Some more CoCo love...

Well, I spent a little quality time with my newly repaired and upgraded CoCo3 yesterday. After reassembling everything, I played Donkey Kong for a while.

Which reminds me, last week when I was playing DK, my daughter and her boyfriend were over. He heard the sound FX coming from my office and came in to watch. I got the expected 'what the hell is that?' remark. I said 'Donkey Kong', what, you've never seen it?' His reply was, 'I've played Donkey Kong on in the SNES, but, it didn't look anything like that.' So I offered him a shot to play it. He never made it to the fourth girder on the first level. He hung his head and exclaimed 'Old skool kicked my ass...'

I told him this wasn't a sappy ass run-and-shoot game you can buy for a PS3 and solve it a few hours. The goal is to see how high of a score you can get, and how many levels up. I then consoled him some before telling him he got his ass handed to him by an 8-bit computer!!!

At any rate, I got out my old Botek serial->parallel converter yesterday, downloaded the manual to my Okidata Microline 320 Turbo dot matrix printer (could not remember how to set the damn thing up to save my life) and played around with TW80 and WordPower3.3. While doing some printing, my 10 year old daughter ran in a asked 'Daddy, what's wrong with the printer?' I said 'What do you mean?' She replied with 'I see it works but something has to be wrong with it to be so LOUD!!!' Kids, they crack me up...


Later,

Brian

Success.....?

About two weeks ago, I ordered a 512K upgrade kit from Cloud 9. The kit arrived yesterday, got it all soldered up (yes, you have to solder the kit yourself) and plugged in (great product BTW). Was all jittery and excited, as with this one updated, I only have one CoCo3 with 128K - the re-pack.

Well, I turned the machine on and realized I forgot to plug in the VGA adapter - rats!!!. So I plug in the VGA adapter and turn the CoCo3 back on, all excited to see that nuclear green screen... and was greeted with a black and blue checkerboard pattern. WTF??? So I turned the power off. unseated and re-seated the upgrade board and tried again. Same damn thing.

After going over the work with a big ass magnifying glass, I discovered that when I removed the 6809 and installed the socket last week, I cut (really broke) a very small motherboard trace. Underneath the 'H' bar on the CPU socket of all places. No idea how I missed it when I was looking everything over before I soldered in the socket. Well, I got that fixed with a piece of voice coil wire (do you have ANY idea how difficult it can be to solder a piece of wire to a trace when the solder point is under some you REALLY don't want to have to remove..???), but, still had the same problem. 

After reseating the memory - again - and a few other checks, I was left with swapping out the 6309, which fixed it.

I doubt anything else was wrong; I ran the included memory test program for about 2000 loops, and played Donkey Kong (great CoCo3 stress tester) for about a half an hour and had no other issues.

Overall, the whole process was maybe an hour, including soldering the upgrade board. This was my first experience with a CoCo3 CPU going bad. Anytime I've had a PC CPU go, the computer failed to do... anything except turn the power LED on. Live and learn...

Holding pattern....

While I wait for the Cloud9 512k upgrade I ordered, I figured I'd operate on another CoCo3 and do a CPU transplant. What was the point of getting 10 63C09EP's if I didn't intend to install them? The very first one I did on my repack (seems like ages ago) I carefully removed in an effort to save the CPU. For what, I have no idea. That could have turned to disaster as the heat was enough to slightly warp the motherboard, so I doubt the CPU survived anyway.

These latest ones I found myself just clipping the damn things out, cleaning everything up and soldering in the sockets. Repack CPU swap time, somewhere around 4 hours. Clip & solder swap time, about 30 minutes. Yeah, don't think I'll be wasting time like that again.

Above is a pic of the bottom of the motherboard, showing the 'cleaned-out' holes. Of course I got the bright idea to take some pics AFTER I had already started to solder the socket in... oh well....

Above are the two recent sacrifices. Didn't dawn on me until I was writing this that the damn things were upside-down in the picture... sigh....

Well, I don't consider myself a vintage computer repairman by any stretch, but, that didn't stop me from getting out my malfunctioning MPI on Saturday and going thru it again. When I last had it out, Gene and I had been diagnosing a strange problem (yes, those seem to follow me...). With a floppy plugged into it, it might load, or might not load a program. Sockmaster's Donkey Kong is a no-go with an error in line 15. I Can get DIR commands to complete, but, they take longer than they should and the drive sounds like it's going to throw an I/O error. The O-90 cart seems to work by itself, can't get any disk access, tho. The REALLY weird part is, with the FDC in slot for, and the speech and sound cart plugged in - does not matter which slot it's in, and the selector on '4', the CoCo boots to a Ext Color Basic 1.1 screen. Essentially a CoCo 1/2 screen. I have replaced the 74LS245 chip and all of the 74LS367 chips. When this problem started way back in '07, I think Robert Gault suggested to check the power supply. I'll be doing that this coming weekend...

Later,

Brian

It's FIXED!!!!

Well, anybody who followed this thread on the mailing list has an idea of the problem with this CoCo3. It seemed to have a fear of floppy drives. After all of the troubleshooting (including a couple miscues by yours truly) we thought we had it pegged to the 74LS138 chip, aka IC9. So, I placed an order with Jameco last Friday for a replacement (and the IC's to build an MPI, that's another story), and that order landed yesterday.

This morning, after my wife went to work, I clipped the 138 off of the motherboard, cleaned the holes and soldered a socket in the chip's place. After cleaning everything up, I inserted the chip, hooked the mobo back up inside the CoCo3 case, made the necessary connections and powered it on. And it worked!!! The two pics below show the floppy light on drive 1 doing a 'DIR' command and loading Sockmaster's Donkey Kong conversion.

Below is the loading screen (for those living under a rock since 2007) of DK. Note the little beige and grey box with lit green LED - Roy Justice's VGA adapter. If you don't have one, buy one. It does not disappoint!!!

The pic below shows the socketed 6309 and 74LS138 chip.

All in all it was a fun learning experience. Anytime I come away from a situation like this and have learned something, I'm happy. Incidentally, how hot should the heat-sink get for Q1? After about half an hour of Donkey Kong, that thing was hot!! I'm thinking about yanking the PS off of this one and using a 90w ATX unit or some other type, maybe something like theotherBob had posted on his now defunct GeoCities page... just for giggles. My repack is powered by a 200w ATX that runs everything except the LCD (which runs off of it's own PS tapped from the 120vac plug the ATX unit runs from), so I think 90w for the CoCo and maybe an internal 3.5" floppy would be fine.

Later,

MCCM

Ordered parts, and update...

Well, I ordered the part today to (hopefully) fix my malfunctioning CoCo3. Gene and I think it's been narrowed down to IC9. Find out in a few days. The CoCo3 that has been flakey had a PBJ 512k memory expansion on it. So, I took it off, clipped C65 & C66 on my 128K CoCo3 and commenced to playing Sockmaster's port of Donkey Kong for a few hours. It was pretty fun.

Along with the part I ordered for the CoCo3, I ordered all of the IC's to make Tom Gunnison's MPI clone, as posted on CoCo3.com. The IC's are dirt cheap - <$20 for them and the capacitors. I'm drawing up a board layout in ExpressPCB right now. Think I'm on my third revision. It's gotta be a double sided board, I don;t see any way around that. I think the cost is gonna be around $125 for 3 6 x 8 double sided boards. I'm trying to find ways to move things around and make it smaller. I think it'd be cool as hell to have it small enough, and dimensioned so that it'll fit inside the existing MPI cases. Then you could continue to use your existing case, or buy one to make your own. The biggest problem I see right now is getting the power and ground planes down correctly... 

One change I am going to make, is going to be a modification to my CoCo's. I'm going to either replace the card slot with a dual row header and just use and existing IDE cable to connect the MPI since I'm using a dual row header on it, or make a male card edge with the header installed on it. Finding a 40 pin male card edge connector has been a royal bitch!!!

I'll be back working on the re-pack again pretty soon. I gotta borrow my dad's Roto-Zip to do some case mods. I thought the Model III case would give me ample room for component placement. I'm questioning that thought, as with Roy's VGA adapter installed, I still have to fit the electronics of the sound system in, and the MPI.

Troubleshooting a perplexing problem...

After all of the new developments on the CoCo scene, I decided to set up a CoCo to play around with. My re-pack is in no condition for this yet, so I pulled my 3 CoCo3's out of the closet and picked one of the 512k models. Wouldn't you know it, it has developed an issue. Seems it doesn't like the FD-502 FDC I have that works fine with my other CoCo3's (and 2's and 1's).

The descriptions of the problem can be found in the MaltedMedia mailing list under the following subjects:

Re: [Coco] Weird CoCo3 problem

and

Re: [Coco] Weird CoCo3 problem - AAAARRRGGGGGHHHHH!!!!

I have checked the power supply and pretty much ruled it out; all of the waveforms match the CoCo3 Service Manual.

I started checking the signals and comparing waveforms and had a few issues with some. For ALL of the pictures below,

execpt for one and it's noted, I set the volts/div and time/div as specified in the CoCo3 SM.

The pictures below are for the guys helping me diagnose the problem (good luck to us all):

This is the waveform of TP2, ECLK - supposed to be a 5vp-p square wave

Here's TP3 - QCLK - again, supposed to be a 5vp-p square wave, and apparently according to the SM, time shifted a bit - these two signals are identical.

Here's TP4 - /RAS - Supposed to be 5v with negative pulses. To save bandwidth, TP5 - /CAS is the exact same waveform. These two signals, according to the SM, are supposed to be virtually identical to each othe - but, negative going pulses and slightly different in timing.

This last one is one Christopher Hawks told me to look at - P3 of IC9. It's supposed to be 5v with negative pulses. However, it's the EXACT same as TP2 & 3, this picture I was using a different Time/Div setting. Pins 1 & 2 of IC9 are identical to P3. P6 of IC9 is a logic HI at 5vdc, P9 of IC9 is close to 2vdc.

Looking at everything as a whole - there's a better than average possibility I'm wrong - but, with all of the malformed waveforms coming from the GIME I'm wondering if that isn't the problem. I've cleaned and reseated it, but, for E and Q to be so different from the pics, and the same signal being on pins 1,2 & 3 of the 138N chip (IC9) it has me wondering. I'm hoping someone else can point me in a direction to either prove or disprove that. Tho the computer is not displaying any of the tell tale signs of GIME problems (graphics issues, sparklies,. etc...)