exxos's Atari, Amiga & retro forum

exxos wrote: 12 Jul 2018 12:58

Forgottenmyname wrote: 10 Jul 2018 21:24 C1 is a little wonky though :P :lol:

I charge more for unique items like that ;)

Seems fair, lol. Hate to think how much the missing.... err custom wired ones will be :D

Great to see all this work seems to have made a very stable booster, well done!

Quick question. Will the new accelerator work with the original dual TOS board, and if so, what mods are needed if any?

rubber_jonnie wrote: 13 Jul 2018 08:39 Quick question. Will the new accelerator work with the original dual TOS board, and if so, what mods are needed if any?

Any of my dualtos boards will work with it. Just this later one is smaller and has the UBE circuit (well will do). So its like a upgraded dualtos board.

The only minor issue is that I think the integrated blitter boards may need a 33R resistors in the ROM CE line to the booster.. but thats easy as there has to be a wire between the 2 anyway. The new dualtos boards has the resistor on the pcb.

exxos wrote: 13 Jul 2018 09:02

rubber_jonnie wrote: 13 Jul 2018 08:39 Quick question. Will the new accelerator work with the original dual TOS board, and if so, what mods are needed if any?

Any of my dualtos boards will work with it. Just this later one is smaller and has the UBE circuit (well will do). So its like a upgraded dualtos board.

The only minor issue is that I think the integrated blitter boards may need a 33R resistors in the ROM CE line to the booster.. but thats easy as there has to be a wire between the 2 anyway. The new dualtos boards has the resistor on the pcb.

Ahh, cool.

The STE I will likely put one of these in has the latest DualTOS board from you, PLUS one of the UBE circuits too:

rubber_jonnie wrote: 13 Jul 2018 11:25 The STE I will likely put one of these in has the latest DualTOS board from you, PLUS one of the UBE circuits too:

:thumbup:

New boards are going to be red, so if you have colour OCD, you will have to get a red dualtos board ;)

exxos wrote: 13 Jul 2018 11:31
:thumbup:

New boards are going to be red, so if you have colour OCD, you will have to get a red dualtos board ;)

Nope, no OCD from my PoV, as long as it works I'm happy :)

So running 2 days now, it decided to die again.. which mirrors the same thing as I talked about earlier https://www.exxosforum.co.uk/forum/viewt ... t=80#p6446

I think not all CPU's can run at 40MHz.. Oddly this CPU started working for a short time, then died again. Its not working in 8MHz mode either. Really need to get some freezer spray :roll:

So in light of this, I will likely build up a new board with something like 35MHz osc and see how things go then. Obviously its not good to keep running at 40MHz if its killing some CPU's.

EDIT:

re-soldered the PCB just to make sure still same issue.. Heatguned the CPU off the board and works in the STE now :WTF: Maybe the heat made the CPU work again :shrug: So will see if I can solder this CPU back on the PCB...

I took out the GAL and all the resistors and basically hardwired the CPU in as a stock setup, and it still does not work..

So I am back to thinking this is sort of bad connection fault again.. Though I do not like getting into these "going around in circles" problems. I guess it is possible these PCBs are simply rubbish..

I got my regular PCB guy to manufacture the final boards anyway. So of course I will do a lot of testing with those. Likely I will have 2 machines running, one with a 40MHz OSC, and another with a 32MHz OSC. And see if the 40MHz one dies or not.

I am still investigating, this seems a bit of a mystery :roll:


EDIT:

Checked bottom of the motherboard to the top of the booster CPU, all connections are fine. So there is nothing other than the CPU in circuit, direct connection as "normal" just doesn't work.

I took the CPU off the booster , back in the MB... works again.. So I will try a new CPU on the booster board.. This really doesn't make any sense...

EDIT2:

OK So I plugged the booster board directly ontop of the CPU in the MB socket.. now it doesn't boot.. hmmm.. there are no shorts on the board from what I can see.. even if there was, doesn't explain 10+ hours before it failed...

EDIT3:

OK so short on PCB between DTACK and BG , must have happened during my last re-solder.. So back on with the CPU again...and.. still doesn't work.. bizzare..

EDIT4:

Tried the suspect CPU in my old V1 booster (32mhz) and CPU came right on.. so can't be the CPU... all connections on the new booster board seem fine.. so what else is there ??

EDIT5:

Put a new CPU on the problem booster board and now it boots! So why does that one CPU work in the MB, work in the V1 booster, but died on the new booster boards.. Even though at the moment its hardwired for 8MHz.. No bad connections, no shorts, just refuse to work..


EDIT6:

So back to working at 40MHz now.. well bit unstable as not got the sil arrays in yet.. and its died again... :pullhair:


EDIT7:

Tried a previous revision board and hardwired that to 8mhz and seems to exhibit the same "fault".

I really think there is something bad on these test boards.. but I don't know what.. Will just have to wait until I get my proper boards and do more tests on those..

EDIT8:

So I found a short on the CPU :roll: So now its running at 40MHz again (though with a different CPU than originally). Though it still doesn't explain why it died earlier when the CPU tested fine in another board.

Think I am just going to put this down to its solar eclipse & Friday 13th.. Good enough explanation as any other...

Had a thought to check the CPU I was having issues under my microscope.. And look what I found..


Basically two pins have not had any solder on them. I don't know why as each pad on the PCB had paste, and I even checked under a magnifier that each pin was sitting in paste.. But the some reason paste did not adhere to the 2 CPU pins, even despite three or four re-solders. This also explains why it worked fine with a totally different CPU (I did not even re-apply any paste on this, just heat gunned the new CPU, so it was not a issue of not enough paste on the PCB).

As to why this worked for several hours without problems then suddenly died, and even past continuity checks is bit odd. Probably there was a few ohms resistance which made it seem okay with my meter, but when running at some MHZ , likely the resistance basically became open circuit. Basically similar kind of issues with bad contacts in PLCC sockets which I saw a lot of also.

So what I will do in future is leave the CPUs soaking in some IPA before I solder them, and of course never assume that CPUs are clean, even on "new" reels. I guess age could be a factor here.. But there does not seem to be any tarnishing on the pins and they look clean. The pins are straight and aligned with the ones next to it. So just something seriously odd where the solder simply would not stick to the two pins :roll:

Well spotted.

Not sure why it would happen but can only assume oxydisation on the pin over the years, not that we can see it on the photo.

Could it be that the cpu and pcb flexed under the sudden application of heat? Reading the instructions on the paste I use talks about raising the temperature in a set procedure.... Not that I ever do? Maybe it would solder better in an oven? Just put it under the pizza not on top.