kludge’s TF534 build

[alenppc]

Does everything go faster with CPU FASTROM or is it mainly the IDE?

Everything, it really makes a massive difference.

[kludge]

Everything, it really makes a massive difference.

I might just have to source at least one 030 with MMU to compare then. I'll have to see it to believe it ;)

[EzdineG]

Everything, it really makes a massive difference.

I might just have to source at least one 030 with MMU to compare then. I'll have to see it to believe it ;)

You shouldn't technically have to. I haven't tested the impact on IDE speed using other methods (all MMU equipped CPUs here), but using a program like skick to remap kickstart to RAM should have the same effect. Some of these work without requiring an MMU.

[alenppc]

I might just have to source at least one 030 with MMU to compare then. I'll have to see it to believe it ;)

You shouldn't technically have to. I haven't tested the impact on IDE speed using other methods (all MMU equipped CPUs here), but using a program like skick to remap kickstart to RAM should have the same effect. Some of these work without requiring an MMU.

Fair point, I haven't really tried skick. You can also load an A600 scsi.device in ram for a modest IDE speed increase.

[kludge]

You shouldn't technically have to. I haven't tested the impact on IDE speed using other methods (all MMU equipped CPUs here), but using a program like skick to remap kickstart to RAM should have the same effect. Some of these work without requiring an MMU.

Fair point, I haven't really tried skick. You can also load an A600 scsi.device in ram for a modest IDE speed increase.

Thanks, guys. I'll play around with it in the coming days to work out if I feel confident selling a TF534 without a MMU capable CPU on it.

What's a good program/game/demo to keep running to stress test it a bit? TestKit? Frontier? Some mandelbrot generator? (The whole construction I mean, not the IDE)

[alenppc]

What's a good program/game/demo to keep running to stress test it a bit? TestKit? Frontier? Some mandelbrot generator? (The whole construction I mean, not the IDE)

Frontier will warm the CPU up a bit. If you can get it to run for a few hours without crashes, you should be good.

[kludge]

Frontier will warm the CPU up a bit. If you can get it to run for a few hours without crashes, you should be good.

Good. It's quite fun to watch still. Maybe not for a few hours, though :)

[kludge]

For anyone wondering (yeah right) the TF534 stopped recognizing IDE devices from one day to the next. I don’t have any spare cables so I’m waiting for a delivery to be able to investigate what’s wrong :)

Managed to repair a Fastlane SCSI Z3 in the meantime anyway :)

[kludge]

I just discovered that my 10uF 1206 capacitors are X5R and not X7R... But that should only be temperature related, and I'm not sure that that would make a difference after running the board for less than an hour. X5R should tolerate 85°C, so I would think that they should be ok to use. I will have to check the 0.1uF 1206 cap. I think I just took something that I had laying around at the moment.

(This project has been put on hold for a while, but I'm slowly working myself back to it. Amongst other things I've bought an Amiga CD32 that I discovered was an NTSC one, so I've started to collect information and components for converting it to PAL...)

[kludge]

So I've dug myself down to the TF534 build once again, and had to backtrack some because I had forgotten exactly what the behaviour was.

Everything seems to work just fine except for the IDE. I've tried about eight different CF cards and a mSATA to IDE adapter. Neither Workbench nor DiagROM finds any of them and just reports that no disk was found. With one CD card and the mSATA to IDE adapter DiagROM just freezes and waits for something while trying to detect the drive.

I'm leaning towards a bad solder connection somewhere. A bit weird as at least one of the CF cards worked for a while, but I've seen weirder.

I've just got to decide if I'm going to reflow this board or build another TF534 and then get back to this at a later date.