New (alt) BIOS for Ultimate 1MB/Incognito

[HiassofT]

Great. Just need McRorie to test with his U-Switch when he gets time

Here's a complete shot in the dark - I don't own a U-Switch myself and don't have my U1MB Atari near me.

 

When looking at the picture of the U-Switch on Lotharek's site I noticed something very strange: it looks like it uses a 74HC04. Unless you drive that with 5V CMOS levels (i.e. high >= 4-4.5V) that's the wrong type - for interfacing with TTL or 3.3V LVCMOS levels one should use HCT types, not HC.

 

Check your boards which chip types you have (on McRorie's picture it's hard to decypher the markings on the top chip) and also check the signal level coming from the U1MB into BIN - if the high-level is around 3.3V and indeed a HC04 chip is used that could explain all sorts of strange problems. Replacing the 74HC04 with a 74HCT04 should fix these.

 

As a side node: from the PCB layout it looks like the inputs of the unused inverters aren't connected to GND - one should not omit this when dealing with CMOS chips...

 

so long,

 

Hias

[flashjazzcat]

Thanks for jumping in Hias!

 

Check your boards which chip types you have (on McRorie's picture it's hard to decypher the markings on the top chip) and also check the signal level coming from the U1MB into BIN - if the high-level is around 3.3V and indeed a HC04 chip is used that could explain all sorts of strange problems. Replacing the 74HC04 with a 74HCT04 should fix these.

Just checked a couple of those I can still get at easily, and they both use HC04. Measured voltage across BIN and it is indeed c. 3.3v.

[HiassofT]

Just checked a couple of those I can still get at easily, and they both use HC04. Measured voltage across BIN and it is indeed c. 3.3v.

OK, now I think it's time to contact Lotharek - using a 74HC04 here is definitely wrong.

 

The datasheet of the OnSemi 74HC04A (the one on Lotharek's picture) doesn't contain a specific entry for VCC=5V, but at VCC=4.5V the guaranteed V_IH is 3.15V and at VCC=6V it's 4.2V. So at VCC=5V V_IH would be around 3.4-3.5V or so - that's at best marginal for a 3.3V signal...

https://www.onsemi.com/pub/Collateral/MC74HC04A-D.PDF

 

so long,

 

Hias

[flashjazzcat]

OK, now I think it's time to contact Lotharek - using a 74HC04 here is definitely wrong.

I've just added a reply to the email conversation he and I were already having about this issue, referencing the foregoing posts.

[lotharek]

Guys, RELAX !

 

IT DOESN`T MATTER the chip type... hct, fo4, ls04..... AT ALL - really...

 

what is matter is LOW logic state that comes from ULTIMATE as BIN (ary) input is pulled high to 5V by 4k7

 

so ultimate makes hc04 input low as it is always HIGH

 

 

 

i can only presume 2 problems : faulty one ( shall i send other ??) or f....ed atari at some point / generally or after mods/

 

 

ps. HIGH LOGIC LEVEL i ment - not that one high : CLICK

Edited February 27, 2016 by lotharek

[+MacRorie]

Guys, RELAX !

 

IT DOESN`T MATTER the chip type... hct, fo4, ls04..... AT ALL - really...

 

what is matter is LOW logic state that comes from ULTIMATE as BIN (ary) input is pulled high to 5V by 4k7

 

so ultimate makes hc04 input low as it is always HIGH

 

 

 

i can only presume 2 problems : faulty one ( shall i send other ??) or f....ed atari at some point / generally or after mods/

 

 

ps. HIGH LOGIC LEVEL i ment - not that one high : CLICK

 

 

As soon as I get time tomorrow, I will test the re-wire, then R&R the USWITCH if it still fails. If it fails after that, well, I dunno. Like has been said, I could turn it off and on from the original BIOS. LED even turned off and on and I giggled like a school boy. However, it could be that the USWITCH failed after that, left the DPOkey in the on state and I never noticed until I installed the new BIOS.

 

It's possible. And, given Occam's Rzor, the most probable explanation. LIke I said, I think I have the extra USWITCH to do the R&R. Will update Sat night/SUn morning.

[HiassofT]

Hi Lotharek!

 

IT DOESN`T MATTER the chip type... hct, fo4, ls04..... AT ALL - really...

I beg to differ slightly. From a design point-of-view the logic family does matter, that's in the 101 of digital circuit design.

 

But from a practical point of few you usually get away with such design errors. The typical (not guaranteed) switching point of 5V CMOS level inputs is usually somewhere around 2V, so 3.3V will usually be detected as high.

 

I also think it's more likely something in McRorie's u-switch went wrong, maybe due to an ESD discharge or something like that.

 

what is matter is LOW logic state that comes from ULTIMATE as BIN (ary) input is pulled high to 5V by 4k7

A pullup won't work if the output signal is 3.3V LVCMOS like on the XC9500XL series as CMOS drives both high and low levels "hard" using push-pull outputs.

 

You can pull-up (LV)TTL outputs - which drive the high level "soft" - and also of course open-collector outputs (which only drive the low level). The latter is the typically recommended configuration when interfacing (LV)TTL and LVCMOS outputs to CMOS inputs.

 

If you already have a pull-up in place on BIN adding a series diode between BIN and the U1MB output - with the cathode on the U1MB end - would be a proper fix. That effectively converts the LVCMOS output into a open-collector type output, and the pullup resistor can pull up the input all the way up to 5V.

 

so long,

 

Hias

[lotharek]

Hiass, of course i agree with You, but it is not the place to delibarate of the purpose.

uswitch can also be switched from AKI , so present design compromises both steering logic.

also, left without any input, toggles stereo on.

 

honestly, there are millons of circuits around us, and i presume NONE is even close to perfect design theory

 

 

anywy - so far there is only ONE ( 1 piece) of uswitch doesn`t work around world reported among hundreds made and delivered..

Edited February 27, 2016 by lotharek

[flashjazzcat]

anywy - so far there is only ONE ( 1 piece) of uswitch doesn`t work around world reported among hundreds made and delivered..

Well... To be exact, one u-switch which is demonstrated to have issues because of the new BIOS. With original firmware, it was assumed to work because software didn't care if M0 was connected to anything or if stereo state actually changed. On the other hand, I have four u-switches and they all work well.

[lemiel]

I am trying to find what is wrong and trying to prepare fresh rom image. (Before visiting someone with more knowledge or just lotharek for help or new device.)

Fresh using Altirra OS and Altirra U1MB Recovery OS.

I set 4kB banks SST flash, flashed sdx, bios, loader, pbi, xl rom, atari basic and cleared empty basic and car slot names (were this short arrow directed to right ► - some special char - could it be just empty default not demanding delete?).

And U1MB bios is not starting from that rom. It does not contain sixteen bytes starting lda 0 sta d5eo at 41ff0 and another places (do not searched all, but just before PBI BIOS there are none of them too) compared to u1mb.rom from Voy's .atr from pigwa with single files mentioned above.

Is flash type written somewhere in this 512k of rom? This my test rom is attached.

Because I choose SST in Altirra option but UFLASH still showed AM29. Now is changed. But I didn't noticed when.

 

UFLASH needs SDX to be enabled in U1MB to recognize device? That's probably why I can't flash it from new rom because default is disabled and these not succesful config writes.

 

And where to see this "boot SDX, and you should see PBI BIOS version number at the top of the screen." Do not see it in Altirra or on many Youtube movies.

 

But SDX sees partitions made by FDISK on CF card in SIDE/SIDE2.

 

Another - is it possible to reset /write by poke (simplest method for me) values to NVRAM?

Could you describe what and where?

 

Will this new loader started from SIDE work without U1MB?

lemiel-test-u1mb.zip

Edited February 28, 2016 by lemiel

[flashjazzcat]

Just tried your ROM in Altirra and it crashed when jumping into the OS since the machine reset vector was completely mangled in that OS ROM slot. This was possibly slot 4, although it was hard to tell without analysing things further since the OS descriptions have been completely edited out:

 

 

Here's the complete 512KB ROM I'm currently using in Altirra, and which is not so different from what I have on real hardware:

 

Ultimate1MBNew.zip

 

The flash type is not stored in the ROM. The flashing software reads the JEDEC flash device vendor and device IDs and does a lookup on a (short) list of common types.

 

Yes: SDX needs to be enabled for UFlash to program the ROM, although it will also work from the built-in XEX loader (which also leaves the bank register exposed and is able to read files from the FAT partition from which UFlash was launched, which is mapped to D1:).

 

To see the PBI BIOS version notice on screen, you need to enable it in advanced settings (it is disabled by default, lest people find it obtrusive).

 

I'll shortly update UFlash to recognize and report the version numbers of the main BIOS, PBI BIOS, and XEX loader. This was not previously possible since original firmware almost completely lacked versioning metadata, but still requires some fiddly code or augmentations to the large tables inside the flashing software.

[flashjazzcat]

Shows how preoccupied I've been with real hardware tests: previous ROM was a couple of versions out of date in all areas. Here's one with very latest firmware:

 

Ultimate1MB Complete (BIOS 0.52, PBI 1.59, Loader 0.41).zip

Edited February 28, 2016 by flashjazzcat

[lemiel]

I again did edited my post adding questions, not thinking you are so fast in responding. Now have learned to not doing that.

In this rom from me I see names I wrote, at 53f10 as in yours (starting with DracOS).

 

OS slot 4 is empty as I DID fresh image from scratch. Only first one is filled with contents. Do they all need to be written in?

1. Altirra U1MB firmware set to 'Altirra U1MB Recovery OS (internal)'.

2. Boot Voy's atr with MyDOS with UFLASH with single rom files.

3. Flash sdx, bios, loader, pbi bios, xl os (in slot 1 only), atari basic (in slot 1 only). Flash empty files in GOS, OS ROM, BASIC and XEGS not used slots.Clear not used OS, BASIC and XEGS slots names, type and flash in names for XL OS and Basic first slots.

4. Dumped U1MB firmware.

Result in #735.

 

I repeated process, without flashing sdx and empty slots. Attached image. Works.

Is OS slot 4 default or I only had it selected accidentaly and was not flashed? And was saved somewhere that it should use slot 4. - attached image.

 

I understand that there can be no SDX in flash, but it must be enabled in BIOS for U1MB being recognized by UFLASH?

And then UFLASH can be executed from MyDOS, xex loader from BIOS or SIDE Cart and will recognize U1MB?

 

I had enabled this PBI banner in settings, do not see it.

 

On BIOS BOOT logo I pressed Escape and there is some garbage text in line under line with 'Press A-O for Boot Drive'.

 

Repeated from edited post:

 

Another - is it possible to reset /write by poke (simplest method for me) values to NVRAM?

Could you describe what and where?

 

Will this new loader started from SIDE work without U1MB or with older PBI bios?

 

 

Sorry for being so tedious.

u1mb-flash-needed-only.zip

Edited February 28, 2016 by lemiel

[flashjazzcat]

I again did edited my post adding questions, not thinking you are so fast in responding. Now have learned to not doing that.

No problem. Regarding PEEK/POKE to NVRAM: it's not that simple. It uses an SPI interface and would require a small test program to be written. This can be done but I need to know what you want it to do.

 

...attached image.

Picture three shows that UFlash has completely misidentified the firmware revision and believes it's operating either with Incognito or the older revision Ultimate BIOS. This is guaranteed to cause utter havoc, although I have no idea what led you to be in this situation. UFlash recognized (the last time I checked) both original and new firmware. The slot descriptions are stored in completely different locations using completely different encoding - hence nonsense on the screen when things are mismatched. An unholy mess, really.

 

I understand that there can be no SDX in flash, but it must be enabled in BIOS for U1MB being recognized by UFLASH?

Although I struggle to envisage a situation in which the complete absence of SDX would be advisable, UFlash does not use the SDX ROM for hardware recognition, no.

 

And then UFLASH can be executed from MyDOS, xex loader from BIOS or SIDE Cart and will recognize U1MB?

If SDX was disabled with COLD /N, it's possible to run UFlash from MyDOS. You can also launch UFlash from the U1MB XEX loader for the purpose of flashing Ultimate. I wouldn't necessarily advice using SIDE's on-board loader for this, although one could always try.

 

I had enabled this PBI banner in settings, do not see it.

Perhaps the BIOS is completely mangled for the reasons described above.

 

Will this new loader started from SIDE work without U1MB or with older PBI bios?

The SIDE build (not updated in a while, but soon) will work without Ultimate. It may work with the older PBI BIOS, but owing to bugs in the very old revisions (i.e. pre-new BIOS), performance is unlikely to be optimal.

 

Sorry for being so tedious.

Not at all. My recommendation: take the most recent 512KB flash ROM I posted in this thread and flash it to the very top entry (the chip name) in UFlash. If you have an external USB programmer, use that, or keep as a backup.

 

The images posted at pigwa were very welcome, but weren't put together by me and contain an unknown collection of firmware revisions I have not had time to test. The most recent 512KB ROM I posted here is at least guaranteed up-to-date and if that doesn't work properly, something else is amiss, since it appears to work everywhere else.

Edited February 28, 2016 by flashjazzcat

[lemiel]

For what this test program? To check if hardware is working properly. not nice thet it is not easy as write and read to/from $D380 UCTL.

 

Image three is from empty rom, for failsafe trying only.

 

SDX disabled by option in BIOS menu - and UFlash not see U1MB. But ok, I understand.

 

For banner, even with your last version, nothing before 'Ultimate clock installed' is shown.

 

Will try this last rom image in programmer.

[lemiel]

Won't check now. Both SST flashes which I bought are dead... Static electricity or programer or anything else? Just probably garbage now. Will check them in SIC! when will get zif adapter. Need to buy another couple.

[flashjazzcat]

Wow... You seem completely cursed. Good luck and keep us updated.

[Panther]

I just finished installing my Incognito. Looking forward to the latest BIOS version. This might seem like a stupid question, but how the heck do I access the Incognito menu? I'm used to the Help key on the U1MB...

 

Edit: Got it figured out! Working well.

Edited February 29, 2016 by Panther

[+MacRorie]

Well... To be exact, one u-switch which is demonstrated to have issues because of the new BIOS. With original firmware, it was assumed to work because software didn't care if M0 was connected to anything or if stereo state actually changed. On the other hand, I have four u-switches and they all work well.

 

Apparently this USWITCH decided to give its life for its POKEY. Did re-wire. Didn't work. R&R USWITCH, using original VCC & GND, works fine. BTW, somewhere in a easily findable location as to how to connect VCC and GND would be just peachy. I am pretty sure I searched high and low to figure out where to connect it and came up short so hooked it to common ground and to JOY (since that is the only EASY place I knew +5V was coming in at).

 

So, tl;dr: it works as advertised. bad USWITCH. Looks like it got toasted at some point. LED still it up on both SPokey and USWITCH so wasn't sure where the fault lay. Like I said, the BIOS switched worked before . . . it is just bad timing that USWITCH dies and I upgrade to new BIOS . . .

 

Thanks EVERYONE for helping out. Now if you could explain how my BBS Message base drive partition took a dumper, I would be most interested . . .

[flashjazzcat]

Thanks for confirming. These things are surely sent to try us.

 

I'll make the Incognito and SIDE builds later. Speaking of hardware issues, I offered to fit U1MB in someone's 1200XL and the saga has expanded to fill all my time, so reluctant is this machine to work with the upgrade. You get one from time to time...

[Marius]

Thanks for confirming. These things are surely sent to try us.

 

I'll make the Incognito and SIDE builds later. Speaking of hardware issues, I offered to fit U1MB in someone's 1200XL and the saga has expanded to fill all my time, so reluctant is this machine to work with the upgrade. You get one from time to time...

 

Hmmm I may have some suggestions what you could do with this 1200xl ....

 

Well... sometimes things are not as smooth as expected, even after a long period of betatesting....

[+DrVenkman]

Speaking of hardware issues, I offered to fit U1MB in someone's 1200XL and the saga has expanded to fill all my time, so reluctant is this machine to work with the upgrade. You get one from time to time...

Oh believe me, I know ...

 

On that note and for what it's worth, when I have a day or two off again I plan to swap out those suggested logic chips in my 1200XL and try the U1MB install there again. Should be fun ...

 

 

Sent from my iPhone using Tapatalk

[flashjazzcat]

On that note and for what it's worth, when I have a day or two off again I plan to swap out those suggested logic chips in my 1200XL and try the U1MB install there again. Should be fun ...

The 74LS swap was a false dawn here, sadly. The machine in question came straight from Hell.

 

EDIT: another thing to try is putting the shielding back (at least the bottom layer). Makes a difference here.

Edited February 29, 2016 by flashjazzcat

[HiassofT]

Hmmm I may have some suggestions what you could do with this 1200xl ....

You mean, like buying some ethanol and following the steps in the instruction manual?

Instructions for the 1200XL and 1000XL

 

Yes, that should work. But always keep in mind what the manual told you:

Once a light it will take 20 minutes for the flame to reach the maximum height, the burn time will be between 6-8 hours and during that time you may need to vent the room by having a window or door ajar.

If you did everything right it finally should look like this

 

SCNR,

 

Hias

[Kyle22]

You mean, like buying some ethanol and following the steps in the instruction manual?

Instructions for the 1200XL and 1000XL

 

Yes, that should work. But always keep in mind what the manual told you:

 

If you did everything right it finally should look like this

 

SCNR,

 

Hias

?