GB GDT-3190 Reads low in all modes...

[knockbill]

It reads .40ADC  + or -   on three different meters I'm familiar with...

[ModemHead]

Oh wow.  That's 400mA which would kill a 9V battery very quickly, and would probably cause something to get hot.

[knockbill]

I'm not sure we were on the same page!!!   I read mA across the battery being used, not in series with the DMM... I thought you wanted to know the current available in the battery which is a dead short across a consumable power supply, explaining the decreasing current readings.

Thinking it thru, seems you want the meter draw current, which now that I've gotten a few lingering business calls out of my head make much more sense!!!!

Here is the accurate current draw thru the GB DMM, set on 20VDC range, test meter in series with battery and GB... 
Meter 1...   .05 on the 20 mA scale
Meter 2...   .05 on the 40 mA scale  (Wavetek meter)

Very sorry for the confusion, trying to accommodate your time/help and dealing with real world BS at the same time doesn't work!!  Got the BS out of the way now, tho!!

[ModemHead]

Yes, a measurement of current in series with the battery and the meter turned on, to see what current it's drawing from the battery.  All I wanted was a confirmation on what you suspected about a "short" (or really an overload) on the power supply of the meter.  But now it appears you're reading 50uA of current draw on the battery, which does not seem like enough to even operate the meter to me.  I would expect it to be around 1 or 2 mA.

For reasons I've explained up-thread, I'm not sure the COB chip is OK.  On the other hand, it is promising that the meter appears to be running and measuring "something", even if it isn't right.

[knockbill]

I agree with you,,, and sorry for not having more knowledge about SS circuits and testing them...
I will continue to try to understand the diagrams you previously posted on the two different HF DMM chips, and see if I can at least confirm that one of them is similar enough to the GB to continue troubleshooting it...
Also,,, as you stated, its curious that the meter seems to make readings on all modes, and they are all low, to a certain %...  I think previous poster Soulman may be on to something with the resistors in the trimmer circuit...  Remember, I tried changing the resistance feeding that ckt originally by replacing  the thermistor with resistors to modify the trimmer voltage,,, it did allow the DMM to read progressively higher but wouldn't lock onto the proper reading...
It can be repaired if the blob chip is useable...

[retiredcaps]

But now it appears you're reading 50uA of current draw on the battery, which does not seem like enough to even operate the meter to me.  I would expect it to be around 1 or 2 mA.

I agree.  The Fluke 70 series lasts 2,000 hours on a 9V battery suggesting a 250uA draw.  Only a handful of meters are in this category of close to 2,000 hours runtime.

At 50uA, that implies the GDT meter will last 10,000 hours.  For a $20 meter?

[ModemHead]

I will continue to try to understand the diagrams you previously posted on the two different HF DMM chips, and see if I can at least confirm that one of them is similar enough to the GB to continue troubleshooting it...

Your photo is too blurry to make out much, but I can see a few things.  The pins for the COB's analog section appear to go across the top.  Most of the bottom pins go to the LCD.  C3, the large film cap, should be the integration cap (Cint).  With the battery out and the GB dial set to any VDC range, you can probably buzz out which COB pins are connected to the COM jack.  The kicker here will be the auto-power-off circuit.  It likely interrupts either the positive or negative supply going to the battery, and may make it hard to find the V+ and/or V- COB pins.

Also,,, as you stated, its curious that the meter seems to make readings on all modes, and they are all low, to a certain %...  I think previous poster Soulman may be on to something with the resistors in the trimmer circuit...  Remember, I tried changing the resistance feeding that ckt originally by replacing  the thermistor with resistors to modify the trimmer voltage,,, it did allow the DMM to read progressively higher but wouldn't lock onto the proper reading...

The PTC thermistor is normally only used in Ohms and Diode-check mode to protect from external voltage being applied at the input jacks.  If in good condition, at room temperature it should have little to no effect on anything.

It can be repaired if the blob chip is useable...

I agree, but I'm not convinced yet that this one is not toasted somehow on the analog side.

[knockbill]

I'll check the top pins against the COM jack and mark them, if Auto Off doesn't hold the ckt open...

[knockbill]

Here's some readings across the top of the blob chip, Left to Right, testing meter set on 200R range, COM lead connected to GB COM jack, POS lead to pins... Another pic of board, maybe clearer?

[ModemHead]

Very good.  Attached is my best guess at the pinout.

You can now do some voltage measurements with the GB meter powered on, with the dial set to any VDC range.
1) V- to V+
2) COM to V+
3) REFLO to REFHI

P.S. 1.7 ohms from COM to V- is very troubling.

[knockbill]

Here's the readings...
V- to V+            6.00VDC
Com to V+         4.93VDC
Ref Lo to Ref Hi .07VDC

[ModemHead]

V- to V+ should be the full battery voltage less whatever small drop appears across the auto-power-off switching device.

The reference voltage is way low, but is derived from the Com to V+ voltage which is way too high.

So none of it makes the slightest bit of sense to me. 

[knockbill]

We figured it wouldn't pan out,,,

Full Batt V was 9.02VDC before starting the tests, it reads 8.55VDC now, indicating to me a major current draw... I checked readings with 3 meters, the Wavetek read ref Voltages at 111+ VDC on auto range, which seemed  rediculous, the other two readings were about the same,,, must be a problem in that meter also, unless the GB is messed up enough to confuse the Wavetek auto range feature...

However, we are basically guessing at the correct pin outs,,, Major lesson here for me was to ignore hamfest DMMs!!! Even the Flukes use hard to find parts...   My eyesight isn't good enough to see the boards let alone the pads and traces!!!  I've had much better luck rebuilding analog test gear,,,

If I run across someone, probably another electrician, that has one of these meters, maybe I can get the readings from it and continue the exercise... Thanks for all the time and effort you put into this, maybe some others got info from it...

[ModemHead]

At the risk of beating a dead horse...   

Full Batt V was 9.02VDC before starting the tests, it reads 8.55VDC now, indicating to me a major current draw... I checked readings with 3 meters, the Wavetek read ref Voltages at 111+ VDC on auto range, which seemed  rediculous, the other two readings were about the same,,, must be a problem in that meter also, unless the GB is messed up enough to confuse the Wavetek auto range feature...

I do agree that probably the current draw on the battery is probably too high, despite the fact that you seemed to measure only 50uA earlier, so that reading is very suspect.  You do need to have reliable test equipment and be familiar with it to get good data for trouble-shooting.  Is it possible the Wavetek was reading 111 *millivolts* and not 111V?  Auto-ranging meters are not that easily "confused". 

However, we are basically guessing at the correct pin outs,,,

I am reasonably confident of the COB pinout, especially if the resistors and capacitors I noted on my drawing actually match the connections on the board, using eyes instead of a slightly fuzzy photo.

Thanks for all the time and effort you put into this, maybe some others got info from it...

I still suspect the COB chip is damaged, but I don't mind going the extra mile to make absolutely sure.  The 1.7 ohm reading you got from V- to analog ground is a big sign that something is terribly wrong.  At this point about the only thing we haven't investigated is the auto-power-off circuit.  I suppose there is a slim possibility a fault in that section could these problems.

[retiredcaps]

Meter 1...   .05 on the 20 mA scale
Meter 2...   .05 on the 40 mA scale  (Wavetek meter)

Hmm, I don't have the Wavetek meter and don't know what your #1 meter is above, but I just noticed that on the Wavetek meter you might/must have to press the yellow button to get DC mA. 

Is your 0.05 reading on AC mA or DC mA?

@OP if you have a DC variable power supply that shows current draw, that is the easiest way to get a reading.  Set your V to 9, and the power supply should show the current draw.

[retiredcaps]

Is it possible the Wavetek was reading 111 *millivolts* and not 111V?  Auto-ranging meters are not that easily "confused". 

Reminds me of a person that I tried to help over at badcaps.net.  OP reported readings of 162V DC on the secondary side of a lcd monitor PSU.  I kindly suggested they measured it incorrectly and wrote detailed instructions on how to do it properly.

OP insisted they did it right.  I said the caps would have all violently exploded with some other things if there was really 162V DC on the secondary side.  OP got pissed at me and decided to "prove" that I was an idiot and posted a picture of the reading.

Sure enough, the multimeter showed 162 ....... mV.  OP negelected to mention the "m" milli.

When I pointed that out, never heard from the OP again on that forum.

[knockbill]

Well, I'm not real confident in the Wavetek I got from a hamfest for so cheap,,, however, I'm not really familiar with it, and  didn't spend the time to digest the whole manual yet,,, best I can do is take a new set of readings and try to photograph them...
Note... the battery lost approx a full volt during testing (4-5 minutes tops??), as some of the readings show a different value...  I used the Wavetek, and did notice the last reading  was in mV, sorry for the confusion,,, guess I'll have to get used to an auto ranging meter, I never was comfortable with them on the job...
I hope these readings are useful,, I double checked them, and they came out close, but declining each time, accounting for the battery load... I don't have a low voltage DC power supply, as I'm normally working with tubes...

2nd pic   V- to V+   7.11VDC
3rd pic   COM to V+ 5.76VDC
4th pic   Ref L to Ref H  129.8mV

[ModemHead]

On auto-ranging meters, just get used to the fact that the units symbol in the display is just as important as the numbers.

--------------
Back to the GB: OK, this is becoming clearer. The unit is obviously drawing excessive current.  Some fault is causing analog ground to be pulled closer to the negative battery as a result.  The question is whether this fault is internal or external to the COB.

To answer the question, I can propose one more exercise with a club and the poor horse carcass.  But only if your eyesight allows!

The COM pin is the fourth one from the left on the top row of COB pins.  It appears to be stitched to the ground plane on the opposite side of the board with a via (the tiny plated-through hole). If you run through that hole with a drill bit that is just barely bigger than the hole it should destroy the copper plating, and thus the connection to the ground plane.  The trick will be to leave enough of the rectangular pad so as not to break the connection from the pad to the COB.

With power applied, check the COM to V+ voltage.  If freeing COM from the ground plane allows it to come up to 3.0V below V+, the COB may not be damaged.  If however, it still reads way high, the COB is beyond repair.

[knockbill]

HA,,,
OK,,, I think I met my match with someone who shares my patience for trouble shooting!!!  I am able to make my own small drill bits and have the proper tools to use them, as I spent a lot of time with an Uncle who was a 1st class watch/clock repairman,,, he left me his tools, and some worth while skills...
   I will give this "trace/pad  cutting/drilling" test a go,, may not get to it till tomorrow tho... I'll make a bit just big enough to ream the plating out of that thru hole...
Thanks for your patience,,, I've been building my 6AS7 PP amp from the test mules into its chassis, for some good Karma, maybe it will rub off on this cheap DMM!!!

[ModemHead]

Well I do like to keep going until I know what the failure is.  The experience may make the next one go faster.

Making your own drill bits must be a nice skill to have!  But before spending too much time, make sure that my assessment of how the pin is connected is correct.  We just want to disconnect it from everything else to see what it does without any influence from anything else...

[knockbill]

Will do,,, I have to disassemble the GB again, I put it back together after the last tests, don't want to lose parts...  I'll study your post, and ream the pad/hole to get rid of the plating,,,  get the reading and post back...

Thanks again, at this point, this is the most expensive 10 buck meter of all!!! May as well see it thru,,, either fix or condemn it, but the learnin'  will be the reward!!!

Now I'm wondering if a HF freeby board will line up with the big display in the GB meter?? It would need the jacks relocated... It never really ends does it!  I've already put a HF meter in the GB holster!

[knockbill]

Results...
Reamed out "rivet" from the opposite side of the board, to not mess up the pad near the chip, I can see the plating is gone half way thru the board, but the 4th pad from top Left still has continuity to the ground plane on the other side,,, maybe a short in the chip?   

Test,,,
Battery has 9.13VDC at start of test, GB set on 20VDC, and turned on... leads connected to COM jack and V+, (7th pin down on Left side of chip)... Reading taken with Wavetek, auto range= 6.00VDC...
I'm about certain I disconnected pin 4 Top from the grd plane on the other side of board, but since it still had continuity the test was invalid, correct?

Sorry for the flash in the pic, it reads 6.00VDC... 2nd pic is the clearest I could get of the reamed hole...

[ModemHead]

To summarize: the idea was to *fully* disconnect the COM pin of the chip from everything else on the board, and then measure the voltage from the COM *pin* (not the COM jack) to the V+ pin, to see if the internal voltage reference would go back to the normal 3.0V.

From looking at the photo, it seemed to me the quickest way to disconnect the pin was to ream the hole (it can be restored easily with a bit of wire and solder.)  I can't see any other copper trace going to that pin in the photo, but you'll have to double-check since you have eyes on it.

If there is still low-ohm continuity from the COM pin to the ground plane and no other copper traces to the pin, then it must be going through the chip internals. A short in the chip as you say.  This is a bad thing, but not too unexpected I guess. I just checked an ICL7106 I have here, the lowest resistance I could measure from COM to any other pin was 4 meg-ohms.

[knockbill]

OK,,, sorry for the confusion,,, the battery dropped a volt and a half as soon as the meter was turned on...  I just measured from the COM pin at the chip to the V+ pin and got 5.55V and declining,,, there must be a short in the chip, as something is dropping V as soon as the meter is turned on, I believe your original suspicion was proved... The chip isn't replaceable in this meter so I guess we figured it out...

I really appreciate all your input, and time with this meter,,,, It fooled me as I am familiar with GB brand, and can't imagine them putting their name on this POS...
I think I have plenty of meters for now,  the Wavetek was a pleasant surprise,  and I also got another CenTech meter about the size of the Wavetek with  many features... I didn't mention it before, but it works very well, and is usable as is...
 I'll try to find the parts for the Flukes as I get the time to re visit them...

Thanks again, I hope some others got some info from this exercise also...

[ModemHead]

I really appreciate all your input, and time with this meter,,,,

No problem, glad to try and help.

It fooled me as I am familiar with GB brand, and can't imagine them putting their name on this POS...

I've posted this before, but I'll post it again for posterity.  If you see a manual-range 2000-count DMM with just two AC voltage ranges, with the lowest one being 200V, you are looking at a very simplistic, low parts count, ICL7106 clone-based design.  It doesn't matter how good the plastic shell looks, or what manufacturer name is on the front, at its core it's the same as the Harbor Freight Freebie.