Myarc cards for sale/repair tips

[+OLD CS1]

House lights dim, the floor vibrates and shakes, and from within the heavy, clunky metal housing comes the sound of "whooooom, whoooom, whoom, whoom, whom, whom, whom, whom, wohm....."

[+InsaneMultitasker]

On 2/5/2022 at 1:17 AM, Restricted-Access said:

I'm sorry... I'm just reading this thread for the first time, and I don't know if this was addressed a few pages from now, but did you check the conductivity on the "outside" of the Myarc shells?  The inside is coated with a material to reduce RF interference.  Although clipping the leads so they don't contact the shell is a good idea, you could also insert a thin, hard sheet of plastic (it would need to be REALLY thin, and not soft enough to be scratched through) to insulate against the conductive coating.  Also, if you are not concerned about RF bleed... you could use some steel wool to strip that coating off!  In any case, it IS a plastic shell... it's just a conductive coating inside to block RF signals. ?

I was repairing/updating a Geneve yesterday (I'll post a bit about that later) and remembered your comment. The card was enclosed in a Myarc plastic shell.  The outside of the shell registered no conductivity on my Fluke with the leads 1/4" apart, at multiple points.   The inside of the shell measured 500ohms - 5Kohms depending on the distance between the leads and the location on the shell.  The resistance levels were 'random' in the sense there was typically much higher resistance the further apart I placed the leads. Pressure also played a part; I could at times increase resistance by increasing the lead pressure (penetrating the coating to the plastic underneath?) in certain areas. 

 

[+InsaneMultitasker]

I have been working on repairing/updating a Geneve that exhibited lockups/crashes after 5-10 minutes of use.  The typical suspect here is the 5v regulators, especially the original "Korea" parts installed on this card.  The owner offered to remove the regulators and caps prior to shipping; I declined, as I prefer to test the card as-is before any work is done.  This is important for two reasons: (1) to verify the failure as reported and (2) to minimize inadvertently masking the original problem.

 

When I placed the card into my system, it started up as expected.  After 15 minutes, I could no longer access any devices.  The regulators were all outputting nominal voltages and had not failed.  Immediately resetting the system resulted in spurious card activity - SCSI LED flickering, Myarc FDC clicking - so I shut down the system for a few minutes.  I turned the system on and went through the same steps with similar results.  This Geneve has a PFM512 chip installed, which includes a 360K flashdisk.  I was able to access the flashdisk even after the bus devices failed, which turned my attention to the '244 and '245 chips that service the external bus. 

 

I replaced the '245 and restarted the system and after 25 minutes, there was no indication of a problem.  I re-installed the original '245 and after 10 minutes, the system again failed. The socket was an original single wipe variety, so I replaced it just in case the old chip's legs were no longer making adequate contact.  At the same time, I replaced the 32K SRAM socket which had evidence of corrosion and resoldered the CE wire to the chip.  

 

With the card's stability seemingly restored, I opted to replace the three voltage regulators at the bottom of the card with DC-to-DC converters and new capacitors.  I installed a new 7805C for the 9938 power circuitry and as I was preparing to replace its 22uF capacitor, I noticed the 12v regulator's 22uF capacitor had started to leak.  I replaced both 22uF/16v capacitors with 22uF/35v caps and tested all five power sources for shorts before attempting a smoke test.  The new regulators and converters worked as expected and the card successfully ran for 2+ hours. 

 

I then replaced the remaining stock electrolytics (330uF, 100uF, 22uF, 10uF) with new caps.   (I install the 100uF at a 45degree angle to provide better access to the RED/ground header).  

 

An Atmel 29c040 PFM512 was installed in the late 90s or early 2000s.  I no longer remember the reason for attaching wires directly to the chip instead of to the socket.  Maybe there was a need to use a standard EPROM such as Tony's v1.00 BOOT eprom?  My old upgrade notebook didn't have any clues, and I decided to leave it as-is. 

 

During my 'final' checkout, I discovered that the joystick port was not working correctly.  The fire button would not register if the stick was moved up,right,left, or right. Diagonals did not register.  Joystick #2 would only move up.  The resistors all ohmed out properly.  Since both joysticks use the same CRU bits, I suspected U203 (HCT04) and the 9901.  Replacing the HCT04 fixed joystick #2's operation but diagonal movement and the fire button were still a problem.  I replaced the 9901 with one of my own and that fixed the movement but not the fire button.  I located another 9901 and now the fire button and all directions work as expected.  I'm guessing not many people used the joystick and/or mouse with their Geneve, which leads me to wonder how many cards have the same issue, including my own!  I touched up the DB9 connector's solder points for good measure.   

 

Anyway, here's what it looks like after the updates.  I intend to shorten and reattach the 29c040 wires and tie-wrap the 32pin socket to the 28-pin socket underneath.  I'll complete a final burn-in before returning the card to its owner.

 

[+arcadeshopper]

I have been working on repairing/updating a Geneve that exhibited lockups/crashes after 5-10 minutes of use.  The typical suspect here is the 5v regulators, especially the original "Korea" parts installed on this card.  The owner offered to remove the regulators and caps prior to shipping; I declined, as I prefer to test the card as-is before any work is done.  This is important for two reasons: (1) to verify the failure as reported and (2) to minimize inadvertently masking the original problem.
 
When I placed the card into my system, it started up as expected.  After 15 minutes, I could no longer access any devices.  The regulators were all outputting nominal voltages and had not failed.  Immediately resetting the system resulted in spurious card activity - SCSI LED flickering, Myarc FDC clicking - so I shut down the system for a few minutes.  I turned the system on and went through the same steps with similar results.  This Geneve has a PFM512 chip installed, which includes a 360K flashdisk.  I was able to access the flashdisk even after the bus devices failed, which turned my attention to the '244 and '245 chips that service the external bus. 
 
I replaced the '245 and restarted the system and after 25 minutes, there was no indication of a problem.  I re-installed the original '245 and after 10 minutes, the system again failed. The socket was an original single wipe variety, so I replaced it just in case the old chip's legs were no longer making adequate contact.  At the same time, I replaced the 32K SRAM socket which had evidence of corrosion and resoldered the CE wire to the chip.  
 
With the card's stability seemingly restored, I opted to replace the three voltage regulators at the bottom of the card with DC-to-DC converters and new capacitors.  I installed a new 7805C for the 9938 power circuitry and as I was preparing to replace its 22uF capacitor, I noticed the 12v regulator's 22uF capacitor had started to leak.  I replaced both 22uF/16v capacitors with 22uF/35v caps and tested all five power sources for shorts before attempting a smoke test.  The new regulators and converters worked as expected and the card successfully ran for 2+ hours. 
 
I then replaced the remaining stock electrolytics (330uF, 100uF, 22uF, 10uF) with new caps.   (I install the 100uF at a 45degree angle to provide better access to the RED/ground header).  
 
An Atmel 29c040 PFM512 was installed in the late 90s or early 2000s.  I no longer remember the reason for attaching wires directly to the chip instead of to the socket.  Maybe there was a need to use a standard EPROM such as Tony's v1.00 BOOT eprom?  My old upgrade notebook didn't have any clues, and I decided to leave it as-is. 
 
During my 'final' checkout, I discovered that the joystick port was not working correctly.  The fire button would not register if the stick was moved up,right,left, or right. Diagonals did not register.  Joystick #2 would only move up.  The resistors all ohmed out properly.  Since both joysticks use the same CRU bits, I suspected U203 (HCT04) and the 9901.  Replacing the HCT04 fixed joystick #2's operation but diagonal movement and the fire button were still a problem.  I replaced the 9901 with one of my own and that fixed the movement but not the fire button.  I located another 9901 and now the fire button and all directions work as expected.  I'm guessing not many people used the joystick and/or mouse with their Geneve, which leads me to wonder how many cards have the same issue, including my own!  I touched up the DB9 connector's solder points for good measure.  [emoji4] 
 
Anyway, here's what it looks like after the updates.  I intend to shorten and reattach the 29c040 wires and tie-wrap the 32pin socket to the 28-pin socket underneath.  I'll complete a final burn-in before returning the card to its owner.
 

Awesome work glad you are able to keep these old beasts alive

Sent from my Pixel 6 Pro using Tapatalk

[+InsaneMultitasker]

For reference, when the Geneve buffer chips are socketed, some of the Myarc plastic clamshells have three nubs that get nice and cozy with the chips and make it hard to close the clamshell.  The middle nub in particular can touch 2-3 leads of the chip. Snipping the tall portion of the nubs at an angle will make it much easier to close the case and removes any conductivity concerns that could interfere with the chip's operation (see post #552)

[Swim]

I seem to recall being told by someone who knew way more about ttl logic than I that the TMS9901 is a chip that actually "wears out" over time. Hard for me to understand how something with no "moving parts" can wear out but when I think about cards that use the TMS9901, the CorComp RS232 for instance, I believe it.

[GDMike]

Capacitors too generally "wear out" age out.

[+InsaneMultitasker]

I have a few body parts that are working their way towards wearing out, too.  lol

[tater1337]

17 hours ago, Swim said:

I seem to recall being told by someone who knew way more about ttl logic than I that the TMS9901 is a chip that actually "wears out" over time. Hard for me to understand how something with no "moving parts" can wear out but when I think about cards that use the TMS9901, the CorComp RS232 for instance, I believe it.

light bulbs have no moving parts....

[+OLD CS1]

4 hours ago, tater1337 said:

light bulbs have no moving parts....

Electrons move.

[+InsaneMultitasker]

On 3/6/2022 at 3:01 PM, Swim said:

I seem to recall being told by someone who knew way more about ttl logic than I that the TMS9901 is a chip that actually "wears out" over time. 

Our cable/ISP was offline nearly 12 hours today so after I finished working, I did some much-needed cleanup of the TI dev/repair room.  With many items put into "safe places", I felt the need to satisfy my curiosity: does my workhorse Geneve's joystick and mouse still work properly?  I inserted the card into the PEB and ran my test program.  Surprise!  Joystick #2 will not move diagonally and moving in any direction inhibits the fire button from working.  I know it worked because this is the same Geneve I used to develop the program long, long ago, and I've played a few games like Barrage since then. 

 

The HCT04 is not socketed, so I'll tackle the repair when I have some more free time.  If the HCT04 fixes the issue (it should!) then maybe it can go into that FAQ @dhe is working on <smiley face>  

 

[+dhe]

I spy a rare Turbo Video Pal!

 

Documenting the joystick troubleshooting will be easy. Have you posted the test program either on Git or 9640News?

[+InsaneMultitasker]

4 hours ago, dhe said:

I spy a rare Turbo Video Pal!

 

Documenting the joystick troubleshooting will be easy. Have you posted the test program either on Git or 9640News?

The joystick and mouse tests are part of CYA.  I have standalone versions of a few of my test programs (that I wrote/used at Cecure) but they would need a bit of tweaking for me to release them.  I don't have time for that at the moment, maybe later this year.

[+dhe]

 

While looking at information on programming PAL/GAL chips, I found this youtube video from Adrian's Digital Basement. The Commodore PAL chip is a bus controller and tends to high failure. 

 

Comments below reveal why this chip's failure rate is higher than normal. I felt this might be of interest given our recent discussion regarding the failures of TMS9901.

 

Anyone notice how rare AMI marked 9901's are?

 

https://www.youtube.com/watch?v=GKyoh_gGTYM

 

[GDMike]

I've got an extra 9901, but I'm not sure if it's real, if it works, or what. I saw it one day for sale, so I grabbed it, I wish I had a way to test without pulling another out of a board. 

[+Ksarul]

5 hours ago, dhe said:

Anyone notice how rare AMI marked 9901's are?

Actually, though they don't show up often, when they do, there seem to be a lot of them. AMI marked 9902s are definitely hard to find though. . .

 

AMI marked 9900s are pretty much available whenever you go looking for them though, usually for very reasonable prices.

[+InsaneMultitasker]

For some time now I've been trying to repair a Genmod Geneve with Memex card.  The final challenge was that the Geneve could not see pages >E8->EB, which are required for now to run MDOS 7.30.  I replaced every chip and traced the circuits as best I could, to no avail.  Recently, I received another Genmod/Memex card for repair; coincidentally, that system exhibited the same behavior! 

 

It turns out that the Memex documentation contains incorrect information about the DIP Switch settings.   SW2, SW3, SW4, and SW5 produce different results, as confirmed by the two systems on my bench and help from @arcadeshopper. Findings so far: 

 

I have been using @9640News  new MEMTEST 2.60 to test the full 2MB memory on both systems with positive results.

[+FarmerPotato]

On 3/6/2022 at 3:01 PM, Swim said:

I seem to recall being told by someone who knew way more about ttl logic than I that the TMS9901 is a chip that actually "wears out" over time. Hard for me to understand how something with no "moving parts" can wear out but when I think about cards that use the TMS9901, the CorComp RS232 for instance, I believe it.

I’ve had grief from the CorComp MPES “cooking” the 9901 (erratic symptoms.)

 

the 9901 data book warns that it will be damaged if a pin is set to output but another source is driving it. maybe just that one pin, but who knows. 
 

This can occur by software abusing it. An SBO instruction to a pin that is normally an input. A trashed R12 that accidentally points into a 9901 before loading some bits. 
 

I don’t know how my MPES’ 9901 got fried, if it was heat or accident. It just happened one day while programming assembly, my game loop slowed to a crawl (updating sprite positions with counters.)  Did not find a bug, same behavior after reverting to earlier code.  Didn’t occur on P-Box system. 

Maybe I can finally solve that mystery, learning more about the MPES in the CorComp thread. 

Much later I heard that the MPES tends to cook 9901s. That I wasn’t the only one to see it. 

 

[globeron]

Two days ago I checked if the Myarc was still running (batteries on the HRD) everything still works fine.  This is the latest OS correct of 4 April 2021? 1 year operarional.   I am a bit hesitating to change now everything works.

[+InsaneMultitasker]

for those of you who remove/insert cards in your PEB periodically, here is a simple tool you might consider investing in. 

 

The timers are plastic, the sand colors are nice and visible, and they are small enough to keep around without being obtrusive.   For me, this is cheap insurance and keeps me from pulling a card from the PEB too quickly.

 

 

[+Torrax]

I just ask my Google Nest Mini to set my timers.

[+InsaneMultitasker]

Thanks in part to @Ksarul and @Shift838 and a few others along the way, the Genmod & Memex card that I have been gnawing at finally seem to be properly working.  There were quite a few challenges along the way, mostly related to poor connections and degraded sockets, and my HFDC's gate array dying along the way.  

 

The most challenging issue was that one of the chip types that I received earlier this year from a reputable source was flakey and/or bad.  I do not recognize the manufacturer symbol. See the attached picture - anyone recognize it? 

 

I had to dig deep into my parts bin to find two used LS02 parts. Installing the first brought stability to the system and the second seemed to have cured the remaining problems.   I ran the system for 4+ hours today, MEMTEST 2.60 showed no errors in 300+ iterations, Memtest v1.1 showed no partial pages, and I was able to assemble MDOS, run the newly-assembled OS, use it to assemble MDOS again, and ran that version successfully.  Not to jinx myself or the cards but I will be happy to send them back to their owner   

 

<-- who makes this chip? 

[+9640News]

And the owner is very much appreciative of your efforts and the contributions of the others that brought this system back to life!

[+InsaneMultitasker]

Memex before/after (and blank board)

  

 

[AwkwardPotato]

6 hours ago, InsaneMultitasker said:

I do not recognize the manufacturer symbol. See the attached picture - anyone recognize it?

It's a symbol found on many bottom-of-the-barrel phony TI parts. Seems that they only bother engraving one half of the TX...