stween's build

[exxos]

At the moment all I get is 5V on all data lines, not the bit pattern from the other thread. Is that meaningful in some way, or does it nonsense?

If GLUE sees all zeroes on every address, and AS,UDS,LDS etc are all low, then you should get ROM2 output on the GLUE going low.. if not, then either you have a bad connection on CPU socket or a bad connection between the CPU and GLUE, namely on the address bus.

Or go around all the address pins etc on the GLUE make sure they are all connected to GND..

[stween]

At the moment all I get is 5V on all data lines, not the bit pattern from the other thread. Is that meaningful in some way, or does it nonsense?

If GLUE sees all zeroes on every address, and AS,UDS,LDS etc are all low, then you should get ROM2 output on the GLUE going low.. if not, then either you have a bad connection on CPU socket or a bad connection between the CPU and GLUE, namely on the address bus.

Or go around all the address pins etc on the GLUE make sure they are all connected to GND..

Ah yeah, all the address lines are showing +5 also. Gonna back this janky wiring job back out and re-evaluate!

[stween]

Alright, I've removed the connecting wires, and checked these lines with no CPU: all 23 address lines, AS, UDS, and LDS sit at 5v.

(With the CPU in place and using the scope: after reset I see AS flips high/low a few times in the first few clock cycles, sits low for a few more, then flips high and stays there. The address lines I've poked don't budge from 5v after reset. I didn't do these exhaustively.)

My current "I'm too tired to think" problem: when I attach the ubeswitch, the clock signals stop dead? Pin 2 on the shifter just reads as a stable voltage, not a clock. Remove the ubeswitch: clock signal. That's new behaviour, because previously I got a nice white screen. Now, nothing. Definitely feels like a couple of steps back, but I'm done for tonight!

[stween]

My current "I'm too tired to think" problem: when I attach the ubeswitch, the clock signals stop dead? Pin 2 on the shifter just reads as a stable voltage, not a clock. Remove the ubeswitch: clock signal. That's new behaviour, because previously I got a nice white screen. Now, nothing. Definitely feels like a couple of steps back, but I'm done for tonight!

Morning epiphany: I remembered that others needed an OSC fix and although I got clocks and a white screen without the fix previously ... now I need the fix if ubeswitch is connected to the monitor port.

Couple hundred ohms between 32MHz and GND on the shifter and I'm back to a white screen!

Back to the address bus: without a CPU, without doing anything else, what should the address lines + the three others read? Mine are all high, but should they all start low?

[exxos]

With no CPU pretty much all signals will be pulled high via the 2.2k bus pull ups.

You have to drive the address bus etc all to 0v. That forces address zero which they glue decodes and drives ROM2 low which drives the ROM CS low which then puts data on the bus.

[exxos]

I also assume you tried jumpering the TOS bank jumper ? I think jumperd is TOS104.

[stween]

With no CPU pretty much all signals will be pulled high via the 2.2k bus pull ups.

You have to drive the address bus etc all to 0v. That forces address zero which they glue decodes and drives ROM2 low which drives the ROM CS low which then puts data on the bus.

Okay, understood. Then I think I'm missing something obvious in this test on how best to drive everything to zero. (All the connected pins are high.)

I also assume you tried jumpering the TOS bank jumper ? I think jumperd is TOS104.

I'm often trying both ways, because I wasn't sure which way is which. (White screen is no jumper; black screen with white border is jumpered.)

[exxos]

Okay, understood. Then I think I'm missing something obvious in this test on how best to drive everything to zero. (All the connected pins are high.)

You just do the bodge wires in the socket as shown in the same post.. That's all you need to do...

[stween]

Okay, understood. Then I think I'm missing something obvious in this test on how best to drive everything to zero. (All the connected pins are high.)

You just do the bodge wires in the socket as shown in the same post.. That's all you need to do...

That's what I expected. But I had all lines at +5v when I expected all-zeros, so I figured I'd pulled in a stray +5 from somewhere. That's when I decided to back it out.

Rudimentary routing question: if the address lines + the three others are +5v and the test chains those together, what should drive them to 0? That feels like a daft question, but a) I never trust my intuition on how electricity flows and b) maybe I'm about to learn that +5v is binary 0.

(fwiw, I may also just grab a cheap logic analyser!)

[exxos]

When you tie all the pins together one wire goes the lot to 0v. I highlighted the pins on the socket, one should be 0v unless I missed it off... But the wire mess works if you follow that anyway.