# Problems with building an ESR meter, part II

Discussion in 'Electronic Basics' started by Simoc, Jul 26, 2006.

1. ### SimocGuest

Sorry for reposting, so I posted earlier a thread "Problems with
building an ESR meter",

http://www.ludens.cl/Electron/esr/esr.html )
Well, now I measured those and got:
Output: 2.65 V
- Input: 2.65 V
+ Input: 2.48 V
Vcc: 4.95 V

So it seems to be oscillating, although I'm a little confused about
that all of them are a little above 1/2 Vcc...but might that just be an
error of DMM in measuring average of pulsating DC?
On DC range, DMM read just 0 mV...but is it AC or DC across the R6?
On AC range, the DMM read 8 mV. Across the R5 I got the same readings.

And btw, if it's AC voltage across the R6, does the DMM give any
reasonable readings, as it's high frequncy square wave? (as AFAIK
usually
the AC range of DMM's is optimized for 50/60Hz sine wave)

How could I continue troubleshooting now?

Simoc

2. ### petrus bitbyterGuest

Well,

The readings of the oscillator fall in the expected range. You may try the
to measure the AC-voltage at the C2/T1 junction. The reading will not be
accurate due to both frequency and waveform but I expect you to find some
2Vac. If you find only some mV at this point I'd become suspicious as to
measure only 50/60Hz hum.

You'll never find any DC across R6 (unless you made a grave wiring error,
have a defective cap or some other serious howler). This part of the circuit
only has very small AC voltages and the 8mV you found, is in the expected
range.

The amplifier, U1B, should have Vcc, so 4.95V as the op-amp share the power
pins. Pin 5 should be at Vcc/2. Take or leave one, maybe two, tenth of a
Volt depending on the accuracy of R9 and R10. For measuring the output
voltage of this amplifier short the test leads and temporarely disconnect C6
at the diode side. Measure between the loose end of the cap and ground.
Should be some 300mV I guess. If no voltage, most likely your amplifier has
gone, but there are still possibilities like bad solder joints, shorts or
other defective components.

After reconnecting C6, you can measure the DC across C7. Should be some
hundreds of mV.

petrus bitbyter

3. ### SimocGuest

Thanks very much for your info. I measured there, and there wasn't any
voltage.
But one fault found...the opamp was bad now! As I took it apart and
wired up for
testing, I got readings of 2V - 3.3V on outputs and inputs when nothing
of them
was connected. So maybe they were just steady DC voltages that I
measured on output
and inputs of the oscillator section of the opamp. When I connected Vcc
to plus
input or gnd to minus input, the output went high, like it should, BUT
when I
connected Vcc to minus input or gnd to plus input, the output went high
too!

Btw, you advised to measure AC voltage at the C2/T1 junction, but isn't
it just pulsating
DC voltage there?

I haven't more TL062 chips, nor I have a component store nearby, so
it'd take less
or more time to get a new one. But I have a working TLC272CP in my junk
box. Would
it work on this circuit? Although probably there is still some other
mystic fault
in the circuit as earlier when I was attempting to troubleshoot it, the
opamp was ok
when I tested it then.
BTW, my DMM is true-RMS, so does that mean that it should give correct
of the effective value regardlessly to the waveform? Although when I
took a look on its manual,
there was rated a frequency range on AC voltage measurement up to
1kHz...but do you think
that it's readings will be reasonable enough to measure these 50kHz
voltages? (although
it of course won't be as accurate)
But like John said, there should be about 200mV according to the text
on the Homo Ludens homepage.

Thanks.
Simoc

4. ### petrus bitbyterGuest

No. Have a look at the schematic. A DC current, pulsed or not, should come
from some place and go to another. Neither the capacitor nor the tranformer
will let through DC. (That's to say you can't get DC from primary to
secundary in the transformer.)
Don't know. Download the datasheets of both op-amps and compare. You can
give it a try anyway.
Can't be sure. The true-RMS may be a disadvantage in this case as it may
contain electronics that does not work at 50kHz at all. A dumb analog meter
using only a rectifier and adjusted for 50-60Hz sine wave will give some
reading up to in the MHz range. (At least mine does.) Totaly inaccurate but
enough to find out whether or not a 4MHz X-tal oscillator works.
200mVdc across R6? I'll eat my hat if someone can make me clear how that has
been done. More likely you'll find it across C7 although I'd expect somewhat
more. Look at the schematic again. Test leads shorted, no battery involved.
Where should the DC current come from to cause that voltage? Capacitors will
not conduct DC so as far as DC is concerned, the circuit ends at C2 resp.
C3.
petrus bitbyter

5. ### SimocGuest

Pulsed DC goes through a transformer. Just steady doesn't. In some
simple SMPSs there is just a pulsed DC fed to the primary. As the opamp
oscillator outputs just pulsed dc (although it gets connected to gnd
when not voltage there), is it the cap which makes it ac, by
discharging itself through the opamp output when it's low? Took a time
to realize that Am I right now?
Yes, but I don't know (I haven't skills to tell from the schematic
that what are the requirements of the opamp in this case. Do you know
how to tell? Then I could check them from the datasheets. Those opamps
have, of course, many differences.
Yes, but then I'm unable to tell if the opamp is incompatible or if
there is other fault in the circuit :-( Is there some way then to find
out by measurements if the inopereability is due to an unsuitable
opamp?
Thanks for the info. I haven't an analog multimeter, but maybe I could
build an ACV meter from a panel meter and rectifier, sometime when i've
time.
I didn't say DC! (Sorry I quoted too much, that messed up on you)
It's AC, of course, but I just meant to say that it should be 200mV, so
8mV is far too low, unlike you stated...

Simoc

6. ### petrus bitbyterGuest

Yes and no. Pulsed DC is not only steady DC but can be considered to have a
"pure" DC part and an AC part. The AC part can pass through a capacitor or a
transformer, the DC part can't. So you'll never find a DC voltage on the
C2/T1 junction. Only AC.

At first glance the TLC looks like good enough though I can't go too deep
into the datasheets at the moment. But you can try it out without problems I
think. Best instrument for looking at these circuits is an o'scope, alas I
guess you have none available But... You can listen! Replace C1 by a 10nF
cap and you can hear the 5kHz by an ordinary cheap earphone. You only may
need to use a series resistor of somewhere between 100 Ohm and 10k to avoid
blowing it. If everything works fine, you can go ultrasone again. Remind,
haste is an enemy. Make sure the power supply is ok. Then, build only the
oscillator, including C2. If you can hear your oscillator, build the Test
leads part, including C3. Sound will be week but usless your earphone is
very insensitive, you will hear it on C3 (test leads shorted.) Next part is
the amplifier followed by the rectifier. Don't forget to replace C1 before
you start ESR measurements.
The oscillator produces some 5Vpp/50kHz triangle. It has an RMS value of
about 1.5V (AC part only, so at the C2/T1 junction.) The transformer ratio
is 20:1 so at the T1/R5 junction I'd expect a 75mVrms which will become
lower when loaded with a 5 Ohm resistance. Measured with a multimeter that's
usually adjusted for 50/60Hz sine, an 8mV looked like acceptable. But I can
of course be mistaken. I guess the 200mV mentioned in the Homo Ludens
homepage is the peak-peak value as seen on an o'scope.
petrus bitbyter

7. ### SimocGuest

Yes, I meant that in flyback-type switchings PSs and regulators, pulsed
DC is used there, having "only the AC-component" being still DC as its
polarity never gets reversed. But that's off-topic since I realized
that the cap, by charging and discharging in turns, makes it "true" AC
(going in reverse direction, too). Was I right?
Agreed, if my cap-making-it-AC -theory was right (So was it?)
I could look the datasheet myself, but the problem is just that the
requirements of this circuit aren't told there ;-)
Correct guess
Oh, I'm impressed! Very nice trick, I have to try it, thanks.
Isn't it square?
Maybe...

Simoc

8. ### petrus bitbyterGuest

You're right for the AC going through the capacitor.

You're right for the square as well. The triangle is on the minus input of
U1A. Which makes my calculations worthless. Hope you nevertheless got the
idea.

petrus bitbyter