Print

[ried]

A while back, I turned my Lisa 2/10 off and removed its rear panel to access the card cage and swap around a few memory boards for testing purposes. I had one of James D's awesome RS232 WiFi Wireless Modems attached to the Serial B port, and left the modem's USB power cord plugged in while I removed the Lisa's rear panel (I know, I know... not smart).

Before I removed the modem and took out the card cage, I laid the rear panel down flat behind the Lisa. I clumsily bumped into it and pushed the panel with its metal shielding underneath the RS232 WiFi Wireless Modem, which contacted the modem's exposed PCB solder joints - while it still had USB power. Tiny sparks and an unpleasant "shhhh" sound for a brief moment before I quickly slid the rear panel back out from underneath the modem. Yikes.

I figured the RS232 WiFi Wireless Modem was toast and decided it's best not trying to use it again. Its replacement arrived and now I find that the Serial B port does not seem to communicate with the new replacement modem, so the Lisa has likely been damaged as well. The Serial A port was not affected and continues to work as normal.

LisaTerminal commands will not register when the modem is connected to Serial B, but they do when connected to Serial A.

Symptom: LisaTerminal does not seem to be communicating with the modem over Serial B anymore. I can type, hit return and the cursor just jumps back to the beginning of the line.

Questions: Am I likely to have damaged the motherboard, the I/O board, or both? Might there be a way to repair the damage in any case?

[jamesdenton]

Sorry to hear about that! I've been meaning to make cases for these things but I'm not much of a 3D modeler and don't have a 3D printer, so it'll have to wait.

Do you have a loopback/null modem cable you can connect between Serial A and Serial B? Or maybe connect SerialA/B to some other device, like a Mac, to see what the behavior looks like between those ports and machines? Or try MacTerminal within MacWorks to see if it behaves differently. I suppose you could've accidentally grounded one or more of the pins, but I'm not sure what the impact would've been inside the machine. There are a handful of 5V pins along with the data pins along the bottom.

When the modem is powered on while connected to Serial B, do you see the startup messages? Or nothing at all?

[ried]

When the modem is powered on while connected to Serial B, do you see the startup messages? Or nothing at all?

If the modem is attached to Serial B, nothing at all... no commands register. If I type "AT" and hit return, the cursor just jumps back to the beginning of that line and sits underneath the "A". If the modem is attached to Serial A, the modem responds with "OK".

I do not have a loopback cable. Would it potentially help me to diagnose the damage?

Update: I swapped the I/O board with another (confirmed good) unit. No change. The damage to Serial B seems to be on the motherboard.

[sigma7]

...
Am I likely to have damaged the motherboard, the I/O board, or both?
Might there be a way to repair the damage in any case?

RS-232 ports are generally very rugged, designed for relatively large voltage ranges (up to +/- 25V vs +5V or +3V for modern interfaces), tolerating an intermittent short to ground, long cables that pick up induced voltages, and so on. But they aren't invincible as you've discovered.

Since it's a Lisa, most electronic failures are repairable!

On the motherboard, there is an interference filter for each serial port which can be damaged. The one for Serial B is RP2.

If you measure the resistance across the part between most pairs of pins that are physically opposed, you should see about 50 ohms. ie. pin 2 to pin 19, pin 3 to pin 18, etc. ... except for the pairs at the ends (pins 1,20 and 10,11), which are directly connected. If you have an open circuit between any pairs then that may be the point of failure.

Another possibility is that a connector contact or circuit board trace was damaged by the high current. I think only the ground pin (pin 7) would be at risk from a 5V source. You should measure close to 0 ohms between pin 7 and the metal chassis.

[ried]

That is very helpful! I shall do some measuring.

[sigma7]

I've been meaning to make cases for these things but...

If you're not in a hurry, I've done some simple designs in FreeCAD for 3D printing, so might be able to help.

In the meantime, fastening some thick paper or cardboard over vulnerable parts of the exposed circuitry helps me prevent accidents with prototypes.

[ried]

On the motherboard, there is an interference filter for each serial port which can be damaged. The one for Serial B is RP2.

If you measure the resistance across the part between most pairs of pins that are physically opposed, you should see about 50 ohms. ie. pin 2 to pin 19, pin 3 to pin 18, etc. ... except for the pairs at the ends (pins 1,20 and 10,11), which are directly connected. If you have an open circuit between any pairs then that may be the point of failure.

Another possibility is that a connector contact or circuit board trace was damaged by the high current. I think only the ground pin (pin 7) would be at risk from a 5V source. You should measure close to 0 ohms between pin 7 and the metal chassis.

I measured ~51-53 ohms of resistance across all of the interference filter pin pairs, with no resistance at 1,20 and 10,11. So that seems to be okay.

On the DB-25 ports themselves, I also measured 0.1 ohms of resistance between pin 7 and the metal chassis, so that also checks out (on both Serial A and B). Out of curiosity I measured resistance between the other pins and the metal chassis (I hope that's okay?), hoping to discern any difference at all between measurements of Serial A and Serial B. On pin 10 4 of Serial B, there are 12.68 k ohms of resistance between it and the metal chassis, but that is not present on pin 10 of Serial A. Might that be a clue? Or is that the way Serial B was designed?

Hopefully all my poking and prodding isn't making anything worse.

[sigma7]

On pin 10 of Serial B, there are 12.68 k ohms of resistance between it and the metal chassis, but that is not present on pin 10 of Serial A. Might that be a clue? Or is that the way Serial B was designed?

Hopefully all my poking and prodding isn't making anything worse.

If you've got the power unplugged and staying away from the video circuitry and power supply, poking around with an ohmmeter is unlikely to damage anything in a Lisa.

Pin 10 of the RS-232 ports isn't supposed to be connected to anything, so double-check that you've got the pin numbering scheme correct...

Looking at the rear of the Lisa, you see the receptacle side of the female DB-25 connector, so the pins are like this:

Code: [Select]

\  13  12  11  10   9   8   7   6   5   4   3   2   1  /
 \                                                    /
  \  25  24  23  22  21  20  19  18  17  16  15  14  /
 
 
A minimal RS-232 connection requires 3 wires/pins

Pin 2 - Transmit data from Lisa
Pin 3 - Receive data to Lisa
Pin 7 - Ground

Sometimes more wires are used for handshaking and modem control.

Assuming you have a 2/10 I/O Board plugged into a 2/10 Motherboard, you should be able to measure low resistance from the I/O Board to Serial B as follows:

Code: [Select]

Serial B
DB 25 - to - I/O Board
Pin 2 - to - U10E-15
Pin 3 - to - U11E-1         
If you turn the power on, the voltage at Serial B pin 2 should go to -5V or so when nothing is being transmitted.

[ried]

If you've got the power unplugged and staying away from the video circuitry and power supply, poking around with an ohmmeter is unlikely to damage anything in a Lisa.

Excellent. Thank you for confirming.

Pin 10 of the RS-232 ports isn't supposed to be connected to anything, so double-check that you've got the pin numbering scheme correct...

Looking at the rear of the Lisa, you see the receptacle side of the female DB-25 connector, so the pins are like this:

Code: [Select]

\  13  12  11  10   9   8   7   6   5   4   3   2   1  /
 \                                                    /
  \  25  24  23  22  21  20  19  18  17  16  15  14  /
 
 

Thank you for that numbered pin layout. Here is my corrected statement: "On pin 4 of Serial B, there are 12.68 k ohms of resistance between it and the metal chassis, but that is not present on pin 10 of Serial A. Might that be a clue? Or is that the way Serial B was designed?"

[sigma7]

On pin 4 of Serial B, there are 12.68 k ohms of resistance between it and the metal chassis, but that is not present on pin 10 of Serial A. Might that be a clue? Or is that the way Serial B was designed?"

Pin 4 of Serial B has a 10K resistor to +5V on the Motherboard, so measuring about 12k to ground is reasonable. Pin 4 of Serial A is driven by a line driver on the I/O Board under control of the SCC, so it will appear different.

So, no clue there... keep looking!

edit: The 10k resistor to +5 on Pin 4 of Serial B is one of the differences between the Lisa 2/10 Motherboard and the 1/5 Motherboards. The Apple 1/5 Motherboards do not have this resistor, so depending on the wiring of the serial cable, a serial device expecting RTS to be high may not work on Serial B of a 1/5 but still work on Serial A. The Sapient variety of the 1/5 Motherboard does included this resistor.