HP (and other) TE power supplies - repair or replace with a new design?

[HalFoster]

All:

I restore old test gear, primarily HP, as a hobby and have more and more been wondering about how to deal with failed - or about to fail any day now - power supplies.  Typically anything I work with is at least 35 to 50 years old - I just replace any and all electrolytics since if they haven't failed yet, they soon will.  Unless something is rare or for aesthetic reasons, I do "functional" restorations - use the best modern components I can with the goal of the equipment lasting another 30 or 40 years and not trying to match the exact component appearance - I restore to **use** not sit on a shelf.  Which brings me to where 90 percent of the repairs are: the power supplies.

The one(s) in question at the moment are in a couple of 3335As that i just acquired.  One has a bad rail, the other works (for now).  IMHO, the 3335A is a truly amazing beast with performance that is very rarely matched today and was obviously designed by very, very talented EEs... who got high and decided to play "what if."  And while they were busy eating chips and designing wondrous things, the new kid got stuck with doing the thermal management for the thing.  Who, when he pointed out that the senior engineers had created a mighty nice oven, was evidently ignored. Because this otherwise impressive feat of engineering was sent out the door with a temperature disconnect on the power supply. Not on a pass transistor or such, oh no. It's there for high ambient temperatures.

High.  Ambient.   Temperatures.

Sigh.

All snark aside, this really struck me that it would just make more sense to do a completely new supply using modern components, keeping the transformer and staying linear, of course.  Which made me think of how many other pieces I have worked on that it would have been faster and cheaper to do the same.  It seems kinda *wrong* somehow, but is it really any different than using a better spec'd modern radial in place of a old screw-in can? Or using an radial in place of a axial?

So, what are your thoughts and opinions?

TIA,

Hal

[bd139]

A wise man plays chess one piece at a time rather than smashing all the pieces off the board and starting again.

On HP stuff, I received a 52 year old HP 6284A a couple of weeks back. That has original parts and capacitors in it, which are all still in spec, and a more than adequate passive cooling and engineering. I would attempt to learn from cases like that and apply to the other things.

Alternatively I tend to accept a 20 year refurb cycle on everything. Is your new power supply going to last as long as the original one did? Maybe not.

[anewmanx]

I don’t see anything wrong with improving the technology. I’ve thought about doing that myself to the one I’m working on as a good affordable replacement to-3 package mosfet that meets the original specs seems to be hard to find.... at least that affordable part is difficult to find.

[Tomorokoshi]

I don't have one of those, but I did look through the service manual. It looks like the pass transistor board, A15, has 4 TO-3 Darlingtons on finned heat sinks that are worth maybe 10W each. Possibly a good 40W into that area with restricted airflow.

I've seen a lot of variation in how HP made supplies. Benchtop supplies might put the transistor in a large finned heat sink, while others put the transistor on the back panel. The ones that seem most trouble some use those finned heat sinks, because they always wind up in the middle of the unit somehow.

Changing the parts while keeping with a linear design probably won't help. It's still going to dissipate all that extra heat. The other thing is that it seems to be from "20V UNREG" only down to "15V", so there isn't a lot of room to play with. What are the "20V UNREG" rails actually at?

Is there a fan in it?

[HalFoster]

There is a fan - a 120 CFM 120V one; it blows inward from one side and pressurizes the interior of the case - the modules all have air inlets and outlets - the air then exhausts through the far side of the case (the sides have perforated panels).  The power supply sits to the near right of the fan intake - the air path must make an immediate right angle, pass through the heat sink fins (and harness wiring that is in the way) and then exit the case.  In the nonworking unit, someone has replaced the fan with one rated at 107 CFM - which would not be an issue except that they did not consider the static pressure requirements.  The original fan probably moved an actual 70 to 80 cfm and that worked.  The new one probably only pushes through 30 to 40 at best - and that follows the easiest path which is straight through with very little making it via the PS heat sinks.  There is a fair amount of headroom even on the +-15 rails - it is designed to work with low (110V or so) line voltage.  It's a straightforward PS - no remote feedback or other integration with the rest of the unit.  The +15, -15, +5 and -5 rails are current limited to around 1.25, 1.5, 2.75 and 2.5 amps respectively; actual current draw averages about half that.  Typical of this time frame, only the 15V rail is adjustable - the others are all based of it's value - and the unregulated voltages are present even when the unit is "Off" as long as it is plugged in; the oscillator oven is powered constantly. Other than the thermal portion of the design, it's classic HP of that era: rugged and simple in a very elegant way; HP electrical designers may do things that seem strange but when they do, it's for a *reason*

The main reason to redo this one is to open up space and improve airflow.  I'll just repair it as-is for now for testing, but may come back later and re-spin the whole thing - I plan on having this one cal'd and I really don't want to have to pay for that twice.  One thing I like to do, especially on something like this, is to change out the existing AC fan for a high quality DC fan; I can get better performance and cut the noise down quite a bit - the stock fan is about 55 DB.  They used mostly AC fans since at the time they were more reliable, but they are a lot louder than DC for a given amount of air moved.  Since I just replaced several capacitors with ones a 5th of the size, it is usually not hard to find a place to tuck a small 2 or 4 VA transformer, rectifier and filtering to run the new fan.  Use the original AC leads to power the transformer.

Hal