Terrible Fire Accelerators

[alenppc]

EzdineG, I am suspecting the same thing. The first thing that came to mind was issues that come up with A600's unbuffered IDE interface when using long cables, except that here I was having them with even very short cables, so I suspect that the SATA converter inadvertently fixes whatever resistance/capacitance issues might be happening. Probably that's also going on with plasmab's CF card adapter.

I have ordered a few different kinds of CF adapters from Ebay but it will be 2 months before I get them from China, unfortunately.

[plasmab]

Originally Posted by alenppc

EzdineG, I am suspecting the same thing. The first thing that came to mind was issues that come up with A600's unbuffered IDE interface when using long cables, except that here I was having them with even very short cables, so I suspect that the SATA converter inadvertently fixes whatever resistance/capacitance issues might be happening. Probably that's also going on with plasmab's CF card adapter.

I have ordered a few different kinds of CF adapters from Ebay but it will be 2 months before I get them from China, unfortunately.

My cf card adapter is dumb. Just wires my cf card directly into the cf card. I can try a longer cable and see what gives


Sent from my iPhone using Tapatalk

[EzdineG]

Originally Posted by plasmab

My cf card adapter is dumb. Just wires my cf card directly into the cf card. I can try a longer cable and see what gives


Sent from my iPhone using Tapatalk

I've only got these two, and they behave identically. Chances are they'd do the same with REV 2. Perhaps I can knock out the pin on the ide cable pulling power to see if it changes anything.

[alenppc]

This is the SATA converter I am using that does not cause issues, plus it gives a 5.6 MB/sec transfer rate.

https://www.startech.com/ca/HDD/Adap...verter~IDE2SAT

[whiteb]

Originally Posted by plasmab

Do you have an FPU? Also what is your CPU version and masks?


Sent from my iPhone using Tapatalk

I think he said (I could be wrong) he had a few of full RC 68030, so they have FPU on Die.

[keith]

Hi was very busy last night, I hope to have some time tonight.

[Mr.Flibble]

Originally Posted by whiteb

I think he said (I could be wrong) he had a few of full RC 68030, so they have FPU on Die.

Nope, MMU on die with the full 030, but the FPU is always via a 6888x on the 68030. The 040 was the first to integrate the FPU

[whiteb]

Originally Posted by Mr.Flibble

Nope, MMU on die with the full 030, but the FPU is always via a 6888x on the 68030. The 040 was the first to integrate the FPU

Ahh okay.

[plasmab]

@alenppc I'll order one of those and test


Sent from my iPhone using Tapatalk

[alenppc]

@plasmab, these are the only ones that work fine on my system though. (Side question, I wonder if this is the same chip you were planning to use on the TF534? Have to take a look.)

Any chance you could try a different CF adapter connected with an IDE cable?
Thanks!

[plasmab]

Originally Posted by alenppc

@plasmab, these are the only ones that work fine on my system though. (Side question, I wonder if this is the same chip you were planning to use on the TF534? Have to take a look.)

Any chance you could try a different CF adapter connected with an IDE cable?
Thanks!

Yeah i'll try a full IDE drive on a standard cable.

[plasmab]

Found this


Sent from my iPhone using Tapatalk

[nogginthenog]

Originally Posted by plasmab

Found this

Just a thought: I wonder if a 40-wire or a 80-wire IDE cable makes any difference?

[alenppc]

Excellent, that one looks very similar to the one I have.

[plasmab]

Some IDE tests..

[ Show youtube player ]

[whiteb]

Originally Posted by nogginthenog

Just a thought: I wonder if a 40-wire or a 80-wire IDE cable makes any difference?

For the Amiga, an 80 core / 40 pin will just act as a 40 pin / 40 core.

The difference is the extra 40 core's act as extra grounding, to counter the extra crosstalk at higher DMA. In Amiga you are running at PIO modes.., no problem.

[whiteb]

Anyway, just watched the stream, interesting problems.

The crystal on the board you used, is that still a 24Mhz crystal and the FPGA's are multiplying up ?

Would it present any problem, using a 40Mhz, with a 40Mhz chip and the 40 Mhz firmware ? I mean would it then stuff up the FPGA's multiplication of the clock from 24Mhz ?

[alenppc]

@plasmab, you rock! Thanks a lot for doing this IDE video. As weird as this sounds, you have helped me solve the IDE problem.

I changed the CF card completely, I changed the INT2 cable, and disconnected the unused OVR wire. The problem is gone. So it seems that I either I had a dodgy INT2 wire or my CF wasn't compatible, but you being unable to reproduce the problem made me consider this and it worked.

Also relocating the scsi.device in fast ram instead of the whole ROM is a good idea and easily doable with loadrommodule. There is even a patched version available online.

Tested the new 24 & 33 mhz firmwares. They work perfectly, with and without FPU, on 25 & 36 MHz configs. Awesome! Gotta try to see if this makes a difference on the A1000.

[EzdineG]

@alenppc

Did you change the 68000 socket on the test machine? Is it possible for you to link me the pins you used to connect to that socket?

I'm if the idea all of the issues I'm seeing are due to contact. I can't do much as nudge the table without getting different results.

Thanks

[alenppc]

EzdineG, I didn't change the 68k socket, still using the original one, I just changed the wire to the INT2 signal on the edge connector. I used a thinner wirewrap type wire instead of the thick wire I was using before.