Repair adventures with a Dick Smith D3800 /Manson EP925/Palstar PS-30 PSU.

[vk6zgo]

Although there already was a thread on the D3800, I was hesitant to "necro post".

https://www.eevblog.com/forum/repair/faulty-dse-d-3800-power-supply-similar-to-manson-ep-925/25/

So "here goes".

Over the past month or two, I have, in a desultory fashion, been attempting to fix a D3800 PSU.

It is a fairly simple design, but has a few weird design features, & mine, at least, looks like the factory all "went down the Pub" & left the rawest Noob to wire it up.

Because of the latter hassle, I hadn't really dived into it, but instead, spent time looking for information on the 'Net.

The voltage setting pot, VR3, had been playing up for a while, so when I couldn't get any voltage control with it, I replaced it-----all good, I now set it to 13.8V & connected my ham rig.

At that point, the output dropped to zero volts!

OK, lets dig around a bit!
I found there was an inline 2A fuse (let's call it FS2) in between the transformer winding powering the LM723 & its rectifier.--------See the Palstar PS-30 schematic on
 https://pa0fri.home.xs4all.nl/Diversen/EP925/ep925eng.htm

This had gone open, & as I didn't have the right value, I temporarily replaced it with a 500mA one.

The thing now powered up OK, & the DMM measured 13.8V or whatever VR3 was adjusted to.

"Coo Callay, Ohh frabjous day!'

Being a bit cautious, I went & bought some 2A fuses, fitted one, & all still seemed to work.
Maybe try a bit of a load?....

I dug around in the overflowing junk, & found a 22w auto indicator bulb.
It lit up, so I  decided to check what current it was drawing.
The Fluke 77 in 10A current mode in series gave a fairly logical 1.7A.

Great, I'll just check the voltage again......."Ohhh bugger!"

That's right, ancient & all as I am, I got caught with a "trap for young players", just like the rankest Noob!

Interestingly, the Fluke seemed OK, but the PSU was "dead as a doornail".
"FS2" was dead again, but on replacement, the supply came up with permanent non-adjustable 24V!

The conventional wisdom would have it that such a condition meant one or more of the 2N3055 pass transistors had collector to emitter short circuits, so I checked that, & they looked OK.

Aaaarrrgggghhh!

And why did"FS2"supply fail------surely it only supplied ic1 & TR6?

Because the connections to TR6 were a horrible "shit mix", I unsoldered them, & tested the TIP31c, which had a collector to emitter short.
If that turned the 2n3055s on hard, it would account for the 24v, but why did the short circuit on the PSU output kill TR6?

“Curiouser and curiouser!” Cried Alice !

At this point, the base connections to the 2N3055s were disconnected, so for want of anything else, I checked both junctions on each of them.

Bingo! the base emitter junction on one (they don't label them, but for convenience, let's say TR4) was very wrong, & measuring it on Ohms it was 445 Ohms.

It seems that this failed, removing the bias on all the other 2N3055s, cutting them off.
The TIP31c was still adjustable, & supplied  voltage to the output terminals via the "TR4's" Base/Emitter short & "R4".

What is a bit confusing, considering the 445 Ohms DMM reading, was that TR6 could supply around 1.7A.
I can only conjecture that the base/emitter "short" went to a lower value with a higher voltage than that from the DMM was.

Anyhow, the PSU output short via the Fluke demanded more than TR6 could give, so the raw 24v from the supply intended for ic1 was now what appeared at the D3800's output.

A few more strange things came up ----Several posters on 'Net forums say some of these PSUs have two thermal switches, one to turn the fan on & off, & one that opens the transformer primary circuit in the case of overheating
The unit on the bench does, indeed have two thermal switches, but the second one opens the transformer secondary of the ic1 supply, & is in series with the fuse referred to earlier.

All the above is the "easy bit"-----now I have to tidy up the horrifically messy wiring, & do everything the way it should have been done!

[m3vuv]

and the point of this post is ?!!

[vk6zgo]

and the point of this post is ?!!

Perhaps that the "accepted" answer to a problem isn't always the correct one.

I have seen multiple posts in different forums, all saying, in part; "If you are getting 24v non-adjustable at the output of one of these PSUs, you have a collector to emitter short in one or more of the pass transistors."

In fact a base to emitter pass transistor short can cause the destruction of TR6, with what appears to be the same symptoms.
Further, before TR6 fails, it will happily supply regulated volts at a very much reduced current capacity, to the output via the pass transistor base/emitter short, leading the unwary to think the fault is fixed.

The poor wiring layout makes it hard to trouble shoot, so is a trap for the unwary.

The second thermal switch has been mentioned in some forums as switching Mains, but I found it was switching that secondary of the transformer which supplies 1c1 (LM723) & TR6.
I have a sneaky suspicion that the messy, tangled wiring, & the quite heavy red wire combined to mislead some folk, leading to the "it switches 240v" hypothesis.

Anyway, "it is what it is", a description of a, albeit uncompleted, repair, posted in the "Repair" thread.
If you don't like it, pass on to something else.

[m3vuv]

I have the same psu,given to me a while ago with no output,ive modded it tho,the 723 was dead,i fitted a socket and a new 723,on the pass transistor side of things  i found there was holes for another pass transistor on each heatsink so i junked the original 3055's and fitted 8x 2n5775 trannys along with 10 turn pots for current limit and voltage settings,also did the fan mod as originaly it sounded like a 747 doing its take off run,also have led backlights and a overvoltage crowbar circuit using a mc3423 ic on a piece of veroboard,its my main psu for my workbench used for everything from powering my had radios to charging batterys also runs my pcb drill well,now a decent psu.

[vk6zgo]

I have the same psu,given to me a while ago with no output,ive modded it tho,the 723 was dead,i fitted a socket and a new 723,on the pass transistor side of things  i found there was holes for another pass transistor on each heatsink so i junked the original 3055's and fitted 8x 2n5775 trannys along with 10 turn pots for current limit and voltage settings,also did the fan mod as originaly it sounded like a 747 doing its take off run,also have led backlights and a overvoltage crowbar circuit using a mc3423 ic on a piece of veroboard,its my main psu for my workbench used for everything from powering my had radios to charging batterys also runs my pcb drill well,now a decent psu.

Sorry to take so long getting back to you, Paul.

You definitely fixed the thing the way it should be done!
At one time, I was thinking of fitting all the guts in a bigger case, with a really big passively cooled heatsink on the back, but such cases are a lot harder to find these days!

I've now lowered my sights to just getting the original thing up & running---after all, it worked OK for years.

About the 2N5775s,-------- I tried looking them up on Google & Bing, & pretty much "got the runaround" on both.
One place was trying to sell me some Russian things as "2N5775 equivalents", which had a weird package I can't remember seeing since the early '60s, & others just showed pictures of generic "2Nxxxx" transistors of various types, so I thought I'd better ask you, as someone who has actually used them.

Are they a different package to the 2N3055s?

Another question:-
Was the original wiring really messy in your PSU?

I've seen various photos, some having neat wiring, & others as horrible as mine----I'm wondering if the former are from the owner tidying things up before the shots, or if they vary a lot.

[m3vuv]

the transistors are the same package as the 2n3055 transistors but rated for 16 amp each if i remember,i brought 10 of them from a uk ebay company only issue i had with them was trying to match them as the beta varied from about 12 to 20,my wireing was a mess too,forgot to mention i have to switch out the overvoltage protection to run stuff over 13 volts,it powers a 40 amp car heater plug relay when it goes over voltage that relay goes o/c,if you get what i mean,didnt fancy the scr rout and have it blow fuses,also triggers a buzzer too in the over volt condition,must say there not the best to work on due to lack of room,73 m3vuv.

[m3vuv]

forgot to mention i used a tip 122 darlington  in place of the original driver for the pass transistors to help the 723 a bit.

[TheBay]

I've put TIP41C's in the last few of these I've repaired and they haven't come back.
Not the nicest of linear power supplies to work on, the whole assembly on the front panel is a faff.
If one of these is going to be used with anything of value, it would be cheap insurance to fit a thyristor based crowbar.

[vk6zgo]

I don't usually necropost, but I was the OP of this one, & I thought I may as well finish the story of this very drawn out repair.
A couple of months back, I finally got the horror back working.

I replaced the faulty pass transistor with a 2N3055 of around the same era, & the driver,Tr6 (Tip31C).

The wiring was tidied up, so that all the wiring from Tr6 emitter to the pass transistor bases was from the main PCB,& the 2N3055 emitter connecting wires to the PCB were replaced with heavier wire.

It seemed to me that having three pass transistors on the bottom heatsink where the air flow was quite problematical wasn't a good idea, so I reduced them to two, with three on the top one.
The thermostat in line with the supply to IC1 was removed & bypassed, & that spot was used for the fan thermostat.
A 1k resistor was wired in series with the thermostat so the fan was run slowly continuously.

The analog voltmeter had died, & anything similar was unobtainium, so a small cheap, digital voltmeter was fitted----- the PSU now looks like it has a bloodshot left eye!


Much of the "fiddly stuff" involved with the above consisted of fitting & refitting the front panel, which was a pain in the bum, as the two LEDs kept bending back, so I glued some stiff cambric spaghetti over the holes to facilitate the job.

At this point, I buttoned the thing up, & tried it. out,
It seemed to regulate relatively well, & the voltage control functioned normally, but the ripple was around 150mv, which seemed a bit excessive, so I had to investigate that.
At that point, the fan died!

A quieter, slightly higher flow rate fan was obtained & fitted, & I then turned to the ripple problem.
The IC1 supply looked to have a fair bit of ripple in its own right.

I fitted a L7824 in place of D13, as per PA0KV's mod on:

https://pa0fri.home.xs4all.nl/Diversen/EP925/ep925eng.htm

The ripple is now minimal.

The PSU is now quite useable, even though it is still a bit of a "kludge" .

[Ed.Kloonk]

Glad to hear you got it going.

I don't know why these cheap PSUs have to be so terrible. I wonder if it's some sort of inside joke.

[bigpimp347]

I have issue with 925 here..
well several..

main one..
voltage good 2.5v to 15v variable..
under load voltage drops to 3v
not looked in to it much but was hoping something simple.

second 925.
powers up most times, other times flick switch no power, then maybe after a few times it will power up and work fine.

third unit Watson PSU similar design internally to 925.
only showing 7v maximum output, i suspect the ic which steps up the voltage to have failed or what ever it is, i know what it does but unsure the way to describe it.

and today a Mydel, no power at all, just blows fuses on power up.
so hoping easy dead short somewhere..


the 925 is my main project,
i repaired one recently at 20v which are easy fix, but no current or below voltage ??

[bigpimp347]

Maybe after nearly twenty years he isn't sat on a foundation licence
and can actual diagnose and repair things without the need to write cheques ??

[bigpimp347]

main one..
voltage good 2.5v to 15v variable..
under load voltage drops to 3v
not looked in to it much but was hoping something simple.

second 925.
powers up most times, other times flick switch no power, then maybe after a few times it will power up and work fine.

third unit Watson PSU similar design internally to 925.
only showing 7v maximum output, i suspect the ic which steps up the voltage to have failed or what ever it is, i know what it does but unsure the way to describe it.

and today a Mydel, no power at all, just blows fuses on power up.
so hoping easy dead short somewhere..


the 925 is my main project,
i repaired one recently at 20v which are easy fix, but no current or below voltage ??

1st one was the AC connection to the bridge wasn't actually connected after an 'amateur' replaced the bridge..!!
2nd still not found the issue..
3rd was the two IC's on the board had failed i assume one had took out the other one drives a pair of transistors which in turn drive the 3055, i can't recall the part numbers, but two IC later and all good..
4th the Mydel was the rectifier..

1st i blew the rectifier myself recently dragging 35amp from the PSU for a few minutes..!!!
although a 50a 1000v was fitted !!

still repaired and all good, using a 3510 and remember not to draw too much from t next time.

my other old Palstar PS-50 is now working but seems under load (35amp) it drops voltage..
now these have an overload protection so that might be set wrong so need to get the HF set out and work up some current to find where it's drops voltage and set accordingly..

fingers crossed..