Myarc cards for sale/repair tips

[RickyDean]

I think the greater understanding of the questions were on page 20: https://forums.atariage.com/topic/279153-myarc-cards-for-salerepair-tips/page/20/#elControls_4802819_menu

 

[RickyDean]

On 4/17/2021 at 4:58 PM, RickyDean said:

Gonna try to do something similar to the 74ls139 for the flash chips, but if not, then a daughter board for the sound chip and 139, then they can be changed at will if there is an issue with them. Shucks, I'm cutting 7.1 inches out of a 4x4 Ford Excursion frame, then going to place a 05 Chevy Avalanche body on it with the Avalanche engine and a HD Chevy 4x4 transmission and Transfer case. Full leather, Navigation, and sun roof. NICE looking 2wd truck as it is. Keeping busy for sure.

About got this accomplished, the Avalanche is running, just have to build an exhaust system. Between working a lot of overtime, the Georgia heat, family stuff, and now the shoulder issue, things haven't progressed as well as I'd like. If I can finish this Geneve while I'm down that'd be an accomplishment.

[+InsaneMultitasker]

5 hours ago, RickyDean said:

I'll have to reread too, but the idea was to move the 138 and 139 from the top of the sound chip for access, and all my chips can be easier to replace, at the same time. Thanks

Ah, ok, that makes sense. Troubleshooting involving the sound and clock chip was a royal pain in the neck once that mod was installed.  

[+InsaneMultitasker]

On 4/21/2021 at 9:59 AM, RickyDean said:

Also, when you do answer those, I have one more: the pins, 13-15, on the 74ls139, that are soldered to the same pins on the 74ls138, on my pfm+, where should they go if I've moved the '138? Should I run wires to the same pins on the 138, in its new location?

I took a quick look at my notebook and datasheets. 

 

The LS138 and LS139 both require VCC (16) and ground (8).  These pins are connected via the sound chip leads in the stacked configuration.

 

LS139 pins 13-15 (input/select) are clipped and tied to VCC pin 16. Pins 9-12 are clipped and NC (these are unused outputs). Pins 1,2,3,4,5 are used for the select logic from connection points. Pins 6,7 NC. Pin 8 ground.  There is a resistor between pins 2 and 8.  There is a solid wire between pins 3 and 8.  [Note: if your pins 13-15 are tied to the LS138, that's probably ok since the LS138 has outputs here but we wouldn't want them tied to the input/outputs and VCC]

 

LS138 pins 9-15 are clipped.  Pins 9,10,11 are tied to diodes for RAM chip select. Pins 1,2,3,6 are inputs from the LS244 and gate array.  Pin 4 is bent to pin 5; pin 5 connects to the sound chip, and I believe it is either GRND or VCC. You'll need to verify.  Pin 7 NC. 

 

A few more rambling thoughts... the instructions I provided to Fabrice were for the single chip 512K mod (that gives you 384K). I don't have any clear instructions for the triple stack and it quickly fell out of favor near the end of the run, since the 512K was much simpler to contend with.  Both the PFM+ and 3x128K upgrades required case modifications, which I had no interest in doing. 

[RickyDean]

14 minutes ago, InsaneMultitasker said:

A few more rambling thoughts... the instructions I provided to Fabrice were for the single chip 512K mod (that gives you 384K). I don't have any clear instructions for the triple stack and it quickly fell out of favor near the end of the run, since the 512K was much simpler to contend with.  Both the PFM+ and 3x128K upgrades required case modifications, which I had no interest in doing. 

Yeh, that's why I decided to go with the 512k sram too. to get the stack down a bit. Thanks, I'll look it over and see where my red wire and white wire go then.

[+InsaneMultitasker]

8 minutes ago, RickyDean said:

Yeh, that's why I decided to go with the 512k sram too. to get the stack down a bit. Thanks, I'll look it over and see where my red wire and white wire go then.

Yep. Just take your time and it should all work out.  And to clarify: the 3x128 and 512K mods are not that different; for the 3x128K, the chip select for each SRAM comes directly from the LS138, versus tying the three outputs via a diode to the single 512K chip CS.

Edited November 9, 2023 by InsaneMultitasker
Clarity...maybe?

[RickyDean]

On 11/8/2023 at 9:13 PM, InsaneMultitasker said:

Yep. Just take your time and it should all work out.  And to clarify: the 3x128 and 512K mods are not that different; for the 3x128K, the chip select for each SRAM comes directly from the LS138, versus tying the three outputs via a diode to the single 512K chip CS.

Ok, think I'm getting there, but do have to ask this question. On the 29C010 on the CS, the bottoms one is enabled from the old socket location, the top one comes from a pin from the bottoms 74ls138 chip from pin #5 on your picture, posted her on Atariage some time ago. Mine seems to go possibly to pin #7 on the same chip. which one is right? thanks.

[+InsaneMultitasker]

2 hours ago, RickyDean said:

Ok, think I'm getting there, but do have to ask this question. On the 29C010 on the CS, the bottoms one is enabled from the old socket location, the top one comes from a pin from the bottoms 74ls138 chip from pin #5 on your picture, posted her on Atariage some time ago. Mine seems to go possibly to pin #7 on the same chip. which one is right? thanks.

According to my notes, the top 29c010 chip's CS ties to the LS139 pin 5.  Pin 7 is N/C.

 

[+InsaneMultitasker]

@RickyDean the stacked LS chips are a LS138 and LS139.  I re-read your question and want to be sure that you realize this as there was a time when Don rubbed the identifiers off of all the chips that were being installed for the PFM and 384K memory mods.

[RickyDean]

40 minutes ago, InsaneMultitasker said:

@RickyDean the stacked LS chips are a LS138 and LS139.  I re-read your question and want to be sure that you realize this as there was a time when Don rubbed the identifiers off of all the chips that were being installed for the PFM and 384K memory mods.

Yes, Mine was one of those from the bottom photo you see. If you look back about a year or so ago I changed those positions, moving the 138 under the 512k chip on a small motherboard and building a small mother board to take the 139 and sound chip apart and make each easy to replace. I believe I was using  @fabrice montupet's method, but making it somewhat similar to what Shift838 had done, only with removable IC's, but utilizing my existing wire framework . From what I can tell the two white wires and red wire I had coming out from  above the location above the R63 go to the pins on the 138 under the 512k chip now, and a red wire from JM3 went to the pin 1 of the 512k. I don't know if I still needed a wire from the position of C128 to wherever that was supposed to connect on the old mod, and I needed to know for sure which pin the CS on the top 29C010 needed to go. It still looks to me like it went on or around the pin 6 or 7 on 74ls138 on the photo of my version, and I couldn't figure out if that wire was grounded or not, would have to be if it had been connected to pin 7, but the pictures aren't the highest quality and my memory of the work back then is not great.

I'm prepping for my shoulder surgery this coming Wednesday and trying to get as much done as I can before I lose use of my right arm for 6 weeks or so. I was hoping to get this baby up and running, so that I could just play on the computers, Geneve included, and watch movies, till I heal up.

[RickyDean]

A@InsaneMultitasker, I have to retract my statement a bit it's pin 8 that is grounded, not 7 on the 2nd level 74ls138 chip in my photo, but as stated it seems that the blue wire coming from the top 20C010 feeds back to that area in the photo, if zoomed up close, it really seems like it ties to pin 8, so maybe it fed under the chip for some reason? It was not detached and lying there as the Geneve would boot up if a red wire, on the back of the computer, that was loose was tied to a +5 volt location, if I remember correctly. I am going to try tying the wire from pin 22 to the 138, pin 5 and see what happens. I've got to find out exactly how this resistor and the red wire are supposed to be connected as it had been disconnected for a while and the computer would only black-screen. I replaced the resistor and believe I need to run it like your photo?

 

At the moment here is the current state of that resistor.

 

Edited November 26, 2023 by RickyDean
added content

[+InsaneMultitasker]

2 hours ago, RickyDean said:

I've got to find out exactly how this resistor and the red wire are supposed to be connected as it had been disconnected for a while and the computer would only black-screen. I replaced the resistor and believe I need to run it like your photo?

Your install might have been a variant of the same theme, where things were shifted around for easier installation.  Re-read my post from November 8 and see if that addresses these last questions, as I tried to map out the connections by chip and pin.

[RickyDean]

@InsaneMultitasker well I haven't seen these screens in a while but I can't seem to go past them. Probably don't have the PFM+ wired right. Have to tinker. Going back to the pictures above should the resistor in the third picture be tied like the the one in the second, your picture? That will be a start.

 

 

Edited November 27, 2023 by RickyDean

[+InsaneMultitasker]

Congrats - you're getting close!  This is indicative of one of two things:  the 384K memory is not present or you need to ensure you have a bootable Os on the disk.

The PFM device looks for SYSTEM/SYS.  If you have the file on disk, select either FDC (for a floppy controller) or H for the hard drive controller.  "D"efault simply refers to the SYS suffix.  Do NOT attempt to write the OS via the PFM screen.  When you are prompted, select No to write, and let it start MDOS. If you can start MDOs, the next steps will be to update the CORE and install the OS with the files in the MDOS 7.40 PFM folder. 

[+InsaneMultitasker]

3 hours ago, RickyDean said:

Going back to the pictures above should the resistor in the third picture be tied like the the one in the second, your picture?

Yes, this makes sense to me.  The via looks to be in bad shape. Some of the jumpers are tied to an internal trace, so you may need to test for continuity before completing the repair.  You probably won't be able to write to the PFM until this is in place.  

[RickyDean]

1 hour ago, InsaneMultitasker said:

Yes, this makes sense to me.  The via looks to be in bad shape. Some of the jumpers are tied to an internal trace, so you may need to test for continuity before completing the repair.  You probably won't be able to write to the PFM until this is in place.  

On JM3, on mine pins 1 and 3 are jumpered with a small piece of wire. The via is shot on pin 2 location. There is a red wire soldered to the pin 1 location along with the jumper wire and I soldered the other end to pin 32 of the 512k stack as that is where I thought it was for the  384k stack from my photos. the resistor that was probably there had been pulled loose during transit years ago and I actually think I asked about it some years ago here on Atariage, when I was first trying to get this working as the original PFM+ and 384K. I had tied it elsewhere thinking that was the location and then tied it at the location in picture 3 before I began the 512k journey. So do I need to find out what layers connected to pin 2 and then find a suitable location that interacts with this pin? To tie in the resistor, I presume, is what your saying in the above statement?

Edited November 27, 2023 by RickyDean
spelling

[+InsaneMultitasker]

I located an older version of my instructions for the Geneve 384K modification.  I am fairly certain there is/was a Chicago Faire release version that I had validated.  And for all I know, I uploaded that version to the forum long ago.  Anyway, I printed the Word doc as a PDF.  And since this is my document, I am blatantly disregarding the 'not for distribution' footer.   

 

Myarc Geneve 384K RAM installation 2006TT.pdf

 

And here is a 25+ year old quick sketch from my Myarc notebook and a trace cut reference picture. 

   

[Shift838]

55 minutes ago, InsaneMultitasker said:

I located an older version of my instructions for the Geneve 384K modification.  I am fairly certain there is/was a Chicago Faire release version that I had validated.  And for all I know, I uploaded that version to the forum long ago.  Anyway, I printed the Word doc as a PDF.  And since this is my document, I am blatantly disregarding the 'not for distribution' footer.   

 

And here is a 25+ year old quick sketch from my Myarc notebook and a trace cut reference picture. 

 

In 2021 I created a new 384K documentation using one of my designed modules taken from some old documentation.

 

 

838-384kModule.pdf

[+InsaneMultitasker]

1 hour ago, Shift838 said:

In 2021 I created a new 384K documentation using one of my designed modules taken from some old documentation

Nice. I intend to install my Shift838 384K memory this month.  I finally located the v1.0 chip and the wires on the same day! The two packets were in different storage areas that I was poking around today. 

[+InsaneMultitasker]

Here's a particularly obnoxious gem of a problem.  This Geneve would periodically stall as if the interrupts weren't being serviced, artifacts would show up on the screen, screen would completely disappear.  Initially, it seemed like a 9938 / vram issue.  The picture below shows as resistor with some of the insulation chipped away in line with the fourth band, nestled tightly against an oscillator/crystal.  Separating the two seems to have stabilized all the video-related issues.  I ran the card for an hour and noticed the keyboard responses were gibberish after it completed full TREE operation, so there is at least one more gremlin to coax out of its hiding place.  (Edit: better,larger picture)

 

[+InsaneMultitasker]

On 1/29/2023 at 1:35 AM, InsaneMultitasker said:

One last thought for those of you with battery-operated devices like the Horizon RAMdisk:  if you haven't checked your batteries in a while, take a moment to inspect them.   While I did not have any leaking batteries, my Horizon Ramdisk 3000 uses alkaline batteries that are close to needing replacement.  I haven't used my Geneve since last summer, so there has been a constant draw for months

As Robert Jordan wrote, "The Wheel of Time turns, and Ages come and pass"... 

On 8/13/2023 at 5:05 PM, InsaneMultitasker said:

The 3000's standard alkaline batteries are still supplying 4.1v after all the system downtime, and as far as I recall it has been 2.5+ years since I last replaced them.  I'll check them again in the fall and replace at the 3 year mark. 

I got busy; fall came and went.  Although the data is still intact, I backed up the disk and 'hard drive' partitions then moved the Ramdisk to a PEB slot that would give me room to change out the batteries with the PEB powered on.  When I pulled the batteries, I discovered that I had written "5-20" on them, meaning they have been in service 3 years and 9 months!    The three old batteries measured 3.8v (total) and the new batteries measure 4.75v. 

 

This following linked article suggests to me that below 1.2v, the battery may drop voltage quickly.    Comparing AA type alkaline batteries with rechargeable Ni-MH batteries / Antonovich.me   Since the 3000 sips power, I could have eked out a few more months but I prefer the peace of mind.  This time, I made certain the installation date-year is visible so that I can more easily check the battery age.

 

I tested the 4000B batteries and it's rechargeables are still going strong.  @Ksarul is there a recommended timeframe for replacing the nicads.

[+dhe]

I'm pretty good at documenting. Labeling batteries is a very good idea.

Something in Florida makes most batteries leak very quickly. I've pretty much cut over to using all lithium batteries where I can, I have yet for them to leak on me.

[+OLD CS1]

1 hour ago, dhe said:

I'm pretty good at documenting. Labeling batteries is a very good idea.

Something in Florida makes most batteries leak very quickly. I've pretty much cut over to using all lithium batteries where I can, I have yet for them to leak on me.

After having years of Energizer batteries leak in every-damn-thing, I have switched to other brands, but mostly lithium and rechargeable.  The worst part is much of the leaking was not happening after the units had been sitting for a long time unused.  My thermostat, for one.

[+dhe]

Heat, Humidity, some residual salt in the air?

I've had batteries leak on equipment, while still providing enough power for the gadget to work. Go figure.

[+InsaneMultitasker]

Here in TX I've had pretty good luck with most batteries, and have stuck with Energizer after having the opposite luck with another common name-brand.  Invariably, it seems to be the remote controls and toys that end up with leaky batteries.  Checking the batteries periodically is good practice.  I've not gone the lithium route outside of the coin cell batteries or devices where they are 'baked in'.