PC Engine PI-TG001 issues

[Fidde_se]

That could just be rosin flux, most in there is lead/tin plated copper meaning it will be green when corroded.

Sometimes when things hit the floor, like when stepping on the cable, there are extreme flex on the board leaving bigger circuits (SMD/SMT) to get loose in their pins (legs), it might look good to the untrained eye but it's not really soldered, going over everything with a microscope might be needed or even resoldering everything just to be sure.

[ScoreAddict]

Any more ideas?

[thesteve]

condition looks about average
got a logic probe? (or scope)
you need to run across the slot pins without the card and see what its doing
almost every pin should read high/low or pulse

[ScoreAddict]

I did some voltage measuring on the card slot!

pin voltage
01 0.0 v
02 2.5 v
03 1.7 v
04 0.0 v
05 0.9 v
06 3.4 v
07 1.9 v
08 3.4 v
09 3.0 v
10 2.4 v
11 4.1 v
12 2.9 v
13 2.9 v
14 2.1 v
15 4.9 v
16 5.0 v
17 5.0 v
18 0.0 v
19 5.0 v
20 5.0 v
21 5.0 v
22 5.0 v
23 5.0 v
24 5.0 v
25 2.5 v
26 3.6 v
27 2.5 v
28 2.2 v
29 2.2 v
30 0.0 v
31 0.0 v
32 0.0 v
33 5.0 v
34 5.0 v
35 5.0 v
36 5.0 v
37 5.0 v
38 5.0 v

No idea if this helps.

[TheOldMan]

No idea if this helps.

Some. At least we know the cart is getting power (pin 38 == 5V), and that /CART is right (0V).
The fluctuations on the address lines probably mean they are changing (if you used a meter, those are average voltages)

It's the block of 5V in the middle that bothers me; Pins 15-23 (excluding pin 18, which is GND == 0V)
should be changing as well. Yet it appears they are stuck at 5V, which is a 1.

Looks like a dead card to me. Or possibly the data lines are shorted together somewhere...

[ScoreAddict]

At least we know the cart is getting power (pin 38 == 5V), and that /CART is right (0V).

Good. At least something.

The fluctuations on the address lines probably mean they are changing (if you used a meter, those are average voltages)

I used a meter.

It's the block of 5V in the middle that bothers me; Pins 15-23 (excluding pin 18, which is GND == 0V) should be changing as well. Yet it appears they are stuck at 5V, which is a 1.

OK. Should I do some more measuring? Any suggestions?

Looks like a dead card to me. Or possibly the data lines are shorted together somewhere...

Since I have two more PC Engine consoles to test (and play) it, I can confirm that the card is in fine working condition.

So maybe it's one of the connection between the main pcb and the card slot?

[thesteve]

i thought that was without card
pins4 and 30-32 concern me

[TheOldMan]

pins4 and 30-32 concern me

That's A16, A13, A14, and A17, right? During boot up it reads page 0, so I could see those being 0V.
Not sure why A15 (pin 5) would be fluctuating though, so it could be an address line problem.

i thought that was without card

Wouldn't you see 0's without the card?
Either way, they should be fluctuating as the cpu acceses the card, if the addresses are changing. (As they appear to be).

My next step would be to check for a shorts between the pins, both with and without a card inserted (but with the machine off).  I know the card slot can get worn, and the connector fingers either don't make contact, or short 2 connections together. After 20+ years, it can get touchy about how the card sits.
Might be as simple as inserting the card to one side or the other of the connector.

If all seems good, its time to get a logic probe, I think.

..................................................................................................
Just to satisfy my curiosity, (and the original reason I replied at all), did your friend really
tell you to 'put a 1' on pin 1? You can tap 5V from the 7805 for that...but I don't recommend it.
Pin 1 (/CD) connects to ground inside the card; without a current-limiting resistor you'll probably blow either the 7805 or something in the card.

I'm not an expert but even I know you don't short power directly to a ground without -bad things- happening..

[thesteve]

the data lines have pull up resistors and remain high without a card

[ScoreAddict]

You know, ORIGINALLY I was told "to put 5V to pin 1". Which - naturally - blew the fuse. Hey, I'm not the one called "Japanese console expert".

Then my "buddy" corrected himself by saying he meant "pin 1 to high".

He suspected the card slot or the cable connections to be the issue. Which are certainly prime candidates on such a used up console. Not sure about a burned out IC though.

[thesteve]

the ribbon cables break on those easily so that is likely
also common are trace and via corrosion (not always visible)
3rd option is chip failure (it happens)

[ScoreAddict]

OK.

Next steps would be to unsolder the card slot-pcb from the main pcb - maybe even the card slot itself? And then use a new set of cables to connect the two pcbs together to see if that was the issue.

Too bad NEC did not go for an integrated solution in this case. Or at least use a decent way to connect the two pcbs together.

[mickcris]

You could ohm out the cable before committing to all that work as that is only one possibility.  I would check it first to make sure if it was me.

[ScoreAddict]

I've "beeped" through all the cable connections already once. But if there's a broken cable, it will not necessarily be detected this way.

[mickcris]

I've "beeped" through all the cable connections already once. But if there's a broken cable, it will not necessarily be detected this way.

Why are you thinking a wire would be broken  if your meter says it isnt?