Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: PetterM on December 07, 2024, 03:37:47 PM
-
Hi,
I'm trying to fix a really nice looking Rev.6a motherboard (512k), no corrosion damage. :o First power on, only give me yellow screen.
Taken out all chips, and give them a round with ipa and deoxide, all looks like new. Still yellow screen.
Measured all voltage, cpu got good clock, reset, and halt stays high.
Don't know the history of this board, but caps will be replaced later, un-soldered a few, and they measured good.
Inserted Diagrom 1.3, but it only give me a strange color output (attaching a short video) (don't have a serial cable yet)
I hope some of you know what this it, I cant find any documentation explaining this.
So now i'm stuck. :-[ Please help me bring back this lovely board to life.
My guessing is that the first bank of chipram is bad, so kickstart/diagrom can't execute code, but that just a hunch. :)
(Chips and sockets are in the mail)
(Sorry for any bad english, norwegian is my preferred language) :)
Petter
-
Yellow screen usually denotes bad RAM. One trick is to piggy back each RAM chip with a good one to find the culprit or culprits.
Good luck.
-
Thanks, that's what I'm hoping for, but it should give me a Green screen.
Yellow is more like "I really don't know whats wrong" message, so it can actually be anything.
I was hoping that anyone knows what the flashing screen output from diagrom 1.3 means. But i cant find a documentations, other than:
"No Chipmem, Fastmem detected:"
Flashing bars on screen when chars is printed
From https://github.com/ChuckyGang/DiagROM/blob/master/DiagRom%20Codes.txt
Could it be that easy?
But then, it could be the Fat Agnus, or some of the memory logic chips, i don't know.
I have to wait for the mailman, bringing me memory chips and sockets. :)
-
My understanding is that a yellow screen from the Kickstart ROM indicates a problem with the CPU. Does your 500 have any trapdoor or edge expansions installed? Check the trapdoor pins for corrosion as well. Have a close look at the JP (jumper) points on the board, as they might show previous modifications.
You'll definitely want to see what DiagROM is outputting via serial, if that works - it's vastly more informative!
-
My understanding is that a yellow screen from the Kickstart ROM indicates a problem with the CPU.
Indeed it does, CPU exception. Green normally points to RAM issue.
-
Hi, just wondering if you've made any progress with your A500, and in particular if you've been able to check the serial output from DiagROM. I looked at your video and I saw a similar colour pattern from DiagROM at one stage while I was troubleshooting my A500, and DiagROM was often reporting CPU exceptions such as AddressError and IllegalInstruction. The cause was one bad RAM chip, though the machine also had a bad CIA B and keyboard issues.
I think it is plausible that memory issues could be causing the Kickstart yellow screen as well, if it results in a CPU exception at boot time.
It's not very well documented, but DiagROM has some useful operating modes controlled by different mouse button combinations at power-on (note that you need to release the buttons before DiagROM checks for stuck buttons). RMB tells it to use the beginning of memory rather than the end, and LMB tells it to use Fast RAM if available (note that no video output is possible in this mode, so you will need the serial connection). You can also combine the two. I found I could run DiagROM without crashes if I installed a trapdoor RAM expansion and ran it from Fast RAM.
-
Hi, sorry for the delay. I have made "some" progress. After i got the serial cable, it was clear the no chip mem was detected ok, and that values written to one address was found on another address. So playing around with the scope, i could see that address line 4 (A4) on the chip ram was pulled low, and if i measured resistance between ground and A4, it was 68ohm (same as resistor pack 201), so its almost shorted to ground.
After snipping out U34 (74F244), i no longer measure 68ohm on address line 4.
My hope is that U34, somehow got a internal short, that pulled down that address line.
So now i waiting for mailman to come with some 74F244. :)
Does this sound even remotely possible, or are my hopes to high? ::)
-
Hi, new update.
Found a broken trace from resistor pack 202 to u34, that give low activity on A0 (adressline).
Fixed that, and now available chipram is 128kb. Huge improvement from 0.
So now diagrom shows this: (se attachment)
Can anyone "translate" this error to me?
-
Indeed it does, CPU exception. Green normally points to RAM issue.
I’m always getting those 2 mixed up…. 🤷🏻♂️ Oh well… No Sheldon’s in to help? Never mind 😆
-
Can anyone "translate" this error to me?
It's showing that the data bus is likely working OK, but test data written to the DRAMs was found at unexpected addresses. That means that you're either got an address bus fault, or one or more DRAMs is bad. Given the previous issues you've seen with one of the address bus lines, I'd suspect the former.
To check the address bus, use the scope to look at the column/row data which should appear all of the time on all 9 address lines, A0 - A8, on each DRAM. See page 4 of the schematic:
http://amiga.serveftp.net/Schematics/A500_schematics/A500_R6_schematic.pdf
Expect to see continuous groups of 4 x low pulses on each of the address lines. If they're missing on one of the address lines, that's the problem. Work backwards from buffer U34 towards the address bus generator (U2) the until you find where the fault is.
Also check that you're seeing column and row address strobes at each DRAM pins 17 & 4.
If that all looks OK, then you've probably got one or more bad DRAMs. It's fairly complicated to work out which one it is in this case, so the simpler option would be to replace all 4 of them.
-
Ok, thanks. This info was very accurate, and to the point. I will try to probe and look for this, and report my findings. I hope I'm find something wrong.
Thanks again.
-
Hi, just did a quick test, was not sure my scope was up for the task, but this is what we are looking for, right?
Will do some more testing later this week.
First attachment was from C16, A4, not A0 as named.
When i probed A0, its only 2 good pulses, and 2 short. So this can be the whole problem?
-
The green waveform is exactly what I'd expect to see. The yellow waveform looks unusual, the half levels suggests there might be a short between A0 and another address line. See what A0 looks like at the input of the buffer (U34 pin 11). It should look the same on the output (U34 pin 9). If the address data looks OK at the buffer input, but it's bad on the output, that's the problem. In which case you're probably going to be looking for a short between A0 and something else.
I just had a quick look at an A500 here, some of the other DRAM address lines have two or three strobes. A0 should have 4 strobes, A1 has 2 strobes.
Also, it would help to perform this measurement with no ROM installed. If the CPU is stuck there doing nothing, all you should expect to see is the constant DRAM refresh cycles. i.e. Constant activity on all of the address lines + column/row address strobe.
-
Hi, is there å place online i can see what pulses is expected from a chip? Can't remember seeing that in any documentation.
I'm quite new with the oscilloscope, so my biggest problem is to know what to look for, what is ok and not.
I will do some more probing during the weekend.
-
The short answer is no. Datasheets for logic such as the 74F244 buffer will explain electrical characteristics of the IC such as voltage limits, switching thresholds, drive levels, slew rate, propagation times, etc. In the case of the buffer, the output function is the same as the input function, so that's easy - the output will be the same as the input. The buffer input signal is generated directly by the DRAM controller (part of U2/Agnus). Meaning you'd need to have a good understanding of the details around DRAM access. I've been an electronics design engineer for 30 years and often find that confusing, so not a good topic for someone new.
Or to explain another way: the easiest way to know what to look for is to measure on a working A500 for comparison, not by reverse engineering the DRAM control logic. Remove the ROMs from both working and defective boards so they're in the same state.
But you're on the right track. As I mentioned above: "See what A0 looks like at the input of the buffer (U34 pin 11). It should look the same on the output (U34 pin 9). If the address data looks OK at the buffer input, but it's bad on the output, that's the problem."
-
Hi, sorry for the long delay, but pictures will come. Just so much other tings to do.
I have 2 issues i want to "check out" before i proceed, and that is, should there a connection from U34-Pin12 to RP202-Pin2?
As picture from Amiga PCB Explorer marks with a dotted line her. Attachment u34.png
And second, should the via at U34-Pin 10 be soldered to anything. Previous this was bridged to Pin10, ground. Attachment Via.jpg
-
I have 2 issues i want to "check out" before i proceed, and that is, should there a connection from U34-Pin12 to RP202-Pin2?
And second, should the via at U34-Pin 10 be soldered to anything. Previous this was bridged to Pin10, ground. Attachment Via.jpg
Just had a quick look at a rev 6 board here, and it correctly matches the schematic, not your PCB reference.
http://amiga.serveftp.net/Schematics/A500_schematics/A500_R6_schematic.pdf
i.e. U34 pin 12 connects to RP202 pin 6. You should measure approx. 68 Ohm from U34 pin 12 to pin 7 of any of the DRAMs. If you measure between all combinations of the address lines (at the DRAMs or on the U34 output side), they should all be a very high DC resistance, i.e. over 100k Ohms. If you see a short circuit or fairly low DC resistance between any two address lines, that's your problem.
U34 pin 10 connects to ground. As with most TTL logic, the 'bottom left' pin is ground (0V), and the 'top right' pin is VCC (+5V).
-
Thanks, did you notice that "via" connection at U34. It was a solder bridge to Pin 10 her, that i cleared up. You think it should be like it is on picture?
They are really close together, so no wonder the previous owner has been a little unlucky with the soldering here.
-
Just had a quick look at a rev 6 board here, and it correctly matches the schematic, not your PCB reference.
http://amiga.serveftp.net/Schematics/A500_schematics/A500_R6_schematic.pdf
i.e. U34 pin 12 connects to RP202 pin 6. You should measure approx. 68 Ohm from U34 pin 12 to pin 7 of any of the DRAMs. If you measure between all combinations of the address lines (at the DRAMs or on the U34 output side), they should all be a very high DC resistance, i.e. over 100k Ohms. If you see a short circuit or fairly low DC resistance between any two address lines, that's your problem.
Hi, i'm just had a "light bulb" moment her, studying Schematic, I'm not used to read this type, most experience with old audio amps. :)
I haven't seen that little number beside the actual pin number, now it's much easier to follow the trace. Felling so stupid here now. :)
Thanks again. Now its time to hunt down broken traces again. And this this time with schematic.
-
Hi, that did it. I found a missing connection between U35-Pin 9 and RP202-Pin 2. That's adressline 8.
I have been so close all the time, but now its all good. 512kb chipram, and diagrom is all green, all the way.
Thanks to you all.
And now, let the mods begin. :)