U1MB + SIDE2 (on XEGS) issues

[phoenixdownita]

This is partially covered in other topics, I just want to "summarize" my experience so maybe someone can help.

 

I have an XEGS and I bought SIDE2 and U1MB form lotharek.

I installed U1MB myself [needed to trim the ICD adapters as on XEGS there's not much space].

 

My U1MB came flashed with the following OSes:

1)HiSpeed

2)Stock XL

3)QMegs

4)Stock XE

 

I did put a jumper on the U1MB XEGS mode pin header bridging the pins, I believe that is what is needed for XEGS mode.

No matter what OS I choose I cannot get to Missile Command, I believe for this to work I need an XEGS os rom in slot 4.

 

Now onto SIDE2.

If I disable SDX and SIDE HW on U1MB I can use the embedded SDX mode with FDISK to prepare the cart, and then put XEX and ATR on the FAT32 portion, then with the switch in Loader mode I can load XEXs fine.

 

now on U1MB + L (SIDE2 switch set in Loader mode):

1) Side HW Off, SDX Off, attempt to start an ATR => no PBI driver/device error [expected]

2) Side HW On, SDX Off, attempt to start an ATR => xxx not mounted error [expected, SDX seems necessary for this]

3) Side HW On, SDX On, attempt to start an ATR => led flashes then goes to basic prompt [NOT expected]

 

because of 3 I cannot flash U1MB as I cannot start any ATR, so I'm kind of stuck.

[i tried both Side HW On settings with and w/o button support, no difference]

 

During all this XEXs always work from any Side loader (either the SIDE2 or the U1MB + L).

 

Even more details, I can only FDISK thru SIDE2 SDX mode (and all settings off for the U1MB).

Otherwise U1MB keeps on complaining that it cannot recognize the device (in this case I have Side HW on, SDX on but I simply don't press L, I just get out of the U1MB setup menu).

 

I believe the issue may be in the U1MB PBI drivers, that are there but have issues maybe in recognizing the CF (and I only have 1 4GB DaneElec to play with), or alternatively in the U1MB SDX drivers that don't play nice with the PBI and fail to mount the ATR, not sure which is which.

 

For completeness pay attention that if you prepare the CF with MyIDE2 the APT partition is not compat with SDX on SIDE2. I noticed that the sizes reported are half on SIDE2 SDX than they are on MyIDE2 SDX [there's a SDX rom to which you can add MyIDE2.sys for MyIDE2] and SIDE2 SDX then fails to format them.

 

So I decided to stick to just one simple FAT32 as all I care are XEXs and ATRs and apprently U1MB should be able to support them both from the FAT32 partition (with SIDE2).

 

Because of the IDE register clash MyIDE2 cannot run as PBI device for U1MB, I know that, but the U1MB extra mem is usable from within MyBIOS just fine, so I was trying to use U1MB as memory expansion + multiOS.

 

What was I trying to achieve?

 

Having U1MB flashed with a XEGS stock bios so I can use my XEGS as normal XEGS (the onboard Stock XE does not seem to work for me) + MyBIOS R2/RE so I can use it with MyIDE2 [i like R16 support] because MyBIOS R2 compat is much better than MyBIOS F2 (soft OS versus rom os).

 

I bought SIDE2 because I thought it would be easy to make it work with U1MB as a way to reflash it but I got stuck.

 

I am now using the same CF among SIDE2 and MyIDE2 and YES I did try a complete ground up FDISK/Format cycle for CF completely from the SIDE2 to make sure only lotharek hardware was involved .... and nothing changed, so I went back to use MyBIOS to do the disk preparation (because it needs a special partition for its own support of mounted ATRs) and simply "share" the CF FAT32 portion among the 2 devices.

 

So to summarize:

U1MB + SIDE2 -> not capable of ATR launch from SIDE Loader, in specific it appears the SDX image in U1MB cannot recognize SIDE2 device (not sure if it may be the particular CF).

 

U1MB stand alone -> not capable of running as a stock XEGS, I believe because the Stock XE OS is not XEGS

 

U1MB + MyIDE2 -> can use the extra mem (I tried commando 320K and TowerToppler 128K file version and they worked), but because can't flash OS then can only use MyBIOS F2 with reduced compat.

 

I contacted lotharek about a week ago with no answer yet, now I ordered an (E)EPROM flasher in the hope that I can use the ROM generator output rom to reflash U1MB and hopefully fix the blunder ..... just not exactly a cheap solution (MyIDE2 FAT32loader allows MyIDE2 self flashing, I wonder if we can have SIDE loader on U1MB to embed a feature like it so as long as SIDE loader can see the file then it can flash U1MB).

I'll contact him again maybe he was just busy with life, I just checked here in case someone as a quick solution, keep in mind I have no access to SIO2PC, SIO2SD or any other device that can simulate real Atari devices, I only have SIDE2 and MyIDE2.

So please do not post anything about lotharek services, it's been too early and maybe he has a solution, for now I'm just being impatient. I like his skills, and the products look fantastic.

 

Anyone else experiencing my issues with U1MB + SIDE2?

If so did trying another CF help in any way?

Edited November 23, 2013 by phoenixdownita

[Kyle22]

I can answer Question #1:

 

You need the XEGS OS in slot #4. Use only the 16K OS part, NOT the full 32K contents of the ROM chip.

 

Try this: http://www.theoldcomputer.com/roms/getfile.php?file=Li9BdGFyaS84Yml0L29zL0F0YXJpJTIwT1MlMjBSZXYlMjA0JTIwJTI4MTk4NyUyOSUyOEF0YXJpJTI5JTVCWEUlMjBHYW1lJTIwU3lzdGVtJTVELnppcA==

 

Atari_OS_Rev_4_(1987)(Atari)XE_Game_System.zip

[phoenixdownita]

I can answer Question #1:

 

You need the XEGS OS in slot #4. Use only the 16K OS part, NOT the full 32K contents of the ROM chip.

 

Try this: http://www.theoldcomputer.com/roms/getfile.php?file=Li9BdGFyaS84Yml0L29zL0F0YXJpJTIwT1MlMjBSZXYlMjA0JTIwJTI4MTk4NyUyOSUyOEF0YXJpJTI5JTVCWEUlMjBHYW1lJTIwU3lzdGVtJTVELnppcA==

 

 

Cool, at least I got an answer to the XEGS mode "incompat", namely I need the XEGS OS.

I guess the dip switch only tells U1MB what OS rom size to expect on the board (16K vs 32K) so it knows if it can use the extra addressing pin on the rom chip, likely also subtle MMU differences, it makes sense.

 

Thanks for the download, I will certainly use it once I get the programmer.

Edited November 23, 2013 by phoenixdownita

[Kyle22]

 

Cool, at least I got an answer to the XEGS mode "incompat", namely I need the XEGS OS.

I guess the dip switch only tells U1MB what OS rom size to expect (16K vs 32K) and possibly subtle MMU differences, it makes sense.

 

Thanks for the download, I will certainly use it once I get the programmer.

I believe the jumper is involved with MMU operation. It is my understanding that ALL OS slots are 16K, with BASIC and games being 8K each.

 

I re-wired my 1200XL as an XEGS, used the jumper, and it worked.

[phoenixdownita]

I believe the jumper is involved with MMU operation. It is my understanding that ALL OS slots are 16K, with BASIC and games being 8K each.

 

I re-wired my 1200XL as an XEGS, used the jumper, and it worked.

It's also the ROM, the chip on the XEGS is a 32K, while for the rest it is a 16K, but they all have the same number of pins.

What happens is that the 32K chip uses 1 unused pin as A14, while for the 16K is simply NC, I've noticed it in the EPROM versions (27C256 vs 27C512), the 27C512 moved the VPP [programming] pin into a different spot and reuses the 27C256 VPP as A14, the ROM have the same pinout, they simply do not support programming [VPP].

 

That is the reason I said U1MB has to know the ROM size if it needs to drive A14 at all [XEGS mode].

Edited November 23, 2013 by phoenixdownita

[Kyle22]

It's also the ROM, the chip on the XEGS is a 32K, while for the rest it is a 16K, but they all have the same number of pins.

What happens is that the 32K chip uses 1 unused pin as A14, while for the 16K is simply NC, I've noticed it in the EPROM versions (27C256 vs 27C512), the 27C512 moved the VPP [programming] pin into a different spot and reuses the 27C256 VPP as A14, the ROM have the same pinout, they simply do not support programming [VPP].

 

That is the reason I said U1MB has to know the ROM size if it needs to drive A14 at all [XEGS mode].

Yes. exactly, that's why I re-wired my 1200XL for XEGS compatibility so I could have a free socket to put a MyIDE internal in. The OS slots in the U1M are 16k, though.

I was trying to make the point that the jumper changes how the U1M interacts with the computer hardware, NOT the size of the ROM to flash to its OS slots.

[phoenixdownita]

Lotharek is assisting me in finding the root cause.

For now except failing to start ATR it all looks normal to him.

 

SIDE2 works as expected, SDX works only when launched from SIDE2, cannot use U1MB onboard SDX in that case.

 

Missile Command access -> set Basic to CAR1 and SDX off -> missile command starts as expected.

 

I guess we are ruling out obvious things. He also suggested me NOT to flash as I have the latest firmware

(given my goal is to have MyBIOS R2 in it I need to flash eventually but until my programmer arrives I'll obviously not touch anything).

 

We'll see what comes out of it, I really wonder if it can be due to some bad soldering or bad contacts in the ICD cable adapters, if it is even possible.

For sure it is bizarre that everything else works but this one feature, maybe it could be the CF but then again it works perfectly for XEXs loading.

Edited November 24, 2013 by phoenixdownita

[phoenixdownita]

An update about there being no further updates so far

 

Lotharek asked me what other mod I have (NTSC XEGS + U1MB is all that there is to it) and then asked me pointers to the ATRs that don't work (given it's all of them is pretty easy, but I pointed to Ultimate2.atr on his site as an example, and all the "Gyruss" and "Mario Bros (v1)" over at atarionline.pl).

 

I have not heard since but it's been only 2 days, again just me being impatient, so I do not know what/if he's investigating or if he has been able to repro the problem on his end and maybe working on a fix, if there's something to fix that is.

[phoenixdownita]

I just received my TL866CS burner.

I will attempt to rebuild my U1MB flash (SST39SF040) and hope that it works.

(Lotharek has not responded to me since he's been made aware of the issue, not sure if he's still investigating or ...)

 

Alternatively I noticed that my SIDE2 has the good old AMD flash chip (AM29F0408), if shit comes to shore I will try to dump its content and flash it instead.

Not sure I can put the SST39SF040 in the SIDE2 but I figured that attempting to swap the chips may just work, assuming my issues are due to the SST39SF040 rather than something else.

 

I'll post something once I get around it ..... definitely it sucks to need to buy an EPROM burner for this, an XEX based solution would be much better given U1MB has plenty or RAM to load the whole 512K of ROM and flasher, and given the fact I can at least load XEX from SIDE loader just fine.

Edited December 1, 2013 by phoenixdownita

[phoenixdownita]

DEFEAT

 

I reflashed from the v2 U1MB stock rom replacing 2 OSes, one with stock XEGS and one with MyBIOS R2.

 

On the good side, both OSes work and myBIOS R2 can see MyIDE 2 just fine and it all works as expected, also XEGS stock OS behaves on U1MB as real XEGS, booting to Basic if keyboard pluggend and booting to Missile Command if keyboard not plugged ... everything as expected there, very good.

 

This ends the nice part, I still CANNOT mount ATRs over SIDE2, it just goes to BASIC (I can hear the "purr" then eventually the READY prompt).

FYI If I press space I can see the "D1" appearing and so on, it's just when I press Enter that I get around 3 sec of "fast purring" on a black background then blue background, then "slow purring" and in the end READY appears.

 

So I'm stuck where I was but with the added extra 2 OSes, I was hoping to restore ATR support on the SIDE2 part.

 

Anyone has an idea of what can be flaky/wrong/misconfigured?

 

I repeat everything else on U1MB appears to be working just fine: OS selection, memory expansion, XEGS game slot, "L" SIDE loader ...

 

I don't even know if the issue is on U1MB or in SIDE2 at this point or maybe SW side as in the SIDE loader or a PBI driver.

 

It's possible there's a bug in SIDE loader on U1MB. I noticed that the one inside SIDE2 and the one inside U1MB are not the same, the difference is obvious once I "updated" SIDE2 with the latest Candle loader with rom support (http://atariage.com/forums/topic/216189-sideside-2-cartdidge-firmware-update/).

The rom support is only there if booting from SIDE2 (U1MB set as mem exp only), if instead I go thru U1MB (SIDE HW on w/button + SDX on) then I press "L" then no rom support.

 

One more piece of info: If I set U1MB as Side HW on w/ button but No SDX, then when I press "L" I go to Basic, is this expected?

 

Help!?!?!

Edited December 1, 2013 by phoenixdownita

[phoenixdownita]

Lotharek suggest I reprogram the U1MB CPLD via a Xilinx JTAG programmer .... not exactly a walk in the park given I don;t have one, there's a parallel port version for around 10$ and I found an old laptop in the house with parallel port.

I am not sure at this point how much of a trouble it's going to be given U1MB JTAG is not even a std 2.54mm pitch connector but a 2.00mm which has its own set of issuse, basically I will be forced to solder cables on the pads rather than use the comnnector .... will see, I don't have the programmer yet anyway so it is not like I am in a hurry.

 

Still so far this experience is far from "plug, solder only 4 signals and play", assuming reprogramming the CPLD fixes the issues of course.

I will report if/when I manage to try yet another "fix" ....

 

To be fair Lotharek as offered to reprogram my U1MB for me by sending it back but it probably costs more in shipping than attempting to buying one of this parallel programmers. If that doesn't fix it though I guess my U1MB has a serious issue with this ATR business.

[flashjazzcat]

Probably a daft question but after you mount the ATR in the SIDE loader (so it has a "D1" next to it), what's the exact next step you perform?

[phoenixdownita]

Press Enter.

[phoenixdownita]

Got a Xilinx USB cable and validated CPLD version with Ultimate-v2.jed file that Lotharek sent me (and also available on website, I did a fc /b on the 2 files and they are the exact same), and it all matched. I really was hoping it didn't as that would point to an old CPLD core.

Anyway I erased and reprogrammed it twice and validated each time the CPLD was erased and once reprogrammed that it matched the checksum for the file and of course it all tallied.

 

Still I can't mount ATR from U1MB + L.

 

Maybe I just got a defective board, but it's hard to tell because it looks in pristine conditions and professionally made.

 

Is there anyone that has my same setup and attest it can actually work?

I mean NTSC XEGS with U1MB v2 + SIDE2, maybe I'm missing something.

Edited December 5, 2013 by phoenixdownita

[candle]

Erm, just read this while travelling to my home town...

You may feel the inresistible urge to kill sombody (me), but..

Xegs does not have pbi support at all, so putting this jumper on, and enabling so called xegs mode leaves you without pbi driver and you wiont be able to boot any atr while u1mb is set for xegs. You can safely usee it in xl/xe mode and side/pbi will work.

I'm sure there are some solid reasons this work like this, but for that i will need to look into my code, and, as you already know, i got no pc atm :/

Sorry for not responding any sooner - sometimes i'm simply overworked

[flashjazzcat]

Erm, just read this while travelling to my home town...

You may feel the inresistible urge to kill sombody (me), but..

Xegs does not have pbi support at all, so putting this jumper on, and enabling so called xegs mode leaves you without pbi driver and you wiont be able to boot any atr while u1mb is set for xegs. You can safely usee it in xl/xe mode and side/pbi will work.

I'm sure there are some solid reasons this work like this, but for that i will need to look into my code, and, as you already know, i got no pc atm :/

Sorry for not responding any sooner - sometimes i'm simply overworked

 

Thanks Sebastian. I almost reached this conclusion in the other thread. It suddenly dawned on me that for some reason the OS mustn't have been observing the PBI device at all.

[candle]

As far as i remember this is due to pinout diffrences between xegs mmu and xl/xe mmu - where standard mmu has mpd signal that enables pbi device rom, the other one has game select signal

[phoenixdownita]

OK later I will try to take the jumper out and see where that goes.

Will this work with an XEGS OS or do I have to use Stock XE/XL OS?

[in XEGS mode the OS made no difference wrt ATR booting]

 

Can anyone explain what the XEGS jumper is supposed to be doing? [we should add it to the documentation on U1MB]

I mean is it there to have U1MB simulate XEGS on std XE/XL or is it there to tell U1MB that the underlying Atari is an XEGS?

 

I guess that caused the confusion for me. I thought the jumper was there to tell U1MB that the underlying Atari is an XEGS and as such compensate for the different ROM size and MMU pinout.

 

 

Regarding why it is the way it is, considering XEGS never had PBI or even the ECI interface it is possible that Atari engineers decided to "drop support", just speculation.

[atari8warez]

I can only speculate on the answers to your questions above and I choose not to do so , however I can tell you one thing, some Atari developers (I am not giving any names) seem to like to keep their users in relative "darkness" about their devices, I am not sure whether this is simply due to their perceived "lack of time", their laziness in writing full and complete documentation, or their desire to keep their customers in dark so that they can continue to be the "gods" they are and watch "helpless users" come back to them time after time and beg for help. Whatever the reasons are, the end result is "no pain, no gain" for the unsuspecting user and they have to live with that if they want their device to operate the way they were supposed to in the first place.

To me there is no excuse for the technical and user documentation to be as poor (or in some cases non-existent) as they are. There are exceptions to this trend of course and I fully appreciate those developers who take the time and effort necessary to document their wares.

Edited December 6, 2013 by atari8warez

[Kyle22]

OK, let me get this straight... The XEGS jumper controls the MMU type? I re-jumpered my 1200XL to be XEGS compatible (one chip, OS, BASIC, Missile Command). I intend to put Internal MyIDE in the other socket. Haven't done it yet because I damaged the MyIDE board by replacing the pin header (broken pins). I know that the existing PBI driver doesn't support MyIDE. I only intend to use MyIDE with SDX, although a PBI driver would be nice for booting .ATR purposes.

 

It works perfectly as - is. Should I keep the XEGS jumper on, or would I get extra functionality by removing the jumper? Would it not work anymore because of the one EPROM (XEGS) setup?

 

I have a MyIDE OS installed as one of the selections on U1MB, although I don't think it's APT compatible. I am perfectly fine with IDE only working in SDX, because that is what I use the most.

 

-K

[flashjazzcat]

You're using the SDX soft-driver Kyle, so it doesn't make any difference whether the machine is an XEGS or has PBI support... and consequently whether the jumper is in place or not. The OP was having issues with the PBI ROM.

Edited December 7, 2013 by flashjazzcat

[Kyle22]

Thanks!

[phoenixdownita]

Nope . This didn't work, only thing it did is XEGS OS if no keyboard attached goes to diagnostic (or turning on with Select or Option pressed and keyboard). I guess Missile Commando is not mapped without XEGS mode jumper and that's it.

 

Now before anyone asks, yes I also reflashed with U1MBv2 stock rom and used "Stock OS" and it is the same (except that the no keyboard just goes to basic as expected), I also tried with copy of the rom on-board the day I received my U1MB which had Stock XL/Stock XE instead of Diagnostics/StockOS [i believe they are slot 2 and 4] and it also made no difference ... this is so annoying.

I must have reflashed 20 times by now , I start to believe the problem is somewhere else.

 

Is it possible that simply XEGS doesn't work with U1MB + SIDE2 and "L" to mount ATR?

Or maybe U1MB should also inject PBI ROM if XEGS mode as otherwise it does not seem to work on real XEGS (oh ... the irony )

 

BTW at least once MyBIOS R2 is on U1MB I found a way to boot the flasher ATR while having U1MB SDX on, I suspect it is a glitch but it works.

Basically I have the flasher ATR set as D1 in MyIDE2 partition space, then turn on Atari (U1MB SDX is set to ON), it locks, then I press reset and it unlocks and boots the flasher, with U1MB SDX off the ATR boots just fine from MyIDE2 but the flasher gets stuck waiting for the "SDX card" ,that's why it needs to be set to enabled on U1MB.

 

I hope Candle gets back his PC soon so maybe he can figure out something, I still have around 99980 flash cycles

 

 

[unrelated note I just found this:

http://www.retrobits.net/atari/pbi.shtml which at the end says:

PBI pin 31 (phi2) should be taken from ANTIC pin 29 instead directly from CPU phi2 output pin 39

makes one wonder ....]

Edited December 7, 2013 by phoenixdownita

[phoenixdownita]

No news on my issues, but it appears that someone has the same combo and it works for him:

http://atarimax.com/flashcart/forum/viewtopic.php?f=17&p=7944#p7944

 

Maybe my U1MB is really busted and it's time to send it back .... lotharek is still assisting but considering Candle has no access to his computer it may be a while before something positive comes out of it.... what a drag.

Edited December 11, 2013 by phoenixdownita

[phoenixdownita]

I got ahold of the ROM that according to http://atarimax.com/flashcart/forum/viewtopic.php?f=17&p=7944#p7944 works fine on 2 Atari XEGS, one with Lotharek's U1MB and one with Candle's U1MB with related SIDE2

Also the poster above stated that both of the U1MBs are set in XEGS mode ON with no issue.

The poster told me that on the Rom there is PBI 0.3.4, the message exchange was thru PMs so it is not on that thread, but I attach the ATR and extracted ROM for interested parties to dissect (the OSes are HiSpeed, MyBIOS RE, OSb, XEGS OS).

Anyway I flashed it to my U1MB and as I already experimented with PBI 0.3.4 it goes to SelfTest , except with OSb in which case it goes into MemoPad (never used OSb so not sure if this is the equivalent catch all state).

As I was at it, I re-flashed and swapped the EPROMs of my SIDE2 and U1MB (an AM29F040B in SIDE2 and the SST39SF040 on U1MB) thinking maybe a timing issue but it made absolutely no difference, everything works on its own right but together they go to Basic with default U1MB rom or Selftest with PBI 0.3.4.

Because I was frustrated I re-flowed and re-soldered every chip and socket on the XEGS motherboard as well as the cartridge connector and quite a number of other components all around and vias too, just to rule out cold joints or the like and of course it made no difference .

At this point it can be one pin/wire of the cables/IC adapters but that one pin/wire must only be used during PBI reboot because the rest works quite well, or U1MB is busted or maybe my CF is the cause of the mishaps.
I have a CF coming in a couple of days (a 4B Transcend) to rule out that component too.

The last hypothesis is that I have a bizarre motherboard version and the SW is having trouble doing the right thing with it.
As always I'll keep everyone posted , I know by now it sounds more like a menace, sorry

ROMXEGS.atr

ult1mb-from-ROMXEGS.rom

Edited December 12, 2013 by phoenixdownita