GW Instek GDM-8034 bench DMM: looking for schematic or calibration notes

[dmmartindale]

I recently purchased a used GDM-8034 bench DMM.  It's a basic DMM with 2000-count LCD display, manual ranging, and 0.5% basic DC accuracy.  It seems to be fully working, and the functions I've been able to test seem to still be in calibration (though I only have an even older Fluke 4.5 digit meter as a reference).  Not bad for $25.

However, it has a slight DC offset error.  For example, measuring a 0.250 V reference on the 2 V range, it reads +0.251 or -0.249 depending on how I connect the leads.  I'd like to trim out this offset if I can.  I have downloaded the manual for it from the Instek web site, but it seems that the manual includes just end-user instructions - nothing about how to calibrate it, no circuit diagram, not even a board layout that identifies what the adjustment pots and cap actually control.

This is in contrast to the manual for the older GDM-8135, which has tables for performance checks and calibration, and a drawing of which calibration controls are where. Now, the GDM-8034 is obviously simplified somewhat compared to the GDM-8135.  The 8135 has 14 calibration adjustments (11 pots and 3 trimmer caps) while the 8034 has only 5 adjustments (4 pots, 1 cap).  But I don't want to go changing the adjustments blindly while trying to figure out what they do, particularly when its calibration seems OK at the moment and I don't have the procedure to recalibrate it properly.

So I would appreciate it if anyone can send me (or point to) a schematic for the 8034 (which would let me figure out what the adjustments do.  Or, even better, the real calibration procedure for this meter.  Note that I do have the 15-page users' manual already, but I need something beyond that to start changing calibration adjustments.

The adjustment points on the 8034 are:
VR401: left side near U401
VC401: near U402 and J401, near left end of U606
VR501: left edge of board, behind input jacks
VR502: adjacent to U501 and near VR501
VR601: between U606 (large chip) and R515 (power resistor)

[Poroit]

The accuracy for the DC Volts ranges on that model are:
+/- 0.5% of reading + 1 digit.

It is in spec.

[dmmartindale]

Yes, I know that the one count error is within the published spec.  And there may not be any way to adjust this anyway; the GDM-8135 does not have any adjustment for this.  But I would like to adjust out this offset if it is possible, and having a circuit diagram would tell me if there is an adjustment.

Plus, having a circuit diagram and/or calibration procedure is useful to keep the meter running in the long term, particularly if some component eventually needs to be replaced.

[coromonadalix]

have you tried at GWInstek ??

other options  is to measure  the trimpots values and try fiddle them ...  witch is a last course of action ...


or trace down  witch pots go the each ic's,  search ic's data sheets and try to figure out the circuit ...

since this is a basic 0.5% meter 3 1/2 digits,  should not be hard ...

[nctnico]

I'd leave it as is. Likely the offset will vary with temperature & humidty as well so trying to adjust will be futile. Fiddling with the adjustments is more prone to breaking something than making an improvement.

[J-R]

Can you post some internal photos?

I say tear into this DMM and have fun.  It makes no sense to try using it given the epic limitations:  2,000 count, manual ranging only, not true RMS, you must use the holes in the front to measure capacitance (can't use the test leads) and even then it can only measure up to 19.99uF...  Then there is the iffy LCD.  So it just doesn't have any redeeming qualities IMHO.

A fun mod could be to add some batteries and a backlight to the display.

[dmmartindale]

Main circuit board from above, front panel on the left:


Left side of the board, where most of the pots and the one trimcap are located:

[coromonadalix]

ok  an classic intersil 7106

you can find many schematics on the web about it, you should understand a bit

vc401 is a frequency compensation, i would never touch it

as other wrotes  it's a very basic meter,  how far do you want to let it be that way   or you still have the urge to fiddle the trimpots ...     i would not play in it ...  too much mechanical switchs etc ...

you seem to expect a lot, out of this meter for 25$ ...

only 2 caps in the psu would need to be replaced if the meter is used 24/24  ...  if not   let it be

[Poroit]

Keep it as a back up and invest in a 5-1/2 digit or better.

[J-R]

I see it doesn't have fuse for the 20A input, so avoid mistakes...

As mentioned by coromonadalix this appears to be a popular DMM IC and I see lots of DIY projects using it with various schematics and pinouts available going back maybe 40 years.  So you should be able to figure this out by doing some research and tracing this specific PCB layout.

I think it's safe to say there is going to be only one DCV adjustment, probably on the 200mV range.  There may or may not be a zero adjust.  The other ranges are going to be fixed based on the input divider resistors.  So making this perfect may not be possible, and not necessary as it's in spec, but fine for tinkering.  If you're able to fully document the adjustments, it could help others down the road.

[floobydust]

I don't think it's calibration to blame. 
OP I think your issue is called "rollover error" which is a difference in reading between +ve and -ve values. It should be less than one count. Check Reference capacitor C603, it should be a 0.5-1uF part. 0.1uF is a bit small and cheap or maybe it is bad with leakage current. Beyond this, for the ICL family, a "rollover resistor" and diode can be added in series between INTegrator OUT and analog COM. Like 1N4148 (cathode to INT OUT) and 100k to analog COM at ICL7106 pin 27. I don't see it on your board. It's in the old Harris/Intersil datasheets.

The multimeter is likely the same front-end as the LED version GW-8035, and GW Instek did white label under a few other brands like Metermaster, Amprobe, Wavetek etc. in your search for a service manual or schematic. It's just hard to guess their model numbers, use image search maybe.
Failing that, it is not hard to figure out the calibration trimpots by tracing out some of the circuit.

[mqsaharan]

... for the ICL family, a "rollover resistor" and diode can be added in series between INTegrator OUT and analog COM. Like 1N4148 (cathode to INT OUT) and 100k to analog COM at ICL7106 pin 27. I don't see it on your board. It's in the old Harris/Intersil datasheets.

Hi floobydust,
Could you please upload the old Harris/Intersil datasheet here that discusses/presents this solution. I downloaded a few old datasheet from different sites but couldn't find it in any of them.
Thanks.

[floobydust]

I can't find the one Intersil datasheet or app note with schematic for it I found last night. It's mentioned as text in others. Was it here... long, slow download over 1,000 pages:
https://usermanual.wiki/Document/1981IntersilDataBook.1668550145/view
http://www.bitsavers.org/components/harris/1991_DB301.1_Harris_Data_Acquisition.pdf

OP on second thought I think you have no rollover problem. "... measuring a 0.250 V reference on the 2 V range, it reads +0.251 or -0.249 ..."
You are looking at the bottom of the 2V range.
Def'n: "Rollover Error -difference in reading for equal positive and negative voltage near full-scale." It's usually 0.5-1 count as a spec.
Intersil attributes problems there due to leakage currents on a dirty PCB or an integrating or ref capacitor with higher dielectric absorption.

[Kleinstein]

The roll over error has in the ICL71xx has mainly 3 components. One is an offset, e.g. from input bias and protection resistors or the divider. A 2nd is the ref. capacitor from leakage and pump out from parasitic capacitance - so rel. high capacitance in a not too large case is wanted. Today possibly even an X7R of maybe some 2.2 µF. The voltage is constant and DA is thus not an issue.
The final part is from the difference in the direction for the comparator. The 4.5 digit version 7135 has some hint's on this with an suggested adjustment for the steps around 0. For the 3.5 digit version this was less of an issue. While the error is still part of the turn over the bigger issue is the DNL around +-0.

It is a very basic meter and chances are it will work for quite some time with no issues. If there is an accuracy problem this would be more like a defect (e.g. old electrolytic capacitor, cold solder joint, dirty PCB or spiders home under the chip). In this case a calibration alone would not fix the issue, but it would need a full repair. Chances are one would than ditch the meter or than look for a schematics or do a reverse engeniering with the help of the typical circuits. Nothing to really worry now as the meter may fail in different ways, like failing display, blow up from a lightning stike, get lost in a flood or fire or get smashed beyond repair.

[coromonadalix]

some sheets


basic circuit,  woth to read
https://electronics-diy.com/ICL7107_volt_meter.php

[floobydust]

Towards calibrating the GDM-8034, I would need clearer pictures and a pic of the PCB bottom to trace out things, the DIP IC markings.
It is generally one trimpot to cal 7106 input, looking like VR601 for 200mV full-scale but I will instead cal the 2V range with it depending on how good the divider resistors are.
One trimpot for Current function (shunt amplifier). Is R606 a shunt?
One trimpot for Ohms source (maybe in conjunction with Q501, it will connect to PTC R517).
One trimpot for AC converter (plus trimmer cap for freq. response).
The beeper I'm assuming the  threshold is fixed.
The "300Hz Point" I guess it has a RC oscillator for something? A bit low for a beep.

Make sure U101 (maybe 7815) does not run too hot and drift. Not sure if it needs a heatsink, the tab is discoloured.
I find it strange they went LCD display it could easily run off batteries. Note these do have a little capacitance to PE ground, the cap by the mains fuse C504?.

[dmmartindale]

Some quick replies for now. I can likely look more closely on the weekend.

- I like the idea of treating this as a secondary meter.  I recently acquired a couple of 4.5 digit DMMs (Fluke 8600A and 8050A) one of which will likely become my main bench meter.

- I looked at the LCD.  It sits pretty close to the display board, connected to it via a pair of rubber zebra connectors.  I don't think there's any place to mount a LED backlight behind it without adding a spacer, which would require a source for longer zebra strips.  On the other hand, a couple of LEDs as a side light looks plausible.

- I found a couple of 7106 datasheets.  That should be enough to trace circuits functionally.  (Though that might have to wait until I retire).

- Yes, 0.25 V and 2.5 V are at the low end of their respective ranges.  That's because the voltage reference I happen to have is based on a 2.5 V LM4132, which is pretty good for 3200-count and 4000-count meters, but forces going up a scale on this 2000-count meter.  However, I've also tried switching input polarity with inputs of about 190 mv, 1.9 V, and 19.9V (near the top of the 3 ranges) and the difference between positive and negative readings is usually 1 count there.  The 7106 datasheet specifies a rollover error limit of +- 1 count so it's within that spec.

- Yes, the 20 A input has no fuse.  I think this was once common; I also have a couple of handheld meters with 10 A or 20 A inputs and no fuse.  Most of the time, I use this to measure AC line current, so there is a 15 A breaker upstream somewhere, but I understand that the unfused input is a real risk with an unfused high-current source (e.g. car battery).  The 200 mA/2 A current input is fused; there is a 5 x 20 mm fuse inside the blocky-looking red input jack housing.

- I found a photo of the interior of GW-8035F, and it seems to be an older generation of instrument.  It has nothing larger than a couple of 16-pin DIP ICs in it - no 40-pin DMM building block chip.  The analog signal circuitry is quite different too.

- On the other hand, the 8034's circuit board has a bunch of unpopulated locations near the right edge (bottom of the first photo): U603 and U604, plus R113-R120.  And there's a 7107 meter chip which has the same functions as the 7106, but its outputs are designed for driving LEDs instead of a LCD.  So I wouldn't be surprised to find a version of this meter with LED display.

- 300 Hz is the frequency used for capacitance measurements in this meter.  Thus the test point labelled "300 Hz POINT" is probably part of the capacitance to voltage converter.  The capacitor jacks are connected to the main board by 4-pin J401.  So VC401 right next to it could be part of the capacitor measurement, or it could be a high-frequency adjustment for the AC converter.

- U101 (main voltage regulator) is a 7808 regulator.  It may look a little discoloured, but it was only slightly warm to touch when I checked it.  The whole meter draws only 0.9 W/1.3 VA when operating.  Of that, 0.6-0.7 W and 1.1 VA is drawn when the power switch is turned off.  The power switch is electrically between the transformer secondary and the rectifier diodes, so the transformer draws magnetizing current even when the meter power switch is off.