Need help dumping the ROM in a failing 80C49 (MCS-48)

[trondl]

Nevermind! All good again. Deep clean with IPA and brush solved it.

[old text]
The weirdest thing just happened:
The Arduino shield doesn't do EA above 4.4V anymore.
It's supposed to do 4.4V, 12V and 18V.

One of the transistors gone possibly?

Schematics here: http://www.mattmillman.com/wp-content/files/mcs48programmer.pdf
[/old text]

[trondl]

Another discovery: EA is actually pulled rather high actually, about 12.9V on my DMM, and there is a spike at the start also.
I've included the screenshot I did to compare EA enable and Reset to check for a dump start before EA was stable.
That may explain why the 8749 has no issues, where the 8049 should have a very close to 12V applied.
Could this be an explanation for the freakout / noisy waveforms?
I could try to lift EA and apply directly from the shields Vin from Arduino, giving a continuous stable 11.5V.

It still doesn't explain why there is no 5V idle on the 8049 Databus though.

I'm tempted to do a basic breadboard version as done here: http://qemulator.blogspot.com/2020/05/dumping-intel-8049-roms.html
Should have more control over voltages and such.

The second MCU from Japan is verified in the unit I'm repairing, so that is a relief at least.

[ooops]
I tried to lift Vin and apply directly to EA.
Now the arduino shield is no longer detected :S :S :S
My impatience is starting to show.
[/ooops]

[trondl]

Nevermind again!
Shield and arduino fully working.
I panicked when all of a sudden the default COM port was changed in the HvEprom app.

[trondl]

Very late update here.
I've been pondering for the last two years about whether the data bus needs a buffer or transceiver in between to get the signals to behave.

I read a bit further in the service manual about how the pins are by default configured and that the data bus is by default an output.

Then a noteworthy observation struck me: The T0 pin (pin 1) is used to configure the MCU in SDE-1000 or SDE-3000 mode (it's bigger brother).
Pull T0 to 5V and it's in SDE-1000 mode, and to ground it's in SDE-3000 mode.

I have a hunch that if T0 is floating at start up, which it most likely is during startup on the programmer, the code will crash as it is in an illegal state, causing weird things to happen.

Am I on to something here?
If my knowledge is correct, T0 is also used to set programming (0V) and verify (5v) mode.
In Intel's suggested ROM verify procedure, T0 is not mentioned, only the EPROM programming procedure mention the T0 usage.

Suggestion:
I could tie T0 permanently to 5V from VCC or the Arduino as that would both set SDE-1000 and Verify mode, which is my goal here.
Worth a try?
In the SDE-1000 it is tied directly to 5V, but should I be on the safe side and put a current limiting resistor of some value between?
In the SDE-3000 it is tied directly to ground.

When dumping, I should get the whole code, not just the code for which SDE mode it is in?

[wraper]

In the SDE-1000 it is tied directly to 5V, but should I be on the safe side and put a current limiting resistor of some value between?
In the SDE-3000 it is tied directly to ground.

You can tie it directly however it's better to use a resistor, say 1k. As it's possible to configure this input as CLK out.

[trondl]

Lifting T0 to 5V did nothing different.
Random probing around, I found VDD to be 4.5V.
I'm tempted to lift and tie VDD to VCC for a stable 5V.
VDD is only boosted during programming.
The Intel datasheet claims VCC = VDD = 5V during verification.
Should be safe hopefully

[trondl]

Tying VDD to VCC helped VDD with getting proper 5V, but no stable data bus yet, apart from a slightly higher idle voltage on said bus(may be a hint to what is going on?).
Either a jump in voltage and/or a buffer / transceiver in between is my next agenda.
I fear putting more than 5V on the Arduino data bus may not be the best idea if a higher VCC / VDD also affects the data bus level.
From what I gather, VDD is powering the RAM and controlling the MCU power mode (idle / standby).

The data bus should idle at 5V as the 8749H does, but as mentioned the 80c49 idles around 1.5V.
Is this a case for a stronger pullup?
No pullups in the circuit as it is direct to the Arduino.
The waveforms looks like what I can imagine a 3.3V device quarrelling with a 5V device looks like.
The Arduino is a 5V device so it shouldn't be the case.
Could this also be an issue where the programmer is designed around HMOS (8749H), not CMOS (80C49)?

Putting a 74245 (HCT or LS?) in between shouldn't be too hard?
The most practical way for controlling the direction is where I may need some guidance though.