Page 6 of 9

Re: STe booster v 1.5 36Mhz

Posted: Mon Feb 11, 2019 4:25 pm
by exxos
The board I have been using is the integrated with the kind and it is the one I am having all sorts of weird issues with.. So I dug out another degraded board tried the booster on that one and still getting all kinds of weirdness. Mostly just random lock-ups for no reason at all.

I moved to a third board, which is a external blitter board, and this one works perfectly fine. So I do not know is wrong with those integrated blitter motherboards.. They just really do not play nice with the booster some odd reason.

So for those who are just having hard drive troubles, the V1 code I posted above could fix that issue. As for intermittent lock-ups, other than bad connections, if this is on a integrated blitter motherboard, I don't know what the fault is with those. Changing the motherboard is the only solution.

I have gone to great lengths to try and figure it out but I'm just going to have to stop selling these boosters as there are just too many issues to debug and go through.

Re: STe booster v 1.5 36Mhz

Posted: Mon Feb 11, 2019 7:20 pm
by ijor
exxos wrote: Fri Feb 08, 2019 12:01 pm As Ijor pointed out with the slow SND DTACK issues on the STFM, he says this issue was fixed on the STE.
I think I should say that I am not 100% sure about that. All I know is I can see the modification and the fix at the STE MCU combo internal schematics (the schematics that Christian recovered). Probably this fix is implemented at the actual chip silicon, but I didn't perform any kind of test on the STE to confirm this.

Re: STe booster v 1.5 36Mhz

Posted: Mon Feb 11, 2019 7:24 pm
by exxos
ijor wrote: Mon Feb 11, 2019 7:20 pm
exxos wrote: Fri Feb 08, 2019 12:01 pm As Ijor pointed out with the slow SND DTACK issues on the STFM, he says this issue was fixed on the STE.
I think I should say that I am not 100% sure about that. All I know is I can see the modification and the fix at the STE MCU combo internal schematics (the schematics that Christian recovered). Probably this fix is implemented at the actual chip silicon, but I didn't perform any kind of test on the STE to confirm this.
That was posted some days ago.. There isn't any sound issues so should be ok, but still could be some "slow" signal somewhere. its something to do with the integrated blitter boards. I'm not going to look into it anymore anyway. Kinda had enough of it all now.

EDIT:

I just tried one of the integrated motherboard, with a different booster, still running from the STE's 32mhz clock.. It booted and ran GB6 twice without issues.. I a cold boot with the keyboard, now it is just frozen on the Atari logo.. No bombs.. No memory test.. Nothing..

I've said before it seems like the fault changes every couple of reboots.. It has to be something to do with the integrated blitter as it is the only difference that I aware of. Like there must be some sort of internal state machine powers up after reset at random. In one state everything works fine, and in another state just malfunctions continuously..

Similar happened a couple of days ago.. Ran 20 loops on GB6, then powered off during the night, powered on in the morning, machine totally dead.

Re: STe booster v 1.5 36Mhz

Posted: Tue Feb 12, 2019 9:15 am
by exxos
I had a thought that I started using Atmel GAL's, so for the hell of it I put back a lattice GAL.. And now the problematic integrated blitter board now seems stable :roll: And yet, the Atmel GAL works fine on the external blitter boards :roll:

Even though speed of the GAL has never made any difference before,RS actually stock the 7ns,15ns,20ns parts, so I'm going to order those and try them and see what happens.

Though this all goes back to some other epic drama from a few weeks ago, where various batches of lattice GAL only worked.. This is why I started buying the Atmel GAL's as they should actually be genuine specification and more consistent with tolerances. Obviously this is a solve one problem and create another type situation :roll:

Re: STe booster v 1.5 36Mhz

Posted: Tue Feb 12, 2019 5:14 pm
by exxos
So the plot thickens.. After endless missing about.. I now have the Atmel GAL working on the integrated blitter motherboard with the STE's 32MHz clock.

Now for the 36MHz clock.. I'm extremely limited by what I can do with this design, the only way to work is if I invert the 8MHz clock. Speed test why is it does not seem to make any difference and seems stable. I'm not really happy about this solution as I don't know if they will be any other side effects to this or not. It has been running at 36Mhz for the past couple of hours now, again with the Atmel GAL and the "problematic" integrated blitter motherboard.

This basically means there would have to be two sets of firmware depending on where the faster clock is generated from. So I will see if I can somehow degrade both firmware is into one GAL, then the user can select which mode to use.

36MHz isn't that much faster over 32MHz other than integer division goes up from 355% to 392%. Will try again 40MHz is for the hell of it to see what happens later..

Re: STe booster v 1.5 36Mhz

Posted: Tue Feb 12, 2019 5:52 pm
by PhilC
Surely you've got to get lucky with this st some point, I mean, myself and RJs ones work fine at 40 as you know.

Re: STe booster v 1.5 36Mhz

Posted: Tue Feb 12, 2019 5:58 pm
by exxos
Forgottenmyname wrote: Tue Feb 12, 2019 5:52 pm Surely you've got to get lucky with this st some point, I mean, myself and RJs ones work fine at 40 as you know.
Is it working with a hard drive though ? RJ's isn't, but its passed millions of loops on GB6 at 40MHz. The hard drive issue should actually be fixed now anyway. Is trying to work out why one GAL works and another one doesn't, and why a separate oscillator is all of a sudden misbehaving.

Re: STe booster v 1.5 36Mhz

Posted: Tue Feb 12, 2019 6:11 pm
by rubber_jonnie
Yeah, my 32Mhz V1 works fine and dandy at full tilt with Gembench6 off floppy or HDD, but the 40Mhz ran all through last night running GemBench off floppy, but won't play with the HDD.

GB just bombs out (2 bombs I think) whenever I try to run from HDD, and as soon as I flick back to 8Mhz it's fine.

Re: STe booster v 1.5 36Mhz

Posted: Tue Feb 12, 2019 6:16 pm
by PhilC
Will they to remember to connect the ultrasatan to it when I'm home later in the week

Re: STe booster v 1.5 36Mhz

Posted: Wed Feb 13, 2019 10:21 am
by rubber_jonnie
Forgottenmyname wrote: Tue Feb 12, 2019 6:16 pm Will they to remember to connect the ultrasatan to it when I'm home later in the week
Mine is a CosmosEx, which is equivalent to US BTW.