DSTB1: testers for rev 2 firmware (requires modding a rev 1)

[Oldskool]

I'm in the Netherlands.
Could maybe also be the combi with the Ultrasatan?
As it seems to crash the Ultrasatan device (I need to reset it as well when crashing).
Could send it to you (you're in the UK I assume?)

[Badwolf]

Could maybe also be the combi with the Ultrasatan?
As it seems to crash the Ultrasatan device (I need to reset it as well when crashing).

It may be a timing interaction with UltraSatan, but UltraSatan is very popular, so I'd want to get it to work with that.

Does enabling blitter cause a crash if UltraSatan is not connected (eg. booted from floppy?)

I'm in the Netherlands.
Could send it to you (you're in the UK I assume?)

Ah. No. Never mind. Brexit once again torpedoes doing things easily. :(

(I had a package to France sent back after three months as the customs declaration wasn't quite done right. I wouldn't want to lose your board in either direction.)

I'll try coding up some new experimental firmwares to test. Perhaps stick to 8MHz mode for now.

Would you mind measuring the resistance between BGACK and VCC on your motherboard (with the power off!) when you have time? I wonder if you have a stronger pull-up than me and it's causing things to tread on toes.

BW

[Oldskool]

You can always come back and exchange the pound for the euro.
Makes things even more easy :)

[Oldskool]

Not sure how you want me to measure the resistance but between pin 12 and 14 on the cpu its 4.7k.

[Badwolf]

Not sure how you want me to measure the resistance but between pin 12 and 14 on the cpu its 4.7k.

That's grand, thanks.

I'll see what mine read (as I know there are different BGACK pull-up strengths out there).

BW

[Badwolf]

Hi @Oldskool,

Just a note to say I've not forgotten this and to thank you for your help, but other things have been in my way recently.

I do intend to return to the firmware for DFB1r2 as soon as I can.

Cheers,

BW

[Oldskool]

Just trying out this Lightning usb device with the dstb1.
Everything works however when I try to set it to 16mhz the ide part of the Lightning crashes. Acsi still works.
Tried all the firmware files.
They all have this issue (if they work.)

8 clkmode2 ide crash @ 16
8 clkmode3 ide crash @ 16
8 clkmode4 does not start at all crashes the st
Clk thru does work at 16 mhz but does not seem to accelerate the system (still at 8 mhz i assume)
Dstb1r2 ide crash at 16
Fastbgo ide crash at 16

The usb part does work. The Lightning has various ide settings but non of them work at 16. At 8 all is ok.

IMG_3389.jpeg

But when you hit 16 mhz

IMG_3390.jpeg

[Oldskool]

This issue seems to be a problem of the lightning device. Only works at 8 mhz. Also tried it with a hbs640.
And the the people of team thunderstorm also don’t support any hardware on the exxos forum. Don’t exactly know what the mean with that. But OK…

[Badwolf]

This issue seems to be a problem of the lightning device. Only works at 8 mhz. Also tried it with a hbs640.
And the the people of team thunderstorm also don’t support any hardware on the exxos forum. Don’t exactly know what the mean with that. But OK…

If the Lightning is mounted 'beneath' DSTB1 then it should effectively be working in 8MHz mode anyway.

However it's not quite that simple: DSTB1 will make some assumptions about what it can get away with timing-wise on the clock switching given it's expecting to be talking to an ST chipset. It's possible [likely?] that's not true with a newer device that probably has its own timing assumptions based on expecting to talk directly to a 68k chip.

The most obvious timing abuse DSTB1 employs is that it allows AS to be asserted immediately and then switches down to 8MHz. that process takes a few CPU cycles, but it cancels out with the normal 68k delay to UDS/LDS on a write or the interleaved RAM access on a read.

The advantage of this is that ST-RAM access is maintained at or near 100% of normal speed, however this CPU clock jiggerypokery could, I imagine, upset something that's relying on the pseudo-synchronous accesses the ST normally provides.

You could try building a firmware that gates assertion of AS until the clock switch has completed. I suspect this would result in a poorer overall performance, however.

To go much further you'd probably have to get a scope or logic analyser on things and work out exactly where it's failing.

Does it work mounted 'on top' of DSTB1? Ie. can Lightning run with a faster clock?

BW

[Oldskool]

No it can’t run with a faster clock already tested it with a hbs640 with just the clock at 28 cache off.
So will just safe it for another project and keep the dstb1 in the old stf….
Fist thought it was a dstb1 thingie but it is not…