Compucorp Sumlock 320G Scientist Calculator

[ChristianFletcher]

Hello Loyd

Yes I think it could be patched but I’m more worried about internal short circuits between tracks. I going to bell it out tonight but also look at the keyboard scanning on a scope. It may just respond to clean but I’m starting to find witness marks on the boards so could be that it’s all be tried before. But I enjoy experimenting.

[ChristianFletcher]

I have been taking a closer look at the calculator today. I cleaned all the edge connectors reassembled everything and it’s actually doing less the when I started. The edge connectors do appear quite corroded even after cleaning.

Taking a look at the power supply board the high voltage generator is working as is some of the logic. Looking at the power supply logic on the power supply board has to be working to run the high voltage inverter. But I am only measuring 3.6 volts across the logic IC pins 7 and 14 which I am assuming is wrong as I was expecting to see 5. The Vcc on pin 14 of the TTL logic is connected directly to the battery. The Battery- connector doesn’t connect to the Gnd of the TTL logic so I’m thinking it must employ some form of low side regulator. But if it does I haven’t identified it. Any idea ?

The information from Tony has been very useful but I’m still struggling to figure it out etc

Thanks Regards Chris

[ortek_service]

Looking at Tony's schematics, Yes it's a rather complicated design compared to more modern circuitry, with everything referenced to effectively a positive-ground (Vss) that might be conventionally +5V or Vcc.

And took me a while to try and see what's happening - and still aven't quite figured it all out - but after some looking through, I noticed that:

Battery positive connects to Vss, so Batt- is around -4.8V
And with no charger connected, Chgr- is shorted to Batt- via the charger jack (otherwise it connects via a diode and a paralleled transistor + few ohms resistors, to control charger current)

A chassis-ground symbol has been used for what has been called ground (and was measured as -4.8V relative to Battery Positive (Vss).
- Which is rather similar to what would be expected across 4 NiCad cells, so I had expected chassis-ground to connect to Batt- / Chgr- at some point to complete the circuit.
The only place I can see where these are connected is via the 'Regulator' circuit's two paralleled-at-DC 2N5193 transistors - which are biased on via the preset's Vsense voltage and the 2N4403 / 2N4401 drivers, which are supplied from Vss and biased by Chgr- (& Vgg for regulation?).
I'm not sure why the connection is via extra windings on the PSU transformer, but I think this may be to provide a boost to the voltage, once PSU is running (with transistors providing active rectification) or some kind of extra regulation feedback.

I haven't found any marked places where supplies go to the logic I's , but I presume Vss connects to their top-right (Vcc / Pin14 etc) and chassis-ground connects to their bottom-left (Gnd / Pin7 etc).

I also couldn't see what the logic IC types were - It could be they pre-date 74 series TTL, and are some older DTL etc. that may well run a bit further below 5V.

So I would first check what voltage is across the batteries, to make sure they are all OK - I have known some vintage programmable calculators that used the battery as a reservoir as PSU (Charger) alone wasn't enough to deliver the peak-currents for its magnetic-strip feed motor.
And then maybe measure all the other supply rails, on the testpoint connector etc. to see if they match those Tony had noted.
I did note that the preset maybe looked a bit dirty, and maybe worth checking its setting / effect of adjustment (noting where it was originally set at)

[ChristianFletcher]

The Logic family is definitely before my time as is 74L rather than 74LS. I have the power supply set to 5 volts and is being used as the battery. I’m not understanding how the inverter can be running if the logic doesn’t have a enough supply as the inverter is driven by the logic. Yes Pin 14 is the + battery voltage.

I haven’t found the preset had any effect unless I set it to the far end when the the TTL voltage suddenly jumps from 3.6 to 6 volts at the far end of the travel

Thanks Chris

[TonyDuell]

It is not at all obvious.... The bit to look at is page 6 of the .pdf

The clock logic must be running for the power supply to work since that provides the PSUDrv signal. But the power supply also powers the logic.

Yes, the TTL logic ics are connected between vss (named for the pmos logic in the rest of the machine) which is battery +ve and ground (the on-its-side 'E' symbol).

The trick is the pnp transistors at the bottom left of the transformer in that sheet. At power-on they are biased on, so the battery is effectively connected straight to that TTL logic. It starts up, the power supply is driven.

Now I _think_ that it's like most SMPSUs. If it's lightly loaded then some outputs will be low. And ground turns out to end up as an output. Have you tried it with the other boards fitted?

[ortek_service]

Originally Posted by ChristianFletcher

The Logic family is definitely before my time as is 74L rather than 74LS. I have the power supply set to 5 volts and is being used as the battery. I’m not understanding how the inverter can be running if the logic doesn’t have a enough supply as the inverter is driven by the logic. Yes Pin 14 is the + battery voltage.

I haven’t found the preset had any effect unless I set it to the far end when the the TTL voltage suddenly jumps from 3.6 to 6 volts at the far end of the travel

Thanks Chris

Well at least it is still TTL - and 74L (like 74S) was a quite short-lived variant of the original 74 series, before 74LS mostly combined the advantages of the Low-power and Schottky versions (although L version does still have some advantages as well as disadvantages over LS) - with 5V (+/- 10%) Vcc.

That does seem rather suspicious, suddenly jumping to 6V at the end of the preset and would indicate that the DC-DC converter is running at that point, if supply is being boosted. And could well indicate that preset track has gone open-circuit.
So it's worth measuring it is continuously-adjusting in its resistance setting / Vsense voltage on the wiper (so probably measuring that whilst adjusting). Probably want to be careful if TTL supply jumps to 6V as that higher than max. operating voltage (but should be just about below absolute max. at least)

[ChristianFletcher]

Ok thanks you have given me something to think about. I was making the measurement with the other boards disconnected so I will try the measurement again. The logic must be working to some extent as it’s responding erratically to keyboard inputs and printing E for I assume error. I also note if i turn the voltage down the whatever garbage is displayed starts to flash. So again I’m guessing that’s the low voltage detect working.

I spent a lot of time cleaning the edge connectors as they had gone green on many of the board edge fingers I’m not convinced about the backplane connection as every time I reassemble it it does something different.

Just had a thought perhaps it needs batteries rather than wiring back to my power supply. If this was a modern circuit I would suspect decoupling issues etc

[TonyDuell]

The 'backplane' is a series of plastic blocks with spring contacts held together with long screws. They connect the top fingers of one board to the bottom fingers of the board above it, so if a signal needs to go to several boards it's passed on by each board in the stack.

You can take it all apart to clean the contacts.

There are some photos of a Compucorp 324 here :

https://www.***********/photos/tony_d...57633917908897

which show how the keyboard, etc, comes apart.

[ChristianFletcher]

Thanks Tony. Those pictures are are great. I have taken it apart and cleaned it but the board edges are pitted so the haven’t come out very well.

I think I’m going to have to take a step back and say I’m having trouble understanding how to measure the supply voltages on the Pmos ICs as I don’t understand what to reference.

If I start by looking at the most easily accessible board which is the Top D-Scan board the TCL 02 I can see pin 1 is VSS, pin 21 is Vdd and pin 22 is vgg. But I don’t find a ground pin ?

I think you mention in your notes that the measurement are referenced to VSS.


If I put the black common of my voltmeter to pin 1 VSS and then red to Vgg I measure - 8.2 volts

If I put the black common of my voltmeter to pin 1 VSS and the red to Vgg I measured -16.6 volts.

I think I’m having a bit of a mental block with this Pmos logic I have to say I always been similar confused with the Germanic love for positive earth lines in grundig radio schematic. Any hints on how to get my head around this much appreciated.

Again I want to look at some of the waveform on a scope what do I use as the reference. I’m running off a power supply at the moment but I have the ground link lifted on the calculator so it’s floating. So in theory I can use any reference.

I figured out what was wrong with the voltage adjustment pot and it has paint on it so the wiper goes open circuit when adjusted to one en. The voltage must jump to 6 when it hits the paint. It’s not actually possible to adjust in circuit as it has a board over the top of the control.

Thanks Chris

[ChristianFletcher]

The Story so far https://youtu.be/2BGoPf0jSAs

[ChristianFletcher]

Sorry some errors in my previous posts from keyboard scanner IC TCL02

If I put the black common of my voltmeter to pin 1 VSS and then red to pin 21 Vdd I measure -7.6 volts

If I put the black common of my voltmeter to pin 1 VSS and the red to pin 22 Vgg I measured -15.5 volts.

From Tony's information Vdd -15V and Vgg -30V Off load

This is about half what Tony has measured which makes me think I am using the wrong reference point ?

I have 5.26 Volts on the TTL across pins 7 and 14 -Thats measured on the edge diagnostic connector.

[ortek_service]

Well it's good to see you've now got the TTL IC's functioning and their supply back up to within their operating range. I think some of this has to be running, to provide clock drive to the switching part of the DC-DC PSU.
Although I'm not sure why you could previously only get 3.6V, with no adjustment (except for going to 6V, when preset wiper was o/c due to paint going off the track).
Can the 5.26V you now get, be adjusted a bit closer to 5V? (Although Tony had measure this a bit low, at just over 4.8V)?

It seem Tony measured everything to the Vss 'Positive Earth' supply rail
So if you connect your voltmeter's red lead to one of the TTL IC's +V pins (14 etc) - or a more-convenient point, after checking it really does connect to this, , and just measure everything else with the black lead, then that should read correct.

It's probably also best to connect 'scope ground to this Vss line, as if you connect it to -5V circuit 'chassis ground', then you don't know exact difference from Vss, if it varies and hence total supply to other circuitry that is fed from Vss and the other negative rails and doesn't use the 'chassis ground' -5V rail.

Also, can you just run the PSU / logic IC's board, without the other ones fitted, to see what the rail voltages are?
Tony had measured these off load, so maybe Vdd and Vgg do vary quite a lot depending on loading (although it seems Vgg connects to one of the transistors that are controlling the regulator, which controls negative supply voltage to the DC-DC PSU's switch, so should form a feedback control loop on that)