PDA

View Full Version : debugging the protect circuit


phe
07-07-2007, 03:12 AM
I'm trying to figure out why the flip flop in IC L on the memory board is not changing state. I've checked the voltage on lines 20 and 70 of the bus when I flip the switch to protect and unprotect. Both lines go high when I flip the switch in either direction (protect or unprotect). With the system powered off, pins 20 and 70 are open (infinate resistance). So there is no short between them.

I've checked resistance across the three pins from the protect/unprotect switch. With the system off, the top and bottom pins of this switch show zero resistance, so something is shorted across those two pins. I would not expect this to be correct.

Anyone have any ideas of further places I could check to find out where the problem might be. I can't image it is a bad switch - that would be nearly impossible wouldn't it? But shouldn't these pins show infinite resistance between them? I'd like to find other signs that point to a bad switch before I desolder and remove it.

Paul

sje
07-07-2007, 08:22 AM
If the two throws of an SPDT switch have zero resistance between them, then there's either a broken switch, bad wiring, or a bad circuit design. I'd say that the switch is by far the most likely culprit.

Get an extra switch or two before starting the desoldering. It's always good to have a spare.

marty
07-07-2007, 05:00 PM
After removing the switch, check the circuit for a short, and if you have not destroyed the switch , check it. If the switch is the problem, them most likely you have dicsovered the problem. If it is still there , then take out the IC, If you don't have a socket on the IC, I would suggest putting one in. If the short is still there, them most likely either you have a short or a pcb short. Sometimes a PCB short can be taken care of with a voltage on one side and ground on the other. I would start with 5V before going to 12V ,, BUT This is a VERY Last resort, I would try every other method I could think of before doing this. I have had to short a High pin to ground to get an internal short fixed, but this was on some Very old Imsai front Panel connections, but the front panel now Works!! :D
Thanks Marty

phe
07-09-2007, 01:02 AM
Thanks for the helpful reply, Marty. I pulled the switch out. The switch is OK. The problem must be in the circuit. I've been visually inspecting for solder bridges and have not seen any. I pulled one of the ICs (I think it was IC W) and there was still a short. I'll be away from home for a couple of days so I cannot try anything new until I return on Tuesday. Anyone have any idea if there is a different IC that I should pull?

Paul

marty
07-09-2007, 03:01 AM
Hi Paul;
I wrote a long letter to you and somehow it got lost, so if you see 2 posts from me this is why.
First, try this, the two wires that go to the switch find the pins on the IC and Carefully bend them up, and attach another switch to them and connect the middle connector ground or to +5V depending on what it needs, I don't have a schematic in front of me. And see if the protect unprotect circuit works.
Also look around the traces that go from the switch to the IC and see if there is another trace shorted to it, unrelated to the protect circuit.
If there is no change them pull the wires to pins 20 and 70 and check them out using either a scope or a logic probe or a volt meter. And see if you can get them to show the proper values. If they work then either you have a problem in the Mother board or the memory board.
Thanks Marty

marty
07-10-2007, 03:01 PM
Hi Paul;
I have some of the circuit in front of me. So connect your other switch the middle pin to ground. there are two pull-up resisters R13, R14, check that they are BOTH 1K Resistors. they then go to a pair of inveters. the protect goes to pin 3 of the '04, and the unprotect goes to pin 1 of the '04, these should invert the signal. So with them being pulled up they should the LOW on their out-put side. on the protect side pin 4, on the unprotect side pin 2. If they are HIGH then either the inputs are not being pulled up, or shorted to a LOW signal or to ground or the IC is not an '04 or not working correctly, (BAD!!) The 8T97 (IC X) should pass the signal through and not alter it at all, HIgh in should be a High out, and a LOW in should be a Low out. Pin 10 and 9 on the protect side and pins 4 and 5 on the unprotect side. Check ALL Traces surrounding any IC or Trace Coming or Going to the 04 or the 8T97 or the solder pins 20 and 70.:o
Thanks Marty

marty
07-10-2007, 03:22 PM
Hi Paul;
After you dis-connect the wire for pins 20 and pins 70, check to see if they work right, If they do then the problem is not in the front end of the circuit. Then take out the memory board, put he wire back in and see if you can toggle them High and Low, if you can't the it is the Motherboard. if it works then most likely it is the Memory board:( and that is for another session. Please let me know what you find !!!:rolleyes:
Thanks Marty

Grant Stockly
07-10-2007, 05:19 PM
If you didn't clip the long wires off that terminate into the motherboard then maybe they are touching underneath.

Look at the bottom of schematic 880-106 for S2. The two switch has two Pull-up resistors, R13 and R14. The switch also goes through an inverter, 04, and then through a buffer, marked as "X" (does not invert).

You should see pin 3/4 on IC'W and pin 10/9 on IC'X when trying to protect

Pin 1/2 of IC'W and 4/5 on IC'X when trying to unprotect
.

What do you see? Take out the memory card when doing this test.

phe
07-11-2007, 12:55 AM
Marty and Grant, here is what I see (with the CPU and memory cards removed):

With the system turned on, IC W pin 1 is HIGH; pin 3 is HIGH; pin 2 is LOW; pin 4 is LOW. These are observed with the switch in the middle position. Pushing the switch to protect gives no change to the above 4 pins. Pushing the switch to unprotect also gives no change to any of the four pins.

The way I read the diagram, these two pins should always be high and one of them should go low when the switch is pushed. Marty, this is opposite of what you said. One of us must be reading it wrong.:eek:

I removed IC W and ran some resistance checks (system off :) ).

Resistance between socket 1 and socket 3 = 0. Resistance across R13 = 500 ohms. Resistance across R14 = 500 ohms. I would expect this if the two resistors are in parallel. Does this confirm my suspicion that somehow the trace that runs from PROT to pin 3 is shorted with the trace that runs from UNPROT to pin 4? If so, I have absolutely no idea where that could be.

:confused: :mad:

marty
07-11-2007, 01:33 AM
Hi Paul;
Yes, it does look that way, as though the two are shorted together. I Think what I was refering to --- was if you bent pins 1 and 3 sideways. so that they would NOT connect to anything, and wire a seperate switch to them, with the middle connected to ground. then with the pins connected to this new switch, they Should respond each going high or low according to the position of the switch. But You are right only one side should high or low depending on your switch position. :D I might have been thinking of pulling each IC high or Low using either a wire connected to ground or to +5V or a Logic Pulser. LOOK at the TRACES with a Very Bright LIght Shining through the Board or use the SUNLight to look through the board. And follow the traces from the switch to IC 'W' , look for a short to another trace.
Take out the Resistors and check the lines again and look underneath them as well. :rolleyes:
OTherwise take out the resistors and the IC, and IF the Short is still there and You have Found NO SHORTS, THEN tie one switch position to +5V and the OPPOSITE IC Would be Connection to Ground , USE A wire and MOMENTARYILY touch the ground wire to where you need to short to, DO NOT HOLD IT THERE FOR VERY LONG !!!! IT might Spark or you might hear a POP !!! Then TEST with a meter for conectivity !!! :eek: :eek:
Thanks Marty

Grant Stockly
07-11-2007, 04:01 AM
Thats why I didn't say remove the processor card. :)

If you don't have the computer "stopped" it won't do anything. (it needs a functioning processor card to get to a stopped state) Notice that the Normally Open sides of the switch are pulled high. It gets the "low" pull down from the Run/Stop FF IC'H. Look in the top right corner of 880-106.

Same thing for single step, examine, and deposit switches.

Grant

Grant Stockly
07-11-2007, 04:04 AM
If Pin 71 on the bus needs to be low for the protect switch to work. Stick your processor card back in there.

marty
07-11-2007, 05:06 AM
Hi;
Grant, I would Agree with you that the low for the Switches comes from the Run FF, But, He still shows a SHORT between the two pins. I just checked my board and I don't show ANY short, Both resistors show 1K each and between the pins on the Switch it reads 2K NOT 500 ohms. So when he brings them Low with the processor board in they both go low and output both go high, not one or the other. Without looking at the board I am only guessing what the problem is. I know that I don't have any other ideas !! :confused: :confused:
Thanks Marty

phe
07-11-2007, 05:47 AM
Problem Solved!!!

There was a manufacturing defect :eek: in the board and two traces were shorted. A bit of work with an exacto knife and I was able to remove the copper bridge between the traces. :cool: I've attached a jpeg of my handiwork.

Now I have a finished Altair. :D Thanks for sticking with me guys to figure it out.

(Gee --- all I have to show for this, from a demo perspective, is a switch that turns an LED on and off ;) )

Paul

Grant Stockly
07-11-2007, 07:59 AM
Yep. Marty was right. I didn't understand the problem. :(

Oh well, at least his system is up and running! :)

marty
07-11-2007, 01:24 PM
Congratulations Paul !!!!!! :) :D :rolleyes: :p
I am Glad that you found the problem and now have a working Board !!! :D :D Now you can get back to writing programs for you machine or watching the Led's blink !!! :cool: :cool: :cool: :cool: