Myarc cards for sale/repair tips

[globeron]

3 hours ago, Shift838 said:

I decided to try and replace the TMS9905 to make sure there was not an issue with the processor.  Still no change and still receive the BAD CPU RAM FOUND message.

 

 

I get this once in a while. My plan is to follow you to replace capacitors, but a bit afraid my board will not work after that anymore

[GDMike]

So very true. I would only do the simple stuff, and let Chris do the risky business, I've not had a survival rate above 1 -

for Multi layered boards., But I don't have a scope either and wouldn't know what to look for  other than obvious, but What a pitty. That these were done like this. And I'm sure during those days, it would be difficult to do it another way.

 

[+InsaneMultitasker]

On 12/23/2020 at 2:11 PM, Shift838 said:

I decided to try and replace the TMS9905 to make sure there was not an issue with the processor.  Still no change and still receive the BAD CPU RAM FOUND message.

 

 

If you have consistent continuity and no shorts across refresh/data lines, it may be time to rule out the gate array. Memory reads/write and refresh tie directly to it.  

 

If the sound and clock chip are socketed, pull both chips out. They aren't necessary to fire up the Geneve.  If the LS245 and LS244s are socketed, you can pull them to almost completely isolate the card from the bus.  

 

On 12/17/2020 at 10:43 PM, Shift838 said:

pin #14 is connected to pin #14 on the DRAM chip immediately to the left or right of itself within chip column #1 and #2

pin #15 is connected to pin #15 of the DRAM chip directly above and/or below itself.

I need to question my earlier comment about some pins not connected across all chips.  I would have thought 14/15 would be connected. I'll pull out my notes tomorrow. I sometimes get those connections mixed up...

[Shift838]

20 hours ago, InsaneMultitasker said:

If you have consistent continuity and no shorts across refresh/data lines, it may be time to rule out the gate array. Memory reads/write and refresh tie directly to it.  

 

If the sound and clock chip are socketed, pull both chips out. They aren't necessary to fire up the Geneve.  If the LS245 and LS244s are socketed, you can pull them to almost completely isolate the card from the bus.  

 

I need to question my earlier comment about some pins not connected across all chips.  I would have thought 14/15 would be connected. I'll pull out my notes tomorrow. I sometimes get those connections mixed up...

This Geneve does not have the PAL, sound, video ram, or LS244/245's socketed.  I believe the clock chip is the MM58274?  It is socketed...

[GDMike]

Could this be coming from the 32K MDOS upgrade? As I checked a chip, it worked out of the box, but after it went in I can't check.

Edited December 27, 2020 by GDMike

[+InsaneMultitasker]

Dabbled in some repair work today.

 

For comparison:  the top socket is new; the bottom socket is original.  Notice the varied gaps between the wipes in the old socket and lack of gaps in the new. 

 

When I removed the chip from the socket I noticed very little physical resistance.  Re-inserting the chip and checking continuity yielded intermittent contact and depending on the insertion pressure, the chip made contact for short periods of time. (The red arrow points to the most problematic connection).    Replacing the bad socket restored the card to operational status.   

 

A good reminder/reason to inspect all existing sockets during rework to look for fatigue, corrosion, single wipes, etc.

 

[globeron]

On 8/18/2020 at 11:16 AM, InsaneMultitasker said:

M76 is the regulator that supplies voltage to the sixteen CPU ram (DRAM) chips.  The three circled regulators and adjacent electrolytic capacitors are those most likely to fail over time, especially without heat sinks.  Even with heat sinks they WILL fail eventually.  Or in other words, adding a heat sink will buy you some time but it is not a long-term solution.

 

There is a fourth 5v regulator near the top of the board and a 12v regulator near top /center that are less problematic. I sometimes leave them as-is depending on the situation.  

 

If you replace the capacitors near the bottom regulators you should also replace the capacitors near the other two regulators.  They are 30-35 years old by now.  A full refresh would include replacing the other electrolytic caps near the neck and interior of the card. 

 

If you complete the work yourself, be sure that each newly installed heat sink does NOT touch the capacitor lead nor the regulator lead.  It is recommended that you test the resistance across each regulator's input and the ground and each output and ground, so that you have reference numbers after installation. If the replacement values are substantially different - or more importantly 0, indicating a short - you must find and correct the problem before placing the Geneve into the PEB. 

 

Some Geneve cards are prone to pad and trace removal upon application of heat and solder. Some more fragile boards are prone to the removal of the inner core while desoldering, often caused by incomplete removal of solder and trying to remove the component from the through-hole too forcefully. If this core is removed, the once-connected inner layer traces may not make contact with the components and/or could present intermittent contact.

 

Moral of the story:  be careful and take your time. If you don't have the right tools, consider reaching out to someone who does.  Studying is a good thing. 

 

 

I replaced these 4 x 5V regulators.  (Not the 12V regulator)., found it difficult to desolder them and get them out (took 2 hours). Now replaced them and it seems the Geneve 9640 is much more stable and not hanging so far.  I need to test it a bit more, but so far it looks okay. @InsaneMultitasker thank you very much for your guidance!   

[+InsaneMultitasker]

23 minutes ago, globeron said:

I replaced these 4 x 5V regulators.  (Not the 12V regulator)., found it difficult to desolder them and get them out (took 2 hours). Now replaced them and it seems the Geneve 9640 is much more stable and not hanging so far.  I need to test it a bit more, but so far it looks okay. @InsaneMultitasker thank you very much for your guidance!   

Good job!   If one or more regulators were at fault, you should see continued stability.  Did you install heat sinks?

 

Removing components from a Geneve is often difficult even with a good solder removal tool.  Based on the pictures, you successfully removed the regulators without removing the internal cores.  One "trick" that can help in the case of regulators is to snip the regulator leads where they meet the component, grab a leg with a needle-nose or similar pliers and apply heat where the leg meats the through-hole until you can easily pull it out (i.e., with no resistance) then remove any residual solder.

 

 

 

[globeron]

19 minutes ago, InsaneMultitasker said:

Good job!   If one or more regulators were at fault, you should see continued stability.  Did you install heat sinks?

 

Removing components from a Geneve is often difficult even with a good solder removal tool.  Based on the pictures, you successfully removed the regulators without removing the internal cores.  One "trick" that can help in the case of regulators is to snip the regulator leads where they meet the component, grab a leg with a needle-nose or similar pliers and apply heat where the leg meats the through-hole until you can easily pull it out (i.e., with no resistance) then remove any residual solder.

 

 

 

 

No heatsinks installed (need to check if they have them on Lazada), but I put thermal grease (grey colour, not white) on the regulators (in between the board). What a relief it is much stable now... still have some hanging issue, but maybe it is the software itself or the HRD3000/Rambo.  Most what I tested is okay.

 

These are my tools used (I am not that professional as others doing all the hardware projects). 

[HOME AUTOMATION]

I see however...

15 minutes ago, globeron said:

...That you are a techno-junkie!

 

[+InsaneMultitasker]

3 minutes ago, globeron said:

No heatsinks installed (need to check if they have them on Lazada), but I put thermal grease (grey colour, not white) on the regulators (in between the board).

Heat sinks are a wise, cheap investment.  The thermal compound improves the regulator situation (as you have already seen)  though experience tells us that without the heat sinks on the lower three regulators, at some point in the future you will again see thermal overloads.  Also, the Geneve board is not happy long-term being its own heat sink as the transferred heat will slowly degrade/char the board and can decrease operational longevity of the adjacent components. 

[+InsaneMultitasker]

On 12/27/2020 at 4:56 PM, GDMike said:

Could this be coming from the 32K MDOS upgrade? As I checked a chip, it worked out of the box, but after it went in I can't check.

The 32K chip is not tested during powerup; the BAD CPU RAM message refers only to the onboard sixteen DRAM chips.  In many cases, bad SRAM (32k or 64k) will result in either no boot or display of the swan with no further activity.

[Shift838]

On 12/23/2020 at 2:11 PM, Shift838 said:

I decided to try and replace the TMS9905 to make sure there was not an issue with the processor.  Still no change and still receive the BAD CPU RAM FOUND message.

 

 

Wanted to post this in case anyone else finds this helpful.

 

I took a trim to InsanMultitaskers Lair as I needed assistance with my HFDC as the hard disk was stuck at 0 when trying to format.  turned out to be a bad socket!

 

But, I also took @GDMike's Geneve with me to see if we could quickly diagnose the issue with the 'BAD CPU RAM FOUND' message at boot up.  I already replaced all the 41256 DRAM's as  more than 1/2 tested bad.  I tested the 9995 which was fine.  Found a loose crystal, recapped the electrolytic capacitors and new voltage regulators.  The Insane one suggested the Gate Array may be a culprit, although he has rarely seen it.  So swapping the gate array got us past the BAD CPU RAM FOUND error.

 

Now, we would see it actually start trying to boot but freeze after attempting to load MDOS.

 

After a few minutes Insane suggested he really felt like it was the LS245, he desoldered the chip and installed a socket with new LS245 and BAM..  all working...  Allowed it to run a bit with my HFDC and DREM unit with no issues.

 

Still some work that needs to be done on the board but at least it's bootable now.  Add another one to the mix of being alive!

 

 

[+9640News]

Glad to hear another Geneve being resuscitated.

 

Beery

 

[GDMike]

2 hours ago, BeeryMiller said:

Glad to hear another Geneve being resuscitated.

 

Beery

 

Absolutely.. and it can only be done by the genius'!! Hats off to another successful fix!! I'm estatic!!! 

[+InsaneMultitasker]

5 hours ago, Shift838 said:

I took a tripm to InsaneMultitaskers Lair

...and you even lived to tell the tale!! 

[+Ksarul]

55 minutes ago, InsaneMultitasker said:

...and you even lived to tell the tale!! 

Now that iiiis scary. . .you must have forgotten to set your intruder trap lines.  LOLOL

[+dhe]

Tim,

 We know you worked at Cecure, so working on Myarc HFDC's and Geneve's are certainly not the only equipment you've worked on.

 Would you say Geneve's and HFDC's are more 'fragile' then other cards or are they just more complicated so more things to break, er uh.. both?

 

[+InsaneMultitasker]

35 minutes ago, dhe said:

Tim,

 We know you worked at Cecure, so working on Myarc HFDC's and Geneve's are certainly not the only equipment you've worked on.

 Would you say Geneve's and HFDC's are more 'fragile' then other cards or are they just more complicated so more things to break, er uh.. both?

 

 

Some Geneve cards in particular are sensitive to flexing and there is at least one batch where the mask, pads and cores are extremely susceptible to heat and will peel away/pull out with little provocation.    HFDC cards are not nearly as bad, though the final batch of cards without the QIC02 circuit had many solder shorts where the mask didn't take properly.

 

Certain component choices including the use of single-wipe sockets, inadequate regulators sans heat sinks,  and the original HFDC transceiver LS31/LS32 chips resulted in premature and unnecessary failures.   In fact, replacing the original regulators, capacitors, and select components almost always yields a card that will run 24x7 for years without a problem.  So much of Myarc's original trouble with the cards could have been solved with slightly better components the first time around....

 

[GDMike]

And just think, how would this card look if it were all SM components today, IF all the components fit that description.

 

[Shift838]

During my visit with @InsaneMultitasker and testing my HFDC, I noticed that my 2nd HDD DATA connector was extremely bent and the cable was extremely hard to remove.  I decided to purchase a few 20 pin shrouded keyed connectors. 

 

I wanted to remove the standard dual pin headers and this would give me a chance to test my new dual solder gun/vacuum sucker.

 

Successful modification.  Dual solder gun/vacuum sucker worked like a charm.  See my finished results below.

 

for anyone wanting to know which one I ordered.  I got it from Amazon and for $130.00 I figured it was a good gamble.  First couple of times I tested it prior to this, it would clog up.  I increased the temperature a bit and when i'm done with using it I clean the nozzle with the nozzle cleaners before shutting it down. 

 

At Amazon:  HERE

 

Old pin headers removed you can see the  holes are nice and clean.  there were a few pins I had to reflow solder to get it complete clear.  This took me all of about 20 minutes to get done.

 

Notice the discoloration of the section of the board that was inside the PEB and the section that hung out.  Cleaned with 99.9% alcohol too.

 

Test Fit:

 

 

Completed and tested:

 

 

 

 

[globeron]

My Geneve (and Horizon3000) seems to be stable now, but still have some issues

and curious if other people have the same experience?

 

but I am not 100% if these are Geneve-card, VDP9938, GPL-mode,

software or video GBS8200 device related (or maybe 1x graphic mode giving issues?). 

I tried to compare it with F18A on my TI (and with emulators, like Classic99, JS99er)

(still need to try MAME Geneve 9640)

 

 

* 80 Columns

   OKAY - In MDOS mode everything seems to be normal with 80 columns

   

* ? - Tetris in a graphic mode give some graphic errors at the right screen.

 

* OKAY - MYART - seems to work okay as far as I see, for both graphic modes to load pictures like SWAN, TIGER, DRAGON, etc.

 

* OKAY - YAPP - seem to work okay (need a DSK1. with YAPP / DSR driver) to start.

 

 

GPL:

* GPL mode

   loaded XBEA modules

   80 Columns

   * OKAY --- INFO  (infocomm adventurer) loads okay for 80 columns (via EA option 5)

 

   * PROBLEM --- EXB80DEMO shows during start-up the 80 columns okay, then it shifts the characters

      (not 100% sure is it the Geneve or the GBS8200 device doing the conversion)

      or something with GPL.

 

      Does someone have the EXB80DEMO working on a Geneve?

      It is on the disk:

80-KOL.dsk

 

   *  MAYBE OKAY? -- LINES;O   (EA3)  then START, seems to be okay, but lines are a bit broken (graphic mode?)

 

   *  MAYBE OKAY? -- (via EA5) FRACTAL   (PAL) seems to be fine with output ?

       

** screenshots in next post (as I need to upload them from my phone)

  

Furthermore, it seems that QUIT does not work  (ALT-9  or ALT-0),   but have to use CTRL-LSHIFT-RSHIFT to go back to GPL mode.

(maybe this is normal)

 

Sometimes the software is kind of locked (or I cannot exit) and

CTRL - LSHIFT - RSHIFT   or CTRL-LSHIFT - DELETE do not work

and I need to turn off the PEB box and reboot from here.

(is that the reason people want to have reset button?)

 

 

 

 

 

 

 

[globeron]

The text is a bit shifted and you see at the lower right corner some pixels missing. When going back to Titlescreen also text is off. Then CTRL- SHIFT-SHIFT  resets the screen and then back to the titlescreen is normal again. Something to do with a video / graphic  mode or software?

[globeron]

There is somehing wromg on my Geneve, VDP9938, GBS8200 ?  as per previous post. 

 

These are the results with MAME emulator of@mizapf

 

[globeron]

Removed and put the VDP9938 chip back.

 

Now in MYART  to load he swan comes back like this.  Is this a memory issue or video chip issue?