Kikusui PAN35-20A 700W linear DC power supply

[coromonadalix]

I'm surprised to see how it was modified like this, to take such lenght ..... maybe in the end to avoid some miss use / temptation of it in the factory ... to whack a small screw driver in it  to play with the trimpots ?

Lucky you,  service manual exists

[JxR]

Lucky you,  service manual exists

True.  Although, I only considered this model because I already had the service manual.  I typically don't purchase any used test equipment unless I have a service manual for that model, just easier that way.

[JxR]

This item has been successfully repaired.  I put a summary of the components needed and the cost to repair on my original post.

Conclusion: It is a good deal for a 700W linear power supply.  The cost and difficulty to repair is low.  This is for the most part a completely analog design.  There is no micro controllers or memory to be found in the unit. The displays are driven by two 3.5 digit volt meter chips that are designed to drive 4-digit 7seg LEDs (https://www.microchip.com/wwwproducts/en/TC7107#additional-features).

The fan is very loud and not temperature controlled.  I may look into a replacement one day.  There is a hum on the transformer when operating, which I believe is normal.  I have load tested it up to my limit of 200W across the entire voltage range and I see no problems.

I have attached a measurement of the output ripple under load, which appears to match the specs in the datasheet.

I also updated pictures of post:
https://www.eevblog.com/forum/testgear/kikusui-pan35-20a-700w-linear-dc-power-supply/msg3060660/#msg3060660
To show you all the required changes.

Just to be clear, the repair is restoring: A1512A Control Board, to the default configuration.  Please check the link to the service manual if you there is anything unclear that I have failed to point out.  You should be specifically interested in service manual pages: 5-7, 5-17, and 5-19.

Please carefully check diagram 5.7.  VR101/VR102 are the MOVs you should consider replacing with components that are appropriate for your line input.

After repair and reassembly, completed pages 6-3 to 6-5 in the user manual (voltage adjustment).  Factory calibration and current calibration were both unnecessary for me personally.

[fanOfeeDIY]

The fan is very loud and not temperature controlled.  I may look into a replacement one day.  There is a hum on the transformer when operating, which I believe is normal.  I have load tested it up to my limit of 200W across the entire voltage range and I see no problems.

The fan looks to be the NMB 4715MS for 100V 50Hz model. Adding some cement resistor in series might help?

You have a great scope. How did you measure the ripple? I would like to do it on my Kenwood/Trio PSU.
Used regular 10x voltage probe for the ripple measurement?

[JxR]

You have a great scope. How did you measure the ripple? I would like to do it on my Kenwood/Trio PSU.
Used regular 10x voltage probe for the ripple measurement?

The measurement was done using an AC coupled differential low noise amplifier with 10x gain (LNA10).  Although it is not ideal for this type of measurement due a low pass filter cutoff frequency of 1MHz (instead of 20MHz).  Connections are through bnc with a banana adapters for the power supply's outputs.  The output of the LNA10 is just connected to the scope.  The scope's input is set to 0.1:1 to account for the 10x gain.

[fanOfeeDIY]

The measurement was done using an AC coupled differential low noise amplifier with 10x gain (LNA10).  Although it is not ideal for this type of measurement due a low pass filter cutoff frequency of 1MHz (instead of 20MHz).  Connections are through bnc with a banana adapters for the power supply's outputs.  The output of the LNA10 is just connected to the scope.  The scope's input is set to 0.1:1 to account for the 10x gain.

Thank you for the information. I did not know about the LNA10. A good preamplifier for the probe seems to have a good price.
I was considering to make a hand made probe similar to the link bellow, to save my money.
https://teledynelecroy.com/doc/build-your-own-power-rail-probe-appnote
I cannot exactly recall my memory, I think saw the similar probe mentioned on the eevblog video.

[JxR]

The measurement was done using an AC coupled differential low noise amplifier with 10x gain (LNA10).  Although it is not ideal for this type of measurement due a low pass filter cutoff frequency of 1MHz (instead of 20MHz).  Connections are through bnc with a banana adapters for the power supply's outputs.  The output of the LNA10 is just connected to the scope.  The scope's input is set to 0.1:1 to account for the 10x gain.

Thank you for the information. I did not know about the LNA10. A good preamplifier for the probe seems to have a good price.
I was considering to make a hand made probe similar to the link bellow, to save my money.
https://teledynelecroy.com/doc/build-your-own-power-rail-probe-appnote
I cannot exactly recall my memory, I think saw the similar probe mentioned on the eevblog video.

Maybe this one (about 32mins in)?


I have no doubt there are things I'm doing wrong in my probing technique as well (besides the 1MHz low pass filter).

The first output ripple pic I showed above was with the outputs floating.

This next pick is with the negative output tied to case ground (which is how the Kikusui performed the noise measurements).

The last pick is with again with the output floating but with 50ohm terminations between the coax and the LNA.

[duckduck]

The power supply has already arrived.  It will need a bit of cleaning and some sticker removal, although nothing too bad (from the outside atleast).  Overall it looks ok so far.

Hey JxR, is that a WHITE wire I see going into the LINE input? Might want to double-check that.

[JxR]

The power supply has already arrived.  It will need a bit of cleaning and some sticker removal, although nothing too bad (from the outside atleast).  Overall it looks ok so far.

Hey JxR, is that a WHITE wire I see going into the LINE input? Might want to double-check that.

Sorry, but I don't remember the color of which wire they used for L/N (and don't feel like taking the beast down from shelf to check).  Although, it doesn't really matter since it was hooked up correctly internally in relation to the power cord.

FYI: I was able to see behind the unit with a flash light, and the black wire is hooked to L, white to N...so standard setup.  I did a continuity check on the cord and the connections before I ever plugged it in.  No reason to trust the connections being correct just based on the color of the wires.

[SloMetalDoc]

Quite fortuitous that I stumble upon this forum and joined just to reply to this post to thank JxR for such an awesome writeup.

I received my PAN35-20A from (almost certainly) the same eBay seller and looks like the same mod's are required.  I ordered the parts you suggested and am reviewing the service manual now - Had a question though...my unit came with a 110V plug and the back of my unit has marking for 100V 50/60hz but has a yellow sticker over the voltage "205". 

I then went searching for the Varistor that you replaced, since I figured this would tell me (among other things) what voltage the unit was wired for.  Not sure where to find it however.

My BS is in mechanical engineering, so the fact that schematics are with symbols rather than little dealies that look like the actual component breaks my brain.   I do a bit of electronics work as part of my day-job - just finished up repairing a dozen machines in my laboratory that were damaged after a deluge of rain-water left me knee deep in water wondering if any of the 480 was still live (shudder).  Planning on using this power supply for random tinkering at home...most immediately some electroplating!

fun times
thanks again for the post!

[JxR]

Quite fortuitous that I stumble upon this forum and joined just to reply to this post to thank JxR for such an awesome writeup.

I received my PAN35-20A from (almost certainly) the same eBay seller and looks like the same mod's are required.  I ordered the parts you suggested and am reviewing the service manual now - Had a question though...my unit came with a 110V plug and the back of my unit has marking for 100V 50/60hz but has a yellow sticker over the voltage "205". 

I then went searching for the Varistor that you replaced, since I figured this would tell me (among other things) what voltage the unit was wired for.  Not sure where to find it however.

My BS is in mechanical engineering, so the fact that schematics are with symbols rather than little dealies that look like the actual component breaks my brain.   I do a bit of electronics work as part of my day-job - just finished up repairing a dozen machines in my laboratory that were damaged after a deluge of rain-water left me knee deep in water wondering if any of the 480 was still live (shudder).  Planning on using this power supply for random tinkering at home...most immediately some electroplating!

fun times
thanks again for the post!

The varistors are attached to the main transformer covered in white silicon adhesive.  They are right below the screw terminals where you need to rewire for 120V.  They have little ring terminal solder tabs that are screwed into the transformer taps.  Anyways just pull them off the transformer after you unscrew them. Keep the solder tabs to reuse.  Good luck on your repair!

[SloMetalDoc]

Awesome, my Digikey came in today, so I just got started with the mods.

One question, the trim pot you replaced... mine has a 520ohm resistor soldered between 2 pins, did you reinstall this resistor?

[SloMetalDoc]

Welp... nevermind.  RTFM.  You already have it marked for removal.

 

[SloMetalDoc]

You stated you replace C34 - but have an arrow pointed at where C43 would go (its empty, but has the green wire fed through from the back side.

So, Did you man replace C34?  Or install the 1uF cap at position C43?

[JxR]

You stated you replace C34 - but have an arrow pointed at where C43 would go (its empty, but has the green wire fed through from the back side.

So, Did you man replace C34?  Or install the 1uF cap at position C43?

Opps, sorry about that.  Where I circled is correct, and is what I should have said was C43.  Positive polarity attached to the bodge.

[SloMetalDoc]

Awesome, I got all that sorted!  That said, I cautiously turned the unit on after rewiring the transformer and let the smoke out from one of the varistors in short order.  I should mention that as-wired originally, it appeared to be for 90-110V - and I did power the unit for a short period of time and did power on.

The old varistors appear to be 130VAC Max.

Well... I typed all this up...and checked the varistor PN I ordered, MOV-14D121K-ND - which appears to be 75VAC max  .  Hmm....might install the old ones to see if I did any more damage.

Fun Times.

[JxR]

Awesome, I got all that sorted!  That said, I cautiously turned the unit on after rewiring the transformer and let the smoke out from one of the varistors in short order.  I should mention that as-wired originally, it appeared to be for 90-110V - and I did power the unit for a short period of time and did power on.

The old varistors appear to be 130VAC Max.

Well... I typed all this up...and checked the varistor PN I ordered, MOV-14D121K-ND - which appears to be 75VAC max  .  Hmm....might install the old ones to see if I did any more damage.

Fun Times.

Right, the old MOVs were meant for Japanese 100V power.  The part I linked in the first post has max VAC of: 150V and max varistor voltage of 264V.  If you order that part the MOV should be fine on 120V.

2x - MOVs appropriate for your line voltage (I used these) - https://www.digikey.com/product-detail/en/bourns-inc/MOV-14D241K/MOV-14D241K-ND/2799114

Please carefully look over page '5-7' on service manual and double check everything once you have the new MOVs.  You need to move the shorting bar from 100V A/B to 120V A/B.  Then move the load wire(L) to 120V B.  The new MOVs should be between 0VA-120VA and 0VB-120VB.

I think I already linked this document, but just in case this is what Kikusui sent me regarding the mains transformer(linked).  I think the service manual diagram is less confusing, but incase it helps...

[SloMetalDoc]

Yep, I was following that page (thanks again for posting the manual!).  I moved the shorting bar from 100VA/B to 120V A/B.

Question - since 120V A/B are shorted with the bar, does it actually matter if the Load wire is connected to 120 VB?

Thanks again!

[JxR]

Yep, I was following that page (thanks again for posting the manual!).  I moved the shorting bar from 100VA/B to 120V A/B.

Question - since 120V A/B are shorted with the bar, does it actually matter if the Load wire is connected to 120 VB?

Thanks again!

I don't see why it would matter personally due to the shorting bar.  It is just easier to say "follow the schematics" so there is no confusion.

[SloMetalDoc]

I'm good with that logic.

I reinstalled the original varistors (PN Z10D201 - 130V) and replaced the fuse - I now have what appears to be a functional power supply!  I probably should scare up a dummy load and test it.  Or maybe just try some electroplating and see how it goes.   

I'm actually surprised nobody has bought that guy on eBay out and repaired them all.  Not terribly difficult - its way nicer than what I need...but what the hell.

Thanks again!!!

[JxR]

I'm good with that logic.

I reinstalled the original varistors (PN Z10D201 - 130V) and replaced the fuse - I now have what appears to be a functional power supply!  I probably should scare up a dummy load and test it.  Or maybe just try some electroplating and see how it goes.   

I'm actually surprised nobody has bought that guy on eBay out and repaired them all.  Not terribly difficult - its way nicer than what I need...but what the hell.

Thanks again!!!

Nice, glad you got it working!  I broke the legs off one of the original MOVs so I never got to use them.  I already had another Digi-Key order in progress, so it wasn't a big deal to just add $0.40 in parts.  The ones I linked were pretty much the exact spec of the discontinued part on the schematic for 120V.  The original MOVs are 'probably fine' if your line voltage is nice and clean, but you are running them closer to their spec'd maximums.

[SloMetalDoc]

Yep, going to get replacements on order today.

Fired the unit up today to check basic functionality and calibration and got some interesting operation. Unit turned on for ~5 minutes and Current limit is set to 0.2-0.3A.  Fluke 287DMM connected to front terminals (Shorting bar and Pos).

Running voltage up in 1-2V increments - no problem till I hit ~15V, when the voltage starts dropping - with the CC Light Flickering to CV.  I think this indicates there may be a short somewhere internal to the power supply.  Any thoughts appreciated!

[JxR]

Yep, going to get replacements on order today.

Fired the unit up today to check basic functionality and calibration and got some interesting operation. Unit turned on for ~5 minutes and Current limit is set to 0.2-0.3A.  Fluke 287DMM connected to front terminals (Shorting bar and Pos).

Running voltage up in 1-2V increments - no problem till I hit ~15V, when the voltage starts dropping - with the CC Light Flickering to CV.  I think this indicates there may be a short somewhere internal to the power supply.  Any thoughts appreciated!

Unfortunately I cannot reproduce any such behavior on my own power supply.  With current limited to 0.1A (and even a bit lower where the limit cannot be displayed) I can go from 0-35V and the measured current stays at 0.0 on the display the entire time.  Even with the current limit just barely nudged to the right from the 0 position, the most I can get is a couple quick flashes of the CC mode LED through the entire voltage range.

It sounds like there is some kind of leakage current somewhere along the output path.  Can you actually see the measured current rise on the display as you increase the voltage?

[SloMetalDoc]

No, I think as soon as it detects over .1A, it switches into CC Mode and drops the voltage. 

I just flipped the unit on to check that, and I was able to get all the way up to 32V, then after a second or two, the voltage started fluctuating.  I'm going to try leaving it on for a little bit, then checking again - Wonder if something is heating up and developing a short...

or some of that dusty floof in there is shorting the output.   

[JxR]

No, I think as soon as it detects over .1A, it switches into CC Mode and drops the voltage. 

I just flipped the unit on to check that, and I was able to get all the way up to 32V, then after a second or two, the voltage started fluctuating.  I'm going to try leaving it on for a little bit, then checking again - Wonder if something is heating up and developing a short...

or some of that dusty floof in there is shorting the output.   

I know I spent about 1.5 cans of compressed air (and some iso) cleaning it out, the thing was filthy (although all my boards looked fine after cleanup).  It is hard to say how hard the power supplies were driven in their original location.  Although based on how dirty they were, they have probably seen quite a bit of use in what was probably a hot and dirty factory type location.

I would certainly clean the unit up if you haven't and look at all the electrolytic caps for any signs of leakage. With as dirty as mine was it was very hard to tell if any of the caps were damaged/leaking.  Personally I would start there.  Clean it up, inspect the caps for signs of bulging/leaking.  Make sure there isn't any corrosion on the board, check all the diodes with your DMM, etc.

PS - While I think it would be unlikely the problem is on the control board, feel free to post some picks if you want me to take a glance at it.

Also, it is a completely analog power supply.  There is no firmware you have to worry about, so what ever is wrong is likely fixable.  But, the unit also has a 30 day return according to the seller.  So...that is something you might want to consider before you spend more money on repairs.  I would feel a bit odd trying to return something I modified, but then again it wouldn't be that hard to put it back into the same state you received it in.

*Pages 5-5 and 5-6 also have some test points to measure in the service manual.