• Please review our updated Terms and Rules here

North Star Horizon power

What we need to do is to find a series current limiting resistor that will allow a reasonable current to flow, but not burn out!

For example a 100 Ohm resistor.

This will permit a maximum current flow of 15 V / 100 Ohms = 0.15 A (150 mA).

At 0.15 A, the resistor will dissipate 0.15 * 0.15 * 100 Watts = 2.25 W. So a 0.25 W resistor would burn out pretty quickly if subject to a short circuit on the motherboard!

See what you have got and calculate the maximum current flow and maximum power dissipation to see what will work. I have some rather (physically) large resistors bonded to a very big heatsink for just this eventuality!

Dave
Thanks. I’ll look into it tonight

I really appreciate the patience all you guys are showing! My knowledge of electricity is terrible. I can follow along on schematics, which is a bit like a programming flowchart, but I failed Ohm’s Law in high school
 
Are the disk drives connected? if so disconnect them.
Also, the 78L12 and 79L12 are voltage regulators. They do not have emitter, base or collector. Look for the spec, on line, for these to show what pins are input, ground and output. Also, this size transistor can have at least 2 different transistor pinouts that I know of.
First disconnect the negative supply. Leave only the +16V supply connected with the light in line.
Report the voltages on the two small regulators? Please use the correct terms for which pins are which.
Dwight
 
What we need to do is to find a series current limiting resistor that will allow a reasonable current to flow, but not burn out!

For example a 100 Ohm resistor.

This will permit a maximum current flow of 15 V / 100 Ohms = 0.15 A (150 mA).

At 0.15 A, the resistor will dissipate 0.15 * 0.15 * 100 Watts = 2.25 W. So a 0.25 W resistor would burn out pretty quickly if subject to a short circuit on the motherboard!

See what you have got and calculate the maximum current flow and maximum power dissipation to see what will work. I have some rather (physically) large resistors bonded to a very big heatsink for just this eventuality!

Dave
I had a 330 ohm 1/4W resistor on hand. It's the smallest I had other than a 100 and 220. With just the +16W power connected I measured +16v on the power supply side of the resistor and 216mV on the other side, and 8mV at the power connection lug on the motherboard
 
Are the disk drives connected? if so disconnect them.
Also, the 78L12 and 79L12 are voltage regulators. They do not have emitter, base or collector. Look for the spec, on line, for these to show what pins are input, ground and output. Also, this size transistor can have at least 2 different transistor pinouts that I know of.
First disconnect the negative supply. Leave only the +16V supply connected with the light in line.
Report the voltages on the two small regulators? Please use the correct terms for which pins are which.
Dwight
Nothing is connected to the motherboard. I don't believe I made any reference to the 78L12 and 79L12s as having emitters, bases, or collectors. I made a reference above to the input and output pins. If I did otherwise, I'm sorry
 
>>> With just the +16W power connected I measured +16v on the power supply side of the resistor and 216mV on the other side, and 8mV at the power connection lug on the motherboard.

I think that confirms to me that you still have some sort of short circuit on the motherboard still.

This may be a short circuit that is present with the power OFF, in which case you should be able to measure it with your multimeter - or it only appears when the power is present.

Let me have a think...

Dave
 
"After repairing the traces ..."

I had an off the wall thought back to a similar situation. A piece of metallic debris in a piece of #5-Xbar equipment eventually vibrated to a position where it bridged between a -48v supply and the rack structure. It essentially vaporized before the 5A fuse blew. We were left with a piece of equipment that was acting in ways that made no sense. It turned out that the vaporized metal had left a conductive coating on what had been an insulator. With a bright enough light you could see a brassy looking sheen on the black bakelite.

Might be time to take a good look at the area around where those traces blew. The metal went someplace. The chances are so slim I hesitated to suggest it. But it is a possibility.

If you get the expected behavior from the two power supplies when they're "stacked" and powering two dummy loads as daver2 suggests, but things go off the rails if they're connected to the system in a similar fashion, then that kinda points to something other than the power supplies.

"There ain't any ghosts in that engine." -- Jake Holman in "The Sand Pebbles"
 
Last edited:
Yes there is a short circuit. First I'd like to make sure the right regulators are on the right power rail. This was the reason I wanted the voltages seen as the regulators. There isn't that much on the supply line sides of the regulators. These regulators have output short protections that can take an output short for some time.
I'd really like to see the voltages on the two regulators. under the conditions I'd stated, with only the +16V and the lamps for protection. While the resistors are good, incandescent lamps are good constant current sources ( feature of a filament in a gas ).
I suspect the power leads may have been swapped or someone installed the 79L12 and 78L12 in the wrong locations. Swapping the input voltages is the more likely case of original failure ( consistent with blowing a trace ).
Measuring is the first step in trying to determine the cause. Please don't get offended by my roughness. I'm just trying to get to the resolution ( we are now at over 100 posts ). I'm 74 years old and grumpy but still know how to trouble shot.
Dwight
 
I took another look at the schematics for the NSH power supply.

Let's have a look at the +16V rail first shall we?

This feeds a number of locations (any one of which could be causing a problem):

1) C3 (6.8 uF tantalum capacitor). Did you replace this or not? If you did replace it, is it the correct way around?

2) 78L12 Voltage regulator. This is only a 'tiny' thing as you know when you replaced it.

3) The two (2) large 7812 Voltage regulators feeding the two disk drives. I assume the disk drives are still unplugged?

4) There are two (2) capacitors after the voltage regulators for the disk drives (C5 and C6). I suspect these should be fine - but (there again) worth checking.

5) R31 - a 330 Ohm 2W resistor designed to provide a sensible load for the power supply. This resistor - in itself - should be enough for a switch mode power supply to operate correctly.

6) The 1488 and 1489 RS232 buffers. I would suggest removing these.

7) The two serial ports themselves. They have +12V and -12V on them (via resistors) for the current loop interface. I assume nothing is plugged into these connectors?

Measure the resistance between the +16V lug and the GND lug of the motherboard using your multimeter. Try the leads first one way round and then the other way round and note the lowest resistance. Because of the presence of R31 (330 Ohms) this should be a MAXIMUM of 330 Ohms. Post the results.

If that does not yield any results I am afraid you may have to remove the three (3) voltage regulators again (the 78L12 and the two 7812s)...

Dwight: I see we cross-posted...

Dave
 
Yes there is a short circuit. First I'd like to make sure the right regulators are on the right power rail. This was the reason I wanted the voltages seen as the regulators. There isn't that much on the supply line sides of the regulators. These regulators have output short protections that can take an output short for some time.
I'd really like to see the voltages on the two regulators. under the conditions I'd stated, with only the +16V and the lamps for protection. While the resistors are good, incandescent lamps are good constant current sources ( feature of a filament in a gas ).
I suspect the power leads may have been swapped or someone installed the 79L12 and 78L12 in the wrong locations. Swapping the input voltages is the more likely case of original failure ( consistent with blowing a trace ).
Measuring is the first step in trying to determine the cause. Please don't get offended by my roughness. I'm just trying to get to the resolution ( we are now at over 100 posts ). I'm 74 years old and grumpy but still know how to trouble shot.
Dwight
With no drives attached and no ICs on the board, just the +16V power supply with the 14V 3.8A bulb: The measurements are +16V on the power supply side of the bulb, 550mV on the other bulb lead, 58mV at the +16V power supply lug, 57mV on the input of the 78L12, and 2.5mV on the output.

The 78L12 and 79L12 are in the correct positions and orientation on the board (I even got in there and photographed them to double check the numbers). C1, C2, and C3 have all been replaced and are in the correct orientation. C4 is original and is part of the +12V power derived from the 78L12. It was working fine under the old linear power supply

I built this machine in 1978 and it worked beautifully for years. It sat in storage between about 1986 and 2020, when I switched it on. It worked nicely for a few weeks and then quit functioning. The old linear power supply was supplying all three voltages but something was causing the -16V line to blow out along the trace to pin 52 on the connectors. Although the power supply was powering the lines I got suspicious about its reliability. That's when I decided to install the three MeanWell supplies instead.

As this thread got going I saw that the 78L12 had blown up, and I also found a burnt out resistor (R25) on the -16V line to the second serial connector. I replaced it, but now I think I should give the area around that row of resistors a good look under magnification. Maybe something's shorting out there.

I'm fine if you're grumpy with me lol. I'm 72 and grumpy because I can't figure out what's wrong with this thing. I've never run into such a trouble shooting nightmare before, and I suspect the solution will be so obvious in the end I'll end up banging my head on the table
 
I am afraid the PSU is not to blame for the -16V track burning out. I had a similar thing many years ago on a Cromemco BlitzBus. A short circuit (from a regulator input to GND on one of the S-100 cards) took out the 'fuse' (aka the PCB track) before it could take out the real fuse!

I can see now how a potential problem 'could' have developed on the +16V rail with your 'home built' heatsink on one of the disk 7812 regulators. It was oversized for the PCB pad. The TO3 case is connected to GND and (with the heatsink being physically oversized) it looks to me as though it was sitting on top of the solder resist of the +16V track. I don't remember you saying anything about a historical problem with the +16V rail - but that failure mode was potentially waiting to happen...

Yes, I think you have to go back to basics with this one. At least it is a two-layer PCB. It is isn't it!

Dave
 
I'd start splitting things in the middle. The 78L12 is most likely thing to be easy to remove. If possible, just remove the output lead first if that doesn't get rid of the short, remove it entirely.
If it does remove the short, next in line are the 1488 and 1489. ( easy to find parts, Anchor Electronics list them for $0.39, shipping not included ). If you don't feel confident at removing them, just clip the leads with sharp diagonals right at the part. Then remove the leads one at a time with tweezers ( I forget if these were socketed ).

I'm beginning to think you had a ground fault.
The resistor on at the RS232 connector is an indication that you had a ground fault failure between your serial monitor and the N*.
It is either something in the N* or the monitor you connected to.
If you don't know what a ground fault interrupter is, find out and get one installed or make up a socket box with one in it at the end of an extension cord.
If your house doesn't have working 3 pin sockets ( with ground ) run a ground wire from your electrical box to you socket box. When I bought my current house, the idiot before me installed 3 prong sockets with none of the grounds connected. I spent two days under the house running ground wires( dirt and spiders, my favorite ).
The 2 resistors are a 560 ohm and a 220 ohm. Either shorted to ground should have only made them really hot and not burned them out.
Dwight
 
Progress! I went back and cleaned the area on the solder side of the board around R25, which I'd replaced about the time this thread started. It was pretty messy but cleaned up nicely with some alcohol and a brush. I also went back and cleaned up around the solder side of the capacitors and regulators where the power leads come in (i.e. C1-C4, 78L12 and 79L12). I did the same cleaning along the -16V trace for pin 52 on the card connectors, making sure all traces of the pin 52 lead were gone and the replacement wires were clear of anything.

I noticed that I was getting very little resistance between pins 51 and 52, even though all the traces seemed to be clear. I flipped the board and started examining the edge connectors and discovered that the pin 51 and pin 52 contacts seemed to be touching on connector #1. I pushed in there and the pin 51 contact seemed to snap back. Three of my 12 card connectors are that type, and I've never liked the way they're built

In the process I checked the resistors at the front of the board. R31, R33, and R34 were all okay, but R32 (330ohm 2w) was only reading about .4ohms, so it's no good. This is the one that terminates the -16V power line.

I risked it and plugged in all the power leads. The +8V and +16V are now working. The pin 51 and 52 lines had apparently been shorted on the card connector contacts. The -16V line still isn't working, but I suspect it will once I replace R32. Oddly enough, it doesn't look damaged

edge connector.jpeg
 
Does anyone want a video of me slapping my forehead? :)
I'll pass on the head slap video. "Stuff" happens and so forth. But I'd like to see one of the machine all buttoned up and running sweet. Kinda like "napalm in the morning." Smells like victory.

;-)

P.S. I'll be turning 73 in a few days. I know a thing or two about being old and grumpy LOL.
 
I'll pass on the head slap video. "Stuff" happens and so forth. But I'd like to see one of the machine all buttoned up and running sweet. Kinda like "napalm in the morning." Smells like victory.

;-)

P.S. I'll be turning 73 in a few days. I know a thing or two about being old and grumpy LOL.
Well, after the mb is sorted out, I'll still need to check out the cards. I've got a Compupro CPU-Z card, a Compupro RAM 17 (64k), and a Morrow Designs DJ-DMA fdc. Hopefully they've survived
 
LOL, as usual, hindsight is 20/20 - so frustrating. We kept testing resistance +/-16v to ground and there was no short. We forgot to test 16v to -16v. And this is one I’ve actually seen multiple times because these two pins are across from each other. I’ll head slap myself right along with you.

Mike
 
LOL, as usual, hindsight is 20/20 - so frustrating. We kept testing resistance +/-16v to ground and there was no short. We forgot to test 16v to -16v. And this is one I’ve actually seen multiple times because these two pins are across from each other. I’ll head slap myself right along with you.

Mike
Haha! I just realized I've been sitting in my home office all day fooling with this thing and forgot to eat, which is unusual for me, given my waistline. I smell supper cooking and need a beer and a shot of Jamison's. The rest of the adventure will have to wait until my Digikey order gets here
 
I want to thank you guys for your help and patience putting up with an electrically challenged noob in finding the solution to this (I hope). I’ll keep you posted as I go along

Again, thanks very much
 
No problem. If it was simple, everyone would be doing it and there would be no reason for us to be here...

The short circuit between the +/-16V rails would certainly have driven the switched mode power supplies daft! It was a good job they have overload protection! You could have done some real damage there...

R32 reading 0.4 Ohms instead of 330 Ohms is (more than likely) a red herring I am afraid. All it probably indicates is a short circuit elsewhere between -16V and GND. Keep looking - you have done an excellent job so far and you are close to the finish line to getting this beast running again...

I suspect - with the short circuit between the +/- 16V rails - you may have inadvertently reverse-biased some of those tantalum bead capacitors. If (say) the -15V supply shutdown, this could put +15V across some of the capacitors (relative to GND) that are on the -16V line. My recommendation would be to recheck the pesky capacitors again. I am also not sure what this would do to the voltage regulators either.

Dave
 
No problem. If it was simple, everyone would be doing it and there would be no reason for us to be here...

The short circuit between the +/-16V rails would certainly have driven the switched mode power supplies daft! It was a good job they have overload protection! You could have done some real damage there...

R32 reading 0.4 Ohms instead of 330 Ohms is (more than likely) a red herring I am afraid. All it probably indicates is a short circuit elsewhere between -16V and GND. Keep looking - you have done an excellent job so far and you are close to the finish line to getting this beast running again...

I suspect - with the short circuit between the +/- 16V rails - you may have inadvertently reverse-biased some of those tantalum bead capacitors. If (say) the -15V supply shutdown, this could put +15V across some of the capacitors (relative to GND) that are on the -16V line. My recommendation would be to recheck the pesky capacitors again. I am also not sure what this would do to the voltage regulators either.

Dave
I'm with Dave. 330 ohm resistors never ever go to 0.4 ohms. You have a short someplace, maybe under the resistor but it is not the resistor's fault.
It is physically not possible for a 330 ohm carbon comp resistor to go down in value more than about 30%. They can go open or just some higher value.
They can be mislabeled ( though unlikely ).
Dwight
 
Back
Top