How about a leaky electrolytic? look for dried out ones, give the cans a quick spin, if it actually rotates it has leaked and dried out.

There aren't any electrolytics in the AC signal path.

Cooling the AD637 causes the displayed voltage to drift but doesn't
clear the calibration error. Power cycling the unit will allow AC
calibration to complete. None of the other parts on the "suspect list"
have any effect.

I've resoldered everything in that path - from the Coto relays through
to the amplifier and RMS-DC converter. No luck - it's still misbehaving.
The readout on the 0.2V setting has dropped 200 counts (0.00200) in the
space of a few minutes. It's reporting "GOOD" for calibration, but I
expect that'll change when it warms up.

There's a Philips 100V / 2.2uF metallised polyester capacitor (C17)
connected between pins 8 and 9 of the AD637 and a second one filtering
the output via a 27k series resistor.

I'm starting to think I'm getting a bit out of my depth... in any case
this is going to have to wait until tomorrow night as I have a to-do
list a mile long and precious few hours left to finish it off!

It looks like most of the circuit sections are linked with passive
components and can be isolated by desoldering one or two leads. Signal
injection and tracing is next on my "try that!" list.

Thanks,
Phil.

On 07/12/12 18:59, m k wrote:

--
Phil.
philpem@philpem.me.uk
http://www.philpem.me.uk/

How stable are the power supplies?

On 08/12/12 07:40, m k wrote:

The analog supplies are fine - the noise level on the +/-15V rails seems
to be on the order of 10 to 20mV pk-pk.
Digital 5V is a little noisy but that's to be expected.

The symptoms have now changed -- the meter is now reporting overload on
all AC voltage and current ranges... I suspect whatever was failing has
now failed completely. :-/

Time to do some signal tracing, I reckon - from the input jacks to the
land of Mordor, where the shadows lie. So to speak...

Thanks,

Phil.
philpem@philpem.me.uk
http://www.philpem.me.uk/

Phil,

However contradictory it may sound, that is actually a good thing.
Continuous
failures are always easier to chase down than your gremlins.

Tracing front to back should bring you to Mount Doom relatively quickly ;-)

Kind regards,
Rob.

Op 13-12-2012 0:10, Philip Pemberton schreef:

On 13/12/12 17:32, Rob Klein wrote:

Indeed!

Well, here's the sum total of this evening's testing:

• Disconnecting LK2 is not a good idea. It appears to connect the A/D
  converter up to the input loop. Disconnecting it kills all the
  measurement ranges (they'll read a constant low voltage).
  I thought it was for disconnecting the drift-correction circuitry from
  the input loop... c'est la vie.

• Lifting one leg of R82 (the 75R input resistor for the RMS-DC
  converter) and injecting 2V pk-pk 400Hz from a sine generator (Zo=50R)
  results in a stable reading of ~0.65V RMS on the 7150 display, the same
  as my Tek DSO reports. The reading is still drifting around a bit.

• All the resistors in the AC input attenuator and opamp feedback loop
  are fine.

• At the junction of RLE, C2 and R10 I'm seeing the input signal with a
  negative-6V DC offset. The output from the opamp is stuck at negative 15V.

This leaves the following suspects:
Relay RLF (an SDS-Relais RS12 metal can relay)
IC18 (DG211 analog switch)
IC15 (LH0062C opamp)
TR12 (WN1001 FET)
TR1 and TR2 (3N163 FET)

I have some LH0062H opamps on order -- the tin can variety, not the DIL
part. Incidentally the 7150Plus PCB is padded out for the DIL part but a
tin-can part has been installed by Solartron and had its leads bent to
fit the DIL footprint. Whatever works, I guess.

In summary: I'm getting closer but I'm not quite there yet.

Adjusting the opamp "offset trim" may be quite interesting -- I don't
think the adjustment procedure is in the service manual!

I'm wondering if it's possible to swap the 3N163s for something else if
they prove difficult to locate. A 3-pin FET wouldn't work (it wouldn't
allow negative Vds like the 3N163 does)... that means it'd probably have
to be an analog switch and some form of level translator.

As for the WN1001 - I can't even find a datasheet for it!
For that matter, the WD460 dual matched FET in the DC loop is also
proving difficult to find even a spec sheet.
I think it's safe to say I'd be pretty well stuffed if the '460 in the
DC amp ever failed!

I did notice a mildly amusing easter-egg on the schematic: the
star-ground on the first floating board schematic is labelled "ROME". As
in "All roads lead to Rome"... Reminds me of the "washer lady" and
"wizard" who were often found in Tektronix service manuals, back when
the schematics were lovingly (?) hand drawn.

Cheers,

Phil.
philpem@philpem.me.uk
http://www.philpem.me.uk/

Adjustment of the offset trim should be trivial - short inputs on an AC range, and adjust offset trim to give lowest possible uV DC reading at TP3.

I just checked the Service Manual and that's exactly what page 3.6 tells you to do (Test 7) - it says to use 20V AC range.

Where did you find LH0062CH at a civilized price!???

Dave
-----Original Message-----
From: volt-nuts-bounces@febo.com [mailto:volt-nuts-bounces@febo.com] On Behalf Of Philip Pemberton
Sent: 14 December 2012 01:44
To: Discussion of precise voltage measurement
Subject: Re: [volt-nuts] Solartron 7150plus - Chasing AC gremlins

On 13/12/12 17:32, Rob Klein wrote:

Indeed!

Well, here's the sum total of this evening's testing:

• Disconnecting LK2 is not a good idea. It appears to connect the A/D converter up to the input loop. Disconnecting it kills all the measurement ranges (they'll read a constant low voltage).
  I thought it was for disconnecting the drift-correction circuitry from the input loop... c'est la vie.

• Lifting one leg of R82 (the 75R input resistor for the RMS-DC
  converter) and injecting 2V pk-pk 400Hz from a sine generator (Zo=50R) results in a stable reading of ~0.65V RMS on the 7150 display, the same as my Tek DSO reports. The reading is still drifting around a bit.

• All the resistors in the AC input attenuator and opamp feedback loop are fine.

• At the junction of RLE, C2 and R10 I'm seeing the input signal with a negative-6V DC offset. The output from the opamp is stuck at negative 15V.

This leaves the following suspects:
Relay RLF (an SDS-Relais RS12 metal can relay)
IC18 (DG211 analog switch)
IC15 (LH0062C opamp)
TR12 (WN1001 FET)
TR1 and TR2 (3N163 FET)

I have some LH0062H opamps on order -- the tin can variety, not the DIL part. Incidentally the 7150Plus PCB is padded out for the DIL part but a tin-can part has been installed by Solartron and had its leads bent to fit the DIL footprint. Whatever works, I guess.

In summary: I'm getting closer but I'm not quite there yet.

Adjusting the opamp "offset trim" may be quite interesting -- I don't think the adjustment procedure is in the service manual!

I'm wondering if it's possible to swap the 3N163s for something else if they prove difficult to locate. A 3-pin FET wouldn't work (it wouldn't allow negative Vds like the 3N163 does)... that means it'd probably have to be an analog switch and some form of level translator.

As for the WN1001 - I can't even find a datasheet for it!
For that matter, the WD460 dual matched FET in the DC loop is also proving difficult to find even a spec sheet.
I think it's safe to say I'd be pretty well stuffed if the '460 in the DC amp ever failed!

I did notice a mildly amusing easter-egg on the schematic: the star-ground on the first floating board schematic is labelled "ROME". As in "All roads lead to Rome"... Reminds me of the "washer lady" and "wizard" who were often found in Tektronix service manuals, back when the schematics were lovingly (?) hand drawn.

Cheers,

Phil.
philpem@philpem.me.uk
http://www.philpem.me.uk/

volt-nuts mailing list -- volt-nuts@febo.com To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/volt-nuts
and follow the instructions there.

Hi,

with several rms-dc converters inside dvm's leads me to find that there is a tradeoff between low level linearity and setting to the lowest levels. check the output at 1% of scale vs the zero offset.

M K