Mac IIcx no chime after board re-cap and startup circuit repair

[phipli]

Yes, I've had that too on an Acorn A3020 (battery leakage)

Upgrading My Acorn A3010 – Stuff & Nonsense

elephantandchicken.co.uk

I know how that feels too. This isn't the worst one that I've worked on, just the only one I've documented.

I'm not 100% confident with the Bomark schematics.

Wise words. They're horrible to read and do contain errors.

I have a dead and a mostly working cx in the house, shout if you're stuck on whether a specific pair of pins are connected, send me the parts and pins and I'll check if they beep through. The non-functional board is just sat on the side in the other room, I wouldn't trust it as glass case because I haven't even started repairing it yet and... well, it has worse damage than yours , but the other one boots and runs (I suspect it has a dry contact somewhere because it seems to crash after a while when it warms up as if a pin lifts as things move a little).

 

[Kai Robinson]

Unfortunately mine isn't socketed so its going to be a pig to get out. 128 pin PGA sockets seem like hen's teeth (or extremely expensive) too.

Peconnectors.com sells them pretty cheaply.

 

[rodders]

OK, I've done a detailed examination of the board (both sides) and while I found a few areas where the solder mask was discoloured, when I scraped it away the trace was good.
I've also buzzed out all of the major components and found one of my repairs was bad (but only a connection to the SCSI socket). I've fixed that but unsurprisingly without curing the boot problem.
I then attached a logic probe and can see the /RESET is asserted for 250ms and the overlay signal from VIA1 being pulled high shortly after it is reset but then nothing.
What is the best way to see if the cpu is doing something? I've tried watching the databus but all lines seem to be pulled high except for a low spike when /RESET gets de-asserted.

 

[phipli]

OK, I've done a detailed examination of the board (both sides) and while I found a few areas where the solder mask was discoloured, when I scraped it away the trace was good.
I've also buzzed out all of the major components and found one of my repairs was bad (but only a connection to the SCSI socket). I've fixed that but unsurprisingly without curing the boot problem.

Good work, it is that hard bit. Tedious and repetitive.

I then attached a logic probe and can see the /RESET is asserted for 250ms and the overlay signal from VIA1 being pulled high shortly after it is reset but then nothing.

Yeah, that sounds like no code is executing. If I have things the right way about, High is the start state, and it should drop low when code starts executing.

What is the best way to see if the cpu is doing something? I've tried watching the databus but all lines seem to be pulled high except for a low spike when /RESET gets de-asserted.

Hum, given the above, we already half have an answer, but what I'd probably do is probe something like A4 and see if there is activity?

Quick question - is your ROM jumper fitted? There is a jumper that selects between internal ROM or ROM SIMM, if it isn't fitted, it will be trying to use a ROM SIMM that isn't fitted.

 

[phipli]

Just double checked, you should have three jumpers fitted on the board total. W1, which enabled the onboard ROM, and W3 with two jumpers, which are always fitted - they're something to do with disabling some CPU lines that were probably used during debugging at Apple.

 

[rodders]

If I have things the right way about, High is the start state, and it should drop low when code starts executing.

I think its the other way round. /RESET is active low so execution should start when it goes high.

Hum, given the above, we already half have an answer, but what I'd probably do is probe something like A4 and see if there is activity?

That was my next step to probe the address lines, but Im pondering what could stop the cpu from starting e.g. a clamped interrupt line.

Just double checked, you should have three jumpers fitted on the board total. W1, which enabled the onboard ROM, and W3 with two jumpers, which are always fitted - they're something to do with disabling some CPU lines that were probably used during debugging at Apple.

All the jumpers are fitted, I even changed the W1 jumper as it seemed a bit loose. I've also tried booting with it off in case I had the logic wrong but no good.

 

[phipli]

I think its the other way round. /RESET is active low so execution should start when it goes high.

I'm pretty sure that the overlay starts high :

...

 

[rodders]

Yes it does and that's what I see. It should be taken low by the Reset handler but that never happens in my case.

 

[phipli]

Yes it does and that's what I see. It should be taken low by the Reset handler but that never happens in my case.

I think we have wires crossed, that's what I was saying. I was using it as evidence that code isn't executing because it never goes low from its reset state of high.

 

[rodders]

One thing that bothers me a bit is the very low resistance between 5v and GND, around 10 ohms.

 

[phipli]

One thing that bothers me a bit is the very low resistance between 5v and GND, around 10 ohms.

I'll measure one of mine, just a mo.

 

[phipli]

One thing that bothers me a bit is the very low resistance between 5v and GND, around 10 ohms.

I get 16R at the hard disk power connector between Gnd and 5v. That is in line with what I've measured on other Macs in the past. They're very low resistance, not sure if it changes when they power on or what. I guess it is only 315mA. 10R is a bit lower, but I'm not sure what the tolerance would be on the board or meters. I was using my not very good meter (the downstairs meter).

Edit, also that was off the battery bombed, unrepaired board. It has just been through the ultrasonic and had caps removed, no repair work.

 

[rodders]

I get 16R at the hard disk power connector between Gnd and 5v. That is in line with what I've measured on other Macs in the past. They're very low resistance, not sure if it changes when they power on or what. I guess it is only 315mA. 10R is a bit lower, but I'm not sure what the tolerance would be on the board or meters. I was using my not very good meter (the downstairs meter).

Edit, also that was off the battery bombed, unrepaired board. It has just been through the ultrasonic and had caps removed, no repair work

Thanks for that. 10R does seem a bit low but that's cheap Chinese ohms!

 

[phipli]

Thanks for that. 10R does seem a bit low but that's cheap Chinese ohms!

Macs can have such a low resistance that some meters will beep continuity between ground and 5v on a working board. It... can make trouble shooting tricky

The first board I repaired I had to get my dad to measure the resistance on his similar machine to check because I couldn't believe how low it was and was scared to power it.

 

[rodders]

Macs can have such a low resistance that some meters will beep continuity between ground and 5v on a working board. It... can make trouble shooting tricky

The first board I repaired I had to get my dad to measure the resistance on his similar machine to check because I couldn't believe how low it was and was scared to power it.

Yes, I get continuity beeps between 5v and GND. It makes finding shorts tricky.

 

[phipli]

Yes, I get continuity beeps between 5v and GND. It makes finding shorts tricky.

Would it help to fit a 5R or 10R resistor to the end of one of the probes? Perhaps just lift it out of the range so you only beeped for real shorts?

 

[phipli]

By the way, did you say you had a logic probe? It might be worth wiring up some of the address bus on a ROM and seeing if you get the expected request for $0. Plus the correct enable signals.

Sorry if you already tried this.

 

[rodders]

Funny you should say that as its what I've just done. At first I was surprised to see that the address wasn't $0 but then I noticed that the select pin (20) was high. Checking GLUE pin 27 it was also high, hmm. However, removing W1 which disconnects the select line from the ROMs I see GLUE pin 27 go low.
It seems that something is clamping the line between W1 and the ROMs. I've checked the pullup, R4 (in circuit) but it seems about right.
Can ROMs go bad?
An alternative approach may be to find a ROM SIMM and try booting from that.

 

[phipli]

Funny you should say that as its what I've just done. At first I was surprised to see that the address wasn't $0 but then I noticed that the select pin (20) was high. Checking GLUE pin 27 it was also high, hmm. However, removing W1 which disconnects the select line from the ROMs I see GLUE pin 27 go low.
It seems that something is clamping the line between W1 and the ROMs. I've checked the pullup, R4 (in circuit) but it seems about right.
Can ROMs go bad?
An alternative approach may be to find a ROM SIMM and try booting from that.

Yeah, ROMs do go bad sometimes. Getting a SIMM is the best answer if you don't mind spending the money. They're a good thing to get anyway because an aftermarket SIMM has a number of advantages...

• They're 32bit clean, which lets you use up to 128MB of RAM without needing software patches.
• They allow you to use up to Mac OS 7.6.1 (but you will need to patch 7.6 and above to do it).
• The have a little built in bootable ROM disk which is perfect for testing a machine (no need to get a working hard disk or floppy disk / drive)

The downside, which you should very much consider, is that most of them disable the RAM test (for speed). This leaves you a little blind when bringing up a machine that has been repaired or in storage for a long time (and has dirty contacts on the SIMMs or whatever).

 

[phipli]

Oh, also, some people sell their old apple ROMs on ebay, any SIMM from a IIx, SE/30, IIsi, IIfx and several other of that era of machines, will work in the IIcx.

Sadly they're soldered in most machines, but sometimes they aren't.