• Please review our updated Terms and Rules here

North Star Horizon power

BTW, whats wrong with the original supply ? Was it faulty or to increase reliability.

I note on one of mine that the 16v line is more like 22v, though given the simplicity, thats not really surprising but it gives the regulators a harder time with a lightly loaded chassis.
 
Certainly seems as if there is some leakage the PSU's dont like and as your not operating them really as designed its possibly not surprising. Really you need a +15/-15V psu that's specifically designed to operate in this mode.

Only thing I can think of now.

Put a 200ohm resistor across each PSU's then connect one +Ve to the other PSU's -Ve and check, then connect this point to the GND connector on the NS motherboard and check.
I'm curious as to how they're not being used as designed. The same set works fine on an Altair 8800c with a 9-slot motherboard, and I've seen them recommended often. What would be a ps that would operate correctly?
 
BTW, whats wrong with the original supply ? Was it faulty or to increase reliability.

I note on one of mine that the 16v line is more like 22v, though given the simplicity, thats not really surprising but it gives the regulators a harder time with a lightly loaded chassis.
I suspected that it might be faulty. When I first powered up the system in 2020, after it had sat dormant for 30 years, it worked for a week or two. After that the system wouldn't work and I began having trouble with the power. The -16v trace blew out on the backplane. I decided I would put in the same switching supply that was working on other systems like the 8800c. I also built an 8800c, and that power supply has worked well for that

You're right about the higher voltage. My original power supply was doing about 11v (8v), 21v (+16v), and -22v (-16v). I understand that was fairly normal for those supplies

I still have the linear supply. I suppose I could reinstall it if necessary (damned thing weighs a ton!)
 
I'm curious as to how they're not being used as designed. The same set works fine on an Altair 8800c with a 9-slot motherboard, and I've seen them recommended often. What would be a ps that would operate correctly?
by 'centre tapping' them. But its interesting that you have it working in a similar setup. I would test as I suggested to see whats bringing them down as looking at the datasheet, it 'should' work especially as GND is floating on the NS board.

Ideally you would use something along the lines of https://uk.rs-online.com/web/p/switching-power-supplies/6031167. This is designed to give a +15 and a -15V supply with a terminal for GND.

(ps not looked any further at that supply to see if its powerful enough, its just an example)
 
Sounds like the power supplies are cycling into self protect mode over and over. A 165 ohm load at 16v is less than 1.6 watts, so a resistive load doesn’t appear to be the problem. Could be a diode short to ground in one of the ICs. Put your meter into diode test mode and see what you measure on the motherboard (no power supplies connected) from the +16v lug to ground (put the red lead on +16v lug), and then from -16v to ground (put red lead on ground lug) .

Mike
Sounds like a good idea, if the shorts after a diode then it will need enough voltage to get through the diode to be seen.
 
Sounds like the power supplies are cycling into self protect mode over and over. A 165 ohm load at 16v is less than 1.6 watts, so a resistive load doesn’t appear to be the problem. Could be a diode short to ground in one of the ICs. Put your meter into diode test mode and see what you measure on the motherboard (no power supplies connected) from the +16v lug to ground (put the red lead on +16v lug), and then from -16v to ground (put red lead on ground lug) .

Mike
I'm getting a reading of 0L on them, also 0L on the 8v to ground
 
So it does not sound like a diode short on the power rails then. Forgive me if I'm duplicating suggestions already tried in this thread, but without a connection to power or to the motherboard, use the resistance setting on your meter to verify open circuit between the DC- terminal and chassis ground, and then between the DC+ terminal and chassis ground on your power supplies.

Mike
 
I decided to pull the ICs and test them on my TL866II-Plus. All the 74 series chips passed. The 4020 (Motorola MC14020B) failed. I can't test the 8251s or the 1488 and 1489s on the device
 
I decided to pull the ICs and test them on my TL866II-Plus. All the 74 series chips passed. The 4020 (Motorola MC14020B) failed. I can't test the 8251s or the 1488 and 1489s on the device
I also checked the power supplies with all the ICs removed. No change: 8v is still okay, +16v and -16v still fluctuate around 0
 
This is what was identified in posts #51 and #53. But, as new information has potentially come to light regarding potential power supply cycling (post #58), it would be good to retest.

The capacitor C47 shouldn’t actually cause a problem (unless it is faulty - short circuit). It is a capacitor, so only a low impedance to high frequencies.

With all of the S-100 cards removed, and the 1488 and 1489 RS232 buffers removed, it should be possible to just connect (for example) the +15V supply on its own to see if that is OK. Then remove the +15V supply and attach the -15V supply.

If the power supplies don’t work on their own, this tells us something.

If the power supplies do work on their own, but don’t work together, this tells us something else.

Thinking ahead, do you have a variable voltage bench power supply with a current limit?

Dave
 
Last edited:
The capacitor C47 shouldn’t actually cause a problem (unless it is faulty - short circuit). It is a capacitor, so only a low impedance to high frequencies.

Dave

I mentioned C47 because there is an option on the circuit board to not have this fitted and instead have a link installed which would mean GND connected to earth which would mean that one of the supplies +15V would have been connected to earth.
 
Gotcha!

Yes, wire link BAD - capacitor GOOD!

Although, given the schematic of the power supply, the output is floating anyhow and the negative rail only tied to earth via a filter capacitor.

Dave
 
Gotcha!

Yes, wire link BAD - capacitor GOOD!

Although, given the schematic of the power supply, the output is floating anyhow and the negative rail only tied to earth via a filter capacitor.

Dave

Yes, I noted that but cant really see another explanation other than some earth leakage as the PSU's work on their own or tied together when disconnected from the NS.
 
This is what was identified in posts #51 and #53. But, as new information has potentially come to light regarding potential power supply cycling (post #58), it would be good to retest.

The capacitor C47 shouldn’t actually cause a problem (unless it is faulty - short circuit). It is a capacitor, so only a low impedance to high frequencies.

With all of the S-100 cards removed, and the 1488 and 1489 RS232 buffers removed, it should be possible to just connect (for example) the +15V supply on its own to see if that is OK. Then remove the +15V supply and attach the -15V supply.

If the power supplies don’t work on their own, this tells us something.

If the power supplies do work on their own, but don’t work together, this tells us something else.

Thinking ahead, do you have a variable voltage bench power supply with a current limit?

Dave

The power supplies all work correctly when not connected to the motherboard. The 8v works okay when connected to the motherboard, regardless whether the other supplies are connected. Each 16v supply, when connected to the mother board by itself (no 8v connected) does not work. If I plug the ground in wire in but not the power wire I get +16v and -16v on the meter connected to the 16v or -16v wire. The power on the two 16v feeds is still fluctuating

This is very frustrating
 
>>> The power on the two 16v feeds is still fluctuating.

This is what I am trying to ascertain. Under what conditions does this ‘fluctuating’ start?

Is it fluctuating with the 15V power supplies not connected to the NSH with either connection, just the GND or what?

Dave
 
by 'centre tapping' them. But its interesting that you have it working in a similar setup. I would test as I suggested to see whats bringing them down as looking at the datasheet, it 'should' work especially as GND is floating on the NS board.

Ideally you would use something along the lines of https://uk.rs-online.com/web/p/switching-power-supplies/6031167. This is designed to give a +15 and a -15V supply with a terminal for GND.

(ps not looked any further at that supply to see if its powerful enough, its just an example)
I don't know, maybe at this stage trying a different power supply for +/-16v might be a good idea. The combination of MeanWell +/-15v supplies and the North Star mb just doesn't seem to be working for some reason. Mouser.com has a Traco TXLN 035-23M3. It's dual output with a common ground connection. Would something like that be worth a try?

Either that or go back to the original linear supply. Sigh
 
Last edited:
>>> The power on the two 16v feeds is still fluctuating.

This is what I am trying to ascertain. Under what conditions does this ‘fluctuating’ start?

Is it fluctuating with the 15V power supplies not connected to the NSH with either connection, just the GND or what?

Dave
It's only fluctuating when the two 16v supplies are connected to the motherboard, as illustrated in #23. It also fluctuates if any 16v supply is attached to the motherboard by itself
 
The MeanWell power supplies seem to work well with a motherboard that is basically a plain backplane, like Todd Goodman's 9-slot board; however, the North Star board has a bunch of voltage regulators, serial and parallel logic, power for disk drives, and other stuff. Maybe that's contributing to the failure in the way it's constructed
 
>>> Maybe that's contributing to the failure in the way it's constructed.

Unlikely to my way of thinking.

Have you tried putting a load resistor on the leads of each 15V power supply as was suggested a while ago?

There is still the possibility that the power supply is OK when there is no load on it, but it becomes unstable with a very low load, but is OK when you get above a certain load threshold.

Connecting an empty backplane is pretty much the same as attaching some wires - they don’t take current from the supply. The motherboard regulators will shunt a small amount of current from the supply to GND to regulate the voltage down. This small load could be what is confusing the PSU.

An S-100 board will only have a regulator on the power lines if it consumes current from that line. It may also be that the tiny current is not enough load for the Meanwell.

In this, purchasing another switch mode power supply may just lead to the same result.

On the machine where they work, do you know what current is being drawn from each rail?

Switch mode power supplies can be quite ‘funny’...

Dave
 
Back
Top