Soooo Back to the TF536...
I think the floppy issues I was having was down to a tiny short on the PLD
mostly it was running fine from the motherboard ROM (TOS206) but it was not running from the flash on flashy clock.. And yet if I used the original motherboard processor (PLCC) it would work fine..
I think what the problem is the current version of flashy was designed to run on a 5 V slot, but the actual flashy port is a 3.3 V port now.. It resulted in two buffers in series. I added a load of pull-up resistors on the 3 V bus, which I'm going to add on the next H5 board I think anyway, as this did seem to solve some sustainability problems to a point.. Though removing the buffers on flashy and hardwiring them over still resulted in the same problem..
For seem to help was having a diode from the TF536 DTACK pin to flashyclock DTACK.. I think the problem was because to buffers were in series, it was not fully getting to 0V making DTACK very "hit and miss"... So I think the logic on the 536 must not be as tolerant to logic levels as 5V stuff on the motherboard is. In any case faulty is down to the buffers on flashy more than anything.
@Icky Is working on removing the buffers to create a proper 3V version for the H5.. the conundrum here is that IO buffers are used as logic level converters between 3.3 <> 5V logic, with these the MOSFET base, they have zero delay which is why I like using them as you can pretty much pretend they are not even in the circuit.. But I need to use one of these on flashy to isolate the lower databus, so I cannot use these particular buffers as a voltage drop across them is causing the above mentioned faults. So using a 3.3 V bus isolator with output enable is the way to go.. But this also creates some delays which I really want to avoid for reasons I don't want to go into right now..
I have been talking to
@Icky Various things over the weekend.. Mostly I was going to do 5 V versions of things like flashy clock and Trudie so they can be used on original machines and indeed the H4 boards. Time and sales wise I do not think it is worth the investment. So I am actually going to ditch the 5 V versions of stuff altogether.. This does not mean we are going to abandon H4 or original machines.
@Icky Is working on a dedicated IO buffer board to do the logic level translations.. Just the H5 boards already has them, so that board will not be needed.
Also my perspective is a problem with 3.3V stuff being on a 68K 5volt footprint.. Is on Trudie I added a crowbar onto it they would not get blown up if someone plugged it into a 5 V port... But with is being pushed for space on the board and that part of the circuit is currently on tested, really should not have to be trying to pre-empt all the dumb things which people will do with it all. So actually considering doing away with the 3.3V 68K footprint on the H5 boards and using our own footprint so people cannot physically plug in the 3 V stuff into a 5 V socket! For original norm machines or any machines with a five old only port, they have to use double buffer board now will have to new footprint on it anyway. So everything will still be backwards compatible but it will save was a whole lot of hassle as well.
- IMG_6208.JPG (334.91 KiB) Viewed 3905 times
EDIT:
So oddly, running TOS206 from 3.3V bus, the TF536 locks up with a row of bombs, but only when I go into the options menu on the desktop... It will otherwise load GB6 and run all tests fine... and yet TOS206 on the 5V bus doesn't do that .. hmmm....
