Repair Racal Dana 4009 / Thurlby Thandar 1905 DMM

[thobo]

Hi all,

I scored an old Racal Dana 4009 described as defect, showing always 88888888. After some search I found out that it is similar to the Thurlby 1905 and I can only find User- and Service manual for the Thurlby. It looks like that 99% is the same and helped me to understand the functions of the circuits. I investigated first the CPU board after checking the power rails. With my new Rigol DS1045z (yeah, it arrived 2 days ago) I saw a lot of noise an all lines on the CPU board, even on the power rails. I pulled it out and used a externel 5V PSU and first the same picture, always 88888888 and a lot of noise on all lines too. Surprisingly there are only 5 47pF Caps and one 47uF Cap for the power lines, I think thats too less. After more investigation on the scematics i found something like a simple watchdog, triggerd by the IRQ from 6522 and the trigger came every 40ms, to often I think. I "disabled" the watchdog and tata the 6502 booted and showed first "cd 00" and after few seconds "error" on the display. At this point i reassembled the board with the rest and powered it seperately and first the same message came, but it changed sometimes and showed the currently measured value. Sometimes I could measure resistance with intermittent display of "cd 00" (but no "error") and i could use the foil keys to switch to 5 1/2 mode. Then i tried to measure the 5V of the board itself, what seems not to work, the CPU does reset.
Does anyone know what "cd 00" means?
I have read out the eproms, i will post it soon for people interested in.

regards
  Thomas

have added the ROMs

[thobo]

After playing with my new Rigol, I found the root cause for the fault. The watchdog holds the CPU too long in reset and there is too less time for booting before the watchdog bites again. I changed the responsible cap with a bigger one and tata, it works. Seems the message of "cd 00" is the ROM version. I added 2 caps for better noise filtering and now it boots back in case.
But it seems I shoot something yesterday, as i hooked up the inputs to the PSU I used for powering the CPU board.
Now I have to digg deeper into the scematics to find out what may happened.

regards
  Thomas

[macboy]

I would recommend pulling then re-seating each of the socketed ICs on the digital board. Sometimes the contacts on those cheap tinned sockets get corroded enough to cause problems. Re-seating will scratch a nice clean metal surface onto the pins and sockets.

This is a nice meter with powerful math capabilities. The ax+b mode can be used (e.g.) to measure very large currents using a current shunt. Measure the shunt's resistance, then use ax+b to translate the voltage across the shunt into the current through it. It also has a 6.5 digit mode by pressing "*" then "5 1/2". The 3-pin header on the digital board has a 2400 baud serial stream that outputs the display contents on each reading interval. I have connected this to an opto-coupler and into a PC for logging (opto-coupler is necessary).

[Vgkid]

I'll second what macboy said. That was one of the problems with my Solartron.

[thobo]

hi macboy,

thanks for the usefull hints 
One of the first things I have done is to reseat all socketed ICs, but I am afraid that all other ICs on the mainboard are not socketed, as far as I have seen.
Do you have a Racal Dana or the Thulby?

regards
  Thomas

[thobo]

Found another issue, seems the pin 11 of the 74LS93 is gone, should deliver a 250kHz signal, but sits at around 1,5V. Ordered 5 new, its hard to wait for 

[macboy]

The logic is powered by 0 and -5V, so the logic +5V/Vcc supply is connected to DMM negative input, Vss is a -5V regulated supply. So, if you connected the digital section power supply +/- to the input +/-, you shorted out the power supply via the internal circuitry.

To answer your earlier question, I have the Thurlby 1905A, with the true RMS option. I don't use it now that I have more modern and GPIB capable DMMs on my bench.

[thobo]

Just a little step forward, I changed the 74ls93 with a new one and now it seems the ADC start to work. But when I connect a positive Voltage I get allways 0V displayed, but when I input a negativ Voltage I get a Value. It shows a Negativ value, but it seems to 1/50th of real Value.
Does anyone have a hint what that could be?
I have not trimmed any potentiometer, it should show correct values.

First picture shows the Input Voltage (in V), the 2nd the Value what it shows (in mV).

regards
  Thomas

PS: attached the service manual

[thobo]

Finaly got it to work, even the 6522 had a broken output pin, stearing the polarity. Ordered a new one via *bay, but from local not from china, expecting better quality and faster delivery. Got it the next (was great), but it was delivered broken and obvious used. Interestingly the seller claims that it was a new old stock, he never sell used parts. He send new one and tata, it works great. But the second chip had scratches underside, looks like it was removed from a circuit.
Does anyone got a new 40 pin DIP fitting into a socket without bending the pins?
I was used to bend the pins from new DIP parts in order to get them into the socket.

The last fault was hard to find, in the schematic (ADC Logic Interface) has the trace from the Pin 7 Logic Conn. to IC12c Pin 9 a gap, just befor the junction to the other trace. After a lot of measuring and non comprehensible results I checked the schemtic with the existing pcb and found out, that there is a connection, which explains the measured results and leads to the faulty 6522.

Now I have a great DMM 

Btw. on the last pic all DMMs shows a Voltage at 1.91V 

[macboy]

Good job on the repair. I don't think I would have found that fault.

On mine, one of the pair of tantalum caps (seen in your photo under the LF351) went bad, overloading the -15 V power supply and causing false readings. I replaced them with modern low ESR electrolytic ones and a small film cap in parallel for good measure.

Don't forget to turn on the filter to get a more useful stable reading at 5.5 or 6.5 digit.

[thobo]

Thanks

Wasn't easy to find, had to do a deep investigation. On the way I replaced nearly all electrolytics, visiualy they looked to be in a good condition, but after some checks with a greate transistor tester (a very nice gadget, I got in the meantime), which could check a lot of different parts, I measured the caps and most of them were bad.

With Filter you mean this ** +5 1/2 Buttons?

[thobo]

Short update, glued the broken case, now its finished.

[shakalnokturn]

Reviving the topic because I've hopefully finished repairing a Racal-Dana version of this DMM.

A couple of other topics relating to the model:
https://www.eevblog.com/forum/repair/thurlby-1905a-dmm-adc-repair/
https://www.eevblog.com/forum/testgear/thurlby-1905-any-good/


Problems:
Won't work, display shows SW version or checksum on power-up but won't respond to input.
Cracks in the case.
Someone previously had a try at it....

Checking power:
Transformer outputs marked 18-0-18 (AC) on service manual measure 22V, 7V winding measures 7.6V, all a little high.
18-0-18 voltages after diode bridge are +24V and -32V, +15V is good -15V is missing.

Replaced both 78L15 and 79L15 regulators for standard TO220 versions, the PCB was a little too heated under the regulators. (The 7815 fits as is, the holes are too small for the 7915.)
Fitted 470µF 35V for C44/C45 where 2.2mF 50V had previously been squeezed in between the mains socket and top shielding.
Replaced the four tantalum capacitors.
Re-soldered all PS diodes directly after transformer.
Replaced ammeter fuse.

This restored the ability to measure.

Next rewired the transformer for 240V, checked +/-15V power rails, readjusted -5V rail for -5.05V on CPU board.

Readjustment:
First noted all the positions of the adjustments on the lower shielding before re-fitting it. (As a side note: The grounding of the face plate is supposedly made through the upper shielding. In my case, due to plastic washers that may have previously been fitted in the wrong places, the front was not grounded.)

Service manual says to use a specific lower case with holes to for readjustment while allowing the electronics to run at closed-case temperature. I just bunged the sides and did the readjustment with DMM upside-down after 30 min warming.

I used a Fluke 5100B calibrator that needs some readjusting itself and Keithley 5100 DMM as actual reference readout. The Racal 4009 was set to hidden 6.5 digit mode with "filter" enabled.

I didn't fool around with any of the "SOT packs" resistor networks assuming the best possible had been done there in factory and that not being trimmers their drift should be minimal.
Only after going through the full SM's procedure once I started thinking that some of it didn't make much sense. (To me at least.)
Some trimmers are used twice in the sequence: A7 for 20k\$\Omega\$ then for 2kV adjustment, A8 for 200k\$\Omega\$ then for 200V this seems like a waste of time, you go to plenty of trouble getting the 20k and 200k Ohms adjustments spot-on then go and trow them off by making the voltage adjustments.
Going by the way the DMM is designed when adjusting the mentioned ranges one must choose between favoring either voltage or resistance accuracy. (I chose voltage.) either that or it's a headache and hard compromise between the two.

Rather than following the manual's sequence these were my steps:
1: Adjust 2VDC zero using A1
2: Adjust 20VDC zero using A2
3: Adjust 0.2VDC zero using A3 (I found there is a slight interaction with step 2, iteration may be needed)
4: Adjust 2VDC gain @1V input using A4 (ADC Vref. for V/A)
5: Adjust 0.2VDC gain @0.1V input using A5 (x10 buffer amp gain)**
6: Set to 2k Ohm range, adjust for 1mA current output on V/Ohm sockets using PR5 (Not crucial)
7: Set to 2k Ohm range, adjust to 1k Ohm reference using A6 (ADC Vref. for Ohms)

8: ?Best accuracy for V or Ohms measurements? For Ohms goto 13

9: Adjust 20VDC gain @10V input using A10
10: Adjust 200VDC gain @100V input using A8
11: Adjust 2kVDC gain @1kV input using A7
12: Adjust 2VAC gain @1V 50Hz input using A11


13: Set to 20k Ohm range, adjust to 10k Ohm reference using A10
14: Set to 200k Ohm range, adjust to 100k Ohm reference using A8
15: Set to 2M Ohm range, adjust to 1M Ohm reference using A7
13: Set to 20k Ohm range, adjust to 10k Ohm reference using A10