16 posts / 0 new
Last post
C128 repair - a community assisted story

I received a C128 with the regular blackscreen.

First off, the 9v (1.4A) fuse in the PSU was dead, so I took the long one from the inside of a dead C64, it's only 1A afair but good enough for testing. Dimensions for the long one didn't fit, so I soldered the bad one with the good one. This is of course while waiting for the new ones to arrive.

I seem to have problems finding 1.4A fuses, am I the only one who has this problem ?

 

Still dead. Testing output clock of the VIC chip, Pin 18 and 23 one is the 2MHz and the other 1MHz. here comes the fun.

 

 

Scope settings 0.2uS/Div  1V/Div, the regular 1Mhz signal shows some ripples it's the upper signal, and the 2MHz signal is 1Mhz.

Now from what I know the C128 works at 1MHz when the screen is on, and at 2MHz when the screen is off. So the signal should be good, but

it would be nice to have this confirmed, and I really don't like the ripples of the 1MHz signal. After watching Mr. Herds repair vid part 2 I could be led to suspect a dried out cap.

(Can anyone confirm that the 2MHz signal really is 1MHz, the schematics seems vague here)

 

Well. So far I've tested power voltages, CPU Reset, Clock. Next is Address lines (I finally got my 40pin IC test clip to do this). I have a dead test cartridge from ebay on it's way.

 

EDIT: I just noticed that the ripple of the 1MHz signal was due to the ground of the probe wasn't attached properly

 

Edited by: tomse on 2012-12-09 04:42
I managed to hook up my open

I managed to hook up my open Bench Logic Analyzer ($50ish). Yes I own cheap equipment, though it's useful, but if anyone want to sponser me a HP Agilent 16500C with probes and modules I wouldn't say no thanks ;-)

 

Well the setup: (Click to get a higher resolution image)

The character rom has been replaced with a Danish one. The original 2364 has 24 pins whereas the 27c64 eprom used here has 28pins, some pins needed rearranging hence the weird setup. There are quite a few external wires that, I have no idea what they are doing. Perhaps it's a late fix or something else? This is a Rev9 board. Does Mr. Herd migth have some insight ?

 

 

I looked through the beginning to a state where it settled into a loop, giving a repetitive result, this was around 5-6 seconds after power on.

So it tries to do something inside B000-C000 This is the Basic HI rom. Now I need to dig up what happens in those areas.

 

As a side note, the Open Bench Logic Analyzer samples until it's memory is full or the amount of samples configures has been reached before sending the data to the PC. This is a cheap piece of equipment that unfortunately doesn't sample from when you press Record till you press Stop, so you only get a small window of samples to work with.

The address line outputs don

The address line outputs don't seem to do me any good. Looking at Ray Carlsen's troubleshooting sheets, the issue might lie in the 8502 CPU.
But before deeming that dead, I must go on a chip select hunt. Which involves in checking the chipselect wire to all the IC's which is handled by the PLA (U11)
I've stopped toubleshooting for today, and hooked up my C64 so I can dump some more tapes to see if they contain stuff that hasn't been preserved yet.

Lol... good troubleshooting

Lol... good troubleshooting first off as the ripples on the clock definitely looked like that particular probe wasn't grounded. Picture the path the signal needs to take using that probe and the distance to the ground for the other probe. Now picture grabbing a coil and opening the loops, the more air between the loops the less mutual inductance.  That is one of the keys to noise suppression and impedance control  is the closer the signal and return are to each other, the more they are in each other's inductance field and so mutually inductive.
I dont remember what the 2Mhz signal looks like particuarly but put your probe on the PHI0 input to the processor (which might be the 2Mhz signal... lol).  If you "back out" the sweep time you should see smears of it jumping frequencies, you can even do triggering tricks like triggering on the sync signals or put it in 80 coulumn mode.
 

Lol... small windows.... that

Lol... small windows.... that is THE trick to the older analyzers, we had windows as small as 64 or 128 clocks sometimes so capturing THE exact moment was tough.

ROMs and EPROMs are chocked full of extra select and enable logic, we were always adapting 24 pin to 28, EPROM to ROM.  The best guide is a data sheet on the device compared to the schematic. 

The main jumper on a C128 was because of a stub on the A10 line if I remember correctly.  There is a whole story behind the jumper if you catch the video.

Try triggering on the chip enable on the ROM in question and then inspect every Address and Data line to make sure it does something valid during the select time.

trigger.PNG

 

Examples of "not true"  shown below.  Note you really need a scope or a real good understanding of how your analyzer treats a problem condition.  I like to couple the triggering of the scope and the analyzer. >:)

Here is an example of when things are true at the right time

quiet true.PNG

Examples of shorted and floating lines look " noisier" in general.

noisey false.PNG

 

blown driver.PNG

 

 

And it could be that the signals are find but the ROM has a dropped bit for example, there the trick is to tell when the processor execution has "left the tracks" (like a train derailing) and can be harder.  I once triggered on a symbol being written to the display memory with a light pen as an example of the tricks you sometimes do to capture the crash with limited analyzer window..

Good luck and also please call me Bil, Mr. Herd is my Dad. >:)

Bil

BTW the picture of the leads all attached to a C128 bring back many memories.

 

 

 

Also if the VIC never gets

Also if the VIC never gets initialized (or the MMU) then the issue would be way early in the Kernal before ever getting to Basic.  That might explain lack of 2Mhz if the circuitry that generates the clock selection logic was never initialized.
Do you ever get Vic /CS?
 

There's definately alot of

There's definately alot of input to work with now. Thanks Bil :-)
I'll check and post more info when I come home from work
and thanks for the correction of the eproms being 28pin and not 32pin (it was a trap to see if you read everything *cough* hehe)- I fixed the posts with the correct amount of pins

I've been busy most of the

I've been busy most of the week, but now I have the time to work on the C128 again :-)

The images below show the Chip selects taken off the PLA - 8721 - U11

This is a zoom-in on the captures, but there is no VIC CS.

 

I setup the logic analyzer,

I setup the logic analyzer, as per Bil's suggestion to read the address lines after doing a chip select on the rom.

The (kernel) Rom is actually always selected, so I triggered on OE, looking at the result i wasn't very successful, and it looks like my Logic analyzer isn't very reliable.

pin 14 was connected to ground. see the spikes it shows from time to time? Probe 15 should be triggered by the down signal on pin 22 of the ROM chip (OE/Read enable which isn't really useful here, but hey.. worth a shot :-)) well on my setup it was triggered sometime, but I did manually trigger it too.

 

 

@Bil - in the "making sense of the chaos" where did you suggest to measure with the scope ?

Hmmm... Let me back up here

Hmmm... Let me back up here and ask a question: How would you fix the system if it were a major chip?  Do you have replacement chips from a donor system?  If so there is no foul in doing a chip swap as the remedy and the diagnosis become the same step. 
BTW dont forget to check the 9VAC fuse, it can affect the Z80 clock which would keep things from working correctly.
Feel the chips, see if any are too hot or stone cold.

Due to the holidays, and

Due to the holidays, and changes of plans in these days, I've removed the C128 from the workbench. What I thought of me having plenty of time to work on this between christmas and newyear, has now been removed. So unfortunately I won't be working on this at all during these days.
 
The original fuse was blown in the PSU, and I replaced it. Now to ask. why did the fuse burn in the first place? After replacing the fuse with a new one, the new one didn't burn.
 
I have a nice laser thermometer bought off dealextreme. I checked all the chips, (measuring in Celcius, for quick reference, water freezes at 0 degrees, and human body is 37 deg) and Ram was ~26 - 27 deg, Z80 and 8502 was ~33 degrees. Other chips were between these temperatures. Measuring happend around 1-2 min after turning on.
 
Unfortunately I have no CBM spareparts to work with other than EPROM's and 4164 ram.
 
Merry christmas :-)

I've gotten hands on a nice

I've gotten hands on a nice C128D model in metal casing.
 
I love this machine, the only thing I can't use it for is dumping tapes using Markus Brenners mtap programs. This is due to the requirements of a X(A/E)1541 cable, where the software conflicts with the internal diskettedrive. But for dumping diskettes using WarpCopy and the 64NIC+ bought off Jim Brains site (go4retro.com) it's just great. Though for dumping diskettes I use a 1541-II as the drive head needs cleaning quite often.
 
now I'm waiting impatiently on MIRKOSOFTS ACE128
 
You can see my newly aquired C128D it in action on : http://awesome.commodore.me/c128/
 
So fixing the model I have in this thread isn't a high priority any more (though at one time I do need to finish fixing it)
 
I've just watched the demo: Risen from Oblivion on it something I can recommend if you're into demos.
 
 
@Bil have you ever considered creating some upgrades for your old retro computers, as in developing new tech with modern hardware to sell to the community (or open source it for that matter)?
 
 
 
 
 
 

I have thought about it but

I have thought about it but since there is currently someone with the rights to Commodore I would have to get permission to do anything.  I was thinking about redoing the Commodore LCD using some modern compnets and the code written back then but again I couldnt use the Commodore name or anything purchased by CommodoreUSA.
Bil

Very interesting. Though the

Very interesting. Though the Commodore name shouldn't be a hinder.
"Bil Herd's LCD revived"
 
Back on topic, I've had to get outside assistance with the C128. I'll try and document as much as possible while we're working on this.
The project hadn't been a high priority due to the fact that I got myself a C128D.
I'll cross my fingers that it's only a part that easily can be replaced.
 
Here's a small goodie if you haven't seen it yet
The only demo I know of for the C128 is Risen from Oblivion
http://www.youtube.com/watch?v=sW4V-ehYFQw
I think I'll have to make a video that doesn't seem to "stutter" as this one does, time to dig out my grabber card.
 
 
 
 
 
 

Re: I have thought about it but

Bil Herd wrote:

>I was thinking about redoing the Commodore LCD using some modern compnets and the code written back then but again I couldnt use the Commodore name or
> anything purchased by CommodoreUSA.

Hey, I like that idea of a reborn LCD! As for name... name... schmame... Admiral (like the old t.v. company), Captain (or in Spanish, Capitan), Commander.... oh, there are a lot of names out there. :)

Wait... there was the Hackintosh (a hacked-together Mac made of PC parts). What about the Hackindore or the Commihack? ;)

That would be some project!
Robert Bernardo
Fresno Commodore User Group
http://videocam.net.au/fcug
The Other Group of Amigoids
http://www.calweb.com/~rabel1/
Southern California Commodore & Amiga Network
http://www.sccaners.org

A few things to think about

Sorry for bringing up an old topic though. 
 
There are three basic genres of rights in intellectual properties: (just in case anyone here doesn't know)
copyright
patent
trademarks
 
There are only two of these genres of rights that matters. Copyright and Trademarks. Copyright duration is a long time. Trademarks are like 20 years like patents except trademarks are renewable. Whatever patent on the C128 and C64 had expired as well as most of the Amiga patents. Those expired over the past 15 years. 
The only issues of intellectual properties is trademarks and copyrights. Copyrights applies to the ROMs so the ROMs maybe a legal challenge from a commercial perspective even to this day. The trademarks are easy... don't use the Commodore logo and name. Then again, Commodore IP holders can lose rights to trademarks if they do not do anything substantial. According to Federal law, if there is no documented active use of trademark that constitutes an active use of trademarks means the company can lose the trademark. Hence, ABANDONED status. This means that another company can assume the rights and gain control over trademarks. The law doesn't reward those who just sits on the trademarks. Seldomly this happens but trademarks can be usurped by another party. 
 
In any case, Bil, you need not be concern about trademark that much if you have no interest in using the Commodore trademark and you comply with fair use provisions of U.S. Trademark laws. As for the patent, they have expired. What that means is the intellectual property holders of Commodore copyrights, patents and trademarks no longer have patent protection or any of the exclusive rights provided by patents to any of the Commodore 8-bit platform patents. Since they have expired, you should not concern yourself with that. Patents are a one time deal... NO RENEWAL rights.
Therefore, the only issue that is most challenging is copyrights and that is the only thing you most likely need to have permission on... that is for the Commodore BASIC, Kernal and DOS ROM stuff.
Other than that, I would say it is a non-issue. 
If you do consider any of this stuff, I would think the only issue you need be concerned about is copyright and MAYBE trademark matters. That depends.
 

Log in or register to post comments