Thread: ESR Meters.
View Single Post
Old 10th Jan 2019, 3:49 pm   #17
mhennessy
Dekatron
 
Join Date: Jul 2007
Location: Evesham, Worcestershire, UK.
Posts: 4,241
Default Re: ESR Meters.

Quote:
Originally Posted by David G4EBT View Post
ESR meters are a 'one trick pony' and in recent years, have tended to be eclipsed by Chinese 'multi-testers' which are so much more versatile.
I agree that the Chinese multitesters are more versatile, but they are not without issues. As has already been mentioned, the Chinese kits fail with a charged capacitor. OK, you could argue that they are cheap enough to not have to worry about that, but personally, I'd rather not consign the whole unit to landfill and have a replacement shipped all the way from China every time I make a mistake with a charged cap.

These Chinese testers are based on the work of a German engineer - it's a shame he rarely receives any credit for his efforts. I believe that he shared his project for non-commercial use only - I don't know for sure, but would be really surprised if the various Chinese suppliers have approached him for permission to commercialise his work. https://www.mikrocontroller.net/arti...ansistortester

Quote:
Unlike Multi-meters, where there are safety issues to consider, and where some enthusiasts strive for a level of accuracy quite beyond what's called for by hobbyists, those issues don't really apply to ESR meters. It's unnecessary to know the ESR value to three decimal places, and it's easy enough to check the calibration of the meter using close tolerance resistors if in any doubt about the result of a reading.
The Peak Atlas meter does not attempt to read to 3 decimal places; it gives readings to the same number of decimal places as my Chinese multitester.

What you are paying for with the Peak Atlas - apart from supporting a UK company, of course - is the discharge function and the support should something go wrong.

It should be noted that Peak Atlas update the firmware for their products fairly often. For example, they have improved the reading of germanium transistors in light of feedback from customers. You will have relatively old firmware in that Maplin DCA55.

Quote:
As a rough rule of thumb, for capacitors in a Switch Mode Power Supply, (which few of us dabble with)
You might think that, but I really wouldn't want to make assumptions about everyone on this forum. There are a lot of TV enthusiasts here, for example. Even if the set lacks an obvious switched-mode power supply between the mains input and the rest of the set, it will have a line output stage, which is very much a switched-mode supply. Many modern switched mode supplies are "flyback converters" - you might see the connection there

Quote:
Just a final point, many if not most electrolytics on transistor radios such as Roberts, Bush and Hacker from the late 50 - 60s, especially physically small ones, will have dried out long ago, and if you put a cap tester across them, will often look like low value resistors. If the radio seems to be working, there's a temptation to 'let sleeping dogs lie' and leave them in place, especially given that they're on SRBP printed circuit boards where it's easy to damage the tracks when de-soldering. However, in some situations, leaving duff caps in place won't just degrade the sets performance, it can destroy the output transistors and seriously shorten battery life.

So, for aging transistor sets, a multi tester will tell you if an electrolytic cap has a high ESR, high leakage, and if it's still functioning as a capacitor or has morphed into a low value resistor. If it's a blue Philips, you might just as well switch your soldering iron on before the test meter. To be fair, what else can we expect of components that are almost 60 years old?

The last picture shows a 680µF capacitor in the emitter circuit of a Roberts R505 output stage. As can be seen, neither ESR nor capacitance came into it - the tester saw it as a 0.53 Ohm resistor. In other words, a dead short from the emitter of the AC187 to ground. It, and it's companion AC188, were getting too hot to touch but survived the ordeal.
As I've just said in another thread, the blue axial Philips capacitors do sometimes fail - and I've read about your one experience with that R505 before - but they have a near-100% reliability record in Hacker and Roberts radios in my experience. I never change them until I have good reason, and it really frustrates me when people do "blanket changes" on them. That is just misguided prejudice.

Yet experience shows that these same caps were "change on sight" in the CRT monitors I used to repair in the '90s. The lifespan of a capacitor is a product of how it's used, as well as the initial quality. Transistor radios are a very easy environment for a capacitor, unlike a modern power supply or a CRT monitor, to pick two examples. It's all about context...

There are plenty of "change on site" caps out there in older transistor sets, such as the royal blue "Daly" types, or the bright red "Elkomold", to name but two. They are guaranteed to be leaking heavily, although their ESR and capacitance is usually OK.

In cool-running transistor radios, leakage is the usually problem - ESR is normally academic, and so is capacitance, given the wide range of tolerances back then.

Unfortunately, a cap with mild leakage won't show up as such on a multitester. I've just done some testing with a 1000uF cap and a resistance decade box - even with 100 ohms in parallel with the cap, the tester still reports the capacitance value and gives no indication that there is a leakage problem. Whereas the Peak Atlas says "in circuit" - it does that until you get above 1k. Still not ideal, but leakage is a tricky one - it must be measured separately, and I'm not aware of many affordable commercial options for that. A bench power supply and a resistor is the easiest way of doing it, but that obviously takes a bit of setup time.

Electrolytic capacitors failing short-circuit is relatively rare, and can happen at any time. Lack of use is more likely to cause the problem than age or brand. Certainly, there is no data to suggest that Philips axial capacitors are more prone to it than others - I must have seen tens of thousands of them over the years, and can't recall seeing one that had started to leak, let alone go short. They usually sit there gradually going high-ESR - depending on the surrounding temperature and ripple current - and depending on the circuits they are working in, they can continue to function really quite happily in that state. When discussing ESR and its importance or otherwise, it's really important to look at the surrounding circuit conditions. It's often not important.

In practice, it is possible to get by without an ESR meter - I managed for decades - by making observations of the faulty circuit. This requires an understanding of the circuit, and some experience, but it's much more intellectually satisfying that random testing or random parts swapping. It's not always viable for mechanical or commercial reasons, but if I can, that's what I still do. The same arguments apply to semiconductor testers too. But not everyone has the ability to do this, and given the low price of these testers - even the Peak Atlas, in real terms - then why not? It's always nice to be able to confirm the diagnosis.

You won't need one if you're only fixing transistor radios, but for something like a tape machine, which runs slightly warmer than a simple radio, then capacitors will fail, and an ESR meter is a useful way to accelerate the fault-finding process - but remember, it's not the Holy Grail; other testing and thinking will still be required. If you can afford it, get the Peak Atlas. The Chinese kits are better than nothing, but get a couple - and glue a strip of metal to the front panel, near the contacts, to provide a convenient way to discharge caps before testing them.

Sorry for the length of this post.

Mark
mhennessy is offline