Fixing the burnt ICs, getting 8k working, and more.
Its been a while since the last post, so theres a lot to update on.
A couple weeks ago we let the magic smoke out of a couple of ICs. We found a lot of cards with bad outputs. We replaced everything we found. After we replaced all of these it started working again. I will post soon with a write up of everything we replaced... there were lots of 7474s and 7400s.
On Friday 7/07/17 we decided to try to get the second 4k stack working. We started up Warrens Tune 3 program and selected data field one with a 6211 instruction.
We took initial measurements with the multimeters and oscilloscope. The two voltages that we needed to keep an eye on were across pins M2-V2 and S2-R2 on slot F2. M2-V2 is the core voltage and S2-R2 is the thermistor voltage.
On initial startup M2-V2 was 20.31V and S2-R2 was 563mV. The initial pulse length for field 1 was 224ns.
We adjusted the pot until we got errors in an attempt to find the center point for the pulse length.
At 267ns we had 50-100 errors on average and at 150ns we had about the same. We centered it at 208ns. There were still a few errors here so we needed to adjust the voltage next.
M2-V2 was 20.28V and S2-R2 was 571mV. M2-V2 should be around 22V.
We adjusted the pot for the core voltage (the rightmost pot on E/F02). One thing to note is that this voltage is the same for both cores. So we have to adjust things with Field 0 too.
Adjusted to 20.82V, zero errors.
Adjusted to 21.3V, zero errors.
Adjusted to 21.82V, zero errors.
Adjusted to 22.3V, zero errors.
Adjusted to 22.94V, zero errors.
Adjusted to 23.4V, zero errors.
Adjusted to 23.8V, zero errors.
Adjusted to 24.3V, zero errors.
We left it at 24.3V for now.
We went back to adjusting the pulse length. The program failed at 290ns and had to be restarted. There were errors at 142ns.
With these new numbers we picked new halfway points. 215ns and 22.3V.
Stopped the program to test on Field 0. (start at 0201 instead of 0200. Copied program to Field 1 to test Field 0).
250ns was the initial pulse length and there were errors at 288ns and 201ns. Centered at 244ns.
We checked the voltages next. No errors at 24.3V or 20.3V. Centered at 22.3V.
The final numbers for our system are 244ns on Field 0, 215ns on Field 1, and 22.3V core voltage.
We used the IR gun and checked the temp of the core and the temp of the room. The core was only about 80F and the room was about 79F.
We ran the checkerboard program D1EC. It ran good with no errors.
We have 8k working!!!
On Sunday 7/09/17 we discovered we couldn't BIN load. The PDP stopped sending or receiving characters. We tracked it down to the skip logic.
Warren put a program in to use for testing.
0200 / 7604 LAS Load AC with Right Switches
0201 / 6046 TLS Load Transmit Sequence
0202 / 6041 TSF Skip on Transmit Flag
0203 / 5202 JMP Jump 202
0204 / 5200 JMP Jump 200
After scoping a few signals, IOC IOP 1 H, IOC IOP 2 H, and preset, we realized I was on the wrong pins. The signals were actually fine, but it still wasn't doing the skip instructions.
Pulling some cards in and out seemed to fix it.
The problem came back even after wiggling the cards.
Monday 7/10/17.
We found a solder bridge on the M216 in M19. It was shorting E1 to high. This was causing our intermittent problem. We fixed the solder bridge. In the picture below it is the two pins on the far right of the top row in the flash light beam.
We checked A35 P01 just to make sure it was working. This is the Immediate Mark signal. It was good with pulses around 1ms.
Tuesday 7/11/17
On Tuesday we played with the OS/8 disk server using some RK05 images Mike had sent us. We tried using OS/8 disk server a couple days earlier and didn't get far. It crashed when we tried to build the system after we removed drivers we didn't need.
Mike sent 4 images. 3 didn't work and one let us get pretty far and actually run DIR on the other "disks" and see the contents of them. It still crashed when we tried to build.
We moved on to some tape dumping.
We noticed that the top drive doesn't always move right and the bottom drive slaps the tape pretty badly.
We swapped the R107s in B11. Those didn't affect anything so we swapped them back.
We swapped the G850s which are the motor/ drive controllers in an effort to see if the problems moved between drives. The top drive worked just fine, no tape slapping and now it starts and stops and moves to the right and left with no problems. The bottom drive now has some motor problems, but its not the same as it was on the top drive. The problem seems to be with the G850 cards. This will be a project for another day.
A couple weeks ago we let the magic smoke out of a couple of ICs. We found a lot of cards with bad outputs. We replaced everything we found. After we replaced all of these it started working again. I will post soon with a write up of everything we replaced... there were lots of 7474s and 7400s.
On Friday 7/07/17 we decided to try to get the second 4k stack working. We started up Warrens Tune 3 program and selected data field one with a 6211 instruction.
We took initial measurements with the multimeters and oscilloscope. The two voltages that we needed to keep an eye on were across pins M2-V2 and S2-R2 on slot F2. M2-V2 is the core voltage and S2-R2 is the thermistor voltage.
On initial startup M2-V2 was 20.31V and S2-R2 was 563mV. The initial pulse length for field 1 was 224ns.
At 267ns we had 50-100 errors on average and at 150ns we had about the same. We centered it at 208ns. There were still a few errors here so we needed to adjust the voltage next.
M2-V2 was 20.28V and S2-R2 was 571mV. M2-V2 should be around 22V.
We adjusted the pot for the core voltage (the rightmost pot on E/F02). One thing to note is that this voltage is the same for both cores. So we have to adjust things with Field 0 too.
Adjusted to 20.82V, zero errors.
Adjusted to 21.3V, zero errors.
Adjusted to 21.82V, zero errors.
Adjusted to 22.3V, zero errors.
Adjusted to 22.94V, zero errors.
Adjusted to 23.4V, zero errors.
Adjusted to 23.8V, zero errors.
Adjusted to 24.3V, zero errors.
We left it at 24.3V for now.
We went back to adjusting the pulse length. The program failed at 290ns and had to be restarted. There were errors at 142ns.
With these new numbers we picked new halfway points. 215ns and 22.3V.
Stopped the program to test on Field 0. (start at 0201 instead of 0200. Copied program to Field 1 to test Field 0).
250ns was the initial pulse length and there were errors at 288ns and 201ns. Centered at 244ns.
We checked the voltages next. No errors at 24.3V or 20.3V. Centered at 22.3V.
The final numbers for our system are 244ns on Field 0, 215ns on Field 1, and 22.3V core voltage.
We used the IR gun and checked the temp of the core and the temp of the room. The core was only about 80F and the room was about 79F.
We ran the checkerboard program D1EC. It ran good with no errors.
We have 8k working!!!
On Sunday 7/09/17 we discovered we couldn't BIN load. The PDP stopped sending or receiving characters. We tracked it down to the skip logic.
Warren put a program in to use for testing.
0200 / 7604 LAS Load AC with Right Switches
0201 / 6046 TLS Load Transmit Sequence
0202 / 6041 TSF Skip on Transmit Flag
0203 / 5202 JMP Jump 202
0204 / 5200 JMP Jump 200
After scoping a few signals, IOC IOP 1 H, IOC IOP 2 H, and preset, we realized I was on the wrong pins. The signals were actually fine, but it still wasn't doing the skip instructions.
Pulling some cards in and out seemed to fix it.
The problem came back even after wiggling the cards.
Monday 7/10/17.
We found a solder bridge on the M216 in M19. It was shorting E1 to high. This was causing our intermittent problem. We fixed the solder bridge. In the picture below it is the two pins on the far right of the top row in the flash light beam.
Tuesday 7/11/17
On Tuesday we played with the OS/8 disk server using some RK05 images Mike had sent us. We tried using OS/8 disk server a couple days earlier and didn't get far. It crashed when we tried to build the system after we removed drivers we didn't need.
Mike sent 4 images. 3 didn't work and one let us get pretty far and actually run DIR on the other "disks" and see the contents of them. It still crashed when we tried to build.
We moved on to some tape dumping.
We noticed that the top drive doesn't always move right and the bottom drive slaps the tape pretty badly.
We swapped the R107s in B11. Those didn't affect anything so we swapped them back.
We swapped the G850s which are the motor/ drive controllers in an effort to see if the problems moved between drives. The top drive worked just fine, no tape slapping and now it starts and stops and moves to the right and left with no problems. The bottom drive now has some motor problems, but its not the same as it was on the top drive. The problem seems to be with the G850 cards. This will be a project for another day.
Comments
Post a Comment