HDL Research lab military power supply LCD replacement

[nightfire]

Just got the result of some spontaneous action from ebay- a linear power supply, military version. Seller says it is basically working, but the LCD displays do not display very well, so it was sold as defective and i got it for a decent price. From the markings inside it has been handled decently, not beaten to death, and built in 1987- some dust inside.

This is something I also want to learn some stuff about linear power supplies, so I probably will ask some questions for the sake of understanding the underlying concepts....

But: As I could not find schematics- here the obvious question: Does anyone know some source of schematics for this baby? Or would it only be the manufacturer itself, that is probably with military equipment not eager to give them out?

Also I would be thankful for some anecdotes or deeper info of this military stuff- some looking showed that the NSN Number has been assigned around 1982 for this Power Supply.

Next weekend I will probably have some time to take care of it- like cleaning it first, getting the dust and stuff out, and then checking the cabling for damages.
After that, I will assess the issues with the LCD displays, and maybe look for replacements. Last pic is the driver PCB from the current display.

[m3vuv]

first thing i would do is replace the biege tantalum caps that are in the last pic,would check all the electrolytics for high esr also and replace if high.

[nightfire]

Disassembled most of it for cleaning- as the PCB are in good condition, only some compressed air and wiping of metallic surfaces were needed.
Did some measurements of the conductivity of the grounding and rough checking of isolation of the high-voltage parts with a multimeter- as I could deem it good i switched it on and checked for stability of voltage at the output.

The 10-turn pots do nicely, very sensitive, but not overly jumpy, LCD display correct values, with the Voltage LCD display there are most letters correct, only some bars in those 7-segment displays seem to be completely broken, other are working fine- so for example the number behind the decimal point would be displayed wrong, like an 8 would bekome a 3...

Voltage output seems stable at open clamp situations, and LCD reading correspond with what my DMM says.

Decided to ramp it up a bit, hooked up my electronic load (rebadged Korad KEL102) and tried to go for a CC load scenario. Put the Power supply to 6V, and went on in 10mA steps.
Current display displayed nicely (without missing bars, at least what I could notice) and displayed a bit higher current than what my electronic load would note.
Current capping also seemed to work nicely, but when I went above 330 mA, no matter how far (to the end of 10 turns) the current limiter was turned, voltage wouild drop to nearly zero, and the electronic load also noticed a max of 330 mA flowing through it.
As this is way apart from the rated current of 1 A, I suspect some defective parts...


So my findings so far:
- LCD display for voltage has issues, so I will have to determine if the display itself is broken or if the managing electronics behind it (on a separate detachable PCB) are defective. As the PCBs behind those LCDs are coded differently with jumpers, I will have to see if this is due to the different stuff they have  to put out, or if they really are the same, so I could do some cross-swap
- The 230V plug is probably not original- and as there is only one fuse on the primary side of the transformer, I really would it like to be on phase
- There are no dedicated buttons or switches for turning the PSU on/off and output on/off- and even after turning it off, voltage at the output will slowly go down over several seconds to nearly zero
- secondary of the transformer puts out several voltages- 8.7V AC are going to the display units at front, at the main PCB  there will arrive approx 35V AC to ground
- at the blue ELNA electrolytic cap i measured about 47V

Further steps now would be to disassemble the heatsink and maybe draw some rough schematics for better understanding and sharing.

[nightfire]

I am working on some hand-sketches for schematics.
First I am to go for the primary side of the transformer, the secondary electronics probably have to wait a bit- it seems that the hex/allen Screws that hold the heatsink are US standard, not metric like my tools...

Anyway, I could need some help with the transformer- it seems that is double-tapped and has multiple voltage levels on its secondary outputs.


 Rough going of wiring as follows:
Cable to wall outlet->single Fuse on line, after that to the switch in the front that switches line and neutral, this goes to a connector board. Nicely done, as it allows for some bridging of connectors. From this board on it goes to the transformer.

This routing board has 12 terminals. I numbered them 1,3,5,7,9,11 for the side coming from the main switch, 2,4,6,8,10,12 for the side facing the transformer.

1: Black wire from Switch/Line potential
3: not connected
5: Bridged to 1
7: bridged to 9
9: bridged to 7
11: light brown wire from Switch/ neutral potential

2: not connected
4: brown wire to transformer
6 grey wire to transformer
8: white wire to transformer
10: black wire to transformer
12: other black wire to transformer


Removed the secondary wiring from the pcb and powered the system up. Voltages measured are like the following:

1/5 (L)-> PE 235 V
3 ->PE 229 V
7/9 ->PE  117V
11  (N) -> PE  0 V


On the secondary side it looks like:

3 connectors coming out. Two of them providing 8.7 VAC as supply voltage to the LCD displays. Each of them is isolated against each other.
The third connector goes to the PCB board, is manufactured from Panduit and looks like an IDC connector type on the cabling side. 6 Pins, one of them PE.
Pins and voltages against PE:
1-orange 4V
2-orange  5,4V
3-black 9,5V
4-yellow 3,9V
5-yellow 15,1V
6- PE green 0V

For curiosity, I checked some voltages against line potential:
1-orange 17V
2-orange  15V
3-black 30V
4-yellow 35V
5-yellow 24V
6- PE green 235V


Conclusions I drew from this:
a) the primary side of the transformer is probably double-tapped to allow adjustment for a supply voltage of whether 110V or 220V (given its date of manufacture somewhere in 1987 according to the date code of the parts)
b) output is not floating (therefore some caution is needed, as a isolation transformer and differential probe for my oscilloscope are still on my shopping list)
c) the transformer is having multiple windings, especially for the secondary side

Question: As this is my first approach to this kind of transformer- could someone point me in the direction for some literature and how to draw it in a nice way in schematics?





Voltages here:

[Neomys Sapiens]

I would not expect much problems with the manufacturer, if it still exists and if they still have something on it. It is not military technology per se, it is not even ruggedized, as you can, for example deduce from the large standing (radial) electrolytic cap. Under rugged use conditions, a no-go due to acceleration forces, as are the unsecured axial elkos. Further, the drop-shaped tantalum caps would be absent (hermetic tantalum caps with metal case and glass feedthroughs instead) and there would not be a plastic DIL on board. MAYBE it has some changes to fit into a test environment where the power supply could come from multiple manufacturers or that is easier to maintain.
So basically you are looking on a minor variation of a commercial lab power supply.

[Neomys Sapiens]

47V ov C1 is a bit high, as it feeds U1 (15V 3-pin type regulator for a auxiliary voltage). You seem to be measuring stray voltages, maybe due to a massive screen winding in the transformer. I would expect the secondaries to have no galvanic connection with the line voltage. Can't tell you more, as your pictures omit all the relevant parts.
You should post a view of the terminal strip, to which the transformer is connected.
Laboratory power supply circuits such as this usually have one or even two auxiliary voltages. Behind C1 are 4 diodes connected probably as a bridge rectifier. Trace the input of the bridge to the corresponding secondary winding. Do the same for the main rectifier, probably D5/D6 connected as a half bridge. then you would have 3 wires going there. The low total power means that it is unlikely that the PS employs a switched tap.
So the secondaries should be either yellow-black-yellow + orange-orange OR orange-black-orange + yellow-yellow.
LM224 is a quad OPA which is doing the control function. Look for power supply circuits based on LM324/224/124.
I can't make out what the 8-pin DIL besides that one is.

[nightfire]

I did some measurements of the traces as far as I could get with the heatsink still attached- have to order some hex/allen key with imperial size first...

Before C1 which is fed from the orange wires I measured 18V AC that go into the bridge rectifier of D1-D4. C1 itself I could not measure, because the heatsink is blocking me out.

The 47V are on C3, the big (approx >2000 µF) electrolytic on the D5/D6 half-bridge.

Better photos will come soon, I also unearthed my old macro lens, and have a new mini tripod (that gets very top-heavy with my Nikon D7100 pointed straight downwards), so there is still something I will have to figure out.
Pin 1+2 (orange) have 18VAC that go to C1, on pin 4+5 (yellow) I measured 69VAC that go to D5/D6.

[nightfire]

Some more pictures, quick&dirty shot handheld with my normal zoom lens, some slight adjustments for exposure and contrasts afterwards- so some parts may not look overly natural ;-)

Measured again the secondary side (open clamp voltage, not connected to PCB):
orange->orange 18V AC, yellow->Yellow 69 V AC, but black to yellow nearly 35 V AC.

The connection board is in the last pictures. Phase/L from Switch arrives at the bottom right, and is bridged via a small jumper cable to two pins left of it- Neutral arrives at the bottom left contact.


The other IC (U3) beneath the LM224 has the following markings:

723 B B
TL431 CP
Pin 1 goes to R6, Pin 7 to C2 minus
So it should be another voltage reference, according to google..


Pictures are resized a bit due to size limits here in the forum, I can do some 100% crops if requested...

[james_s]

Those displays look like they're just off the shelf digital panel meters, you can buy new ones that will work but you'll have to look and see if there are any that will directly fit in the openings. With old displays like that it's usually the zebra strips that need cleaning.

[nightfire]

After getting rid of the inbus screws (by the way: Whether my "normal" metric inbus tools nor freshly ordered imperial size fit- I had to get creative with screwdriver and tweezer...

below are two better resolution pictures of front- and backside of the PCB.

[max-bit]

I don't think this power supply is "military", I don't see layered (or active) armor. Besides, it has no defense systems (passive or active)
 

[nightfire]

47V ov C1 is a bit high, as it feeds U1 (15V 3-pin type regulator for a auxiliary voltage). You seem to be measuring stray voltages, maybe due to a massive screen winding in the transformer. I would expect the secondaries to have no galvanic connection with the line voltage. Can't tell you more, as your pictures omit all the relevant parts.
You should post a view of the terminal strip, to which the transformer is connected.
Laboratory power supply circuits such as this usually have one or even two auxiliary voltages. Behind C1 are 4 diodes connected probably as a bridge rectifier. Trace the input of the bridge to the corresponding secondary winding. Do the same for the main rectifier, probably D5/D6 connected as a half bridge. then you would have 3 wires going there. The low total power means that it is unlikely that the PS employs a switched tap.
So the secondaries should be either yellow-black-yellow + orange-orange OR orange-black-orange + yellow-yellow.
LM224 is a quad OPA which is doing the control function. Look for power supply circuits based on LM324/224/124.
I can't make out what the 8-pin DIL besides that one is.

I finally found some time to unscrew the heatsink and do some reverse engineering to try to draw some schematics for further debugging and discussion.
As this is the first time for me doing this, I expect them not to look very pretty ;-)

Anyway, the assumptions above seem to be present in this power supply. I disconnected the transformer from the terminals, and basically on primary there are 2 coils that can be combined or separated- probably for being able to do 220V/110V back then. No galvanic connection to the secondary side, as far as my multimeter is concerned...
On secondary side there are 2 outputs:

One with 2 orange cables leading to the bridge rectifier (D1-D4 and then to a voltage source), the other is the main rectifier on a half-bridge, here we have yellow-black-yellow going on.

From what I have seen (and measured) so far, the circuit behind the orange cables regulates the MOSFET Q1 (an IRF130) that is connected to the half-bridge rectifier, that is fed from the yellow/black/yellow cables.
The other IC that is hard to see beneath the (U2) LM224N is an TL321CP with additional "723 B  B" marking, that is marked as U3. After reading the datasheet, this is another voltage regulator with 0.5% accuracy, that can be adjusted via external resistors, and is connected to the LM224N.
Here i have to spend some work, because the traces are not that easily traceable.
And the (-) of the output is grounded/earthed=has a connection to PE of the primary side.

Questions here: As this seems to be a quite standard setup for a lab power supply: I would appreciate some tips or links to existing schematics with a similar setup, so that I can draw some ideas of how to arrange the parts fittingly...

[Neomys Sapiens]

On this forum were several proposals/ideas for lab power supplies discussed, which use the quad OPA LM124/224/324. Look at those for a start. I think there was one in some early appnote too, but I don't know where. The voltage regulator does probably compare the reference with a fraction of Vout derived via the voltage pot.. The current regulator compares the actual current (from shunt resistor, can be part of PCB) with the current pot. Another one could be used to amplify the current signal.
remark: LM124/224/324 are functionally identical, only differing in the temperature range. LM224 is the industrial version, while in most circuits the cheaper commercial 324 is used.

[tooki]

FYI, for when you don’t have imperial/customary/SAE hex keys around: try a Torx driver. I don’t know if it’s coincidence or if Torx sizes are based on inches, but they’ll often work to drive them.

[nightfire]

After lots of measuring and tracing the connections (sometimes routed under an IC..) and admiring some detail work like areas probably intended for shielding that are grounded I managed to break the parts of the PSU in 3: Transformer part, Power circuit and control section.
As I am not very familiar with CAD systems, I decided to draw this in good old manual fashion...
Attached are Schematics for the transformer side and for the power circuitry, the control section is a bit more complex and i have to figure out how to draw it without cluttering too much- the LM224N has all 4 Opamps in use...

On the bright side: The LCDs that were originally reported as faulty are ok now, some disassembly and cleaning of the zebra strips with IPA and contacts for the Zebras with a fibreglass pen worked magic.
Also whilst further diagnosing the output problems I blew the fuse, and after that I could measure (in-circuit) that Diode D16 has a resistance of 0.00 Ohms...
So some parts are on its way to replace the fuse (I will go for 6,3mmx32mm fast blow fuse of 630mA) and the diode. As the Motorola MR501 is not available anymore, a 1N5408 is supposed to take its place...

[nightfire]

Aaand the culprit is:


This Resistor measured 0,3 Ohms after desoldering and measurement out-of-circuit... The neighbour, Diode D16 was fine, at least for my multimeter....
*sigh*

Anyway: Looks like a wire-wound resistor, RCL 3%, T2B-79, 33 Ohms, from the datecode below probably manufactured in 1981...

[alm]

Are you sure that resistor should be 33 Ohm, and not 0.33 Ohm? According your schematic it is in series with the output with a diode in parallel, so above 20 mA D16 would be forward-biased and conducting. What would be the point of this resistor in a 1A power supply? While if it were a 0.33 Ohm resistor, the voltage drop at 1A would be only 0.33 V, well below the forward voltage of the diode, and dissipation would be a reasonable 0.33W. Then it could be a current shunt used to monitor current.

Q1 looks very odd in the schematic, by the way. Could it be that you have gate and source switched?

[nightfire]

Good Point- Upon closer inspection, the resistor really exhibits some point before the value ;-)
It was late last night and after some frustration de-soldering experience  I did not look that closely...

Anyway, measured the value again with hooks and came to 0.345 Ohms, what my U1272A thinks, cables and hooks alone measure 0.025 Ohms- so that would be 0.320 Ohms for the resistor and therefore the value would be plausible and still within spec after 40 years of age.

For the Mosfet: Yes, as I head to measure the board from the downside and do some mirroring in the head, it certainly can be that I accidentally at some point mistook the pins for each other- thats the downside of trying for the first time to reverse engineer some schematics. But: I wanted to learn the hard way about things in that Power supply, so I already learned some valuable lessons.

[daisizhou]

At present, the current of this power supply 0-1A seems to be a little too small.
3-5A is more suitable for daily use, not practical unless it is very accurate

[nightfire]

At present, the current of this power supply 0-1A seems to be a little too small.
3-5A is more suitable for daily use, not practical unless it is very accurate

I also have another (switching) power supply that does the heavier lifting, if needed.
The reason to buy this Power supply were a bunch of arguments:
- linear power supply, so hopefully low ripple and noise at output- I want to do some experiments with circuits that measure the duration of camera flashes or PWM signals with photodiodes
- By doing this and try to repair (or at least perform maintenance) on such gear I will learn lots of stuff, from the experiences of other forum members they share, or my own faults
- The stuff I want to do with this PSU are usually below 1A, like providing some camera gear with energy for filming longer videos, doing experiments regarding charging batteries or powerbanks with USB connectors etc.

[nightfire]

Regarding the central mosfet Q1: I was able to get my hands on some NOS via ebay and just measured the old mosfet out of circuit against some other IRF130 that seems to be real NOS, no signs of it built in.
From https://pdf1.alldatasheet.com/datasheet-pdf/view/68112/IRF/IRF130.html it is a N-Channel mosfet, and from the pin assignments I would conclude that, lying on the back (TO3 packaging) when the pins point towards me, round case part to the bench, the pins are on the leftmost side, then the upper pin (1) is source and the lower pin (2) is gate, case is drain.
(What irritates me a bit is that in the datasheet it is also mentioned that IBGT use pin 1 as gate and 2 as source...)

Did with my Agilent U1272A and some probes with hooks the following tests:

Diode Test:
                old                    NOS counterpart
Gate -      Source +     open      open
Gate +      Source -     open              open
discharge
Drain -      Source +     0,5378 V   0,534V
Gate +      Source -     open              open
Drain +      Source -     0,004 V      0,02V
Drain -      Source +     0,004 V      0,02V


Questions: Do i mix something here? From what i read and assume, this should be the typical and expected behavour of a N-Channel mosfet?
What does the difference in voltage mean? That the NOS is in worse condition that the one that was in service for some time?

[james_s]

What difference in voltage? Am I missing something? It looks like they're identical to within a few thousandths of a volt, that's identical for all practical purposes. You can have much larger variations just due to normal manufacturing tolerances.