ESR Meter Prompted By David's (G4EBT)

[percival007]

Hi All,

i was prompted to build one of these meters after reading David's Thread here:-
https://www.vintage-radio.net/forum/...ad.php?t=54367
I have started another thread as i didn't want to 'hijack' his.

For some unknown reason (even to me!) i chose one of the designs David rejected! This one:-
http://www.qsl.net/iz7ath/web/02_bre..._esr/index.htm

I have built it and it does work and has helped me fix a JVC DVD Recorder Power supply!

However it has a strange quirk!

If i short-circuit the test leads to 'zero' the meter (the LED lights doing this also, usefully indicating the short) the meter can be correctly zero'd. However if i then put a (good) capacitor of say 680uF across the test leads the meter goes past F.S.D.!!! But, the LED does not light indicating a good cap, not short circuit though. I think this is due to the circuit detecting a DC anomaly when actually presented with a short circuit.

So, the LED part works well and will be a usefull indicator, but i have to use the known good 680uF cap to 'zero' the meter. However, now when the tester is presented with a short, the meter does not quite reach F.S.D. (but the LED does light!). This all i can live with and as i say it has helped me fix a unit.

I wonder though why it is doing this? I have emailed the Italian fellow who posted the circuit but have not received a reply as yet.

All the best,

James.

[David G4EBT]

It looks a neat circuit and does have the advantage (as does the one I built) of not needing a little transformer to be wound, but as I said in my posting, the only reason I rejected that design was that it didn't have a PCB layout.

But then for anyone who doesn't want the hassle of making a PCB, ESR meters aren't critical about the layout, and can be built on matrix board or a tagstrip. Very cheap to build too, and indespensible. Even if you have to buy a meter movement, as I did, it's well under a tenner it total.

For the LED to light up, TR3 needs to be turned on, so if the LED doesn't light up, it means that TR3 is staying off. Why that happens, I've no idea. All ESR merters seem to work on a similar on a similar principle - the application of a low AC voltage at a frequency of about 150kHz to the capacitor being tested.

Might be worth checking the voltage on pin 7 of IC1, and the base of TR3 with the LED lit up, and when it's not lit up. I guess you could try substitiuting the IC and transistors to see if that makes a difference.

An ESR meter soon proves its worth.

Modern physically small capacitors seem to dry out and cause problems - often the reason for dimming of displays of aging microwaves, VHS recorders etc, yet if tested on a capacitance meter, may seem ok, even though they have a high ESR.


Best of luck with it James!

David

[jimmc101]

I think the LED is behaving exactly as it should, it indicates a DC Short.

Without this facility a shorted capacitor would be indistinguishable from a good unit (using the meter alone).

I must learn to read the original post more carefully

It might be worth double checking your circuit, a short across C4 could lead to the symptoms you describe by allowing the DC component across the capacitor under test to add to the meter reading.
Another thought, the LED draws about 10mA; is your battery voltage sagging when the led is ON?

Jim

[percival007]

Thanks for the replies chaps.

David, sorry my explanation wasn't great was it?

The LED part is working correctly, it lights with a dead short and doesn't with a good cap.

However, the meter as a whole reads 'less' (current-wise (500uA F.S.D.)) with a dead short than when a good cap is tested.

For example, with a 680uF cap, i have 'zero'd' the meter (no LED lit=good), then if i place a dead short across the test sockets, the meter reads not quite F.S.D (actually reads 460uA compared with the 500uA of the 680uF cap) but the LED does light (that's a good thing) indicating a short circuit.

Jim, no, no short across C4 and the battery voltage remains a good 9V even when the short circuit test sockets is measured.

I've checked my 'build' again and made sure components are right values.

As i say, i can live with this, it does enable me to test the ESR of caps, it is just a little strange.

Cheers,

James.

[jimmc101]

OK, I think I understand the problem now...

The drive to the bridge is 0, 4.5v.
With nothing connected the voltage at the junction of R9,R10 (Va)= 0, 97mV.
Similarly the voltage at the junction of R11,R12 (Vb)= 0, 97mV.
The differential gain of IC1c and associated resistors is 47 (R15/R13)

Now with the terminals shorted Va = 0 and so the output of IC1c swings 0, +4.55v but the supply voltage is only +4.5v and the output of the TL084 can only get to within about a volt of the supply rail.
In other words IC1c saturates and the meter reading is lower tan it should be.

With a large cap across the terminals Va = 48.5mV DC and so the output of IC1c swings +2.275v, well within its capabilities and thus giving the correct meter reading.

To cure the problem the output of IC1c must be reduced,alternatives are..
a} Reduce gain, increase R13 & 14 to 1k5 or reduce R15,16 to 33k
b} Reduce drive to the bridge increase R9 & 11 to 1k5
c} Reduce output of the bridge decrease R9 & 11 to 15R (this will affect the scale)

Any of these will reduce the swing on IC1c output to +3v,-1.5v max. under all conditions

(I would also suggest adding a diode from pin 8(A) to pin 9(K) of IC1d to give its output an easier time...it's driven hard into saturation when its input is negative.)

Not very well explained, I hope you can make sense of it.

Jim

[percival007]

Jim,

you're a genius!! Many thanks for taking the trouble to look at this for me, that has sorted it perfectly.

I have added the diode too.

I had tried reducing R16 which was starting to address the problem but have put that back to 'as was' and have increased R's 13 & 14 as i had more 1k5 resistors in the pot than anything else!! Doing so has had the desired effect.

If you're ever up this way pop in and i'll buy you a pint!!

All the very best,

James.

[jimmc101]

Glad to help.

(For IC1c to be a true differential amplifier R16/R14 must equal R15/R13)


Why is it that, no matter how many times you look at a post, you can miss even the most obvious mistakes?

The final alternative should of course have read..

c} Reduce output of the bridge decrease R10 & 12 to 15R (this will affect the scale)

The perils of cut and paste

Jim

[David G4EBT]

Sorted!

You can now e-mail to the Italian designer again James to let him know that between yourself and Jim, you've de-bugged his design! Seriously, he might appreciate feedback because it's evident from Jim's erudite diagnosis of the design quirks, that others who have built it - or may wish to in the future - will clearly have the same difficulties.

Maybe he doesn't speak English, in which case (assuming you aren't fluent in Italian!) you can type it in English and convert the text to passable Italian (or any other language come to that, and do the same in reverse) at this link:

http://uk.babelfish.yahoo.com/translate_txt

Now that the ESR meter is tickety boo, you'll have to guard against an irresistable urge to test every cap is sight, if only out of curiosity!

have fun!

David

[percival007]

Hi David,

you're right, i have tested most electrolytics in the workshop already!!!!

I will email the Italian fellow with the design changes, but it was all Jim's technical Know-How that sorted it so i shall give him the credit. Thanks for the link to the Translate page, that may be the reason he didn't/hasn't answered. Although his web-site does have an English option, but this could be 'produced' through a Translater as he may not have any English himself.

Best wishes,

James.

[ex 2 Base]

Originally Posted by David G4EBT

Sorted!

Hello David, I have ready your comments about the ESR meters and wondered if you would send me details of your design as I would like to make one. I am a real novice when using this site so hope I'm not infringing rules in anyway. Ted

[percival007]

I'm sure David will reply but if you follow the link ( https://www.vintage-radio.net/forum/...ad.php?t=54367 ) i mention in my first post on this thread it will take you to David's original thread about the making of his Meter. Full details can be found there and David has included a link to a site which has schematics and full details there.

Best of luck, you will find it invaluable once built!!

James.

[Kat Manton]

Hi,

Originally Posted by jimmc101

c} Reduce output of the bridge decrease R10 & 12 to 15R (this will affect the scale)

When I was evaluating ESR meter designs earlier this year, I built (well, breadboarded) this one. I found it worked better with R10 and R12 reduced (from 22R) to 10R. I think I made a few other tweaks while refining it but can't remember what (and eventually rejected the design anyway.)

Maybe I should've mentioned something at the time? Though, I don't think anyone else here had built one at the time so I didn't bother.

Glad it's now working well enough for you (it just didn't work well enough for me, though with a few tweaks, it's one of the better ones I evaluated.)

I liked the ability to determine whether a cap is s/c without having to perform another test and this feature ended up in my own design. (This was also abandoned as it was taking too long to refine it and turn it into something repeatable.)

Regards, Kat

[jimmc101]

James I owe you an apology...

In my second post I suggested adding a diode between pins 8 & 9 of IC1d to stop the OP-AMP being driven hard into saturation.
I'm afraid this is nonsense, the extra diode discharges C6 back into the output of IC1d and stops the circuit operating as a peak detector.

You will have to remove this diode, sorry.

Jim

[percival007]

Crikey, no need to apologise Jim. I have removed the diode.

Kat, would you reduce the values of R's 10 & 12 as well as the 'mods' Jim has suggested?

(With the diode now removed) the meter seems to work very well indeed. What improvements will the change you suggest make? Slightly more sensitive at the very low Ohmic values? This is the only (slight) disappointment i have, but it still works very well now, thanks to Jim.

I have attached some pics of the meter.

James.

[jimmc101]

As Kat's not around at the moment I'll reply if I may.
The scale of the meter is non linear with no deflection for open circuit, FSD for short circuit and a mid scale reading of 22 ohms (= R10 & 12).
Very similar to an AVO 8 on the ohms divide by 100 range.
This does mean that the minimum discernible resistance is about 0.5 ohm.

Reducing R10 & 12 to 10 ohm will change the mid scale reading to 10 ohms but R13 & 14 would need to be reduced to 820R to maintain the same output.

Can't go much lower without increasing the drive current (by reducing R9 & 11) which will cane the battery.

Jim

[percival007]

Cheers Jim,

i think i will leave well alone now. It is working great (If it's like the AVO 8 then how could it be improved!!).

I think i will need it to fix a bust Power Supply i have at the moment (see another thread of mine).

Thanks for everyone's help on this and especially you Jim.

All the best,

James.

[Kat Manton]

Hi,

Originally Posted by jimmc101

Reducing R10 & 12 to 10 ohm will change the mid scale reading to 10 ohms but R13 & 14 would need to be reduced to 820R to maintain the same output.

This is primarily why I was playing around with resistor values - to attempt to expand the right-hand end of the scale a bit.

Originally Posted by jimmc101

Can't go much lower without increasing the drive current (by reducing R9 & 11) which will cane the battery.

IMO this is the advantage of the designs which use a transformer; a low voltage (to allow in-circuit measurements) at fairly high current is needed. I was using the 'detection/measurement' bits of this circuit with a 'drive' circuit of my own devising which included a 100:1 transformer. Total current drain of the whole thing was less than 10mA.

Regards, Kat

[Radio_Dave]

It looks like somebody else has done some development to this design, including a pcb http://kakopa.com/ESR_meter/ I just wish I found it before attempting to make mine on matrix board!

David

[David G4EBT]

Very well written paper, which gives a good explanation of what ESR is and how an ESR merter works. It emphasises that the ESR of an aging electrolytic capacitor rises much more than does the capacitance, which means that simply measuring the capacitance on a capacitance meter, may, and often does, infer that it's in good condition. EG, if it should be say 8uF and measures as such. However, it may neverthelss have a high ESR which makes it useless. That's why an ESR meter is so useful, especially given that one can be built so cheaply and easily.

What an excellent PCB layout!

Just a couple of fairly obvious possible pitfalsl for the unwary:

1) The PCB layout as printed will need reducing to some 35% to be to scale,

2) As shown, the layout is the reverse of how the track layout needs to be.

That will be obvious when comparing the X-Ray image of the component layout, but I won't be the first to more than once have made a neat, and sometimes complex board, only the find that it's back to front! An elegant but useless monument to my creative incompetence!

If you're unsure, it's always best to print it out onto thin paper, place the track layout over the component layout, hold it up to the light and see which way around the mask needs to be. If anyone wished to use 'iron-on' techniques (eg, a laser printer onto photo paper, then ironed on, and immersed in water till the paper comes off), the layout as ironed on will be the correct orientation.

Hope that helps.

David

[David G4EBT]

The actual size of the PCB at the correct scale is just 1.5" wide, and 1.25" high.

When I copied and pasted the article into MS Word, rather than just printing if off from my browser it came out the correct size. Very neat!

David