Datron/Wavetek 4910 restoration

[ManateeMafia]

I used the same batteries on my 4912.

I had some bad tantalums on the main board so I removed the battery tray and the frame around it. I could replace the parts without removing the main board.

[nikonoid]

I used the same batteries on my 4912.

I had some bad tantalums on the main board so I removed the battery tray and the frame around it. I could replace the parts without removing the main board.

Thank you, MataneeMafia. I like this idea. I will give it a shot tomorrow. I am making a rule not to work on anything very sensitive or high voltage after 1am

If/when you have a chance, can you test if pins 1 though 5 on the back of the unit are all tied together? I am now using the pin numbering from Dayton’s manual below:



PS. I am considering 3D printing plastic brackets for PS-621 to bring them up to size of original batteries.

[ManateeMafia]

I will try and give it a shot. It isn't located in an easy to reach location as my pyramid of 10v references has the Datron at the hardest to reach location. 

I seem to recall all 5 supplies were independent and have separate ground paths. Maybe transit mode switches all grounds together? J601 should also read the same as J103.

[ManateeMafia]

I think your links on the references are causing this. They are the "AVERAGE" links and are probably all connected. If they are, disconnect them and see if this changes for you.
The links tie all LO outputs together.

[nikonoid]

Pyramid of 10v references.... oh my!

I already tried transit switch yesterday: 5 battery channels all had common ground regardless of switch position.

I will try to undo “AVERAGE” links tonight. Thanks.

[nikonoid]

ManateeMafia, you were right on the money with averaging mode links causing all negatives for all 5 battery channels to be connected together. Thank you.

Yesterday I took the unit almost completely apart.
Luckily there is no visible damage to the unit. 3 electrolytic capacitors show minor leakage. I was going to replace them all anyways.







On main PCB there is a large red bank of micro switches. Would anyone know what they are for?

[nikonoid]

Another short update.

I replaced electrolytic capacitors on main board. It was just is time: in few places electrolyte started to eat into the board under the cap, but no traces had been killed yet. One of old capacitors was shorted out (channel 4) it killed this VP0808M transistor.



I will have to find and buy a substitute. For now the cell 4 is out of power.

Meanwhile with new capacitors unit did power up without problems. On power up all lights for cell 4 were off and for cells 1-3 I got:
1) “battery” led solid red (no battery connected)
2) “temp” led solid red and it would change to solid green if I let unit warmup and pressed reset button at the back
3) “average” led is solid green regardless of me putting cell in average mode or not. This is strange.

Voltage outputs on individual cells were all under 0.5v, and readings from average, buffered, 1V and 1.018V did not make any sense at all. Sometimes hung else must be wrong beside blown transistor.

With help of ManateeMafia and Pipelie (thank you, guys), found that DB15 connector that leads into the cell is providing 5V power correctly, but PWM signals on pins 11-14 are missing. I get 5V on pin 14 and 0V on others. There is a high frequency noise there but it is very small (millivolts).

This led me to the module below. Unfortunately I have a hard time opening it, as power pins (red and black wires) seem to be soldered directly into the PCB on inside. Would anyone know if there is a way to open this without desoldering pins? Thanks.

[Pipelie]

Hello, Nikonoid
you have to desolder the pins.
upload a photo  for your reference

[nikonoid]

Thanks. These pins were quite difficult to carefully remove.

Here is my photo to compliment yours.
These is possibly a missing component and also one IC does not look original. So this had possibly been worked on before.

[nikonoid]

And we finally have 10 volts!!!!



Keithley 2001 is not a meter to judge Datron 4910 by. I just had it handy, since I used it while repairing 4910 and it was already warmed up. It may not be very precise, as I calibrated it against another calibrated K2002, but it is pretty good. While not scientifically significant, there is is something very satisfying in seeing six zeros there considering neither K2001 or 4910 were adjusted for the occasion.

Four individual channels read about +/- 1ppm of average. Buffer, 1v and 1.018v all seemed spot on. All in all a good result so far.

These are the parts that were bad:


Capacitor shorted then transistor overloaded and blew up. Independently 8mhz oscillator was bad.

I replaced original oscillator with ACO-8.000MHZ-EK
I hope it is good enough.

There is still more capacitors to be replaced, cosmetic cleaning and battery compartment to work on.

I will be uploading high def photos to XDEVS server soon.

[kj7e]

Nice work!  And I'd say 0.8ppm is pretty dang good.

[TiN]

Oscillator got tired of oscillating and become stable  .
Good job, looking forward for proper measurements.

Perhaps would be interesting to compare cells in differential mode to establish relative stability.

I'd join suggestions to replace all tantalum/electrolytic caps everywhere.

I will be uploading high def photos to XDEVS server soon.

Thanks. I think we can expand unfinished 49xx article to better state, using your feedback as well.

[nikonoid]

Finally all the caps are here including all tantalums. Time to recap.

Meanwhile, in two weeks measuring with 3458a the 4910 had been rock stable: The biggest move I saw was 0.2ppm when ambient temperature changed by 2C. Running autocal on 3458a brought it back to spot on. Otherwise normal noise periodically goes by about 0.05ppm up or down.

I had to power down the 4910 for an hour. When I powered it back on, it was within 0.2ppm of old value in 15 minutes. After 4 hours it got back exactly to the value before power down.

So far I am VEREY IMPRESSED with 4910.


While waiting for caps here is a preliminary design of the 3D printed bracket to convert commonly available batteries to old size.





Special attention had been paid to provide mechanical support compatible with original enclosure, while keeping as much surface area open as possible for proper dissipation of heat.

[CalMachine]

I like the battery solution!  Good idea   

[nikonoid]

I am attaching the STL model for the battery holder that I used. It adopts new SLA batteries PS-621 to the size of the batteries originally used in Datron/Wavetek 4910/4911/4912.

It allows to use original hardware for holding batteries and all original cables, while providing enough thermal air flow to dissipate heat for battery charging. 

[Echo88]

Regarding transit-switch S601 in the Datron 49XX:
Its used to switch the complete device into transit-mode and i cant find any contemporary switch that could be used as a substitute for this switch, is it a special construction? 
Page 123/124 in the service manual shows it: https://xdevs.com/doc/Datron/4910_4911/Datron_Wavetek_4910_4911_sm_service_manual.pdf
https://xdevs.com/doc/Datron/4912/set%201/IMG_1296.jpg Middle left side show the switch-pins: 8 rows a 12pins corresponding to Pins A1-12...H1-H12
Its only capable of switching between two positions: normal and transitmode, by opening and shorting its respective pins A1-A12...H1-12.
An example of this is visible on page 124 top right side: During normal mode the switch-pins f4-e4 is open and f3-e3 is closed, changing to transit mode switches the open/closed switches. This double SPST-scheme is used also to control the power supplies for the cells and the DC-input doubler/tripler.
The BOM lists only NSF as the manufacturer and no real partnumber.
https://www.elma.com/-/media/products/products-files/08_datasheet_e.pdf could the switch be (in theory, just wanting to understand it) be replaced by such a standard configurable rotary switch and a suitable pole-configuration?