dead TT motherboard repair attempt

[jookie]

I've replaced the TTVIDEO chip and the other clocks (16 MHz, 4 MHz, 2 MHz) are now generated (they weren't before).

After soldering new 68030 to the board, I can see 2 low pulses on XROM1 (TOS EPROM CE signal), which looks like the new CPU is alive. HALT signal is always H, but there are 2 low pulses on BERR signal, according the TT service manual:
"If there are logic low pulses, some component is malfunctioning and the PAL as location U112 is generating the error. Verify the clocks and replace these components to verify them."

So now I'll have to see what could cause the U112 to generate the BERR pulses, then investigate those chips.

tt030_signals.jpg (64.42 KiB) Viewed 3521 times

[jookie]

Only 2 accesses to XROM1 looked suspicious, I've looked at the data and the data signals were damped (1.77 V max instead of 5V). As all the data signals were damped, I assumed it would be caused by chip which has all the data pins connected, and that would be:
- Funnels (U205 and U209)
- DCUs (U901 and UA01)
- CPU (but that was just replaced, so hopefully not this one)

I've picked a data pin on EPROM - D1. I've cut a D1 pin on Funnel (U205) - no change, then desoldered pin on DCU 1 (U901) - no change, desoldered pin on DCU 2 (UA01) and now the D1 signal can reach 5V levels. I've removed that DCU (UA01) and now all the data lines are fine (going fully to 5V). I've fixed the pins on U901 and U205 back.

Now I can see more EPROM access on the XROM1 signal:

xrom1_more.jpg (86.49 KiB) Viewed 3517 times

...but there's still BERR pulses afterwards. I'm stil waiting for DRAM chips from China - at least one DRAM (U512) was showing much lower resistance than others and has been removed. Once I get that DRAM chip in place, I might hook up the diagnostic cart to see if something appears on the serial port.

[jookie]

The DRAM chips arrived today, I've inserted the missing DRAM chip (U512) into socket, checked the RASx and CASy signals, one of them wasn't moving (CAS2). I've found a missing connection on CAS2 (on a U202 chip, which I've replaced) and fixed that. Now the TT boots into (exxos) diagnostic cartridge menu. Current state of the machine:

• the machine doesn't boot into TOS (black screen), but boots fine into diagnostic cart menu, so CPU, memory and video chips are OK. This might be to still missing UA01 chip, which I've removed few days ago, because it was blocking / dampening the data bus
• doesn't boot at all with FPU inserted, boots into diagnostic cart menu without FPU (so FPU is broken)
• boots fine with the ST RAM extension card (2 MB), so it passes RAM tests with 4 MB ST RAM total
• the machine doesn't boot after power on, requires a RESET button press to boot (and also requires a reset button replacement)

So my TODO list now is:
- investigate XRES and XPOR after power on
- order and solder DCU (UA01)

tt_diag.jpg (233.28 KiB) Viewed 3449 times

[frank.lukas]

- investigate XRES and XPOR after power on

PowerGood signal check too ...

[jookie]

PowerGood signal check too ...

You're right, that will be the cause of the power-on reset issue. As I bought this TT as motherboard only (no case, no PSU, ...), I'm using MeanWell RQ-65B instead of original PSU, and the PowerGood (PG) line is not present on that PSU, so I've only got a 330 R pull-up to 5V line, but that won't do the power-on reset. Is there anything that is usually used to create the PG line? If not, then I would use MCP100-475DI/TO (microcontroller supervisory circuit) and it should do basically that.

[frank.lukas]

Why not use for repair a old AT PC PowerSupply or ATX with Adapter ...

[jookie]

Why not use for repair a old AT PC PowerSupply or ATX with Adapter ...

Yes, I could do that, but I don't have one around, and I already have the RQ-65B, so adding something (e.g. MCP100-475) for the PG line s an option.

[Steve]

Hey Jookie,

Just looking at that MCP100-475, so is that basically just a kind of voltage tolerance detector? So if the voltage isn't within spec it doesn't output anything?

I saw the specs say:
Overvoltage Threshold: 4.75 V
Undervoltage Threshold: 4.5 V

That's pretty cool. Is this what the real 'power good' line in the TT psu does? I've always wondered. I guess technically you could just route 5v directly to power good and it'll work, but miss the point of it?

On further recollection, I remember my TT psu 'blew' once and outputted 7v over the 5v. Lucky my TT survived. But I guess the power good line didn't do its job in my experience here...

[jookie]

Just looking at that MCP100-475, so is that basically just a kind of voltage tolerance detector? So if the voltage isn't within spec it doesn't output anything?

Yes, if the voltage is below the threshold, it outputs the reset signal:

MCP100 is a voltage supervisory device designed to keep a microcontroller in reset until the system voltage has reached the proper level and stabilized. It also operates as protection from brown-out conditions when the supply voltage drops below a safe operating level. The MCP100 will assert the RESET whenever the voltage on the VDD pin is below the trip-point voltage

The reset line is asserted (active) for ~350 ms, so after a power on there should be at least this amount of reset signal active.

Is this what the real 'power good' line in the TT psu does?

Yes, that should be it. Compared to MCP100 chip the whole XPOR (power-on reset) in TT is split to two parts:
- the TT PSU does the comparing of 5V line to some threshold value (see 'PSU_PSM-5341.pdf' - Phihong TT PSU schematic) - the PG line
- the U103A (LM556 dual timer chip) then does the duration of XPOR signal for around 240 ms.

I guess technically you could just route 5v directly to power good and it'll work, but miss the point of it?

That's what I have now, but the TT doesn't boot when powered on (it needs additional reset button press). With all the right voltages from PSU it could work after power on, but the CPU might be in some weird state, so the XPOR line needs to reset the CPU after this power on. With direct 5V connection to PG line this reset doesn't occur.

On further recollection, I remember my TT psu 'blew' once and outputted 7v over the 5v. Lucky my TT survived. But I guess the power good line didn't do it's job in my experience here...

The PG line doesn't have the power to protect the computer from such over-voltage situation, the best it could do is to reset the computer if that happens, but that won't protect the chips from death from the over-voltage. (Plus most of these PG lines don't check for over-voltage, only under-voltage).

[jookie]

The results from diagnostic cart tests:

• RAM: ok
• Keyboard: ok
• OS ROM: ok
• MIDI: ok
• Serial port - ok
• SCC: loopback passes on Port A, B, D - ok
• Color: ok
• Real Time Clock: ok
• VME: No VME board found in system (which is correct)
• Audio: Yamaha channels OK, DMA sound doesn't get through LMC1992 (but sine wave is present at the input of LMC1992), but both Yahama and SNDSHIFTER work, so shouldn't cause BERR issues
• Floppy disk: no floppy connected (yet)
• SCSI bus: still missing DCU (UA01) on motherboard, gets stuck on 1st test
• Printer/Joystick Port test: probably OK, parallel port data goes out of Yamaha, BUSY signal toggles. Getting P1 Busy line error but don't have parallel port + joystick ports cable (test fixture).
• DMA port test: D0 - timeout, fail at cycle: 00000001, but don't have DMA test fixture

The things I'm still missing on motherboard are DCU (UA01) and FPU (U101). My guess is that those two missing chips shouldn't cause the TT not booting (black screen) to TOS, I'll verify this assumption once those chips arrive. If that would be the case, I need to figure what else I can check.