Anyone got a STE with DMA issues ?

[exxos]

The other option might be to build one of those STM32-based kits & try one of those instead...but are they likely to show problems as much as the U/S do...?

I don't know how flaky they are to start with. At least with the ultrasatan it is basically tried and tested by a large number of people.

I did look on evilbay but the only STE there is a motherboard and I think it was £75+. So not really worth buying it just on the off chance it does not work properly :lol:

I do have a untouched machine or two here which I already tried and did not fail. So not really sure what is going on there. If I get chance over the coming days I will see if I can just heat up a board little bit and see if it fails.

[exxos]

Been doing some soldering on a new STE test machine. I gave up trying to get DMA issues so decided to deliberately cause them by installing a IMP DMA. it did not disappoint :lol:

It would load stuff seemingly fine but as soon as I created a folder it trashed the entire drive. So I'm going to start running the existing hard drive test software to see if it trips up firstly...

IMG_0081.JPG

IMG_0082.JPG

[exxos]

FFS! You cannot use HD11 on a 512K machine ? Seriously ?! :roll:

IMG_0085.JPG

[exxos]

So it's failing on writes all the time it seems when trying to partition in HD11. Seems to setup fine but after reboot the partition is gone. It passes jookie's DMA test program fine 99% of the time. Even the new program I wanted to test :roll:

Anyway, will start doing the usual fixes to see if the IMP DMA can be made to work. Though in the past I think I gave up with it and advised never to use IMP DMA in an STE.

Will spend tomorrow on it but I'm not going to spend huge amounts of time on it again.

[exxos]

So it's pretty odd with the IMP DMA. Using a HC CPU makes no odds. Though adding pullups on the IMP bus regardless of value seems to make things worse. One of the control signals could be slightly out causing issues. Its hard to tell. Someone with a LA would have to get the whole thing connected to to compare the IMP to a -38 DMA , or whatever works, to see what is actually different. There is no use doing any further tests because just about everything and being tried previously anyway.

Anyway, I would just suggest not using a IMP DMA in a STE at all. Annoying as I can't seem to simulate the write faults with jookies DMA test program when it will trash the drive easily on desktop. Mostly it was the -38 (but sometimes the "good DMA") which would malfunction without a HC CPU. But I cannot find a STE currently which faults with a -38. Anyway, I am not going to pursue this any further now.