DATEL Voltage Calibrator DVC 8500, Repair [SUCCESS]

[HighVoltage]

I bought this broken DATEL Voltage Calibrator DVC 8500 and just received it.
Starting this thread before I start the repair.

This seems to be a very nice little calibrator with good specs:
+/- 19.999 V
+/- 25ppm accuracy

First thing to notice is a broken mains cable, it is like falling apart to dust.
May be I will upgrade it to a real mains connector.

I did find a nice Datasheet, even with calibration instructions, here:
http://www.calibratorsinc.com/images/DVC-8500_DATA_SHEET.pdf

The lever switches are moving with difficulties.
Does anyone here have experience with these lever switches.
I don't think they can be taken apart.
May be I can clean/grease them from the front, hmm ... will be interesting.

I first need to get a tiny torx to open it.
 

[HighVoltage]

Ok, next step in repair....
Installed a 3 pin mains connector with capacitor across the mains and a new fuse holder.
This looks much better already.
Now comes the electrical output testing.

[HighVoltage]

The output voltage was all over the place and not even one one value was correct.
Obviously the switches got corroded or contaminated or just old.

So, I drowned the complete front end of the calibrator in to pure alcohol and moved the switches probably a hundred times each.
A lot of dirt came out to to the front.
Then I had the alcohol evaporate in an oven for a few minutes at 50 degree C

And then I filled the switches with my trusted "super contact cleaner, with poly phenyl ether" and moved the switches another hundred times each - at least. Then I took compressed air and blew the access lubricant out.

And ... look at this...
All values are working perfectly.
It looks like I do not even need to do a calibration.

But I might have to exchange the potentiometer since it jumps the values a little. But this is an odd little 50k Pot with a tiny shaft.
Ok, time to take some measurements and histogram for each value, to see how stable it is.

Too bad there is no full schematics available for this little unit.
It has a 10.000 V Ref output in the back.
I will take some statistics measurements there as well.

[Kleinstein]

The circuit does not look that complicated. So reverse engineering should be possible, if really needed.
The transformer looks quite large for just a calibrator.

[HighVoltage]

The transformer looks quite large for just a calibrator.

Very true, especially since I am using a 100mA Fast fuse and it holds perfectly.
I will make some measurements on the transformer output to see what the real power is.
 
The potentiometer needs to be exchanged for sure.
But besides this, it works perfectly.
I have it hooked up to my 34461A and will collect output data over night with histogram

[HighVoltage]

kept the calibrator on over night and this is the result.
It was sitting on the bench without housing.

The 34401A shows 20.000V and does not move at all.
The temperature changed by about 3 degrees during the night.

This is not a bad result for such a small calibrator that is so old.

[HighVoltage]

All back together. Installed new foam for the front bezel to not be loose towards the panel.
Looks really good and works extremely well.

Tested in all ranges and all of them agree with my trusted Agilent 34410A.
It is a really nice LM399 based reference.
I will keep it running for a while and take some long term stability readings.

Anyone interested in such a small calibrator, I can only highly recommend getting one.
They are cheap when broken and easily repaired.

[dom0]

The transformer looks quite large for just a calibrator.

Very true, especially since I am using a 100mA Fast fuse and it holds perfectly.
I will make some measurements on the transformer output to see what the real power is.

They specify 10 watts (a bit much, probably only needs a fraction of that). Due to the compact assembly it might be that they used a larger core than necessary and a lower flux than possible to reduce heating of the transformer and stray fields.

[HighVoltage]

I just bought another one of these little calibrators.
This one was sold as dead for $50, so a real bargain if I get it repaired.

And ... the only problem it had was a broken mains cable.
I will also install a mains connector on this one.
But I probably need to attach a big warning label for 120V only, since this one came from the USA

Here are some pictures.
This one looks newer than the previous one.

[HighVoltage]

And some more pictures.

When I am done with the repair, I want to run a few comparison test's between both of the calibrators.

[dr.diesel]

I scored one of these cheap on eBay as well not too long ago, power cord in the same condition!  Neat little device, still on the bench awaiting repair.

[HighVoltage]

Great, that you got one, they are indeed neat little calibrators.

They really put attention to detail:
The white cable that runs to the front power switch is even shielded to the housing.

The second one I got is also working perfectly, just needs a little calibration and is warming up.
I also installed a plug in the back for the power cable and it works and looks really great.
A new fuse is also installed, with a 100 mA fast glass fuse.
Here is a picture of the mod:

[dr.diesel]

^ Nice mod.

From the date codes mine is from around 1988, hopefully nice and thoroughly aged by now. 

[dr.diesel]

Ok, you inspired me.  Cord replaced, but no other changes, so no contact cleaning, no calibration, factory caps etc.

It's now warming up (inside enclosure, front panel 2-wire test) and I'll trend it against my K7510 overnight. 

[dr.diesel]

The low spike at about 1700 was caused by a 3 degree F temp drop when storms rolled in last night.

8500 was set at 10v.

[HighVoltage]

Nice that you got it to work.
Can your K7510 output the statistics of this plot?

When you wiggle the small levers just a little in different positions, is the reading stable?
I had to clean the switches in both of my units.

After I am done with the calibration, I will run both of them against my meters.
With your data in addition, we can get a good uncertainty of these calibrators.

[dr.diesel]

I'm going to clean it up, check/replace the caps etc, then I'll do a proper test with statistics.

[dr.diesel]

Ok, as arrived from eBay, only mod was the cord replacement.

[HighVoltage]

Your calibrator seems to be very stable.
It probably needs a little correction with the dial right in the middle.

Great screen shot of your Keithley 7510.
Very nice that you also get the Pk/Pk measurement in the statistics screen.
I am missing that on my Keysight meters.
 

[HighVoltage]

Here is a measurement of my 120V AC mains unit after about 4600 readings
At 10V setting of the DATEL calibrator I seeing about +/- 10 uV up and down
This unit is not fully calibrated, that is why the dial is turned a little clockwise to be at 10.000 V output

[dr.diesel]

Happen to find a schematic laying around the internet somewhere?

Mine has two 470uf 63v axial caps bodged in on the power supply side, wondering what the originals were?

[dr.diesel]

Either of your units happen to still have the original caps?

[HighVoltage]

Schematic is not available, as far as I have searched.
Your caps look original. Both of my units had the same bodged / extended wires on the negative side.
I will take another picture tonight.

Did you find good replacement caps?

[dr.diesel]

Wow, OK, that surprises me.  My plan was to return it to factory condition for a thorough before modification test run, oh well.   

Since I have two of these in my spare bin, think I'll use them:

https://www.digikey.com/product-search/en?keywords=565-4030-ND

Factory caps measured good at ~80mOhm.

[HighVoltage]

Your Digikey link does not work.

Here are two pictures of my first unit.
This is a 230V transformer EU edition.

[dr.diesel]

Sorry, my bad, fixed above.

[dr.diesel]

Finally got around to replacing the caps and attempting to clean the switches, and all I did was make it worse, 

Last night's run:

308uV - Peak to Peak
56uV - Standard Dev

I used the same method and dunked the face of the instrument in IPA, toggling the switches, got lots of dirt out in the process.  I'd really like to take the switches apart to properly lubricant, but on this unit that would really be a PITA.  I can see over 1mV of deviation just by touching the most significant digit switch handle.

[HighVoltage]

Finally got around to replacing the caps and attempting to clean the switches, and all I did was make it worse, 

Last night's run:

308uV - Peak to Peak
56uV - Standard Dev

I used the same method and dunked the face of the instrument in IPA, toggling the switches, got lots of dirt out in the process.  I'd really like to take the switches apart to properly lubricant, but on this unit that would really be a PITA.  I can see over 1mV of deviation just by touching the most significant digit switch handle.

Great that you got out the dirt with IPA.

But make sure to add some kind of high quality contact lubricant, like I did and then blow the excess with some air pressure.
(Some lubricant will exit the back housing)
After that, you should not see any changes anymore, when you touch any of the switches.

I repeated this process of cleaning and lubrication a few times, until the reading was stable for all switches in all positions.
Now I can touch the switches even harder and the values stay stable and do not move anymore.

[USMC_Spike]

@ High Voltage Can you tell me, when you soaked your front panel
pots in ISO, did you fill a pan with ISO and put it face down in it
and let it soak?

Or did you pour it in from the face?

Same with the MG chemicals too
after you were finished.

@ Dr. Diesel What was the logging meter for lack of a better term that
you used that gave you a read out of the measurements and dates?

Thanks,

Spike

[dr.diesel]

@ Dr. Diesel What was the logging meter for lack of a better term that
you used that gave you a read out of the measurements and dates?

Thanks,

Spike

Keithley 7510.

[HighVoltage]

@ High Voltage Can you tell me, when you soaked your front panel
pots in ISO, did you fill a pan with ISO and put it face down in it
and let it soak?

Or did you pour it in from the face?
Spike

I placed it face down in to the cleaning alcohol until the air had completely escaped.
And then I moved the wheels a lot and I mean a lot until all the dirt was out
In between I blew it out with air pressure

This might sound brutal but it worked very well.
And based on the dirt inside the alcohol, it was very effective

I repeated this procedure with a second unit and had the same success.
Both of them are working perfectly right now.

Good luck


 

[USMC_Spike]

This is good information for me and probably many other people.

I've had one of these for about a year now and it's never worked
properly, nor had I found good readable information about it.

With your postings I'm confident now that I can get the problems
with my unit worked out. 

[Assafl]

This is good information for me and probably many other people.

I've had one of these for about a year now and it's never worked
properly, nor had I found good readable information about it.

With your postings I'm confident now that I can get the problems
with my unit worked out. 

Yes - BTW They do have a data sheet which has calibration information. They do not publish schematics.

They also do exist as a Company (called Calibrators Inc.) and while they don't seem to answer their official email address they answer their phone and do stock spare parts.

I called them and replaced a bad LCD (25$) on a DVC-350a I got for 15$ (listed as-is for parts) on ebay. Altogether I spent about 80$ including shipping to Israel. My device is from the GE days (mid 80s) - Have never seen a company that makes the same electronics for 30 years. I wonder how they source the old parts?

[Messtechniker]

Have never seen a company that makes the same electronics for 30 years. I wonder how they source the old parts?

I have seen special time related manufacturing agreements with subcontractors
(hopefully) guaranteeing supplies for 10, 20 or even 30 years.

[USMC_Spike]

And sometimes with companies, depending on the company, they have their contracted manufacturer make a lot
of parts and they warehouse them.  (expensive).  The audio company Audio Research Corporation comes to mind.
I bought some new replacement knobs for a pre amp that hand not been manufactured since the 70s.

Can you post their phone number I will need two things, the bezel and one of the small vertical holders
for the bezel...someone man handled mine and sheared off the bezel pin at the top along with
cracking the bezel.  Unscrew from the bottom, rotate forward from the bottom and the
bezel rotates and lifts out.

For the first time since I've had the calibrator, it works.  Who every handled it, had one ground
not connected and the ground from the pot floating.

Can you post the phone number? Or did you use the phone number listed on the  data sheet?

I've got to order two pieces.

Thanks,

Spike

[Assafl]

I think you are right and that they actually stock all these very old parts. Quirky QIP (zig zag DIP) multiplexers... Or they buy from Rochester...

The phone number I have is: 401-769-0333

[USMC_Spike]

Thanks Assafl, appreciate the phone number for them.

Also I didn't know they made even smaller calibrators.
The DVC-8500 is pretty small, but I saw pics of the
DVC-350a and that looks to be pocket sized.

Can you post a link to a data sheet for that?

Thanks,

Spike

[Assafl]

Handheld:
http://calibratorsinc.com/images/DVC-350Adatasheet.pdf

Benchtop:
http://calibratorsinc.com/images/DVC-8500_DATA_SHEET.pdf

Top-of-the-line: (but never saw any references to it)
http://calibratorsinc.com/images/PDI_9500_Data_Sheet.pdf

[IconicPCB]

What might the attenuator resistors be ?
specification and supplier?

[HighVoltage]

What might the attenuator resistors be ?
specification and supplier?

They are probably custom made.
I have looked and found nothing, when I repaired my first unit.

DVC-350A
There is one on eBay Germany right now
But it seems to have a broken display and is too expensive
http://www.ebay.de/itm/290660105128
See pictures below

DCV-9500
I have looked for a long time but have never seen one for sale
Well, I have not even seen a real picture of it.
But it has some impressive figures:

Part number DVC 9500
• UNPRECEDENTED Calibration Accuracy < +/-0.0005% error or < +/- 5 PPM
• Less than 3 PPM / deg C temperature drift
• 13 types of Thermocouples output simulations
• Both a +/-20 Volt and +/-2 Volt Range ( programmable bipolar operation )
• Direct TOUCH PAD INTERFACE FOR INTUITIVE HUMAN INTERACTION
• Compact rugged aluminum case that takes up very little precious bench top space
• Digital Output Vernier control down to 1uV steps
• Settable limit controls to protect devices.
• Automatic standby function protects device under test
• Six (6) Manual/Automatic set points per output range
• Programmable Dwell time between set points
• Tellurium-Copper binding posts reduce thermal EMFs to <1uV
• Source 25ma Voltage output with 13 thermocouple ranges

[Assafl]

I got a DVC350 for about 20$ with a totally black display. Verified it worked - and purchased the LCD for 25$. Shipping to Israel doubled the price.... My DVC8500 is stuck in customs for a fortnight. Sorry I have to wait....

As for the DVC350 - it doesn't use a resistor ladder: It uses an AD1403A 2.5V regulator and a CMOS multiplying D/A (an Exar MP7614) to create the output voltage steps. It then uses various op-amps to get the voltage to get the voltage to a 1.2 or 12V range (or for Hex 1 or 10V range), and to operate the current limit. 

Further control of the display is an ICM7211NIPL, an EPROM, and an Intersil CDP1802A CPU and various multiplexers I think from Harris in QIP packages (I could not find them anywhere).

Power supply is a RC4193 (I think when Raytheon still made chips) with a support array of components (as switching power supply circuits were back in the day).

Overall the DVC350 is pretty linear - if quite noisy (my comparison is unfair - a saturated standard cell - a boat anchor but stable as a brick). The last digit drifts quite a bit as well. But I find I use it quite a bit more than I expected. Whenever I suspect a dubious reading I use it to validate my setup is okay. Were it to generate accurate AC it would be even more useful. 

[Arjan Emm]

I also bought a datel 8500 half a year ago. The cord was also really bad, replaced is it first. Switched it on and measured all digits perfectly with my fluke 187 and the cal nob at 0. But after 10 minutes of playing with, it went nuts, voltages are all over the place. It now gives only negative voltages and only roughly close to the settings, 20-30% off. For example +19.999 volt gives -24.8V and -19.999 gives - 24.8V.
The switches feel really smooth.
Any ideas what could be wrong? I find it really hard to troubleshoot a device without a schematic.

[Assafl]

Did you check the LM399 to see that it has the correct voltage?

Also did you check that the shorting bars are well tightened in the back?

When something goes wild like that it usually is a contact or something the blew up (I did have a tantalum give up the magic smoke in the DVC350).

[Arjan Emm]

Thanks, i didn't measure the lm399 it self, but the 10 v reference on the back measures 12.4 V. 
the shorting bars are very tight and the contacts look clean.
There seem to be many tantaliums in there, does it make sense if they fail after 10 minutes before being powered off for probably years?
Edit: The lm399 measures 7.0250V on the fluke, not 6.95.

[Vgkid]

The 7.025v seems good. With that 12v instead of the 10v look into anything involving the reference amplifier. Also how do the voltage  supplies look?

[Assafl]

Received mine, replaced the cord (kept a 110V cord as it can be damaged if plugged into 220V) - the cord disintegrated because of the hot transformer. The rest of the wiring is PTFE so could care less if it is hot... The only other damage is the knob cap (for the potentiometer).

It is right on calibration (much less than 100uv deviation). No output changes touching or wiggling the levers. It does need 10 minutes or so to settle some initial drift. After that the 6th digit (on the HP34401A) is pretty stable at the 10V range +- 1 digit. I'll let it settle overnight.

What's up with not having a tapped voltage transformer for 100, 110 and 220V??? Is it a mil-spec thingy?

[Assafl]

What might the attenuator resistors be ?
specification and supplier?

They are marked R1 R2 R3... Also marking is 2307 (the two large resistors have the same marking). The resistors seem to be potted on one side. So maybe wirewound? The switches are gold flashed levers from Cherry. Potentiometer is a Claro USA.

Chip codes are 86 (the LM399) to 1988 (for the 714 and transistors). So 1988 could be the manufacturing year - albeit - If they keep long stock then chip dates may not reflect manufacturing date...

Since the "big" electrolytic caps from Japan (Nichicon) are end mounted axial, they also have the extensions.

The transformer is mounted on two metal spacers to the PCB - it was completely loose and pushed some of the TO220 sideways (but no damage was done).

Apart from the 110V cable which seemed to have lost the plasticizer (it became somewhat sticky on the outer section and the insulation was falling apart on the inside section - usually means the plasticizer is gone) the rest seems to be almost mil-spec to the letter - I say "almost" as the transformer mount is not mil-spec and the TO220 heatsinks are not anchored, and the end mounted electrolytic caps have 2-sided adhesive mount but would usually need a shock mount. Also not having a manufacturing date code probably means no real mil-spec...

[Assafl]

Thanks, i didn't measure the lm399 it self, but the 10 v reference on the back measures 12.4 V. 
the shorting bars are very tight and the contacts look clean.
There seem to be many tantaliums in there, does it make sense if they fail after 10 minutes before being powered off for probably years?
Edit: The lm399 measures 7.0250V on the fluke, not 6.95.

I just measured - the 10V on my unit measures 10.00004V. It is within the uncertainty of the HP34401A.

[USMC_Spike]

My DVC-8500A is getting better.  It reads close to what it should be after
cleaning everything and then plugging the wireing (grounding) back in.

So I almost have it dialed back to zero...but it wanders about 0.000,010
on my 34401A. 

My main filters are Orange Siemans from 1978 date code. Most of the
resistors on the main board are carbon comps, sadly.  The others are
looking like bulk foil, behind the cherry switches and then along
the side next to cal pots down to the power inlet are other bluk foil
resistors. 

Anyone know about the front panel pot?  How would a precision 10 turn work there?
or a conductive plastic vs. cermet?  I'm almost sorry now that I touched the R46 zero
offset pot....

--Spike

I"ll have to read the brochure again as I'm not to sure how to send a voltage...

I guess for a clean voltage reference, I could use my HP339A and set it for 60 Hz at
the 19.999 voltage? and I'd hook it up where is what I was wondering?

Any guidance appreciated.

Thank,

Spike

[Assafl]

So I almost have it dialed back to zero...but it wanders about 0.000,010
on my 34401A. 

I wouldn't expect it to be accurate enough to match (let alone be better) than the HP34401A. Although it uses the same voltage reference - the LM399 - the HP34401A uses a combination of precision components combined with procedures (e.g. the auto zero) to compensate for the rest of the system. The Datel is good - but I wouldn't expect it to be as good - just look at the specs.

I"ll have to read the brochure again as I'm not to sure how to send a voltage...

I guess for a clean voltage reference, I could use my HP339A and set it for 60 Hz at
the 19.999 voltage? and I'd hook it up where is what I was wondering?

Do you mean for calibration? They suggest using a DC reference voltage and using it with a DVM connected as a null meter to ensure the voltage is identical. So you'd connect the - together (- of Datel and - of DC source together) and connect the + of the Datel to the DVM + and the + of the source to the DVM.

You are basically looking for 0v on the DMM (so that the voltages are identical on the Datel and the reference). The advantage to this scheme, is that if you use your HP34401A - you can use the full precision of the 120mV range (100nV resolution). Which is nice but can drive one nuts as the numbers will flutter about with noise and drift).

If you have a 34401A I am not sure that is needed. Their calibration instruction require 100uV resolution - since you can utilize the long precision 6.5 digit (or better - use the computer link and get about 7 digits) mode that gives you the required resolution at full scale. You may want to use a voltage such as 11.999V instead of 19.999V to take advantage of the higher resolution of the HP34401A at that value (10uV at 12V range and not the 100uV at the 120V range). Of course calibrate the zero at the mV range. Let it warm up at least 15-30 minutes. It does drift a bit at power up.

[HighVoltage]

So I almost have it dialed back to zero...but it wanders about 0.000,010
on my 34401A. 

I wouldn't expect it to be accurate enough to match (let alone be better) than the HP34401A. Although it uses the same voltage reference - the LM399 - the HP34401A uses a combination of precision components combined with procedures (e.g. the auto zero) to compensate for the rest of the system. The Datel is good - but I wouldn't expect it to be as good - just look at the specs.

Very true..
The DCV-8500 is rated a 4 1/2 digit reference.
But then I was also surprised to find out how stable the reference really is.
Here are some measurements, taken over 19 hours with a 34461A
The drift is mostly contributed to the change of temperature during the night.

The short term drift of the DCV-8500 is claimed to be +/- 25ppm and that is about the same as the 34401A
So, one could probably use it as a short term transfer reference for even more than 4 1/2 digits.
 

[Assafl]

HighVoltage - what is the short circuit current (current limited) on your 8500?

Mine only gets to 50mA but the spec says 70mA. So I wonder if the sense resistor (or threshold) is off.

[Assafl]

For those that have issues with their 8500 I measured the voltage on the internal Test points relative the COM terminal.

1. 6.938V
2. 10.000,00V
3. 15.91V*
4. 31.54V*

* Voltage on TP 3 & 4 depends on the output voltage settings. Voltage set to +00000mV.

[USMC_Spike]

Assafl, HV:

Very good stuff.  I then must have misread the instructions on the DVC-8500.

I think your units are better equipped wtih resistors that my unit, they are mostly
I think 1/8 W carbon comp resistors...but the drift I think I saw was more to do
with Temp Co....but still I'd rather not have cc resistors though out.  It is what it is.

I was thinking of using either a 10 turn pot for the 46 pot or a 10 turn pot for the
vineer...I have 5K, 10k, and 20k ten turn pots, just not a 50k, with a smaller
shaft.. I think 3mm. 

Thanks for the explanation, Assafl and the data HV.

I'd like to upgrade to the 34465, or 70...but haven't. 
I also don't have any computer controlled devices yet
and need to figure out how to do it.

Thanks

Spike

[Assafl]

Few mistakes and a semi preliminary circuit analysis:

TP1: 10V to COM
TP2: 6.938V to COM
TP3: 32.65V to 40V supply ground (NOT to COM!)
TP4: Currently 31.54V to COM (when set to +0mV).

There are two power supplies:
Rear supply (transformer leads 3&4): 35VRMS, rectified to ~41VDC gets reduced (two power transistors, zeners and resistors) to 24V for the LM399 heater and to 32.65V (TP3). Use the capacitor - lead to measure TP3.

Forward supply (transformer leads 5&6): 42VRMS, rectified to 53VDC. Not traced these yet.

Bottom right side - a few resistors - only one of which is connected. Why? Is it trimming? Spares (as fuses)? Never seen that before. These resistors seem to work in conjunction with the bottom multiturn - the full range trimmer. So my guess is that entire bottom right circuit is setting the full range.

The 714 underneath the trimmer sets the balance between the +19v and the -19v.
The 714 under the lever switches (to the left of the single turn 0v balance) sets the zero volt point. My guess is it uses the 53v as +-19.999v and a floating zero (as opposed to a dual polarity supply).

[HighVoltage]

HighVoltage - what is the short circuit current (current limited) on your 8500?

Mine only gets to 50mA but the spec says 70mA. So I wonder if the sense resistor (or threshold) is off.

Short circuit current on my two 8500:
First unit (230V) is 51.2 mA
Second unit (110V) is 51.8 mA

So, it seems your sense resistor is alright.
It takes about 1 min to stabilize the short circuit current on both of them.
 

[HighVoltage]

For those that have issues with their 8500 I measured the voltage on the internal Test points relative the COM terminal.

1. 6.938V
2. 10.000,00V
3. 15.91V*
4. 31.54V*

* Voltage on TP 3 & 4 depends on the output voltage settings. Voltage set to +00000mV.

Here are the same values of my two units:

TP1. 7.2850 V
TP2. 9.9999 V
TP3. 15.000 V
TP4. 2.8849 V

TP1. 7.0488 V
TP2. 9.9999 V
TP3. 15.444 V
TP4. 2.9994 V

All settings to 0V and "minus" output polarity selected

[HighVoltage]

Here are some pictures of the resistors of my units

[Assafl]

They are also not connected. Nice.

[Assafl]

HighVoltage - TP 3 must be measured to the ground of the 40V supply.

[Assafl]

Here are some pictures of the resistors of my units

Just saw this again - this time not on the phone. What the heck are they doing here? They seem to pick only one resistor out of the ladder - and not clear for what purpose. The values of the resistors double for each step of the ladder.

However, the resistors are not the same values between the units. Not even close.

Mine are (from the bottom) 155k, 77k, 38k, 19,5k, unknown, 2.12k and 10k. Look next to the resistors and you'll see a wire tying ONE resistor to ground. In your devices it is the bottom 640k resistor (red resistor picture) and it is the 2nd from the bottom 78k in the Blue resistor unit. For my device it is the 5th from the bottom resistor (about 10k but the value is underneath the resistor hence unknown).

My hunch is that these resistors create the 10V out of the 6.9V zener in the LM399. So these resistors will be picked for tempco and - more importantly - for the same tempco (hence always the same type of resistor). Why install 4 unused resistors? Don't know...

I am still a bit unclear how it works. Flipping the +- Switch seems to flip the LM399 from being referenced to COM at the Anode to being referenced to com at the Cathode (i.e. a +6.9v or a -6.9v supply). The 10V reference then multiplies that + or - 6.9v voltage to a + or - 10V Voltage. This same signal (B pin on the lever board) then seems to get the pair of MPSU05 and MPSU55 (NPN and PNP) power transistors (in the power supply section) biased which then seem to set the floating COM voltage at the proper level. To ensure the symmetrical nature of the "flip" - the two uA714 have their Offset adjust connected to the trim pots.

 

 

[HighVoltage]

However, the resistors are not the same values between the units. Not even close.

Mine are (from the bottom) 155k, 77k, 38k, 19,5k, unknown, 2.12k and 10k. Look next to the resistors and you'll see a wire tying ONE resistor to ground. In your devices it is the bottom 640k resistor (red resistor picture) and it is the 2nd from the bottom 78k in the Blue resistor unit. For my device it is the 5th from the bottom resistor (about 10k but the value is underneath the resistor hence unknown).

Yes, I have been wondering about that too.
Especially since they are in parallel connection and have huge difference in value.
May be done to fool people like us?
Would not be the first time I have seen that.
 

[Assafl]

However, the resistors are not the same values between the units. Not even close.

Mine are (from the bottom) 155k, 77k, 38k, 19,5k, unknown, 2.12k and 10k. Look next to the resistors and you'll see a wire tying ONE resistor to ground. In your devices it is the bottom 640k resistor (red resistor picture) and it is the 2nd from the bottom 78k in the Blue resistor unit. For my device it is the 5th from the bottom resistor (about 10k but the value is underneath the resistor hence unknown).

Yes, I have been wondering about that too.
Especially since they are in parallel connection and have huge difference in value.
May be done to fool people like us?
Would not be the first time I have seen that.

Well - they are always the same type and series. So my guess is tempco. Also, it doubles as the adjustment for 10V (which does not have a trim pot) - being wired in parallel with the main 10V/6.93V dividing resistor (which is the 5th resistor) - your DVC8500 needed less trimming than mine - hence the 1st or 2nd resistors were picked - while I had the 4th. My guess is that they solder them in, let them age a bit (or relax from the soldering heat) and then select which one's get as close as possible to 10V.  I think the 19.999V is set by the 7th resistor - also of the same batch/type - but with the Burns 10 turn trimmer (so with compromised tempco).

The 10V is guarded, as is the 'C' pin of the lever switch board. So my hunch is that the 10V reference (out the back) is at least an order of magnitude better than selecting 10V through the front. However, without access to a guarded COM - it is still not as stable as the 34401A's implementation of the LM399 (nor was it designed to be).

The floating/flipping COM (your TP4 measurement was for a "flipped" COM - hence you measured 2.88V for - voltages while I measured 31.54V for + voltages) is I think the main limiting factor in getting to uV accuracies. It will always "float" away from perfect uV level symmetry.

The flipping COM also makes the circuit substantially harder to decipher as even the reference point changes (53V ref, 41V ref, - voltages COM and + voltages COM). Each gives an obviously different reading (that really threw me off the first time I tried to take TP readings). Another obfuscation is that the uA714 OP Amps have pin 5 connected. But pin 5 is an NC pin. So one looks for substrate connections, etc. NADA. Perhaps it was meant to support different opamps (like the 741 which has pin 5 as Null)?

I think the mix of 741 and 714 is due to supply voltage. 741 only allows +- 15V input (and therefore handles the COM voltage) while the 714 allows full +-22V (hence handling the 19.999 range).

Still in the dark on how and where the current limiting happens. The current limit LED is operated by the TO92 transistor next to TP4. It seems to work in conjunction with the TO220 next to the output binding posts.

I admit that this circuit is proving to be taxing on my reverse engineering skills - unlike the DVC-350 which is rather easier to decipher albeit it has 3 boards, a CPU, display and parts that have no available datasheet... C'mon - no ovens to control, no batteries to charge, nada. 3 opamps, a handful of transistors and a strange double sided layout. Geesh!

As for deciphering the lever switch: A - COM; B- (+ or -) 10V; C- Guarded (output?); D- Out Sense(?huh); E,F,G,H- Zener Flipper (COM and 6.934 volts as TP1 and TP2). Pondering over why are they guarding point 'C' (and decoupling it with two huge capacitors)? And why does the lever circuit need the sense terminal to be routed back to it???


   

[wictor]

Hi,

I have been doing some reverse engineering stuff lately. Here is "raw" schematic of Datel 8500 without front panel dividers. Most of the components are placed similarly how they are on the board and that's why it's a bit messy.

[HighVoltage]

Hi,

I have been doing some reverse engineering stuff lately. Here is "raw" schematic of Datel 8500 without front panel dividers. Most of the components are placed similarly how they are on the board and that's why it's a bit messy.

WOW!
Thank you for putting the work in to this. I am traveling and will have a closer look, when I am back in the lab

[Assafl]

Thank you for reverse engineering the 8500. I think you may have reversed the TP1 and TP2 (TP2 is the 10V reference and TP1 is the LM399 reference Voltage + or - 6.9..VDC).

DIV_E, DIV_F, DIV_G and the DIV_H (COM) set the polarity by flipping the anode/cathode of the LM399 between COM  and U1. It uses a crossover configuration switch (2x2 matrix).  This then in turn flips the 10V reference from a -10V (for positive voltages) to +10V (for negative voltages).

This gets amplified by the 714 to the +-19.999V.

Does Q13 miss a link to +V2?

Have you figured out what do all of the transistors on the right do?

[wictor]

Thank you for reverse engineering the 8500. I think you may have reversed the TP1 and TP2 (TP2 is the 10V reference and TP1 is the LM399 reference Voltage + or - 6.9..VDC).

DIV_E, DIV_F, DIV_G and the DIV_H (COM) set the polarity by flipping the anode/cathode of the LM399 between COM  and U1. It uses a crossover configuration switch (2x2 matrix).  This then in turn flips the 10V reference from a -10V (for positive voltages) to +10V (for negative voltages).

This gets amplified by the 714 to the +-19.999V.

Does Q13 miss a link to +V2?

Have you figured out what do all of the transistors on the right do?

Thanks for comments! Maybe I have done too much reversing lately  and reversed some points... I'll check those.

[wictor]

I think you may have reversed the TP1 and TP2 (TP2 is the 10V reference and TP1 is the LM399 reference Voltage + or - 6.9..VDC).

OK, now it starts to go interesting... My board has TP1 and TP2 located as yours, TP2 closer to board edge. But voltages are really reversed. My test point voltages relative to COM:
dials -00000
TP1: 10.000
TP2: 6.888
TP3: 16.201
TP4: 2.9644
dials +00000
TP1: -9.997
TP2: -6.888
TP3: 16.811
TP4: 31.201

[Assafl]

It might be my mistake. I'll check.

[wictor]

I did some cleanup of the schematic. Here is also image of my board.


Reason for repair:

[Assafl]

You  were correct about TP1 and TP2.

So the "extra" resistors are for trimming the 10V reference voltage. Funny that they never seem to use more than 1 trim resistor.

Excellent work reverse engineering & cleaning up the schematic. Makes much more sense now!

A question - how does the output voltage loop close around U3?

 

[Assafl]

Okay - U3 was a puzzle to me. I think I know how it works. So a bit of a circuit description:

The inputs to U3 are Div_C and COM_Sense (two "grounds"). The output of U3 is the output voltage (+ the voltage drop of the buffer transistor).

How does it work?
Div_C is the "wiper" of the divider.
Div_D is the OUT_SENSE line. It has the output voltage on it.
Div_B is the reference voltage: -10V for positive voltages and +10V for negative voltages (opposite polarity).

The ratios of the divider are calculated so that the voltage at Div_C nulls, when the correct output voltage and an opposite polarity reference voltage are applied to the divider. U3 - a uA714 (OP07) is responsible for nulling Div_C by creating the proper output voltage. C6 1uF and C7 0.1uF (the two big metalized film caps connected to Div_C) are there to probably ensure the output reacts fast to dynamic loads (given that the entire system is littered with RC pairs for stability).

As an example, for a selected +10V output, the "Wiper" is centered, and when U3 settles the 10V output (on Div_D) and -10V (on Div_B) cancel each other out.

Also, since U3 tries zero Div_C - it does not really matter if OUT needs to be a source or a sink (the NPN or PNP buffers are active). Of course as long as the current is within compliance (+-60mA or so).

The ability to have the 10V at one polarity and the 19.999V at the opposite polarity is the reason V and V2 are floating relative to each other. Note - Since both are regulated to 33V (minus the BE forward bias voltage of the MJE340) - I don't really understand why the RMS voltage of the V and V2 windings are different. The power supplies consist of a current source (2x 2N2907) driving a 33V Zener (1N4752) buffered by an MJE340 power transistor. V2 power supply is a bit different in that Q7 detects the V2 current flow (mainly U3) and lights the Overload LED.

Also - as per the previous posts, DIV_E, DIV_F, DIV_G and the DIV_H (COM) connect to the polarity lever and set the polarity by flipping the anode/cathode of the LM399 between COM  and U1. It uses a crossover configuration switch (2x2 matrix).  This then in turn flips the 10V reference a negative 10V (for positive output voltages) or to a positive 10V (for negative output  voltages).

Since the 10V and the main divider are the most important in terms of TC - the resistors used are part of matched sets - hence the weird markings on the resistors. (HighVoltage has a red set of resistors from these guys: http://prpinc.com/products/leaded-metal-film-resistors/pr-series-leaded-ultra-precise-metal-film-resistors). I don't know who makes the other style resistors. I also don't understand why the values are always different (and thus requiring a different R32-R37). Perhaps they are trying to match the Tempco of the lever resistors?
I also have no idea who makes the lever divider resistors. They look like PWW resistors, edge sealed with epoxy. The switches are Cherry but I cannot find them in any catalog. 

COM level set: Q3 and Q4 using two diodes D18 and D19 respectively (I think they are ~3V Zeners) - at 3V below TP4 (for negative voltages) and at 2.8V above -V2 for positive voltages. TP1_10V is the trigger that causes Q3 and Q4 to switch the voltage. 

Don't know what the difference between COM and Div_A is (and if the common ground is in the divider - which would make sense).  Div_a and Div_H measure like a short...

[USMC_Spike]

This is all really good stuff Assafl, wictor, HiVoltage.

Wictor, looks like someone was in there changing the
resistors or pulling and replacing a bad one as that
is the first I've seen pics of them out of alignment
with each other on the board.

I thought I posted pics of my unit but I guess I didn't.
School hit also so I am just learning about solid state devices
and the other class is instrumentation and measurement.
The I and M should be a better class but the guy teaching
it seems to like the challenge of figuring out the answers in
his head instead of teaching the process to properly figure it out.
The class doesn't use a text book, it has some self written text
from another teacher.

ON the good side, I found an actual older textbook that teaches it
better than the prof and has better examples too. 

I appreciate the effort from y'all and I'm certain it will benefit
many folks for years to come too.

[HighVoltage]

Wictor
That was a great cleanup of your schematics, Thanks so much.
I will have a closer look with one of my 8500 open.

[wictor]

This is all really good stuff Assafl, wictor, HiVoltage.

Wictor, looks like someone was in there changing the
resistors or pulling and replacing a bad one as that
is the first I've seen pics of them out of alignment
with each other on the board.

Thanks! I don't know about those resistors... The solder joints seems to be original. Anyway, you are right that they are not so nicely aligned, maybe I have monday piece 

Wictor
That was a great cleanup of your schematics, Thanks so much.
I will have a closer look with one of my 8500 open.

Thanks for good pictures and repair tips! I'm getting help from this forum so I'll try to help others too. I'm not an electrical professional, but I wan't learn more and that was also one reason to reverse engineer the unit. When I was doing the schematic, I was doing the layout also to verify all the nets are correct. I'm quite confident that nothing is missing now except the divider itself. But if you find any mistakes, please let me know.

I think Assafi has done the biggest work trying to understand the operation of all the parts. I may publish the whole Kicad project(schematic, pcb, footprints etc), but first I would like to have it "ready" and probably have better names for all the nets and power rails etc.

I have tried to calibrate the unit and the common mode adjustment seems to be quite coarse. Maybe multiturn trimmer would be one improvement for this unit.

Wictor

[Assafl]

I may publish the whole Kicad project(schematic, pcb, footprints etc), but first I would like to have it "ready" and probably have better names for all the nets and power rails etc.

I have tried to calibrate the unit and the common mode adjustment seems to be quite coarse. Maybe multiturn trimmer would be one improvement for this unit.

Wictor

It is meant to be a 4.5 - 5.5 digit calibrator. To get to real 6 digits will need at least better operational amplifiers. There is a tendency to be stuck in an endless loop calibrating this (hence their recommendation to not calibrate it unless absolutely necessary). What I found works well is when you back and forth two pots (e.g. zero and zero balance or range and range balance) to calibrate the pot 3/4 way to desired outcome and then do the other pot. It tends to close in faster on the calibration.

Short term, however, it does seem able to do six digits (I was able to use it to prove my 34401A's reference was noisy using the Datel).

I do have some suggestions for better clarity and naming of the nodes (I didn't want to be pushy - but since you said you are interested in doing this anyway....):

1. Power supplies: There are two 33v power supplies in the Datel. You mark 1 as +V and -V. The second is TP4 and -V2 (+V2 you call the name of the rectified voltage, inconsistent with +V1).
      a. So the suggestion is to call the Emitter of Q12 "+V2".
      b. V1 and V2 are identical (except 1-2 resistor values & the overcurrent detector formed around Q7) - so draw V1 and V2 supplies with identical layout in the schematic - it is easier to read.

2. TP1_10V - I'd call this 10V_ref or Ref_10V or something like that. TP1 is just one thing that is attached to the 10V_ref and is inconsistent.

3. The feedback around U1 (Comprised of inverting voltage TP1_10V through divider network of R32 and R33-R37 & R39 and noninverting voltage DIV_G through R1) is unclear. 3k01 R_3011 is actually misleading where it is as it is a bias resistor for the LM399... 

4. DIV_E-DIV_H - I'd draw the matrix switch. It flips the LM399's zener between COM and DIV_G. (That in turn - via Q3 and Q4 - slide V2 supply to the opposite polarity - all D1-D2-D3 seem to do is ensure the diode is always properly biased). Perhaps Q3 and Q4 deserve a nice area near the 10V buffer.

5. AFAIK - DIV_A is not related to DIV_B DIV_C and DIV_D. It is the same as COM and should be labeled COM. The connection between Div_A and COM happens at the big pad under the transformer.

6. DIV_B C and D should be drawn as a potentiometer so people understand the functionality. Here is how it works:
Between Div_C and Div_D - 8.139K resistor (fixed)
Between Div_B and Div_C - Variable. At 0V - open (infinity). At 10V - 8.139k resistor. At 1mV 89.7k resistor.

Hope this isn't confusing...

-al

 

[Assafl]

More pictures.



Each BCD wheel has 4 resistors and a decade resistor. Two large resistors are the reference 8.139kOhm (R1) resistor and the identical 10V range resistor (R2).

Since the reference 10V is of opposite polarity, at -10V reference the output would have to be 10V for Div_C to be at COM level.

All resistors (except the reference resistor and the three decade resistors at the bottom) tie to X (Div_B) or Y (Div_A) as per the following truth table:

     1 2 4 8
0   Y Y Y Y
1   X Y Y Y
2   Y X Y Y
3   X X Y Y
...
9   X Y Y X

Most significant decade is just connects R2 in series to Div_A or to Div_B.

     1 2 3 4
0   Y .......
1   X .......

 

[wictor]

Thanks Assafl! I'm happy that you have time to clarify things. I'll do some modifications based on your comments, but right now I don't know when... I'm travelling and busy for some time in the future.

[Assafl]

No hurry. At this point 99% is done and the rest is so that it takes less time for others to decipher (and for us in a few years when we inevitably need it).

Or you can send me the diagram and I'll try to learn KiCad.

[worsthorse]

Wictor and Assafl... nice work on your respective 8500s.  I just picked one up and I am wondering if you ever got any further on the documentation project? In any case, all the info here will help a lot with my repair project, thanks for putting it all up here.

[HighVoltage]

I have just my two  DVC8500 and I was surprised how stable it stayed in calibration since the repair.

@worsthorse
What are the symptoms of your faulty DVC8500?
 

[rtekal]

Hi,

Thank you all for the detailed analysis of the DVC 8500. I have a issue with the calibrator output going out of spec beyond +-3 Volts, but is in spec at +-19 Volt. The zero and full scale adjustments are in spec. I noticed that output deviates on the negative side after 9 volts and goes positive gradually till 18 volts. The max deviation is around 2.5 mV.

I suspect the values of the decade resistors have deviated.  Even at zero output setting, the resistors show a parallel circuit values. Has anyone got the values of the decade resistor mounted along the thumb wheel switches? They don't seem to carry any value markings.

[HighVoltage]

Have you cleaned your contacts ?
I don't think the values of the resistors will deviate outside of allowed tolerances over the years.
But small dirt and corrosion on the contacts will have a significant influence.

Best to clean the contacts in such a way as I have shown early in this thread.

[worsthorse]

I have just my two  DVC8500 and I was surprised how stable it stayed in calibration since the repair.

@worsthorse
What are the symptoms of your faulty DVC8500?

Sorry... just saw this. Still awaiting arrival of my 8500. The first one was "lost in warehouse" and never showed up. I just found, and bought, another. Should be here next week. So no known faults. Just want to be able to calibrate and repair it if/when needed. The only documentation I've surfaced on the interwebs is the Calibrators Inc doc with a general description and basic calibration procedure, so I am following this thread with interest.

[mwhamster]

The power supply filters are rated at 450 uF, 63 volts.  I measured the power rails and they are running at 63 and 50 volts.  It is bad practice to run components at or near their rated voltage.  I replaced mine with 100 volt units.  Also check for loose connections on the slip-on jumpers, especially near the power switch.

[Ciffino]

Hi everyone!
I'm Umberto, an Italian new user of this blog.

I want to thanks all of you for informations and electrical diagram I found here. I have one unit of this simple calibrator, and just few days ago it died without a reasonable effort. I suppose problem in a tantalum, but I want to understand how it works before try to repair. The electronic design of the device is not as I thought, I found it "a little" atypical. To help me understand how it works, I will probably redesign Wictor's great and wonderful work to make it more linear and easily intelligible.

see you soon, sincerely,
Umberto

[HighVoltage]

Hello Umberto, welcome to the forum.

Yes, the design of this little unit is very different from normal electronic design.
I hope you can repair your unit well.

Always look for dirt in the switches, as I have shown in this thread.
Sometimes this dirt prevents any output at all.

 

[Ciffino]

I'm analyzing the electronic design ......I really cannot understand if the designer was a genius or a madman...... but probably the first, if is me who are too traditional in my projects.....

in a few time I will post schematic in a different aspect. Now i have to disassemble switches.... Work in progress

[Ciffino]

Update:

i found zener D6 broken. Now I have to understand why it failed ... I replaced it but it heats up a lot, it seems to me that there are other problems ...

mumble.... mumble.....

[Caliaxy]

I got one in working condition - quite impressed! I'm not sure what I am more proud of - my new (to me) voltage calibrator or my meters. 

[HighVoltage]

Hello Caliaxy
Welcome to the DVC 8500 club

Did you have to clean the switches or are they working perfectly.

I bought a second one of these little references and it had corroded / dirty switches
as well as my first one and I had to go through the cleaning procedure again.

[Caliaxy]

Thank you!

The switches seem to work fine. The transitions are "noisier" (show higher over/under-shoots) when switching between some positions than others - like  5<->6 and 6<->7 in the attached screenshot (the steps are 1 mV). I guess this is totally expected, not necessarily a sign of the switch getting dirty and, even if so, really not an issue as long as the output is stable once switched to the desired value.

It looks like mine came from somebody who actually used it and took good care of it. I was expecting to have to do some cleaning, but I don't think I'll bother (if it's not broken...).

I was reluctant to buy such an old voltage reference, but reading your posts here encouraged me to do it. I am happy I did it.

[eDan]

I am curious if anyone is still involved with these DVC-8500?
I am looking for a good, (manufacture's) schematic if anyone has managed to find one.
I have done a bit of reverse engineering in order to come up with a replacement method for the
damn switches. The solution includes the possibility of computer control of the voltage(s).
It is perhaps a fools errand, but if there are enough interested fools around I will continue the work.
cheers Dan

[HighVoltage]

I am curious if anyone is still involved with these DVC-8500?
I am looking for a good, (manufacture's) schematic if anyone has managed to find one.

Only the one by Wictor
See Reply #63