Ultimate 1Mb Upgrade

[flashjazzcat]

So - have I missed a step? Should I see new menu options? Can I safely flash the new firmware?

The Xilinx (JED file) update merely adds the extra board logic necessary to support the newer feature set (RAM under CCTL, etc). The actual 'firmware' menu is held in ROM (on the PLCC flash chip), so you still need to upgrade that.

[spookt]

The Xilinx (JED file) update merely adds the extra board logic necessary to support the newer feature set (RAM under CCTL, etc). The actual 'firmware' menu is held in ROM (on the PLCC flash chip), so you still need to upgrade that.

 

Thanks Jon - I guess that's what I'm getting at. Having flashed the new JED am I safe to go ahead and use your BIOS flasher? I saw screenshots of another 'Candle style' menu with captions of 'v2 JED' which had a different colour banner and extra options. So I'm guessing that the JED itself doesn't add those ?

Edited May 20, 2019 by spookt

[flashjazzcat]

I'm not aware of Candle ever changing the colour scheme of his menu banner; perhaps the screenshot showed an NTSC machine (in which case the banner would appear to be a different colour). None of Candle's BIOS setup menus required the extra RAM provided by the V2 JED anyway: only my firmware requires it (this allows far better integration of the various firmware components than was possible before).

 

Anyway: a JED update doesn't visibly change anything at all until you actually update the firmware, and you can go ahead and do so safely now. You might as well go for the July update on my website, since the (now rather copious) user documentation focuses entirely on that firmware generation now, and the user experience is somewhat improved. I'll be releasing another significant update shortly which brings a lot of further improvements.

[spookt]

 

So after initially asking the question six and a half years ago in December 2012 I finally updated the JED and the BIOS in that very same Ultimate 1Mb in my 600XL! Many thanks to folks here for the repository of wisdom that it is, and of course Jon whose wonderful alt BIOS I'm now running. (Small donation incoming Jon!)

 

Glad to say that reflashing seems to have cured some odd behaviour I was seeing using anything above stock memory with this unit when it came out of storage, so it looks like corrupted flash was to blame.

Edited May 20, 2019 by spookt

[flashjazzcat]

Excellent - many thanks!

 

Glad to say that refreshing seems to have cured some odd behaviour I was seeing using anything above stock memory with this unit when it came out of storage, so it looks like corrupted flash was to blame.

This is becoming a recognised issue, possibly confined to older boards, although at this point I'm not sure. I have an 800XL here with an older U1MB, and the CPLD actually suffered 'bit rot' which resulted in the upper half of the 1024K of extended RAM actually repeating over the first half (which is to say, one of the higher order bank selection bits was ignored). This caused the thing to brick every single time I tried to flash it via UFLASH. Eventually I tried a CPLD flash and everything worked perfectly after that.

 

The same thing happened with another U1MB (again, older, black PCB) which was sent in for diagnosis. I went around in circles with it (it would lock up when UFLASH started reading a ROM file) until I reprogrammed the CPLD, which fixed it.

[spookt]

Excellent - many thanks!

 

 

This is becoming a recognised issue, possibly confined to older boards, although at this point I'm not sure. I have an 800XL here with an older U1MB, and the CPLD actually suffered 'bit rot' which resulted in the upper half of the 1024K of extended RAM actually repeating over the first half (which is to say, one of the higher order bank selection bits was ignored). This caused the thing to brick every single time I tried to flash it via UFLASH. Eventually I tried a CPLD flash and everything worked perfectly after that.

 

The same thing happened with another U1MB (again, older, black PCB) which was sent in for diagnosis. I went around in circles with it (it would lock up when UFLASH started reading a ROM file) until I reprogrammed the CPLD, which fixed it.

 

 

This is a board from the first run I believe. I have it installed with one of Dropcheck's MMU adapters which added the resistors which IIRC were incorporated into later production runs. I'm going to give it some decent testing over the course of a few days, but it's already 100% more stable than it was.

[flashjazzcat]

Brilliant stuff, and thanks again for hitting the tip jar.

 

As I mentioned in the email I just sent you, if you're into testing stuff, I have pre-release copies of the new update here.

[chevymad]

I'm having a bad chinese delivery weekend. Ordered 2 things wrong and wasn't home to sign for my new PIA chips either. Anyway, I ordered a xylinx jtag programmer to finally upgrade my original Candle U1mb. It came and it's a Digilent highspeed jtag smt2 copy instead of the dlc9g. Finally found the correct Impact software and built my cable. It finds the chip and identifies it no problems.. But then there's no option to program. Then I found in the help section that programming and configuration of 9500xl series is not supported by digilent cables. Looking on the web I found some software called Adept for the digilent cable. They say that with this you can program the 9500xl series. So I d/l that. This software won't accept .jed files only svf. So I use Impact to create an svf file. I then use Adept to burn the file to the U1mb. It finishes and says programmings complete. Yay! So I turn my atari back on and use Uflash to upgrade the entire rom. And it's a brick. I've tried reprogramming the firmware several ways a few times. No good so far. Just black screen. I even tried pulling the chip and flashing the bios in case that's what went wrong and that completes successfully, but no love.

 

So any other ideas before I go and buy another programmer?

Edited May 26, 2019 by chevymad

[spookt]

Edit - never mind. I reread your post.

Edited May 27, 2019 by spookt

[chevymad]

I found another warning that says these chips aren't recommended to be flashed with an svf file because svf files make a single pass, whereas .jed loops and makes multiple passes to burn the chip. Looks like I need the programmer I thought I was ordering in the first place. Amazon and Ebay's list of alternative sellers has gotten me a few times lately with devices that look the same but aren't. Like I now have a pair of 6522ap chips instead of 6520aps as well.

Edited May 27, 2019 by chevymad

[wt808]

I would recommend getting the DLC9G and following the proven procedure with the JED file. Using one-step SVF in iMPACT, I got a "Programming completed successfully" even with no programmer plugged into the computer!

 

On another note, here's an alternative, no-solder JTAG cable made with a 2x5 2.0mm IDC female connector. This uses the 2x7 2.0mm cable assembly included with the DLC9G kit and you can still use it for 2x7 headers.

[chevymad]

I can't win. Recieved the dlc9g tonight and hooked it up. DOA. Can't be found by software and even the LED stays dark. Talked with tech support and new one allready on it's way again.

[chevymad]

Replacement cable came today. Showed life signs but Impact still couldn't discover it. Trick was to uninstall the cable in device manager. Reboot into safe mode and erase the windrvr6.sys file manually. Then restart and Reinstall the drivers from the xylinx install files. Next time I fired everything up the LED came on green and impact discovered the cable and chip. Flashed v3 jed and I'm unbricked. Another original Candle v1 U1mb is now all up to date with FJC's firmware!

[_The Doctor__]

Not that it matters, but you could put a an led that light up red, green, orange (or yellow) in place of the current led, or mount one near to and pointing at the led window plastic and illuminate it that way... no holes and very simple....  I think the sofware XEGS manager should still function in XEGS mode as well, and any software that was XEGS aware (was there any?) should still manage to swap in the XEGS rom suite. It is a bit of a duplication of functions though....

 

***omg, yeah this is a necro response to something in ancient history, need to pay better attention to the date!***

Edited April 18, 2020 by _The Doctor__

[blacka013]

Hi,

 

I have a Xilinx DLC9LP so I can update my early Ultimate 1MB. The problem I have is the leads that came with it, they are all 2.54mm connectors.

 

Does anyone know where I can source the 2.0 mm connectors from? I've been looking online and can't find any.

 

Hope you can help.

 

Regards

[Faicuai]

18 minutes ago, blacka013 said:

Hi,

 

I have a Xilinx DLC9LP so I can update my early Ultimate 1MB. The problem I have is the leads that came with it, they are all 2.54mm connectors.

 

Does anyone know where I can source the 2.0 mm connectors from? I've been looking online and can't find any.

 

Hope you can help.

 

Regards

Here, these will make your life a lot easier (and cleaner / better-looking, too):

 

https://www.amazon.com/gp/product/B07FM72VBF/ref=ppx_yo_dt_b_asin_title_o04_s00?ie=UTF8&psc=1

https://www.amazon.com/HiLetgo-5x40pcs-Breadboard-Assortment-Arduino/dp/B077X99KX1/ref=sr_1_1?dchild=1&keywords=2.54+to+2.0+dupont&qid=1588001553&sr=8-1

 

I personally got the first set. But both will work.

 

[blacka013]

Hi Faicuai, thank you for the info, much appreciated. I can get on and update this early board.

[marcokitt2000]

Hello,

 

I read when update bios to 3.10 must all so update xillinx chip 9500 serie to versie 2? And where to dowload?

The normal site is down or gone.

I have the tools ( i updated the megaspeedy to 1.10 use the is plate).

 

 

 

Gr.

Marco

[BillC]

15 hours ago, marcokitt2000 said:

I read when update bios to 3.10 must all so update xillinx chip 9500 serie to versie 2? And where to dowload?

FJC attached v3 to post# 110.

[marcokitt2000]

Thnxxxx

[+Stephen]

OK - scanned all 6 pages.  I have a TL-866 Mini Pro USB programmer.  I assumed I had a bad flash chip - my black PCB Candle U1MB would not re-program the flash chip.  So, I bought a new one (AMIC AM29040BL).  My Mini-Pro doc says it will handle an AMIC 29040B.  However, it does not appear to be reading the chip correctly, and it will not write to the chip (nor the new one).  I can't imagine having two bad flash chips.  At this point I am thinking the Mini Pro is not compatible with this particular Flash chip.

 

However, this begs the question.  Why would my U1MB not be able to write to the original flash?

[candle]

1) sdx got disabled

2) flash chip is write-protected by the bios (option)

 

no other options i'm affride

most of flash chips uses standard JEDEC algorithms for programming, so if you have external programmer, check if it allows to skip signature/chip id checking before programming

commands (erase, program) are generic, so any supported chip on the list will do

 

[flashjazzcat]

9 hours ago, Stephen said:

So, I bought a new one (AMIC AM29040BL).

Buy an SST 39SF040. The AMIC chips were complete junk in my experience; all the AMICs in my original Candle U1MBs died within 1-2 years and all my boards now have SSTs in them.

[TGB1718]

@flashjazzcat Hi Jon, any news of the U1M firmware for SIDE3, can only see V3.1 on web site

 

thanks

[flashjazzcat]

54 minutes ago, TGB1718 said:

Hi Jon, any news of the U1M firmware for SIDE3, can only see V3.1 on web site

Not yet. My entire week last week was thrown into total chaos for reasons I'll be happy to explain in private if anyone is curious, and the run of miserable luck shows no signs of letting up. I'll publish the SIDE3 U1MB firmware as soon as I'm able to focus sufficiently well to actually get it prepared for release. Sorry for the delay.