CTPCI not seeing Falcon IDE.

[Dal]

Yesterday, I thought I'd install a spare CTPCI I have here into my Falcon. I don't want to use any PCI Cards, just want the high speed IDE.

CT60 ABE/SDR is at 7F/7D
I flashed the CTPCI with the latest Jedec (1M)
TOS Boot version upgraded to v2.01

IRQ pad soldered to U25 pin1.

However during boot, both IDE tests return "magnetic media... Error".
If I remove CTPCI, boot from internal IDE is OK.

Any ideas?

[wongck]

IIRC You can set to use either the falcon IDE or the CTPCI ide in the CPX.
Give the settings a try.
Sorry Falc in the cupboard, FB in it's place, so I cannot take a look at it.

[Dal]

Hmmmm I just see New Boot /Old Boot (what's the difference?!?) options.

There are two IDE channels there 0.0 (internal primary?) and 1.0 (CTPCI primary?).

These aren't switchable per-sé but you can effect the order so that the one you wish to boot from is polled first. I can't really see anything else in the CPX that looks relevant.

I also reset the NVRAM in case this has an effect - it didn't.

The strange thing is, if I just have a drive connected to the internal Falcon IDE and boot, it seems to be present on both IDE channels but I see an error (see pic). So I think there is something squiffy with the IRQ. I tested continuity between the through hole on CTPCI before the IRQ pad and U25-pin1 and it seems ok. Perhaps the IRQ isn't changing state properly...

image.jpg

[Dal]

Just tested with a logic probe (probing at the through hole on CTPCI before IRQ pad). The line stays low throughout the boot process. Does anyone else have a means of testing this (alanh perhaps??)

[Dal]

Bounced it past Rodolphe and he reckons there should be a section in the CPX about IDE, but I just don't see it on mine. Was another CPX released after v2.01? I don't see anything on Didier's pages about it, but maybe it's something that he distributed 'unofficially'.

[Dal]

Solved.

The option to toggle between Falcon and CTPCI IDE only shows when the CT60 boots with the CTPCI active.

The Default setting when flashing appears to be CTPCI meaning I couldn't boot from my disk on module mounted on the Falcon's IDE connector.

In the end, I had to create a boot floppy with WDIALOG, COPS.ACC and CTCONF.CPX in order to sent the value back to Falcon. Now I can get a drive set up for the CTPCI and prepare it so I can safely switch between the two busses.

[wongck]

Glad you sussed it out .... hiding settings when certain condition not met really messes up someone's thought.

Sorry was not able to check back on the status, was a busy weekend for me as usual.

I also have a floppy boot disk with the CTCONF.CPX for this purpose, as I was playing with the different configurations/testing HDD speeds sometime back. In my case I also had hddriver, qed plus some other tools on it. Need to compress them to squeeze into 1.4MB floppy.
(I used Xcontrol there so no need wdialog)

[Galvez]

Solved.

Now that is solved could you benchmark the speed difference between both IDEs (Falcon and CTPCI), when I did it I was quite disappointed.

Transferring 100 MB file with dd command.

CT60 clock at 86 MHz
MiNT:
IDE (CF Flash, CTPCI IDE port) -> RAM = 704 kB/s
IDE (CF Flash, Falcon MOBO IDE port) -> RAM = 669 kB/s

[wongck]

Did not try with CF previously... here was what my results were back in 2012 when I tried it with normal HDD.
http://www.atari-forum.com/viewtopic.ph ... 62#p201637

and with normal cache (i.e. no 5s write cache wait) giving better writes speed
http://www.atari-forum.com/viewtopic.ph ... 25#p201684

[paul92706]

Hmmmm I just see New Boot /Old Boot (what's the difference?!?) options.

There are two IDE channels there 0.0 (internal primary?) and 1.0 (CTPCI primary?).

These aren't switchable per-sé but you can effect the order so that the one you wish to boot from is polled first. I can't really see anything else in the CPX that looks relevant.

I also reset the NVRAM in case this has an effect - it didn't.

The strange thing is, if I just have a drive connected to the internal Falcon IDE and boot, it seems to be present on both IDE channels but I see an error (see pic). So I think there is something squiffy with the IRQ. I tested continuity between the through hole on CTPCI before the IRQ pad and U25-pin1 and it seems ok. Perhaps the IRQ isn't changing state properly...

image.jpg

Hi Dal,
Sorry to bring old post up, but i am having same issue, i tried everything you explained? Do you remember anything else you did? Thanks

[Dal]

Sorry Paul - I can't remember exactly what I did other than what was stated, however I did quickly abandon the idea.

You must have the latest ABE/SDR installed on the CT60 and latest CTPCI flash too. Then you need Didier's TOS 404 flash version 2.01. For me, this is where I encountered an issue as I have the SuperVidel which requires flash 1.03c. Added to this, even if Didier's flash v2.xx was SV-friendly, stacking the CTPCI+SuperVidel+EtherNAT on a CT63 means compromising on speed to make it stable.

For me, I gave up on the idea of using the CTPCI and sold it. As Galvez found, there's not much improvement in speed in any case.

[paul92706]

Yes i have latest ABE/SDR firmwares installed. Where can i find that Version TOS flash? And yes i agree the CTPCI IDE speed is not much different from Falcon IDE, if i can get it working i will use Falcon IDE.

[mikro]

Yes i have latest ABE/SDR firmwares installed. Where can i find that Version TOS flash? And yes i agree the CTPCI IDE speed is not much different from Falcon IDE, if i can get it working i will use Falcon IDE.

https://github.com/mikrosk/ct60tos

[paul92706]

Delete

[paul92706]

Thank you mikro looks like i have TOS drivers v1.02 Beta 10, this is wrong huh?!

[wongck]

You need to boot up with CTPCI and then go to the CT60 CONF.CPX ( what is that called... I have not been using since a long time now) and set it to either use falcon IDE or CTPCI IDE.

The speed of IDE is almost doubled.
Note your IDE device must be understand the next PIO level.
Our device can be overloaded so your performance may not be seen ( as in CF / SD convertor).

[paul92706]

Ok i got the HD working with TOS 2.01 but now i don't have access to CTPCI/Radeon?! This strange, radeon is detected at bootup , but no output on radeon, anyone know why?

[1st1]

What is TOS 2.01? TOS 2.x on Falcon?

By the way, I would love to have an addon card for my Falocn which includes SuperVidel, CT-IDE (part of the CTPCI) and network card.

[shoggoth]

For me, this is where I encountered an issue as I have the SuperVidel which requires flash 1.03c. Added to this, even if Didier's flash v2.xx was SV-friendly, stacking the CTPCI+SuperVidel+EtherNAT on a CT63 means compromising on speed to make it stable.

I'm not sure the SuperVidel strictly "requires" flash/TOS 1.03c, but there are some severe bugs/features in later revisions which I didn't fancy implementing workarounds for. The correct fix is in the CT-TOS itself, not the SV XBIOS.

[shoggoth]

Double post, sorry.

[Dal]

What is TOS 2.01? TOS 2.x on Falcon?

By the way, I would love to have an addon card for my Falocn which includes SuperVidel, CT-IDE (part of the CTPCI) and network card.

When we talk about TOS 1.03c /TOS 2.01 for CT6x, we are referring to the version of Didier's distribution. The actual TOS version is 4.04, but modified by Didier for the CT60.

1.03c is considered to be the more stable version but 2.01 brings additional functionality (like the support for additional IDE channels).

[paul92706]

Ok i got her working, i needed CTPCITOS.BIN for my setup, its been along time since i messed around with flashing, so bare with me.