Print

[compu_85]

I'm repairing a high hour 1.8a Datapower Lisa supply. I think the original fault was a shorted dual diode, however the plastics on the standby power transformer were so crispy that when the owner shipped it to me, the transformer's plastics turned to dust, it came loose, and ripped the wires off

Looking at the schematics, this transformer is a bit more integral to the design than with the Apple supply; it provides some feedback to the later stages of the supply... not just 5v standby.
If someone has an extra one of these they'd like to offload, I'd be interested!

I could try finding the ends of the wires and re-attach it, but at the least there would be fewer turns remaining after I'd done this.

Thanks,

-Jason

[sigma7]

I'm repairing a high hour 1.8a Datapower Lisa supply.

.. the standby power transformer ... came loose, and ripped the wires off

Looking at the schematics, this transformer is a bit more integral to the design than with the Apple supply; it provides some feedback to the later stages of the supply... not just 5v standby.

The standby power transformer doesn't do as much as it may seem...

It provides power for the +5STBY linear regulator and the PSU's daughterboard, and a signal for "line sense" which is used to detect loss of A/C power to indicate "power good/fail" to the Widget.

As the parameters for these aren't critical, a readily available transformer should be a fine substitute.

+5 Standby is regulated by a 78M05, max supply of ~35V, max current 500 mA
The SG3524 on the regulator daughterboard has a max supply of ~40V, max output current 100 mA (possibly x2 for the two outputs).

So, I'd suggest a 32 to 36 V center-tap transformer rated at 700mA or so. The original has two primary windings to select 110/220V, which seems like a useful feature to retain.

As a starting point, consider Triad VPL36-700 BUT NOTE: I have not verified this will fit in the available space, nor that the power rating is appropriate!

There are some PC mount transformers in this voltage/power range that might even be drop-in replacements... I suggest establishing the minimum voltage/current specs and looking at those if time permits.

I don't know what line voltage the brownout adjustment is set to trigger at; I think this would only be a consideration if you are using an actual Widget. Someone should check the setting on an original DataPower... volunteers?

[stepleton]

I do have a 1.8A PSU, but I'm not sure how you would check the brownout adjustment: would you use a variac? Or I suppose you could remove the card and power it up separately somehow?

[sigma7]

I do have a 1.8A PSU, but I'm not sure how you would check the brownout adjustment: would you use a variac? Or I suppose you could remove the card and power it up separately somehow?

I think a variac likely is the way to go since the objective (in this case) is to determine how to adjust the setting if a different standby power transformer is used. (Which would output a different "line sense" voltage).

If no variac is available, a good-enough option may be to measure the AC Line voltage at the same time (more or less) as the DC voltage at the wiper of the brownout adjust trimpot.

(Hopefully we can get both types of measurements for a few PSUs to determine if the no-variac method is sufficiently accurate)

Regardless of method, I imagine the measurement should be done with a full load on +5STBY.

eg. 500mA @ 5.7V would require 11.4 Ohms @ 2.9 W, so one might try two 22 Ohm @ 2Watt resistors in parallel or four 47 Ohm @ 1Watt resistors in parallel as the load.

[sigma7]

[Regardless of method, I imagine the measurement should be done with a full load on +5STBY.
eg. ... two 22 Ohm @ 2Watt resistors in parallel

Initial results with two PSUs at hand (configured for 115V):

Transition voltage (+/- 1V) for /PowerFail using a variac...

With 11 ohm load  on +5 STBY
 PSU A: 113 VAC
 PSU B: 120 VAC (more than the local line voltage)

With 22 ohm load
 PSU A: 94 VAC
 PSU B: 96 VAC

With 56 ohm load
 PSU A: 82 VAC
 PSU B: 84 VAC

With no additional load
 PSU A: 76 VAC
 PSU B: 77 VAC

PSU A sn 18282
PSU B sn 15139

So, very strong dependence on the +5 STBY load. Such a big effect that I imagine the spec was to detect complete power loss rather than brownout.

If someone else observes something similar, then I suppose we can say the setting doesn't need to be precise.

(However, this presents a mystery in why is there a trimpot at all?  I suppose the regulator board was designed to be used in other PSUs as well.)

edit: Note that /PowerFail is an open collector output, so you'll need a pull-up resistor to +5STBY to observe it easily.

[stepleton]

Got it. I have a friend with a variac; I might be able to check myself in a week or so, but not sooner. I would be surprised to find a big difference, and if I did, I'd most likely suspect a misconfiguration or problem with my own PSU than an issue with the two of yours.

[sigma7]

The May 1983 version of the Hardware Reference Manual says the +5 STBY is rated at 100 mA.

So possibly this was the "official" load rating.

I'll update the readings above for the 56 ohm load case for completeness.

[compu_85]

I shipped a spare supply I had to the friend I was making the initial repair to.... and my supply suffered the same fate! Transformer broken off of its plastic supports.

Now I have 2 to fix....

-J