian
Posts: 290
Joined: Thu Feb 28, 2019 5:02 pm

Re: Capacitor ESR testing

Post by Maximilian »

Thanks!!
User avatar
rubber_jonnie
Site Admin
Site Admin
Posts: 10462
Joined: Thu Aug 17, 2017 7:40 pm
Location: Essex
Contact:

Re: Capacitor ESR testing

Post by rubber_jonnie »

Just been reading this as I'm keen to improve my understanding of ESR, primarily as I have a Behringer mixer with a faulty output channel and the damn thing is full of caps that I want to check.

I know low ESR is a good thing, because the less power the cap dissipates the better, but what are good values?

Having done some reading, it appears that ESR is related to test frequency, but a reading of between .1R and 1R is good, though because of the frequency aspect I'm finding it difficult to get a kind of 'stock' answer.

Of course there may not be one, and it may depend on the frequency you use to test, so all a little confusing. And of course as @exxos has said, temperature is a factor too.

I do have a small multi component tester, and so far, the caps I've tested from the mixer are averaging about .4R, so my question would be, does that seem to be a bit high for ESR, or does that range of .1R to 1R seem reasonable.

I just looked at a Peak ESR meter and they're a bit on the expensive side for the use I'd get from it, so I might be stuck with the multi test tool. For now!!
Collector of many retro things!
800XL and 65XE both with Ultimate1MB,VBXL/XE & PokeyMax, SIDE3, SDrive Max, 2x 1010 cassette, 2x 1050 one with Happy mod, 3x 2600 Jr, 7800 and Lynx II
Approx 20 STs, including a 520 STM, 520 STFMs, 3x Mega ST, MSTE & 2x 32 Mhz boosted STEs
Plus the rest, totalling around 50 machines including a QL, 3x BBC Model B, Electron, Spectrums, ZX81 etc...
User avatar
exxos
Site Admin
Site Admin
Posts: 23488
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: Capacitor ESR testing

Post by exxos »

As you can tell nothing is ever simple, "good" when ? Good 30 years ago or good now ? Axial caps tend be be even worse ESR regardless of age. There is no "stock answer" as there is just so much variation with everything.

Physical size of caps is also a factor. Rolled up caps, as most are, you have a resistance across the entire metal foil. So when you get taller caps, the area is larger and resistance is lower.

Also higher voltage caps like 450v 47uf areas are typically 1R these days. They could have been more like 10R or higher 30 years ago. This is why you need data sheets, as with no frame of comparison it's almost meaningless to test them. But also in terms of lower voltages, 16v caps to 25v caps of the same value, the higher voltage caps tend to have better ESR values.

So I can't really give any general rules unless the original data sheets were found to see whats they are supposed to be when they were new. But that is mostly meaningless also unfortunately.

I guess worst case, caps should always be below 0.1R for 1,000uf type caps. But these days 0.02 - 0.05 is the normal, at least for quality caps. But this is for known recent caps as this is generally what good ones are. But not all caps are created equally. Just look up a 4,700uf 16v cap from several manufacturers and you will see what I mean.

But like I said before, you can have a 30 year old cap which can measure better ESR then new caps. Not because they are better, but because they are failing / shorting out internally, they could have been 0.4 when new, started to fail, and now measure 0.1R, so you can make a huge mistake in thinking it's good when it's not.

So ESR meters are basically useless unless you know what specs the caps are supposed to be in the first place. But even if the data sheet matches what you measure, it still is not really a valid test. The cap can still be faulty even "testing fine".

Again if the cap ages / dries out, it increases from a nominal 0.1R for example to 0.2R, then you can know something isn't right. BUT years later when it starts failing internally and drops back down to 0.1R.. You would assume it's perfectly fine., but it could have been going faulty for years and the ESR would not give you "the truth". Measuring the actual value can give indications also. But with tolerances 20% it's not so simple to go on that either. Which is all I explained above.

It's also generally a false assumption that ESR will increase with age as it's also based on a faulty assumption they tend to dry out. They only dry out if they are leaking in general. Overall they go faulty with a lower ESR values than it should be. Again as I explained above.

So the only thing you can really do is study the circuit they are being used in. Like PSUs, I don't go on what all the meters say, just look on the scope, check the ripple, if you're seeing more than like 0.2v or huge spikes, then things are failing. Same with the video caps on the STFM cause faults. Same with caps in the reset circuit etc etc. Some circuits may well function fine even with a really bad capacitor. It depends what it's doing.

I've known equipment which fails with changing for newer caps, this is why I never say "recap" without good reasons. I've seen people accuse me of "jumping on the recap bandwagon" as it's a de facto standard these days, but I never actually did that. All the caps I say change have been backed up with a lot of hard graft research.

Like the PSU and video circuits, I've personally tested and diagnosed the circuits and swapping them is a good thing. A lot of people recap with doing no investigations at all. Most people these days just replace caps as they are old, which is understandable. But most haven't done any research into what they are doing. A low ESR cap has a higher turn on inrush current and it can kill poor PCB traces which can't handle the extra turn on load for example. It's also why my PSUs have a slow ramp up voltage as it protects equipment better.

If you're talking about audio caps, it's all a fail IMO to even use electrolytics in signal paths. You tend to get a lot of hiss. For some bonkers reason audio designers keep using them when we've had better alternatives for years. Most audio line caps tend to be in the order of 0.47 - 22uf low voltage caps. Just take them out, stick in ceramics or polyester caps. Caps in the PSU areas of audio equipment will be better with newer caps as they will give a much better stable voltages etc.

TL;DR
Just ditch the new age toys and diagnose things properly by real world testing. Just make sure you research what caps you will swap them out for as there are a lot of cheap garbage caps out in the market sadly.

Edit:
When I go up in the loft again (likely soon) I'll look for my avo meter and see if it helps with cap testing. The old timers where I used to work refused to even use digital meters!! There's food for thought right there as well.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
User avatar
rubber_jonnie
Site Admin
Site Admin
Posts: 10462
Joined: Thu Aug 17, 2017 7:40 pm
Location: Essex
Contact:

Re: Capacitor ESR testing

Post by rubber_jonnie »

Thanks so much for all that info @exxos , definitely food for thought, and there are so many factors at play here, which is why I was confused and asked in the first place, because I know you have a lot of expertise.

And this is why I think there are no clear answers when you research the subject. I guess the other problem is, there have been so many different makes and no brand makes, how are you supposed to find a specific datasheet?

For example, the brand in my mixer is Xunda. Good luck with finding a specific datasheet for that brand and value!!

I suppose you could take the approach of sticking to known reputable brands like Nichicon, United/Nippon Chemi-con, Rubycon, Panasonic and Samxon. At least these are the brands that doing some research says are reputable.

With regards to my mixer, all the caps tested so far are potentially bad, however, I still have to trace from the output socket back to the source, because it could be one of the ICs that has failed and not a cap at all, so I need to stick to the plan and do that, because it may work just fine as is and the capacitors aren't actually an issue. If I do replace any caps, it will be for a decent brand.

I had noticed that there are 5 outputs

Th higher inrush current for low ESR caps is also very interesting, and something I hadn't thought about.

Interesting you mention the Avo, we used these in the RAF along with Meggers for testing ignition harnesses, and it reminded me my Dad has a meter, possibly an Avo, that I'll have to look into next time I visit.

Back to work for now, then tonight I shall trace the board and see if I can find a schematic. It is interesting to note that there are actually 5 outputs (Phones L&R combined, control room out and main out, both separated L &R) and 5 caps that seem to be associated with these, so they could at least be something worthwhile and easy to change with a decent brand.
Collector of many retro things!
800XL and 65XE both with Ultimate1MB,VBXL/XE & PokeyMax, SIDE3, SDrive Max, 2x 1010 cassette, 2x 1050 one with Happy mod, 3x 2600 Jr, 7800 and Lynx II
Approx 20 STs, including a 520 STM, 520 STFMs, 3x Mega ST, MSTE & 2x 32 Mhz boosted STEs
Plus the rest, totalling around 50 machines including a QL, 3x BBC Model B, Electron, Spectrums, ZX81 etc...
User avatar
exxos
Site Admin
Site Admin
Posts: 23488
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: Capacitor ESR testing

Post by exxos »

rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am And this is why I think there are no clear answers when you research the subject. I guess the other problem is, there have been so many different makes and no brand makes, how are you supposed to find a specific datasheet?
Exactly. It's like trying to calibrate a car engine when you don't even know what the timings are supposed to be in the first place.

rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am I suppose you could take the approach of sticking to known reputable brands like Nichicon, United/Nippon Chemi-con, Rubycon, Panasonic and Samxon. At least these are the brands that doing some research says are reputable.
This was the topic the other day about reputable brands. But "who" says they are reputable? There are a lot of reputable brands who also make garbage. Just because "everyone on the net" says some brand is reputable, is it really? How many of those people actually looked at datasheets? Or is it more of "Panasonic make loads of stuff, so they must be good" type of thing.

In case of Panasonic, good brand of course, I use them, but their smaller values of caps are pretty bad compared to a company like Rubycon. Higher values EPCOS tend to be better (and cheaper!) . You have to do the legwork.
rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am If I do replace any caps, it will be for a decent brand.
There are a lot of car tyres which are reputable brands, but are they all created equally ?
rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am Interesting you mention the Avo, we used these in the RAF along with Meggers for testing ignition harnesses, and it reminded me my Dad has a meter, possibly an Avo, that I'll have to look into next time I visit.
Digital meters are "high impedance" . That was one of the sales pitches back in the day. "better" for in-circuit testing as they would not try to power up the circuit like AVO's would. So people kept on using them, other than the old timers. They used the AVO as they were low impedance, they would load stuff better and give better results (assuming you removed the part from the pcb, which you should be doing anyway) . I pretty much used a DMM during all my work days. Even though they pretty much told me not to for several years :lol:

Being a n00b I didn't listen until one day old bob told me to test a transistor with a DMM to see if it was faulty or not. Diode test shows good results, resistance showed good results, "yes bob is good".. With a big grin on his face old bob told me to test with a Avo, it was totally shorted out, like zero ohms. Its one of those lessons you don't forget.

So while I still use a DMM, I never take it as gospel even if it says a diode / transistor / capacitor is good. If things test good and I think they are not, I change them anyway. As we used to say in our old workshop "if in doubt - chuck it out". In fact I just use a DMM to see if transistors are shorted or not and generally when things fail, they fail short, but I never trust a meter 100%. I would use the AVO, but its a case of "too big to use". Most of what I do is just measure volts and amps with DMM's. Most other things I use a scope.

In fact take this "argument" that everyone tests voltages with a meter, see 5volts, things are fine and good right? That's what everyone will start with and end with. Right there the meter is lying to you. Connect a scope and you can see 1V ripple (meters are generally not fast enough to detect ripple) and several volts spikes, again meters are not reliable. You have to know when to use and not use equipment and know its limitations. This only comes with years of experience. You can have 5V on a PSU and it blows fuses or trips out and people don't understand why. Get the scope out and you see several volts spikes which trip the PSU's protection. But people's knowledge will stop at "5V = good" and then they will start taking advice from random people and go down the rabbit hole of insanity.
rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am so they could at least be something worthwhile and easy to change with a decent brand.
:chairsmack:

Honestly, Lately I just think the whole internet is just full of armchair experts. It seems to boil down to who produces the most convincing article factual or not. If people get convinced enough, everyone follows and it becomes "fact". Information is so bad it's not even worth seeing "what the word on the street" is anymore. I quickly looked for "do meters lie" articles and didn't find any. I rest my case. If someone published such a article, I would be more inclined to trust other stuff they are reviewing.

Manufacturer technical docs are the only decent source of information. They don't always get things right, but that's where you need to look for information. I think it was Texas Instruments who did a very nice article on scope probes and how proper grounding is needed etc. I never saw anyone else even taking that advice or publishing more tests on it. Though I just don't tend to look anymore anyway.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
User avatar
rubber_jonnie
Site Admin
Site Admin
Posts: 10462
Joined: Thu Aug 17, 2017 7:40 pm
Location: Essex
Contact:

Re: Capacitor ESR testing

Post by rubber_jonnie »

exxos wrote: Wed Dec 01, 2021 12:42 pm
This was the topic the other day about reputable brands. But "who" says they are reputable ? There are a lot of reputable brands who also make garbage. Just because "everyone on the net" says some brand is reputable , is it really ? How many of those people actually looked at datasheets ? Or is it more of "Panasonic make loads of stuff, so they must be good" type of thing.

In case of Panasonic, good brand of course, I use them, but their smaller values of caps are pretty bad compared to a company like Rubycon. Higher values EPCOS tend to be better (and cheaper!) . You have to do the legwork.
exxos wrote: Wed Dec 01, 2021 12:42 pm There are a lot of car tyres which are reputable brands, but are they all created equally ?
No they aren't (I've had some crap tyres too!!), but if we don't make some sort of a choice, then we won't buy any. I've spent quite a while looking at what are considered 'reputable' brands rather than some random brand from eBay.

This is info I've pulled from various sources whilst trying to avoid the AEs as we can call them ;)

:headbang:
rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am Interesting you mention the Avo, we used these in the RAF along with Meggers for testing ignition harnesses, and it reminded me my Dad has a meter, possibly an Avo, that I'll have to look into next time I visit.
exxos wrote: Wed Dec 01, 2021 12:42 pm Digital meters are "high impedance" . That was one of the sales pitches back in the day. "better" for in-circuit testing as they would not try to power up the circuit like AVO's would. So people kept on using them, other than the old timers. They used the AVO as they were low impedance, they would load stuff better and give better results (assuming you removed the part from the pcb, which you should be doing anyway) . I pretty much used a DMM during all my work days. Even though they pretty much told me not to for several years :lol:

Being a n00b I didn't listen until one day old bob told me to test a transistor with a DMM to see if it was faulty or not. Diode test shows good results, resistance showed good results, "yes bob is good".. With a big grin on his face old bob told me to test with a Avo, it was totally shorted out, like zero ohms. Its one of those lessons you don't forget.

So while I still use a DMM, I never take it as gospel even if it says a diode / transistor / capacitor is good. If things test good and I think they are not, I change them anyway. As we used to say in our old workshop "if in doubt - chuck it out". In fact I just use a DMM to see if transistors are shorted or not and generally when things fail, they fail short, but I never trust a meter 100%. I would use the AVO, but its a case of "to big to use". Most of what I do is just measure volts and amps with DMM's. Most other things I use a scope.

In fact take this "argument" that everyone tests voltages with a meter, see 5volts, things are fine and good right ? That's what everyone will start with and end with. Right there the meter is lieing to you. Connect a scope and you can see 1V ripple (meters are generally not fast enough to detect ripple) and several volts spikes, again meters are not reliable. You have to know when to use and not use equipment and know its limitations. This only comes with years of experience. You can have 5V on a PSU and it blows fuses or trips out and people don't understand why. Get the scope out and you see several volts spikes which trip the PSU's protection. But peoples knowledge will stop at "5V = good" and then they will start taking advice from random people and go down the rabbit hole of insanity.
Well, I know what you mean, 'It's a DMM so it must be right?' Nope and it is limited to just volts being wrong?

What I wrote here https://www.exxosforum.co.uk/forum/viewt ... =91&t=4886 shows we are relying on things that aren't necessarily accurate. I was using that £5 meter to work on a BBC whilst I was in a rental as my 'good' tools were in storage and didn't want to spend much. I just stopped believing it and binned it off because it led me down the path of the PSU being bad, when actually it was the DMM reading low. I now tend to use the 'scope for most things as it is on my workbench permanently and I can see what you mean about ripple.

Same is true of the £29 oscilloscope I bought for the same reason, I was desperate to fix this damn BBC (Also dumb with desperation!!). Did it do me any good, nope, it didn't have the scope (Oscilloscope joke there :) ) to do what I wanted, which was to measure frequencies. I could do the low stuff, but the rest was a mess, even TTL stuff was just awful.
I may do a comparison with my main scope and use my frequency generator to generate some nice source frequencies and waveforms.

Even connecting a bench PSU to the scope (Should be perfect right? Hmm, maybe not), ripple is evident there at the levels you mention. Not terrible, but present nonetheless. DMM just doesn't see it.
rubber_jonnie wrote: Wed Dec 01, 2021 10:22 am so they could at least be something worthwhile and easy to change with a decent brand.
exxos wrote: Wed Dec 01, 2021 12:42 pm :chairsmack:
I'll take that one on the chin. I'm kinda thinking out loud here, as I sometimes do.
exxos wrote: Wed Dec 01, 2021 12:42 pm Honestly, Lately I just think the whole internet is just full of armchair experts. It seems to boil down to who produces the most convincing article factual or not. If people get convinced enough, everyone follows and it becomes "fact". Information is so bad its not even worth seeing "what the word on the street" is anymore. I quickly looked for "do meters lie" articles and didn't find any. I rest my case. If someone published such a article, I would be more inclined to trust other stuff they are reviewing.

Manufacture technical docs are the only decent source of information. They don't always get things right, but that's where you need to look for information. I think it was Texas Instruments who did a very nice article on scope probes and how proper grounding is needed etc. I never saw anyone else even taking that advice or publishing more tests on it. Though I just don't tend to look anymore anyway.
Honestly, I always get the datasheets etc and go by what they say because they are the truth from the manufacturer, but capacitors just seem to be such a terrible black hole of nothingness, unless of course you're buying new.

I'm trying very hard to ignore 'opinions' because as you say, there are a lot of AEs out there. It's often quite hard to decipher the guff from the good, and in people terms the 'good from the evil'!!!

As with most things I've learned over the years (particularly recently) is to not take the first thing you find as gospel.

Wow, took me quite a while to write this, better go do some work!!
Collector of many retro things!
800XL and 65XE both with Ultimate1MB,VBXL/XE & PokeyMax, SIDE3, SDrive Max, 2x 1010 cassette, 2x 1050 one with Happy mod, 3x 2600 Jr, 7800 and Lynx II
Approx 20 STs, including a 520 STM, 520 STFMs, 3x Mega ST, MSTE & 2x 32 Mhz boosted STEs
Plus the rest, totalling around 50 machines including a QL, 3x BBC Model B, Electron, Spectrums, ZX81 etc...
Higgy
Posts: 400
Joined: Tue Apr 23, 2019 8:05 pm
Location: Somerset

Re: Capacitor ESR testing

Post by Higgy »

I recently changed roles at work and now get involved with lots of old industrial electronics (need to post about PC/104 and 800Mhz '486's').

We have some 'optical distribution cards' that were not working. The control cabinets they live in are really warm. I changed the caps and now the cards are working well.

I bought a Peak Electronics ESR tool and the old (mid 90's) caps (Marcon 100uF 63V) were reading 53uF and ESR of ~7R.

I tested a load of caps from my computers and the Peak Electronics and my home cheap eBay unit gave pretty similar results. Those Sharp X68000 PSU caps certainly get fudged with heat!

@exxos what temp roughly was your hot tests? Thanks
User avatar
exxos
Site Admin
Site Admin
Posts: 23488
Joined: Wed Aug 16, 2017 11:19 pm
Location: UK
Contact:

Re: Capacitor ESR testing

Post by exxos »

Higgy wrote: Fri May 26, 2023 11:37 am @exxos what temp roughly was your hot tests? Thanks
I'd guess I would have used something like 50c.

The problem with a lot of power supplies is they tend to put electrolytic's right next to heatsinks which can reach 85c. It dramatically reduces their life. Going back to my old job, we just used to replace everything with 105c caps and had much less failures. A tiny bit of heat over time can build up particularly in unventilated power supplies.
https://www.exxosforum.co.uk/atari/ All my hardware guides - mods - games - STOS
https://www.exxosforum.co.uk/atari/store2/ - All my hardware mods for sale - Please help support by making a purchase.
viewtopic.php?f=17&t=1585 Have you done the Mandatory Fixes ?
Just because a lot of people agree on something, doesn't make it a fact. ~exxos ~
People should find solutions to problems, not find problems with solutions.
Post Reply

Return to “PSU (power supplies)”