1983 Aztec PSU needs to "warm up" before working mystery.

[Bob-O-Rama]

I have a number of AT&T 4425 Teletype Terminals which I have been able to restore to working condition. For all of these, when I first got them, presumably after many years of being unused, they had a weird phenomenon. After being off, they would not POST, or after being let sit with power on would spontaneously POST a minute later and work fine. After this warmup period, the terminal could be power cycled without incident, or left off for a few minutes, etc. After a few months of use this phenomenon diminished - except if I stored them in a very cold area for a while, then its perhaps a 5 second pause.

I now have another one of these ( not powered up in 10 years at least ) which does the same thing, After the PSU "warms up" you can power cycle repeatedly and it come on immediately. To see if this is an issue with the logic board or CPU, I have swapped a "hot" and "cold" PSU between chassis, and the issue follows the cold PSU. I've looked at them endlessly for cold joints, etc. Nothing stands out. Output voltages seem normal before and after POST.

One thing I did notice is that the PSU makes a faint chaotic, high pitched squealing noise when its misbehaving. This is from the PSU not the video section, its not the flyback. Eventually this disappears and then machine POSTs normally and the analog section / flyback is powered on. I used an audio spectrum analyzer and can see this "noise" in the 10 - 20 Khz and higher spectrum - and once that stops, 5 seconds later POST completes, you get a nice solid tone from the flyback and a much, much fainter noise from the PSU itself.

I have not observed any differences in voltages from the PSU before or after post. There is a "DC-OK" lead from the PSU which seems to provide +5V shortly after power on.

So after all that ... what would cause this "warming up" phenomenon? I have to think its mechanical - like a cracked solder joint. But I've looked endlessly.

 

[NeXT]

I typically lean on cap issues when a power supply seems to be audibly noisy on a cold start but it settles down after a few minutes of use.
It may not be "noisy" electronically but the capacitance values will drift off with time and lack of use. The semi-technical term for the caps drifting back into spec with use is "reforming" and is usually a temporary measure.

 

[mdog69]

Bootstrap capacitor? The one that kicks a switched mode PSU into life when power is first applied. If you have a circuit diagram it's a bit easier to find - usually fed by an out of place resistor which makes you think "wots that there for?"

 

[Bob-O-Rama]

Bootstrap capacitor? The one that kicks a switched mode PSU into life when power is first applied. If you have a circuit diagram it's a bit easier to find - usually fed by an out of place resistor which makes you think "wots that there for?"

Prompted by your reply, I found an Astec schematic for a similarly designed early 1980s PSU. ( http://www.bitsavers.org/pdf/astec/Astec_AA11320_38W_AA12090_65W.pdf ) The document also included general troubleshooting for the"flyback" circuit and speaks to the specific issue you mention. Thanks!

-- Bob

 

[Hugo Holden]

These sorts of supplies don't have anything like a boostrap capacitor. They have a fairly standard oscillator circuit driving the primary, it starts (should start) the moment the line power is applied. Then, when the secondary voltages start to build up, that is compared to a voltage reference on one of the DC outputs, that is fed back typically via an isolating transformer (or an Opto-coupler) to the oscillator to reduce the duty cycle of the primary drive, and it stabilizes the output voltage.
If there is any issue with any electrolytics on the primary side, or where the feedback circuit is developed on the secondary side, it can result in abnormal modes of oscillation. I would not suspect the main filter caps though, initially at least. If the supply was playing up, it would be worthwhile to perform a full re-cap of the small electrolytics (except the large main filter caps that are usually ok) and use good Nichicon 125 Deg C rated parts so you don't have any more bother with it.

In some more complex psu's a small line transformer is used to supply power until the IC driver circuits stabilize and the supply's own output takes over.

When the main filter caps fail, the whole arrangement starts to behave as a low frequency giant relaxation oscillator, because their terminal voltage collapses with initial load, and then it recovers.

 

[mark0x01]

I would also check the resistor(s) that provide the startup power feed (R2 in the scematic referenced.) as they can go high with age.

 

[Hugo Holden]

R1 and R2 are 4 Ohm NTC resistors (called thermisters) that limit the peak charging currents of the four filter capacitors C5,C6,C7 & C8. These are negative temperature coefficient types. They start out at room temperature with a much higher resistance, and as they heat up with current flow the resistance drops. The application of them is to reduce the surge currents into the capacitors C5 to C8 at turn on, which, without them would be extremely high, because the source resistance of the line power supply is very low, the, forward resistance of the diodes in the bridge rectifier is very low, and the filter capacitor represents a very low impedance to a rapid change in voltage across their terminals. With the NTC resistors, the peak (transient) current, at turn on, could be many tens of amps. You can think of these NTC resistors as giving a "soft start".

 

[Radix]

I would also check the resistor(s) that provide the start up power feed (R2 in the schematic referenced.) as they can go high with age.

As Mark0x01 said above:
+1 for checking / replacing the 150K R2 on the diagram - seen that part fail in numerous switch mode supply designs over the years...