Continue to Site

Welcome to our site!

Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

  • Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Making an esr meter.

Status
Not open for further replies.

vlad777

Member
Making an esr meter.


This is the schematics suggested to me in this thread:
https://www.electro-tech-online.com/threads/theory-and-repair-of-lcd-monitor.130015/
https://ludens.cl/Electron/esr/esr.html
**broken link removed**

This is my experimental implementation:

**broken link removed**

Transformer I made is autotransformer with 290 turns with a tap at 15 turns.


I shorted the leads input and set output current with trimpot to 100uA.

0R - 100uA
0R5 - 91uA
1R - 84uA
4R7 - 45uA
10R - 18uA

Measurements are taken with DMM with DC setting at 2mA.
So the values are probably mean value.

Also I measured output voltage at pin one of TL062 and I get 2.55V,
which is mean of square-wave 5V.


My question is about bogus readings when I measure with AC setting on my DMM.
I am attempting to measure output voltage of the transformer to figure out
should I rewind it.

But first to verifie I measured voltage from pin 1 after C2.
With DC setting on my DMM it measures 0 - OK.
With AC setting on my DMM,it measures 6.05 V -NOT OK

Since square wave integral is pi/2 bigger then sine integral,
I tried to do this 6.05*2/pi and it is 3.85 which also looks wrong.

So how should I interpret readings from my DMM when AC and
should I rewound my transformer?

Many thanks.
 
Last edited:
Glad to see the thread inspired you to build one of these.
Your meter might not be reading correctly due to the frequency being measured, not all meters go above audio frequencies.
Not saying your circuit wont work, however the current levels seem quite low, my meter chucks out 1/2 an amp or more with a short, but due to the tx its only around 200mV.
Glad to see a tlo62 not an '082.
I threw my tx together, the turns ratio is well out, and the lay is terrible, still works fine.
Are you sure the tx doesnt have an air gap?
 
Last edited:
Are you sure the tx doesnt have an air gap?

Yes sure, it doesn't.

If you get 500mA (as in 500000uA) then something is definitely wrong...


Edit:

Max output current (trimpot=0) is 0.5mA.

Is mA what you ment?
 
Last edited:
Yup 1/2 an amp,
however I was talking about the test current not the output current to the meter, that would have been ridiculous, but so is a test current through the cap on test as the output is 200mV and even with a dead short there is a 10r resistor to ground so the theoretical max test current would be 20mA.
Draw from the battery is around 30mA.

Now I've thought about it a little more I understand what your question was, the o/p of the circuit is rectified and smoothed, so you need to measure dc current with your meter, I'd just test a few low value resistors and see what values you get, the output isnt linear so you'll have to approximate, a calibrated analogue meter is quicker to read.

Seeing as your still in the prototype stage and you want to use a dmm to view the o/p have a look at this, double check it but I think the o/p is linear so your meter will read directly in ohms, dont think its much good for in circuit testing:

**broken link removed**

I seem to be having many blonde moments, I think it must have been moving house.
 
Last edited:
Hi,

A few things dont look right about this circuit.

1. The gain of U1B looks too high. Small resistances might saturate the output.
2. Cap C6 has no resistor in series with it. The output of op amps are not intended to be used like that.
3. Cap C7 should be smaller, or C6 larger. Only a fraction of the energy from C6 can get to C7.

What it looks like might help:
1. Lower the gain of U1B unless you intend to measure very high ESR values.
2. Place a 100 ohm resistor in series with C6.
3. Try lowering C7 to 0.01uf. Depending on the load with the pot and meter this may not work, and then we could look at raising C6 instead, or else we have to live with a decrease in sensitivity.

This of course assumes that the transformer is working properly.
 
Last edited:
I was looking into gain, because I do get saturation. (I just lower R8.)

TL062 has 50 Ohm resistor on its output (but maybe I should add another.)

If values of C6 and C7 are the same, each should get half the energy (not counting voltage drop on D2).

Why should I increase energy on C7 ?
 
Last edited:
My thoughts agree with you al, i did lower the gain, the 062 can handle a short but its not good practice, i dont know why the diode to ground on the o/p of the op amp is even there as the op amp cant go below ground anyway.
 
I was looking into gain, because I do get saturation. (I just lower R8.)

TL062 has 50 Ohm resistor on its output (but maybe I should add another.)

If values of C6 and C7 are the same, each should get half the energy (not counting voltage drop on D2).

Why should I increase energy on C7 ?


Hi,

More energy transferred to the output the more output voltage per unit ESR resistance, thus the sensitivity is increased. This isnt a critical issue though like the gain, which is more critical.

For the gain, here is the basic analysis..

The square wave generator generates a wave from 0v to +5v.
The transformer divides that down by a factor of 20, to 0.25v, or plus and minus 0.125v.
Given a low resistance we get all that across the second 10 ohm resistor, so that's plus and minus 0.125v.
The op amp U1B is biased at 2.5v, and the power supply is 5v, so we've got less than 2.5v headroom on the output.
That op amp gain however is 39, and 39 times 0.125 equals 4.875, so it becomes obvious that somebody made a mistake on the gain
because 4.875+2.5=7.375 and that voltage exceeds the op amp positive voltage supply so the op amp will saturate.
Furthermore, since the resistance of the ESR and second 10 ohm resistor reflects to the output as a simple ratio, given an ESR of 0.1 ohms we'll still see 99 percent of that output (given a higher supply voltage) at the output so that's still 7.3 volts which is still way too high.

Given the op amp output characteristics we'd have to look up the max output. Then we could decided what the gain should be, although there may be some variation in the transformer operation which attenuates part of the signal so maybe just short the ESR and adjust the gain such that the output of the op amp is within range of it's ability.

There could also be a problem using a TL062 op amp because the output swing may not make it up or down high enough even with lower gain. You'll have to do a test for linearity, but a rail to rail op amp would work better, or else increase the supply voltage to say 20v.
The output swing data for that op amp isnt very explicit so you'll have to test it, but be aware that if you build another one of these testers it may not work even though the first one works.
 
Last edited:
be aware that if you build another one of these testers it may not work even though the first one works.
Apart from constructional errors, why would this be?
 
Hi there alec,

Because of the variation in manufacturing of the part that is evident from reading the data sheet. One part may be able to swing somewhat close to the supply rails while the next package of the same exact part number may be 0.5v lower, and that could make the difference. It could be even worse than that but that's just an example. The real problem is the power supply voltage is too low for this op amp part. The voltage swing isnt guaranteed to accurate standards so may vary from part to part.
 
One part may be able to swing somewhat close to the supply rails while the next package of the same exact part number may be 0.5v lower, and that could make the difference.
Even allowing for component tolerances there doesn't seem to be anything super-critical in the circuit? The calibration process should take care of the tolerances. I agree the voltage supply is on the low side, which leads to lower resolution than would be possible with, say, 12V. But is an accurate measure of ESR really necessary at hobby level? I would have thought a rough 'good or bad' indication was probably sufficient for many practical purposes.
 
Hi,

The kind of variation we are talking about here is critical to the operation of the circuit, not just a simple tolerance issue really, unless you consider 500 percent tolerances to be within spec.

When this thing should be measuring 0.1 ohms ESR, it might look like 0.5 ohms ESR, and everything between 0.1 ohm to 0.5 ohms would look like 0.5 ohms. That's because if the output saturates we loose linearity and that cant be tolerated. If we get lucky, the transformer will be so imperfect that it will itself attenuate and help to prevent early saturation. But if not, we have to lower the gain, but that means a loss off range which means less accuracy too. This is especially bothersome to me because a simple increase in supply voltage would help to fix this problem without any other changes.

We havent looked into the effect of slew rate and bandwidth on the results either, which would be interesting seeing as how the square wave is coming in at almost 90kHz and the bandwidth of the op amp is only 1MHz and the original design calls for a gain of almost 40, which right there tells me this op amp is not a good choice or else something else has to change. Im betting that with a better part and/or supply voltage this thing might work pretty well. But without some changes it may not work at all giving bad results.

We can go over these particulars i guess just to see what effect they have but it would take a little time to do.
 
Last edited:
Status
Not open for further replies.

Latest threads

New Articles From Microcontroller Tips

Back
Top