• Please review our updated Terms and Rules here

Finding out what killed my North Star Horizon

Ttpilot

Experienced Member
Joined
Nov 8, 2020
Messages
100
Location
South Dakota USA
About a year ago I fired up my ancient North Star Horizon. It had sat since about 1986. Lo and behold it worked...for a couple of weeks, after which it would no longer boot. I finally got myself a multimeter and started checking voltages on the motherboard (it's been 40 years since I've done this sort of thing, so I'm pretty rusty). The voltages from the power supply to the motherboard were all good. The voltages powering the floppy disks were also good. I got about +10v on bus pins 1 and 51, +20v on bus pin 2, and 0v on bus pin 52. Looking at the motherboard power supply schematic, I'm guessing one of the tantalum capacitors (C1 perhaps?) has failed. I've included the schematic. I'm looking for advice from folks who are more experienced than I am, and I'm guessing that's most of you

Thanks in advance
 

Attachments

  • IMG_9657.jpeg
    IMG_9657.jpeg
    171 KB · Views: 11
Last edited:
It could be a number of things, so double check starting with the transformer, is there AC voltage present between pins 5 and 6 (approx 20V - I'm guessing yes since you have +20 at bus pin 2) and 6 and 7 (approx 20V) and 5 and 7 (approx 40V)? Is there a good connection from the transformer pin 5 and the rectifier? Same for pin 7 and the rectifier. Also check for continuity from the transformer pin 6 to ground. Try disconnecting the C1 tant to see if that fixes it. If not disconnect the 79L12 regulator for the -12V line. If that doesn't help, disconnect the electrolytic marked 8,900 (uF) and see if there is voltage from the negative terminal of the bridge rectifier. Check for cold solder joints too.
 
As cruff has stated, follow the voltages from the transformer secondary (pins 5, 6 and 7) through the bridge rectifier to the reservoir capacitors and see what is missing.

If C1 has gone open circuit, the -16V line will not be at 0V. If C1 has gone short circuit - it will probably have blown itself to bits!

Dave
 
I agree, unlikely capacitor C1 shorted, but it could have shorted taken out the diodes in the bridge rectifier feeding it or caused the transformer to go open. Likely either both of the two diodes in the bridge that generate the negative supply is open, or half the transformer secondary winding is open, or there is a fractured / broken connection in that circuit, possibly more likely. So as advised check the transformer's AC output and check the diodes in the bridge. Since the secondary transformer windings are fairly thick copper wire, seldom they go open. In my Sol, they were crimped to spade terminals with poor crimping and went open that way. I'm not sure how the transformer wires are terminated in your computer. Generally they can exit the transformer as insulated copper wire, or they can be terminated inside the winding to pvc covered flying leads with a solder join there, then they can be attached in various ways to the pcb, there are a number of places where a connection can fail.

With the power off you can check for continuity between terminals 5, 6 &7 of the transformer as well as check the bridge diodes.
 
Thanks for the input. I'll try those steps. I took the motherboard out for a look. The trace for bus line 52 shows heat damage and has been blown off the board at one point. Very depressing
 
S-100 bus pin 52 is the -16V line - so could be related to your problems...

If the track is burnt there could have been a short circuit between the -16V line and either 0V/GND (if an onboard regulator went short-circuit from input to 0V/GND) or +8V (if there was a short circuit between pins 51 and 52.

Either way, if you don't have a fuse, the backplane PCB track is the fuse :-(!

My Cromemco Z2D has a 15A fuse between the power supply and the backplane. In reality, a direct short circuit will probably still take out the PCB track! That is how I got my first S-100 system. It came from the Byte Shop in Birmingham and was a Cromemco 21-slot Blitz Bus. It had a burnt out +18V or -18V track that had been repaired with wiring. It worked fine though (unlike my current system)...

Anyhow, back to your issue, I also had a bridge rectifier go faulty. Two of the internal rectifier diodes went open circuit and a lost a supply rail. I suspect it just died - and wasn't pushed...

Dave
 
It's hard to tell from your description what section of the -16 VDC PCB trace vaporized. Look at the blown PCB trace. One part of the Trace
will go to the Negative lead of C1. So, likely that part of the Power supply was supplying the current (Voltage) to blow the trace. The other
section of the blown PCB trace likely goes to the SHORT (load) that caused over current. If R32 was shorted or near Zero OHMS, you would
have 16 = I * 330 Ohms where I = 48.4 ma. If R31 was 1 or 2 OHMS then the Current would be 16 AMPS (1 OHM) or 8 AMPS (2 OHMS).
If you OHM out that PCB Trace to GROUND, how many OHMS of Resistance do you have? It's also possible that this Section of the PCB trace
could be feeding the 79L12 Negative 12 VDC Regulator, so don't over look it. (If you OHM from the blown PCB trace to the Input of the
79L12 Negative 12 VDC Regulator, you should see ZERO OHMS or nearly zero OHMS.)

Where else is the -16 VDC used in the computer? The Schematic doesn't show us that.

Once you find and fix the problem, just use a piece of heavy stranded copper wire to solder to existing PCB Pads to repair the PCB.
#18 should be good for 15 AMPS, and under NORMAL conditions you will have less than 50 ma current through it. I'd probably
use #20 AWG wire.

Larry
 
Last edited:
>>> As far as I know that is its only use

S-100 bus line 52 supplies -16V for use by the S-100 cards you have installed...

For example, if you have a DRAM card using 4116 DRAMS, these require a -5V supply and this will (invariably) be provided by an on-board LM7905 (-5V regulator) fed from the -16V line (pin 52). Around the regulators you will usually find the dreaded tantalum bead "smurf grenades". If one of these goes short circuit (especially on the input side to your regulator) this will short the -16V bus line to 0V/GND and "puff" - firework bus track. Most often the tantalum bead explodes - but not always.

Dave
 
....tantalum bead "smurf grenades". If one of these goes short circuit (especially on the input side to your regulator) this will short the -16V bus line to 0V/GND and "puff" - firework bus track. Most often the tantalum bead explodes - but not always.

Dave

Yes, it is interesting, if the Tant capacitor shorts to a very low Ohmic value of a fraction of an Ohm, the power transferred to it by the high current is lower. If the Tant shorts to a higher resistance, more power (heat it generated in it and it burns ) and if it just goes a little electrically leaky over some 10's to 100's of Ohms, it might barely get warm if at all. The maximum power that can be transferred to the leaky capacitor as heat occurs when the resistance of the capacitor equals the internal resistance of the power supply, at this point the supply would be forced down to 1/2 of its off load voltage. It is like the DC circuit equivalent of the old AC impedance matching problem in audio work and power being transferred to a load being the highest when the load impedance matches the output impedance of the amplifier.

Likely if the shorted Tan cap is intact and not burnt, it is a very low Ohmic short and that would have resulted in enough peak current to vaporize the circuit track, but there are other explanations for the short, regulator, physical shorts etc, but I would guess since it was working and failed spontaneously, a shorted tant cap is on the top of the suspect list.
 
Hugo,

Very inciteful (as usual),

Assuming the power supply doesn't shutdown (or crowbar) as a result of the excessive current flow - then the graph shows a rapid increase in tantalum power dissipation when the resistance increases from 0 Ohms up to the internal resistance of the power supply. The tantalum power dissipation then gradually decreases as the tantalum resistance increases further.

That's two bits of knowledge I have captured today!

Thanks,

Dave
 
Likely if the shorted Tan cap is intact and not burnt, it is a very low Ohmic short and that would have resulted in enough peak current to vaporize the circuit track, but there are other explanations for the short, regulator, physical shorts etc, but I would guess since it was working and failed spontaneously, a shorted tant cap is on the top of the suspect list.

Wow, that’s great info—well above my knowledge of how these things work lol. In any case, I’ve ordered replacements for all 6 on the board
 
The problem appears to be with some board plugged into the S-100 BUSS on Pin 52 (opposite Pin 2). So, have you OHM'd from Pin 52 to Ground,
then removed each of the Boards Plugged into the S-100 BUSS, one at a time? That should get you to the Board that has the problem, which is likely
to be a RAM IC that has the -5VDC shorted Internally to a + VDC.

REF: http://www.s100computers.com/General Images/ieee696spec.pdf

Mass replacing tantalum caps may not fix the problem. You need to be on the correct S-100 Board, then find the problem on that board.


Larry
 
Sorry I haven't posted back on this for a while. While looking for information I got enthused about building a new system. With any luck, I'll soon have a functioning Altair 8800c computer. I plan to try out my old boards in it to see if they're functioning. It's been a fun flashback to when I built the North Star

Anyway, I've attempted patching the trace for bus line 52 a couple of times. When I check the voltages for the line, there is a spark, and trace blows at another spot. Here's what I've got so far:

At the Power supply:

Transformer
pins 5-6 16.1v
pins 6-7 16.2v
pins 5-7 32.4v

At the power leads on mb

3 (-16v) -22.04v (Note: these four are all as they should be according to the North Star manual)
5 (+16v) +21.8v
8 (+8 v) +11v
9 (+8 v) +11v

At the S-100 bus pins

1 (+8 v) 11.04v
2 (+16 v) 21.2v
51 (+8 v) 11.03v
52 (-16v) -22.1v (only for slots 11&12, rest are dead, blown trace again)

Again, these voltages seem to be normal for North Star systems

At disk drive power supplies

D1
(+ 5v) +5.9v
(+12v) +12.4v

D2
(+5v) +5.02v
(+12v) +12.03v

I'm going to patch bus line 52 down the whole mother board; there are several scorched looking areas that probably won't hold. Once I've done that maybe I can move forward. Right now I don't dare put a board on the motherboard

I'm mystified as to why putting the multimeter on the line to check the voltage would cause a spark and a blown trace. Perhaps that's a function of the line being damaged along its length
 
Last edited:
Yes, you need to find where that line is shorted, just continuing to patch it up will just continue to damage it further. Check for conductive debris that might be trapped in the S100 card connectors down the line from slot 11.
 
A spark is indicative of a high current flow.

If your multimeter is correctly set to measure Volts, it should have a high resistance (much greater than 20 k/Volt) so that shouldn’t cause much current to flow.

To confirm: You have NO S100 cards plugged into the motherboard and everything appears to be ‘normal’ (i.e. no death and destruction) when not measuring any voltages at all.

Question: What multimeter are you using to measure the voltages and how are you configuring it to take the readings?

I also agree with cruff regarding checking for conductive debris.

Dave
 
A spark is indicative of a high current flow.

If your multimeter is correctly set to measure Volts, it should have a high resistance (much greater than 20 k/Volt) so that shouldn’t cause much current to flow.

To confirm: You have NO S100 cards plugged into the motherboard and everything appears to be ‘normal’ (i.e. no death and destruction) when not measuring any voltages at all.

Question: What multimeter are you using to measure the voltages and how are you configuring it to take the readings?

I also agree with cruff regarding checking for conductive debris.

Dave

There are no boards in the system. As I stated, the voltages all appear as expected, even line 52 on the slots that come before the blown trace

I’m using an AstroAI TRMS 6000 set to volts dc

Thanks for the advice from you and cruff about checking for debris in the slots. I’ll do that
 
That multimeter is fine. It has a 10 MOhm input impedance, so should have a negligible effect on the power supply!

Providing your negative (black) probe is connected to COM on the meter (and this is connected to the 0V/GND of the power supply) and the positive (red) probe is connected to INPUT on the meter and the meter is switched ON and on the correct Voltage dc range before taking a measurement, you should be “good to go”.

The only other thing I can think of is that something is accidentally shorting out when you take a measurement. But where?

Dave
 
Not in a North Star, but a 4 slot S-100 motherboard in one of my systems I bought in the past few years. I found that plugging in a card (power off, but AC plugged in) would pop the ground fault circuit breaker for my shop. That had me thinking for a bit. I could repeat the problem trying a second time after resetting the ground fault. Turned out that the power leads were under the motherboard and pressing down on the board shorted AC power to ground. Dumb design and soon changed by me. I'm surprised that it lasted 40 years before arcing.
 
At the power leads on mb

3 (-16v) -22.04v (Note: these four are all as they should be according to the North Star manual)
5 (+16v) +21.8v
8 (+8 v) +11v
9 (+8 v) +11v
In which manual did you find these Volt specifications? I have looked through all my manuals but found nothing. I am talking about the "increased" voltages.
 
Back
Top