Hi All

So on "BOTH" displays digit 5&6 13&14 are both dead when I turn the machine on.

After about 20 seconds digit 6&14 on BOTH displays is up working, bright and crisp.
After about 90 seconds digit 5&13 on BOTH displays is up working, bright and crisp.

The fault moves with the MPU so not the displays or chassis!

MPU

1J1/6 - Strobe 3 - both Displays
1J1/5 - Strobe 4 - both Displays
1J2/6 - Strobe 11 - both Displays
1J2/5 - Strobe 12 - both Displays

So I replaced U44 no difference and then U51 on the MPU but again no difference. U41 I already replaced to fixed another issue.

The resistor arrays on the MPU SRC 7, 8 & 9 all measure ok.

No acid damage.

I have checked the male header connections and all seem fine.

Once it's working you have to leave it off 2 hours before you get the issue again.

Thoughts? What am I missing?

See pics - PIC 1 TURN ON - PIC 2 - CIRCA 20 SECONDS - PIC 3 CIRCA 30 SECONDS - PIC 4 CIRCA 75 SECONDS (ignore strobe 9 as under internal test mode) - PIC 4 CIRCA 90 SECONDS
20191011_115059 (resized).jpg20191011_115120 (resized).jpg20191011_115211 (resized).jpg20191011_115233 (resized).jpg20191011_115320 (resized).jpg

Question - the resistor arrays are 4.7k 470pf. Can these be replaced with straight 4.7k array's (no capacitance) and use a 9 pin instead of 10 pin???

I'm wondering if one or two of them are not dodgy for that quick period under load until they stabilise. Probably the capacitance side of things

Thanks Grumpy. Have ordered a bunch

Put in new SR7, 8, 9 and no change (just resistor types - no caps)

Jumpered the outputs of the PIA to the inputs of the Decoder and the outputs of the decoder for the offending strobes to the applicable SR array just to ensure it's not a track/trace issue under load = no change

Cleaned the header pins = no change

Tomorrow going to jump from the SR Array to the header pins and again replace the decoder... desperate measures but I really can't see what else is wrong here. The scope measure the outputs on all pins 100% as soon as you turn on yet these offending strobes still need to "warm up" yet the fault is on the CPU as it moves with the board... weird!

Any other ideas???
Displays (resized).png

What is strange is that:

Strobe 3 & 11 warm up together
Strobe 4 & 12 warm up together

There must be a sign in that? I thought it must be an ABCD inputs type (track / trace) thing thus adding the jumpers and replacing the Decoder...

The 74154 table shows there is no state any two strobes should be off at once never mind 4 of them lol ...
74154_function_table (resized).png

Yes reflowed all of them right in the beginning...

I agree - that's why i did the jumper wires just to be sure it wasn't a faulty track under load or something.

There has to be something in this as they are all different outputs on U44 so why do:

Strobe 3 & 11 always warm up together?
Strobe 4 & 12 always warm up together?

will get a pic done

20191023_131922 (resized).jpg

20191023_131931 (resized).jpg

U41 is for the left display segments. They are all working good. U51 is strobes and Right Display Segments. U44 Strobes...

Back on the bench:

Scope connected to strobe 3 & 4 - as soon as your turn on the signals are there. 4.4v from the pull up resistors but the data is there from U44 immediately and yet the displays have to "warm up" but the fault moves with the board (not displays)... it's extremely strange!!!
20191023_141028 (resized).jpg20191023_141036 (resized).jpg

AND the other machine I tested the board in Dr Dude - has LED displays! They need so little current to turn on and yet they had to "warm up" as well...

No - different board. Just using Pool Sharks for testing it

Told you I get the weird and wonderful boards! lol

I'm totally out of ideas. Done everything. Even just upgraded the ROM's to Ver7 just to tick off it's not them in any way...

Got back to this board...now listen to this lol

It's got worse. Now digit 1,2,3,4,9,10,11,12 is out till "warmed up"

So I focused on the "pattern" of them coming back on. I checked the binary counter U29 which gives a "sequence type output" and which connects to the display PIA and thus the display decoder.

It seemed to be measuring fine on the scope. Outputs Q7 Q9 Q3 giving circa 256, 512, 930 (930 a bit out of sequence but same as the other board). But decided to change it anyway but made no difference. So even though all gates giving the right outputs according to inputs decided to shotgun U35, U32, U36 & U14 which all play a part on the out of U29 on the way to the PIA. Made no difference with new IC's so put the original parts back in again.

Was totally confused as that was everything in the display circuit changed and no difference at all.

So decided to experiment.

U35 Pin 8 measuring 918 micro seconds - diagram says 1ms - so within spec...

U29 Disconnect pin 14 Q9 - Board wouldn't boot as IRQ affected - reconnected pin
U29 Disconnect pin 13 Q7 - Displays WORKED FINE!!!!! - U35 Pin 8 measuring 800 micro seconds - reconnected pin
U29 Disconnect pin 12 Q3 - Displays WORKED FINE!!!!! - U35 Pin 8 measuring 600 micro seconds - reconnected pin and displays buggered again unless you leave now about 5 minutes before all are up and running!

So if I take out pin 12 or 13 the displays work immediately - no warm up required! (I have only connected the displays on the CPU so haven't checked if there is an affect else where)

WTF?????

2019-11-09 (resized).png

So the output of U35 also goes to U9 and U9 is sound related so not used on SYS C. So I'm thinking if I disconnect Q7 (effectively disconnecting W14) which then gives 800 micro seconds output (closest to 930us) and displays working... Shouldn't damage/affect anything else???

BUT this also bugs me because I'm basically modifying the board to get the displays to work fine...

Thoughts???