Author Topic: HP6024A PSU restoration project  (Read 4030 times)

0 Members and 1 Guest are viewing this topic.

Offline nctnicoTopic starter

  • Super Contributor
  • ***
  • Posts: 27483
  • Country: nl
    • NCT Developments
HP6024A PSU restoration project
« on: November 08, 2013, 12:31:33 am »
I was looking for a PSU on Ebay when I saw a couple of these for sale in 'for parts' condition:


These PSUs can deliver up to 60V and up to 10A within a limit of 200W. Because these are switching power supplies they weight next to nothing. Which is a major reason to buy them. After some haggling I got 2 for $100 each and some shipping costs and some import duties. Because I'm getting old these looked more recent to me then they really are :palm: Their year of manufacturing is 1987. OTOH my HP6002A from 1982 is still going strong.

Upon unpacking the distinct smell of cat-pee entered my nose telling me at least the capacitors are going to be a problem. Each had a red tag card. One said 'fluctuating output voltage' and the other said 'blown fuse'. Both cards where dated around 2000 so these PSUs have been sitting on a shelf for over 10 years. The first order of business is changing them over to 230V which requires setting some switches and removing a wire. Opening is easy. Just remove 4 screws and the cover slides open.



The PSU consists of a main board and a control board. On the main board there is (from left to right): input rectifiers, primary MOSFETs and capacitors, down programmer (to sink current for short periods) and low voltage transformer and a connector for the control board. The control circuitry and PWM logic is on the control board. The control board in close-up:


The one with the fluctuating output voltage still seemed to work OK. The problem in the one with a blown fuse was easely spotted:

The diode was a short and had cooked the PCB quite badly. Time to order some capacitors and diodes... After I changed the diodes and the fuse the power supply with the 'blown fuse' got back to life but the output didn't look clean and it made noises. Time to pull out the differential probe and take some measurements. Measuring the voltage acros the capacitors on the primary side resulted in this image:

Yes, that is a rectified 150V sine wave. The capacitors are completely dead indeed!

So time to change the capacitors. I decided to replace all the electrolytic capacitors on the main board. The new capacitors take about 4 times less volume than the old ones:


After the job was finished the main board looks like this:

In this picture I have circled the most magical part of this power supply. The power supply itself is a forward converter. The transformer next to the green circle is the transformer of the forward converter. A forward converter can be seen as a buck converter with a transformer pasted between the primary and secondary side so many formulas are similar. What makes the converter in this power supply special is that it can operate over a wide range of output voltages. The magic is the little inductor inside the green circle. The inductance of it varies with the current through it. With increasing current the inductance gets lower (a gradual core saturation) which is exactly what is needed to accomodate a wide output voltage range.

At this point I had two working power supplies but I wasn't satisfied. I wanted to make the fans temperature controlled to reduce the significant amount of noise these gritters produce. There where two options: replace them with DC fans which are easy to control or keep the old AC fans. Because I didn't want to risk overloading any of the DC rails I choose to keep the 115V AC fans. I first tried the circuit described here: http://www.circuitstoday.com/temperature-controlled-fan-regulator but that didn't work at all. In the best case I got the opposite function (fan going slower with increased temperature). This probably has to do with me using ceramic capacitors because the simulation showed the circuit should work (sort of). Anyway, I had enough of messing around with this circuit and remembered I wanted to use a solid state relay in a different project anyway. Lets throw a microcontroller at it  >:D I came up with this diagram:


I don't think I ever used a microcontroller with only 8 pins and I even have one pin to spare. Woohooo! I use a 100k NTC as a temperature sensor since I already got them for the other circuit. In order to synchronise the firing angles I used the AC from the rectifier (from which I also take the 12V) and feed that into the MSP430 with a zener and a protection resistor. I made a PCB out of it with mounting holes which match some empty mounting holes in the casing:

Some heatshrink tubing isolates the less pleasant to touch areas and hot-melt glue keeps the wires in place. Using a single sided PCB with the surface mounted components facing inwards makes sure any metal object poking into the casing can't cause a short or an otherwise dangerous situation.

On this picture you can also see the PU based sealant I used to fixate the big capacitors.

Time for testing... I mounted the temperature sensor on the same heatsink where HP put the thermal cutout switch. After some testing the fan still didn't kick in. :wtf: It turned out the MOSFETs at the primary side aren't getting hot at all. Only the heatsink with the rectifier diode and the downprogrammer where slightly warm. So I decided to stick the temperature sensor to the heatsink of the rectifier diode and down programmer. The software starts turning the fan on when the heatsink is warmer than 35deg. Celsius and is at full power when the heatsink reaches 50 deg. Celsius. These should be safe values. I tested the power supply at 10A for a while and the fan only runs very slowly at this load level. It seems the fan is way oversized!

Happy? Not quite yet. I spotted another problem in one of the power supplies:
\
That doesn't look original. Worse it is not even a 10-turn pot! I ordered a replacement from Ebay which may take some tinkering to fit (still need to do that). The original ones seem to be no longer available. Another problem was a missing foot so I ordered some HaPpy feet:


As the last step I added some handles for easy carrying:


All in all a neat project which resulted in two nice working power supplies. Another thing I might add is a switch on the front which sets the output to 0 by shorting the potmeter. This is handy to turn off the power supply without needing to use the mains switch.
« Last Edit: November 08, 2013, 12:52:25 am by nctnico »
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline megajocke

  • Regular Contributor
  • *
  • Posts: 191
  • Country: 00
Re: HP6024A PSU restoration project
« Reply #1 on: November 08, 2013, 04:07:33 pm »
Nice work. :)

I also have a HP6002A with the same output range, but I bet there is a significant difference in weight...

When adding temperature control for the fan, did you check the temperature of other components of the supply over different parts of the operating range? The diode loss is for example relatively low at high output voltage even at full power though other components are still highly stressed. Output capacitors experience a very high ripple current in this power supply because of the flyback* topology, and the switching transformer also has losses which are still high at a low output current/high voltage combination. But on the other hand, the supply is designed for full power operation even at 55 degree C ambient so might not need much fan cooling at normal room temperature...

I like the choice of a small microprocessor for fan control. :) I also used an 8-pin IC to give my HP6002A a temperature controlled fan, but I used a low-tech solution of a TL071 opamp and replaced the fan with a 24 V DC unit because the old one was starting to go bad.

* Yes, it is actually a flyback topology. This can be determined from Section 4-10 of the service manual where the operating principle is described. But it is also possible to deduce this from the schematic diagram as there is no freewheeling diode, which is required in the forward topology.

The inductor you encircled in green is more likely a saturable reactor that is used to soften the reverse-recovery of the secondary rectifier diode rather than a "swinging choke" that can be used to extend the range of continuous current mode operation in a forward converter. I found this document from Toshiba with a short description of how it works:
http://www.toshiba.com/taec/components/Generic/How_Amorphous_Magentic_Noise_Suppressors_Work.pdf
 

Offline robrenz

  • Super Contributor
  • ***
  • Posts: 3035
  • Country: us
  • Real Machinist, Wannabe EE
Re: HP6024A PSU restoration project
« Reply #2 on: November 08, 2013, 04:18:37 pm »
Nice job, better than new! :clap:

Offline nctnicoTopic starter

  • Super Contributor
  • ***
  • Posts: 27483
  • Country: nl
    • NCT Developments
Re: HP6024A PSU restoration project
« Reply #3 on: November 08, 2013, 05:19:11 pm »
Nice work. :)

I also have a HP6002A with the same output range, but I bet there is a significant difference in weight...

When adding temperature control for the fan, did you check the temperature of other components of the supply over different parts of the operating range? The diode loss is for example relatively low at high output voltage even at full power though other components are still highly stressed. Output capacitors experience a very high ripple current in this power supply because of the flyback* topology, and the switching transformer also has losses which are still high at a low output current/high voltage combination. But on the other hand, the supply is designed for full power operation even at 55 degree C ambient so might not need much fan cooling at normal room temperature...

* Yes, it is actually a flyback topology. This can be determined from Section 4-10 of the service manual where the operating principle is described. But it is also possible to deduce this from the schematic diagram as there is no freewheeling diode, which is required in the forward topology.
You are right. Its just very hard to believe they used a flyback topology for a 200W lab power supply which should have a low ripple output. I still have a feeling there is much more to this power supply than the service manual describes. I have not been able to find an accurate diagram of this particular power supply. Besides that the part numbers for the MOSFETs are also kept secret.

Anyway, I didn't skimp on the capacitors. I have used low ESR types. Actually I spend about €50 on the capacitors for each power supply.
Quote
I like the choice of a small microprocessor for fan control. :) I also used an 8-pin IC to give my HP6002A a temperature controlled fan, but I used a low-tech solution of a TL071 opamp and replaced the fan with a 24 V DC unit because the old one was starting to go bad.
I put a similar mod in my 6002A  ;)
There are small lies, big lies and then there is what is on the screen of your oscilloscope.
 

Offline nssouza

  • Contributor
  • Posts: 14
  • Country: br
Re: HP6024A PSU restoration project
« Reply #4 on: August 21, 2020, 06:35:04 pm »
Hi! I'm new in this forum, and I want to know if you have a page 7-6 of the service manual, that is missing in all PDF I downloaded from internet (all pdf files is downloaded from KEYSIGHT site, and it is missing... Exactly the most important diagram is missing! thank you
 

Offline siggi

  • Regular Contributor
  • *
  • Posts: 218
  • Country: ca
Re: HP6024A PSU restoration project
« Reply #5 on: August 28, 2020, 07:12:08 pm »
Hey there, it looks like Artek Manuals has a scan: http://artekmanuals.com/manuals/hp-manuals/. Not free, but not insulting price either.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf