CT60/CT63 will be available soon again

[foft]

If changing the software to add support I guess adding 29F160 support would make sense.

At the moment it accepts any manufacturer of bottom device code 29f800. Ie 0x2258. It looks like the old code even checked the manufacturer.

For top codes I think the flash erase might need to take account of the different block sizes as well as changing the id.

Top device codes
– M29F200FT: 0x2251
– M29F400FT: 0x2223
– M29F800FT: 0x22D6
– M29F160FT: 0x22D2
Bottom device codes
– M29F200FB: 0x2257
– M29F400FB: 0x22AB
– M29F800FB: 0x2258
– M29F160FB: 0x22D8

Also is there a reason that there is no pull request with 1.05 changes for the GitHub repo for insane’s changes? It’d make it much clearer and easier to patch.

[mikro]

Who do you thing they could claim a refund from? How on earth could they blame sw tool developer? Really, I fail to see a logic here.

You misunderstood. Now you are the guy there are going to scream at. ;-) Because you overlooked that fact. If you delivered the software ALREADY without that check, you'd "destroy" user's CT60 and the only way to fix it would be to hunt someone to make a new software release (or return the device to you).

[mpattonm]

Fear not then, I have been adulting for long enough, I bear consequences for my own actions.

[foft]

Fear not then, I have been adulting for long enough, I bear consequences for my own actions.

So, is there a plan? Should we send boards back and you'll change the flash chip?

I think mine could use a check anyhow. Aside from the wrong flash chip it doesn't start properly in 060 mode from cold. 3.3v and 5v are stable prior to coming out of reset normally. Though exxos helped me extend the reset pulse (and a pull-up on reset) to be sure. I've checked with a second PSU too.

[mpattonm]

So, is there a plan? Should we send boards back and you'll change the flash chip?

Yes. I will provide solution as soon as it is tested.

I think mine could use a check anyhow. Aside from the wrong flash chip it doesn't start properly in 060 mode from cold. 3.3v and 5v are stable prior to coming out of reset normally. Though exxos helped me extend the reset pulse (and a pull-up on reset) to be sure. I've checked with a second PSU too.

No problem, just send it back in for secondary inspection. But please understand that from your description, the issue you are seeing is most likely caused by mainboard instability and/or PSU performance rather than CT60 itself. Of course I will be happy to inspect both, should you decide to send them along. Perfectly functioning and maintained mainboard and PSU is a must.

[Orion_]

mpattonm > I finally got my CT63 to boot, only after booting in 030 mode then switch to 060 and reset. I will try to replace the C7 with a 47uF and restore the 33ohm resistors on R216, R221 & R222 to see if it fix the cold boot problem.
my question is, how do I modify the frequency of the 060 ? When I try to modify via the Setup, save settings, it doesn't change after reboot, and neither when I use the latest version of the CPX CT60conf
also, do you have a tip to fix the fan you provide with the CT63 on top of the 060 ? I added some thermal paste but it doesn't fixed the fan firmly

[GCoons]

mpattonm > do you have a tip to fix the fan you provide with the CT63 on top of the 060 ? I added some thermal paste but it doesn't fixed the fan firmly

In this case you need to use thermal glue and not thermal paste. Thermal glue will set after a relatively short time.

[sunaiac]

mpattonm > I finally got my CT63 to boot, only after booting in 030 mode then switch to 060 and reset. I will try to replace the C7 with a 47uF and restore the 33ohm resistors on R216, R221 & R222 to see if it fix the cold boot problem.
my question is, how do I modify the frequency of the 060 ? When I try to modify via the Setup, save settings, it doesn't change after reboot, and neither when I use the latest version of the CPX CT60conf
also, do you have a tip to fix the fan you provide with the CT63 on top of the 060 ? I added some thermal paste but it doesn't fixed the fan firmly

I used this : https://www.amazon.fr/gp/product/B07MPBFRRH

[foft]

No problem, just send it back in for secondary inspection. But please understand that from your description, the issue you are seeing is most likely caused by mainboard instability and/or PSU performance rather than CT60 itself. Of course I will be happy to inspect both, should you decide to send them along. Perfectly functioning and maintained mainboard and PSU is a must.

Many thanks. I might wait a few weeks to see if there is a solution found for this one, then if not send you the whole lot (ct63+falcon+psu) to check!

3.3v and 5v power rails look stable and smooth on the scope, the reset pulse is good and the expansion header is clean (and works with dfb1).

Since it’s runs after a few resets it must be connected ok and have working power. It does this when warm too, so is not a cold solder issue.

I’m wondering if it’s a software issue. I might try the logic analyser on the expansion to see what it is up to… ie if the 060 is hanging or running something.

[Orion_]

I read back a bit because I missed some post and now I understand why the Frequency isn't retained.
The Flash chip I have on my "centurion tech CT63" is a Micron M29F800FT.
I found the datasheet and wrote a patch for flash060.prg (see below)
But, what bothers me is that this flash chip is in Top Boot configuration whereas previous CT6x flash are Bottom Boot configuration.
This doesn't seems a problem for flash060.prg since there is a table for each page, but ct60conf.cpx uses the last page of the chip for the parameters, and the last page on bottom boot flash chip is 64Kbytes, whereas on this new M29F800FT chip the 0xF0000 page is 32Kbytes (followed by 8k, 8k, 16k page)
I dug a bit into ct60conf.cpx source code, and there is a "PARAM_SIZE equ (64*1024)" which might be a problem.
From my understanding, it cause a problem if you modify the CT60 configuration more than 512 times.
If you modify the source so that PARAM_SIZE is 32*1024, it won't work either because some defines/equ are using this address to compute the base base address (which is wrong if different than 65536)
Also, the new patched ct60conf would be compatible only with this M29F800FT flash chip (otherwise, lots of modification of the sources are required)
For now I didn't recompiled ct60conf.cpx, I just made a binary patch to change the Flash chip ID of the AMD one to the Micro one.
Before making a new ct60conf.cpx recompiled from source, I'm waiting for some thought about this by more experienced coder.

devices:
; ...
dc.l 0x000122D6, micron_M29F800FT
dc.l 0

micron_M29F800FT: ; Warning: Top Boot configuration ! (previous flash are Bottom Boot conf)
dc.l FLASH_ADR+0x00000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x10000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x20000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x30000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x40000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x50000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x60000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x70000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x80000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x90000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xA0000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xB0000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xC0000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xD0000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xE0000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xF0000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xF8000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xFA000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0xFC000, FLASH_UNLOCK1, FLASH_UNLOCK2
dc.l FLASH_ADR+0x100000, 0, 0

[fidzen]

No problem, just send it back in for secondary inspection. But please understand that from your description, the issue you are seeing is most likely caused by mainboard instability and/or PSU performance rather than CT60 itself. Of course I will be happy to inspect both, should you decide to send them along. Perfectly functioning and maintained mainboard and PSU is a must.

Many thanks. I might wait a few weeks to see if there is a solution found for this one, then if not send you the whole lot (ct63+falcon+psu) to check!

3.3v and 5v power rails look stable and smooth on the scope, the reset pulse is good and the expansion header is clean (and works with dfb1).

Since it’s runs after a few resets it must be connected ok and have working power. It does this when warm too, so is not a cold solder issue.

I’m wondering if it’s a software issue. I might try the logic analyser on the expansion to see what it is up to… ie if the 060 is hanging or running something.

I would say this is due to reset circuit on your Falcon. I have had the same problem with both CT63 and CT60E.
You have to replace some components in the reset circuit to get rid of this problem. So I doubt it's due to a problem
with the new CT60.

[Orion_]

Here is my patched CT60conf for M29F800FT flash chip, use at your own risk (it might not be safe if you change configuration too often)

ct60conf_M29F800FT.zip

[foft]

No problem, just send it back in for secondary inspection. But please understand that from your description, the issue you are seeing is most likely caused by mainboard instability and/or PSU performance rather than CT60 itself. Of course I will be happy to inspect both, should you decide to send them along. Perfectly functioning and maintained mainboard and PSU is a must.

Many thanks. I might wait a few weeks to see if there is a solution found for this one, then if not send you the whole lot (ct63+falcon+psu) to check!

3.3v and 5v power rails look stable and smooth on the scope, the reset pulse is good and the expansion header is clean (and works with dfb1).

Since it’s runs after a few resets it must be connected ok and have working power. It does this when warm too, so is not a cold solder issue.

I’m wondering if it’s a software issue. I might try the logic analyser on the expansion to see what it is up to… ie if the 060 is hanging or running something.

I would say this is due to reset circuit on your Falcon. I have had the same problem with both CT63 and CT60E.
You have to replace some components in the reset circuit to get rid of this problem. So I doubt it's due to a problem
with the new CT60.

I checked the reset circuit and the power rails on the oscilloscope. They look absolutely fine. Does it rely on some special reset behaviour?

[fidzen]

Many thanks. I might wait a few weeks to see if there is a solution found for this one, then if not send you the whole lot (ct63+falcon+psu) to check!

3.3v and 5v power rails look stable and smooth on the scope, the reset pulse is good and the expansion header is clean (and works with dfb1).

Since it’s runs after a few resets it must be connected ok and have working power. It does this when warm too, so is not a cold solder issue.

I’m wondering if it’s a software issue. I might try the logic analyser on the expansion to see what it is up to… ie if the 060 is hanging or running something.

I would say this is due to reset circuit on your Falcon. I have had the same problem with both CT63 and CT60E.
You have to replace some components in the reset circuit to get rid of this problem. So I doubt it's due to a problem
with the new CT60.

I checked the reset circuit and the power rails on the oscilloscope. They look absolutely fine. Does it rely on some special reset behaviour?

Measuring is good, but have you tested the fixes for the reset circuit?
What type of power supply do you use for your CT60?

[foft]

I would say this is due to reset circuit on your Falcon. I have had the same problem with both CT63 and CT60E.
You have to replace some components in the reset circuit to get rid of this problem. So I doubt it's due to a problem
with the new CT60.

I checked the reset circuit and the power rails on the oscilloscope. They look absolutely fine. Does it rely on some special reset behavior?

Measuring is good, but have you tested the fixes for the reset circuit?
What type of power supply do you use for your CT60?

Yes absolutely. I increased C7, which changes almost nothing (confirmed in a Micro-Cap simulation too). I also increased C1. This one directly changes the length of the reset pulse. Finally I tried a pull up on the XRESET too.

I'm using a CHIEFTEC INDUSTRIAL CDP-085ITX (spec data sheet: https://www.chieftec.eu/_getfs.php?tb=p ... &fs=fs1_en). I also tested it with the ATX power supply from my PC, which is a lot beefier, with no change.

I also tried it on my other Falcon.

Finally I tried another 68060, just in case.

Any suggestions of what else to check are welcome, I studied microelectronics (though just work in software) so should be able to do most tests. I found the schematic online but could really use the HDL for the CPLDs to dig in properly.

[foft]

I was mulling the problem while washing up and wondering if it could be something like:
i) 68030 does not release bus properly for some reason in my system (Note that in my other system I tested in there is a CT1 which may be causing issues).
ii) Relative order of 3.3v vs 5v power up

[Orion_]

Just a thought: why does the flash utility have to check for _any_ chip ID at all? I mean it is up to the builders to assemble CT6x with compatible flash chip in terms of capacity, I/O timing, voltage tolerance etc. I would certainly understand the situation utility fails due to some timeout during r/w/v operations, but beats me why it does have to check for chip IDs.

Well, in the case of my CT63, you used a Top boot configuration flash chip, which doesn't have the same page size and order than the Bottom boot configuration flash chip used in all previous CT6x.
In that case, not checking the Flash ID would cause the flash to fail due to Page size and order difference.
so, just changing the ID in the tool won't help, we need to rebuild all the tools and the CT60 Boot ROM to check for the correct page size where the parameters are store.
That is really annoying that we the customer need to do the work ourselves because you made a mistake in the flash chip type.
There is no way to change your CT63 frequency as is, the CTCM you charge extra money for being useless unless someone release a ROM and tool patch.

[Rustynutt]

No problem, just send it back in for secondary inspection. But please understand that from your description, the issue you are seeing is most likely caused by mainboard instability and/or PSU performance rather than CT60 itself. Of course I will be happy to inspect both, should you decide to send them along. Perfectly functioning and maintained mainboard and PSU is a must.

Many thanks. I might wait a few weeks to see if there is a solution found for this one, then if not send you the whole lot (ct63+falcon+psu) to check!

3.3v and 5v power rails look stable and smooth on the scope, the reset pulse is good and the expansion header is clean (and works with dfb1).

Since it’s runs after a few resets it must be connected ok and have working power. It does this when warm too, so is not a cold solder issue.

I’m wondering if it’s a software issue. I might try the logic analyser on the expansion to see what it is up to… ie if the 060 is hanging or running something.

I would say this is due to reset circuit on your Falcon. I have had the same problem with both CT63 and CT60E.
You have to replace some components in the reset circuit to get rid of this problem. So I doubt it's due to a problem
with the new CT60.

Yes, age old issue. Think at first users were having to reset Falcon a number of times to boot. This is before the 060 was dreamed of (ok, who knows what Rodolphe dreams ).

Changing reset circuit capacitors, AIRC, allows for power supply to ramp up before Falcon starts to boot.
This was seen in most all accelerators boards and replacement power supplies.

[mikro]

That is really annoying that we the customer need to do the work ourselves because you made a mistake in the flash chip type.
There is no way to change your CT63 frequency as is, the CTCM you charge extra money for being useless unless someone release a ROM and tool patch.

Come on, that's a bit harsh. Yes, Pavel overlooked this issue (as he doesn't use the Atari flash tool) but that's about it. Mistakes happen. It's infinitely easier to release a new flash tool than manufacture/upgrade new CT63.

Releasing a new version of the whole trinity (cpx, flash tool, tos image) can be done within a weekend for anyone willing to take a look (as you've done already).

If you still feel so outraged, you can always put up your CT63 for sale, it will disappear within 10 minutes.

[Orion_]

There is no way to change your CT63 frequency as is, the CTCM you charge extra money for being useless unless someone release a ROM and tool patch.

Come on, that's a bit harsh.

Well, that's not harsh, that's just a fact If people are not watching this topic, they might never be able to fully use their CT6x/CTCM, and given the high price of the product I would have excepted a bit more testing and quality assurance.
Anyway, I have the knowledge to make a patch and I will gladly do so to help people, I'm happy that more people can get a CT6x because so then more people can enjoy 060 productions (that I too make and will continue to make in the future hopefully)
finger crossed for a Supervidel repro too

I contacted Didier M about this patch, hopefully he can help me validate if the patch is ok and not harmful before a release.
Patching the CT60 boot ROM might be a bit more difficult and tricky, I don't want to brick my CT6x.
I'm not sure if it is required though as it's just "reading" data from ROM, not page specific related.

[saulot]

Boot rom can be patched in 030 mode as far I remember, so if something wents wrong you can reporgram it. Honestly there were some ct60 hardware revisions along the way and after this long time it's hard to remember what version of a tools should be used for given hardware version.

[foft]

Boot rom can be patched in 030 mode as far I remember, so if something wents wrong you can reporgram it. Honestly there were some ct60 hardware revisions along the way and after this long time it's hard to remember what version of a tools should be used for given hardware version.

I tried running the flashing software in 030 mode and it says 'CT60 not found'. So how does one go about this? Can it be done by JTAG or does the chip need removing to flash?

[alexh]

finger crossed for a Supervidel repro too

I was under the impression that a Supervidel v3 was in development? If so, it doesn't seem economic to make any more Supervidel v2's? I maybe wrong.

[sigmate]

How did you get this rejoicing impression?

[Orion_]

mikro > are you able to recompile ct60conf.cpx ? Because I'm trying using PureC but I can't get it to compile, I get some errors, I'm not used to PureC.
Here is my patch for the param.s file.

Code: Select all

PARAM_SIZE equ (64*1024)		; keep this for compatibility of the address computation in following defines/equs
PARAM_SIZE_M29F800FT equ (32*1024)	; The 0xF0000 page is only 32k on Top boot block configuration of the micron_M29F800FT flash chip

NB_BLOCK_PARAM equ (PARAM_SIZE_M29F800FT/(MAX_PARAM_FLASH*4))
SIZE_BLOCK_PARAM equ (PARAM_SIZE_M29F800FT/NB_BLOCK_PARAM) 

...

devices:
;.....
	dc.l 0x000122D6, micron_M29F800FT
	dc.l 0

fujitsu_mbm29f800ba:
amd_am29f800bb:
st_m29f800db:
micron_M29F800FT:
	dc.l FLASH_ADR+0xF0000, FLASH_UNLOCK1, FLASH_UNLOCK2	

That should work for other Flash type too, it's just that it won't use the full 64k, only 32k (which allows up to 512 modification (instead of 1024) before erasing the page anyway)

For the flash060.prg patch, I think this should do it (but I didn't test because I can't compile it either)
Here is my patch for the flash.s file.

Code: Select all

devices:
	; ...
	dc.l 0x000122D6, micron_M29F800FT
	dc.l 0

micron_M29F800FT:	; Warning: Top Boot configuration ! (previous flash are Bottom Boot conf)
	dc.l FLASH_ADR+0x00000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x10000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x20000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x30000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x40000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x50000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x60000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x70000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x80000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x90000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xA0000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xB0000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xC0000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xD0000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xE0000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xF0000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xF8000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xFA000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0xFC000, FLASH_UNLOCK1, FLASH_UNLOCK2
	dc.l FLASH_ADR+0x100000, 0, 0