Myarc cards for sale/repair tips

[+Schmitzi]

1 hour ago, InsaneMultitasker said:

You have the standard Geneve memory + 384K memory upgrade.   The PFM device is separate from the 384K RAM and is programmable flash EEPROMs.  The PFM+ is two x 128K Atmel 29c010 chips, as identified    Hooray, it worked!   Your PFM+ is a candidate for updating and loading MDOS 7.xx when released. 

 

ah thanks ! So this is 512KB Geneve BaseMemory + 384KB upgrade = the shown 896KB (917.404 Bytes)

Plus 128KB FlashRAM on the 2 Atmels (for quicker boot only)

and NO 32 KB SRAM (that I often read about) ?

 

 

[RickyDean]

8 minutes ago, Schmitzi said:

 

ah thanks ! So this is 512KB Geneve BaseMemory + 384KB upgrade = the shown 896KB (917.404 Bytes)

Plus 128KB FlashRAM on the 2 Atmels (for quicker boot only)

and NO 32 KB SRAM (that I often read about) ?

 

 

The 384k sram has that 32k built into it. So you could see it as 32k + 352k, though it is all contiguous memory. The Geneve has a 32k sram chip next to the eprom, then you would add the extra 32k on top to run the newer mdos's, but the 384k sram does that and more.

[+Schmitzi]

23 minutes ago, Schmitzi said:

 

ah thanks ! So this is 512KB Geneve BaseMemory + 384KB upgrade = the shown 896KB (917.404 Bytes)

Plus 128KB FlashRAM on the 2 Atmels (for quicker boot only)

and NO 32 KB SRAM (that I often read about) ?

 

 

 

10 minutes ago, RickyDean said:

The 384k sram has that 32k built into it. So you could see it as 32k + 352k, though it is all contiguous memory. The Geneve has a 32k sram chip next to the eprom, then you would add the extra 32k on top to run the newer mdos's, but the 384k sram does that and more.

 

 

ah OK Now I see on an old flyer that the standard base memory is 640KB RAM,

and 640KB + 352KB + 32KB is 1024KB and not the 896KB (917.404 Bytes) my CHKDSK shows on the screen above...

So there must be some more that I dont know, or I dont have the full base memory ?

Edith: Or is the 128KB FlashRAM added/shown by CHKDSK ?

 

 

 

[RickyDean]

9 minutes ago, Schmitzi said:

 

 

 

ah OK Now I see on an old flyer that the standard base memory is 640KB RAM,

and 640KB + 352KB + 32KB is 1024KB and not the 896KB (917.404 Bytes) my CHKDSK shows on the screen above...

So there must be some more that I dont know, or I dont have the full base memory ?

Edith: Or is the 128KB FlashRAM added/shown by CHKDSK ?

 

 

 

If you take 512k (the dram chips on the left side of the board) from 640, you have 128k, which I believe, is the video memory, which is the 4 dram chips under the video chip. I think the first 32k sram chip next to the eprom is used by the cpu. Nor sure where it comes into the total memory picture, but to use later versions of mdos, it was necessary to stack another 32k on top and add a wire. The 384k mod takes all this into it's construction. The Insane one can explain it better probably.

[+InsaneMultitasker]

27 minutes ago, Schmitzi said:

 

 

 

ah OK Now I see on an old flyer that the standard base memory is 640KB RAM,

and 640KB + 352KB + 32KB is 1024KB and not the 896KB (917.404 Bytes) my CHKDSK shows on the screen above...

So there must be some more that I dont know, or I dont have the full base memory ?

Edith: Or is the 128KB FlashRAM added/shown by CHKDSK ?

 

 

 

don't believe old fliers  

DRAM = 512k (the 16 chips on the LED side of the board)

SRAM = 32K stock;  can be expanded to 64K or 384K

videoram = 128k stock, can be upgraded to 192k

 

512DRAM+128K Videoram=640k.  Programs don't run from video ram.  The Geneve flier was in my opinion as misleading as some of the texas instruments /4a fliers that counted ROM, GROM, etc. 

37 minutes ago, RickyDean said:

he 384k sram has that 32k built into it. So you could see it as 32k + 352k, though it is all contiguous memory. The Geneve has a 32k sram chip next to the eprom, then you would add the extra 32k on top to run the newer mdos's, but the 384k sram does that and more

Nailed it. 

 

 

[+9640News]

48 minutes ago, Schmitzi said:

 

ah OK Now I see on an old flyer that the standard base memory is 640KB RAM,

and 640KB + 352KB + 32KB is 1024KB and not the 896KB (917.404 Bytes) my CHKDSK shows on the screen above...

So there must be some more that I dont know, or I dont have the full base memory ?

Edith: Or is the 128KB FlashRAM added/shown by CHKDSK ?

 

And the reason you don't have 1024KB but have the 917K ram, is the following.

 

On a stock Geneve, you have 64 x 8K pages which are what we call "physical" pages that can map into the mapper registers as pages >00 to >3F.  That's the first 64 x 8K pages or 512K.

 

Next, on the stock Geneve, you have 32K ram which is another 4 x 8K pages.  They are 4 of the physical pages from >E8 to >EF.  Not sure which of the 4.  If someone added the 32K ram expansion, then they have 8 of those 8K pages with physical pages >E8 to >EF all available.

 

The PFM provides pages >C0 to >EF or 384K ram.  Here is where you see the stock ram and PFM ram overlap.  I think it was determined and confirmed with some testing by others, the original 32K stock ram is still active, so really, the PFM only provides 384K - 32K or 352K.

 

Now while the chips are 512K, the memory in physical pages >F0 to >FF are ROM, not RAM.  The BIOS on a stock Geneve sits at physical page >F0 and >F1 and physical page >FF is used as a "null" or "dead" memory page that is not writeable.  With both the Memex and PFM, anything in the Physical page range of >F0 to >FF is not writeable.  The @InsaneMultitasker would need to confirm what is on the other pages that are dictated by what the PFM writes.  I think it is the PFM Flashdisk, but not completely certain.

 

If you add a modified Myarc 512K card, or a 504K Memex, I think they both use physical pages >80 to >BF with page >BA being mapped out as it is unique page to access the PEBox Card DSR's and chips. In addition, page >BC is hardware accessible, however the MDOS operating system removes that memory page from useable memory as if you have a Rave Speech adapter in your system, or some other more rare cards, page >BC gives mapped access to the speech synthesizer and would cause memory errors if you were to have a program loaded in that memory space.

 

Pages >40 to >7F are only accessible if you have a GenMod system.  With a GenMod system, it also opens up pages >C0 to EF like the PFM does.

 

Everything I mentioned above deals with pageable 8K memory blocks called "physical memory".  When we start talking task pages, then we are talking about the current memory map from address >0000 to >FFFF.

 

In MDOS mode, the mapper is at addresses >F110 to >F117.  You can if you so desire, map something like page >32 into >F110, >F111, F112, >F113, >F114, >F115, >F116, and >F117 effectively having each 8K bank identical  Make a change at >0500, and the change would be replicated at >2500, >4500, >6500, >8500, >A500, >C500, and >E500.  Or, you could map pages >32, >33, >34, >35, >36, >37, >38, >39 into the range >F110 to >F117 for 8 x 8K or 64K of contiguous ram.  The only thing to be careful here in MDOS mode is the TM9995 chip does have it's only RAM memory one needs to be careful.  That memory resides in a portion (not all) of the >E000 to >FFFF range.  Consult a TMS9995 manual for more details.

 

When a user loads a MDOS program, MDOS assigns pages to accommodate the file, and only it's file size.  So, if you load a program and it fits into memory and does not extend beyond address >2000, then only TASK page 0 at >F110 will be assigned in addition to a "default" page >13 (I think that is the physical page number) in task page 7 at >F117 for the >E000 to >FFFF memory range.  This is upper memory range is always assigned so the user can have access to workspaces.  If you need more memory beyond the immediate file size up to the top of that current task page 8K bank, then you need to use the Memory XOP to request more memory and map it in.

 

Now, when you go into GPL mode or "TI-99/4A Emulation Mode", the Geneve memory mapper locations move from >F110->F117 to address range >8000 to >8007.  The GPL interpreter also does some special mapping to give access to additional banks of memory at >6000 to >7FFF for cartridge space, plus all the GPL cartridge bank memory.  This is where the Geneve Gate array does a lot of its magic.

 

Anyways, this is a bunch of information.  If you program, it is probably worth it to copy this text and print it out for reference if this is new to you.  If you don't program, then you just read a message and are probably sitting here in a daze.  If you are in a daze and want to understand things more, then I highly suggest getting Bruce Hellstrom's BHDMV program as you can watch MDOS "at work", and have some limited access to playing with the memory map.  I used Bruce's programs years ago to learn some of the underlying abilities to multitask as well as to decipher how to page the MDOS Command Line Interpreter into a program.

 

Enjoy.

 

 

 

 

 

[+InsaneMultitasker]

1 hour ago, 9640News said:

he PFM provides pages >C0 to >EF or 384K ram.  Here is where you see the stock ram and PFM ram overlap.  I think it was determined and confirmed with some testing by others, the original 32K stock ram is still active, so really, the PFM only provides 384K - 32K or 352K.

 

Now while the chips are 512K, the memory in physical pages >F0 to >FF are ROM, not RAM.  The BIOS on a stock Geneve sits at physical page >F0 and >F1 and physical page >FF is used as a "null" or "dead" memory page that is not writeable.  With both the Memex and PFM, anything in the Physical page range of >F0 to >FF is not writeable.  The @InsaneMultitasker would need to confirm what is on the other pages that are dictated by what the PFM writes.  I think it is the PFM Flashdisk, but not completely certain.

The 32K stock ram is removed, as is the 64K (stock+upgrade) as the 32K chips cannot physically co-exist with the 3x128K (or single 512K) chip in the same socket.  What was determined is that the ram beyond the 64K is not FAST ram as once thought. 

 

Pages F0-FF are write-able, however, in all current incarnations there is an EPROM or EEPROM that responds to one or more of the pages.  The PFM chips are write-able from banks F0-FF, however, one must first initiate a write sequence to program the chip. 

 

[+dhe]

Thanks Beery,

 That post is a keeper, and one I think, I'll need to read multiple times.

  It really shows the hardware potential of the Geneve - I think two good examples are Genmod and Lou Riccio's Window's Demo.

  Just like most people think of the power little TI, as that console sold in K-Mart, they had no ideal of the potential hidden in the architecture that could go up to 640K with a RAM card, and be expanded in to infinity with a PBOX.

   We can only hope, that one day, FPo will have a zero wait state, full memoried out Geneve for us to use!

[+InsaneMultitasker]

8 hours ago, 9640News said:

And I will be testing my system tonight, and hopefully making the commit to update to the CORE V7 version.

 

Core version 8.01 has been sent your way with the updated command line utilities.  Good luck.   I've successfully updated two 29c010, a 29c040a, and a Winbond 29c040.

 

I'll hunt around for my ZIF socket tonight so that I can use it to hot-swap the 29c040 that I corrupted and with luck, reprogram it with the Geneve. Unorthodox process, alas, the 040 is the only chip type that hasn't been tested with the new core utilities and is the variety installed in the wild  

 

[+9640News]

21 hours ago, InsaneMultitasker said:

I found a spare PFM+ stack and installed it (replaced the existing chip stack) into my working Geneve.  Everything worked fine except for the PFM identification.  As best I can tell, the LS139 is interfering with the ID sequence when the flashdisk chip is selected.  This makes little sense except that the LS139 Select line may be flip-flopping the chips during the operation.  The fix for my Geneve was to force the 9901 to select the system chip prior to the ID sequence.

 

I've attached the identification program.  I would appreciate it if a few of you with the PFM512 or PFM+ could run it then share the resulting text, along with the type of PFM in your Geneve. 

 

PFMPROD  (tifiles format)

 

Edit: I updated the core standalone utilities and proceeded to upgrade the spare PFM stack from v3.0 to v8.0 then updated the installed MDOS from 2.21 to 7.22 - so far all is working as expected.

Tim, your updates worked fine on my PFM+ system, and I now have MDOS all 136K bootable.  Fast boot times.  That one line of code for a floating CRU bit fixed everything.  Who would have thunk two different Geneve's configured with the same hardware would have behaved differently.  

 

PFM Flash Disk burned to the Eprom without issue as well and the procedures with your notes worked perfectly.  I've got them printed out for the future!!!!

 

Great work.

[+Schmitzi]

12 hours ago, 9640News said:

And the reason you don't have 1024KB but have the 917K ram, is the following.

 

.....

 

Anyways, this is a bunch of information.  If you program, it is probably worth it to copy this text and print it out for reference if this is new to you.  If you don't program, then you just read a message and are probably sitting here in a daze.  If you are in a daze and want to understand things more, then I highly suggest getting Bruce Hellstrom's BHDMV program as you can watch MDOS "at work", and have some limited access to playing with the memory map.  I used Bruce's programs years ago to learn some of the underlying abilities to multitask as well as to decipher how to page the MDOS Command Line Interpreter into a program.

 

Enjoy.

 

 

 

great deal, thanks a lot Beery

 

 

And as I always need pictures, here is a picture from what I´ve gathered now

 

 

 

 

And as I always need a list, here´s a list. Maybe I should list my lists ? 

 

 

 

Maybe you all can have a look if all mods are given as written down here ?

Maybe there are more PFM-mods in the list than existant ? (I mean the big FlashDisks)

And maybe the Contributors are not complete ?

Sorry, this part of my list is very old, from its first time

 

ACK ACK ACK

 

 

 

 

 

[+InsaneMultitasker]

@Schmitzi  see picture for a few changes/comments.  

 

[+InsaneMultitasker]

On 4/2/2021 at 8:15 PM, 9640News said:

Tim, your updates worked fine on my PFM+ system, and I now have MDOS all 136K bootable.  Fast boot times.  That one line of code for a floating CRU bit fixed everything.  Who would have thunk two different Geneve's configured with the same hardware would have behaved differently.  

 

PFM Flash Disk burned to the Eprom without issue as well and the procedures with your notes worked perfectly.  I've got them printed out for the future!!!!

 

Great work.

Thank you for subjecting yourself and your Geneve to the test     This afternoon I was able to resurrect the Atmel 040 that I corrupted which coincidentally jump-started my work on the PFM utilities.  I took a picture of my adhoc Geneve 9640 programmer.  <grin>   It took me a few minutes to remember that I needed to add another 32-pin socket to raise the height enough for the ZIF socket to clear the adjacent 64K memory stack.   And then it only took 5 times to get the 040 into the ZIF without the Geneve protesting via lockup that I was hot-swapping chips whilst still running!    The good news is that the programs all work with the 040, so that question is now answered.

 

[+Schmitzi]

10 minutes ago, InsaneMultitasker said:

@Schmitzi  see picture for a few changes/comments.  

 

 

This is what I did read off from my card

Maybe it´s because it´s like this piggy thingie mounted on top of the ones that you figured here ?

So I can add this as the original chips to my sheet, "under" the mod-chips ?

 

 

Edit: Here:

 

 

 

Turned around:

 

[+InsaneMultitasker]

Just now, Schmitzi said:

 

This is what I did read off from my card

Maybe it´s because it´s like this piggy thingie mounted on top of the ones that you figured here ?

So I can add this as the original chips to my sheet, "under" the mod-chips ?

 

What I documented in the picture is accurate.   Labels can be misleading.  Sorry, I do not know what you mean by your last sentence and "under" the mod-chips...?   

[+Schmitzi]

Just now, InsaneMultitasker said:

What I documented in the picture is accurate.   Labels can be misleading.  Sorry, I do not know what you mean by your last sentence and "under" the mod-chips...?   

 

yeah I think I confused it? because imprints is upside down.

 

"Under" means there are chips under the ones that you can see or at least I thought that it is "piggy back/stapled"

The pics are old an when I tokk them I did not know that the perspective from the side may be important.

I only have 1 pic where I can maybe see if chips are "stapled". Sorry my englitsch is beta

So I have to dismount the system to see if this are sockets or other chips under the one I have written down

 

This is the one pic, not taken from top but from side view:

 

 

 

 

[+InsaneMultitasker]

Just now, Schmitzi said:

"Under" means there are chips under the ones that you can see or at least I thought that it is "piggy back/stapled"

The pics are old an when I tokk them I did not know that the perspective from the side may be important.

I only have 1 pic where I can maybe see if chips are "stapled". Sorry my englitsch is beta

So I have to dismount the system to see if this are sockets or other chips under the one I have written down

Yes, there are chips underneath -- they are the same type of chips as what you see on top.  There is only one "modification" per socket. 

Your Geneve has the 384K SRAM upgrade (3 chips of 128K stacked) and PFM+ (two 29c010 Atmel).  Five total chips; two total modifications. 

 

[+Schmitzi]

 

10 hours ago, InsaneMultitasker said:

Yes, there are chips underneath -- they are the same type of chips as what you see on top.  There is only one "modification" per socket. 

Your Geneve has the 384K SRAM upgrade (3 chips of 128K stacked) and PFM+ (two 29c010 Atmel).  Five total chips; two total modifications. 

 

 

 

ah OK, I just confused/swapped my own chips, they are this way:

 

 

And the "M5M5256P" and the "27C128" are correct (not swapped) and they came from this logo pic I found:

 

 

 

 

[+Schmitzi]

 

yup, hope it´s better

 

(best learning for me, to make changes on such things by hand, for 1.000 times)

(it´s like my girlfriend has to tell me 1.000 times to take the trash out)

 

[RickyDean]

Ok!! Back in business, got my diodes in and started putting my little plan into action. Got this far, just need to connect wires to the pins of the sockets and then, if all goes well, sram!!. I am doing this to be able to replace the 512 sram or the 134 chip if one or the other fail. Plug and repair.

[RickyDean]

O @InsaneMultitasker, I am close to getting my sram project finished, but have a question or two. Then when finished I will continue getting the flash chips operational. In Fabrice's schematics it states " Solder the blue wire from pin #3 of the 74ls138 to the pin #40 of the gate array (AMD signal). But in the 512 doc from Chicago Faire it states "Pin 3 (LS138) to Gate Array pin #42 [Blue]. I have a red wire going to pin 40 not 42, that went to pin 3 (ls138). discrepancy or another way? Also the 1k resistor does it go between 20 and 28 on the 28 pin original sram location? And if you look at some of my prior pics, Geneve backside, the red wire was tied to a 1k resistor at pin 27 along with a white wire. Does that belong there? Thanks for helping.

[RickyDean]

[+dhe]

Wow, that takes soldering a couple of jumper wires to a whole new level.

    Can you say DAUGHTERBOARD! ?

Anyone that can build that, and rebuild an engine is a true renaissance man.

[RickyDean]

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.

[+arcadeshopper]

On 1/30/2021 at 8:14 PM, InsaneMultitasker said:

Yes, though the preferred course of action is to replace it considering the potential force applied to remove it. Is the through-hole core intact?

Ok well I can't really tell, so I did a continuity test to the resistor above from the back of the hole and it beeped, so I jumpered it up to the resistor.. 

 

Geneve boots, but after loading system/sys it hangs up  tried both from floppy with boot rom 1.0 and boot rom .99 and tipi.. tipi shows it loads system-sys fine and then nothin

boot rom 1.0 menu works.. so video/keyboard/whatever ram it uses is good I assume 

 

no LEDS light up at all.. cept the ones on the genmod switch box

 

Ps: it took me this long to try.. scary geneve