Agilent 66332A Dynamic DC Source 20V/5A Repair and mods

[Berni]

So i got my hands on a Agilent 66332A. While it looks like a run of the mill rack mount power supply it is actually a SMU. This means that apart from being a power supply it also has reasonably accurate readout (Not a terribly impressive 4 1/2 digits here) and has the capability to sink current to act as a electronic load. However it is limited to positive voltages only. It is possible to get true 4 quadrant operation if you have option 760, this adds a relay to the output that allows it to swap its output terminals around so its not the most elegant way or doing it. But these things can be found on ebay for under 200 bucks!

Mine however is a 120V unit so it had to be a case of "Don't turn it on, take it apaaaart" in order to swap some internal connections to turn it in to a 230V unit. And since i am taking the lid off i might as well take some photos.








The transformer has quite a complicated rats nest of wires on it with no indication of what is what so you do need to consult the service manual on how to wire it up for 230V or vice versa.

[Berni]

Now you may have noticed the word "Repair" in the title. This is because the unit was being sold as not working and indeed when i powered it up all you get is a transformer buzz. After a bit of looking around i quickly noticed that a tantalum cap was looking rather toasty on the digital board. As you can see its pretty much black and since tantalum caps tend to blow in to a short circuit this would be a pretty good reason why the unit doesn't want to turn on.






After a quick pray to the electronics gods i desoldered the capacitor in hopes that that it didn't blow due to the supply rail being too high. Put everything back together, powered it on and HEYO it boots up fine with no errors. Checked its output with a multimeter and it was working perfectly! That was a pretty straightforward fix.






While it does work without any capacitor there its still better to replace it. None of the same footprint tantalum caps in my parts collection had a 35V voltage rating so i instead put in a 10uF 63V electrolytic cap with the leads bent for SMD mounting.






I also found another issue with the unit in that the knob on the front was really lose and wiggly, but it seamed to still work. For that i pulled off the front panel and once the board was out the problem was pretty obvious on visual inspection. Looks like someone gave the knob a really good knock in order to rip the top part of the encoder off. This looks worse than it actually is. I was able to bend the tabs back in to place and i also added a drop of super glue to each one to hopefully hold it a bit better. After that the knob felt as good as new and worked perfectly.

[Berni]

At this point the unit was in fully working condition, but before i close the lid there is one more thing on the list. I want to use it as a bench unit so front pannel terminals are a very useful thing. Luckily there is a option for this so there already are prepared cutouts in the front panel for the connectors. All we have to do it punch a hole in the front panel sticker and mount a pair of binding posts on there. For wiring them in there is a connector at the back where i ran a nice thick 2,5mm2 wire.







Once i had the unit all back together i gave it a calibration. It was very good on voltage out of the box but the current was off by a little bit. Its a pretty easy process where you just punch in some numbers in to the front panel. Here is the final result. The only problem with this unit is that setting currents under 2mA becomes inaccurate, but this is normal.

[Macbeth]

It's essentially the same as my HP 6632B, but with some extra commands in the firmware to turn it from a SYSTEM DC POWER SUPPLY to a DYNAMIC MEASUREMENT DC SOURCE. 

[TheSteve]

I have a 6632B as well, from the pics they do look the same, likely just a firmware change. I also added front mounted binding posts.

I got mine for 25 bucks as it displayed "bad int" when powered up. Of course it was wired for 220 VAC and the seller tested it on 110 VAC. I re-jumpered the transformer and it has been great ever since.

[BFX]

I also use wires for sensing and different terminals configuration
I have also this wiggly knob on my second unit, waiting for repair in near future Thanks for hint 

[gslick]

I have a 6632B as well, from the pics they do look the same, likely just a firmware change. I also added front mounted binding posts.

My Agilent 66332A serial number MY4300xxxx has firmware version A.01.03 in an M27C202-70K6 OTP EPROM (2-Mbit, 128Kx16) with the part number 5080-2513 etched on it.

Attached is a binary dump of my 66332A firmware version A.01.03 if you are curious enough to try loading it on your 6632B. I also have a couple of 6632B and haven't gotten around to trying that myself yet.

[TheSteve]

Awesome, I'll likely try it at some point.

btw, my 6632B info: HEWLETT-PACKARD,6632B,0,A.01.04

Comparing the boards they certainly look to be the same with the same components installed. I just took a picture of my 6632B mainboard specific details.

[KJDS]

I've repairs dozens of wiggly knobs on 6632B supplies.

It takes very little force to do displace the top part of the encoder, it's one of the few design issues with them.

[Berni]

That is interesting. I would have expected there to be extra circuitry in there to provide the current sinking functionality. Or do 6632B supplies also sink current if you ask them to?

I forgot to take a photo from that corner of the board and the unit is already put together and on a shelf.


Oh and on a side note i just got a HP 6624A 4 channel system PSU. But due to my stupidity i bought one that fails self test. Turns out channel 1 is dead. It looks like its a complex internal circuitry fault that Dave would love to do a repair video on. I haven't got any further in my limited time due to how unnecessarily complicated the guts of the thing are, but if there is interest i could document the repair process a bit.

[TheSteve]

Yes, the 6632B does sink current very nicely. You can even set the max current unlike some other Agilent supplies.

[AndyC_772]

Indeed, they're really excellent pieces of kit if you're in the habit of designing products that use rechargeable batteries. Try to 'charge' a standard bench PSU and you'll be in a world of hurt. On a 663xB, you just get to see your charging current and terminal voltage displayed on screen

[Berni]

Yes, the 6632B does sink current very nicely. You can even set the max current unlike some other Agilent supplies.

I looked trough the manuals a bit and they indeed say they can sink. Only difference i can find to the 66332A is that this so called Dynamic source can play back 2048 points at up to 64KHz sample rate. That's easily something that might as well just be a software thing.

I am surprised they don't put something special in the name to tell you its a supply with sink capability. As far as i know such functionality is quite unusual to see in a regular PSU.

Oh and i also noticed that the heatsink gets very warm when you leave the output just turned on but without any load. So they must be using a pretty high bias AB output stage on it.

[nctnico]

Indeed, they're really excellent pieces of kit if you're in the habit of designing products that use rechargeable batteries. Try to 'charge' a standard bench PSU and you'll be in a world of hurt. On a 663xB, you just get to see your charging current and terminal voltage displayed on screen

Indeed. I have the 66332A little brother 66311B and it occured to me it is perfect to simulate a battery in an upcoming project.

[AndyC_772]

I am surprised they don't put something special in the name to tell you its a supply with sink capability. As far as i know such functionality is quite unusual to see in a regular PSU.

They put something special in the price!

[Berni]

They put something special in the price!

Hence why i tend to get things that say HP or Agilent on it from Ebay.

[AndyC_772]

Sure, me too! 3x6632B, 1x6633B and 1x6634B at last count...

[PTR_1275]

I just got my 66332a delivered and took it out of the box. It wasn't packaged the best and the knob has taken a bit of a hit. I took the unit apart to change the transformer taps for 240VAC, took the front panel apart and forced the knob clips back together. It is holding, but not the best and I would prefer to replace it.

Looking through the service manual it is saying the front PCA has no user replaceable parts on it. Do any of you have a part number for the rotary encoders you have used as replacements?

The power supply also seems to be a bit on the noisy side (hum wise, not electrical noise wise) and it seems to be coming from the big filter caps right beside the transformer. I will investigate a bit more and see if anything needs replacing. It has date codes from late 99 / early 00, so it is 16 years old.

Out of curiosity, has anyone flawed a 6632B power supply with the 66332A firmware and seen if it works?

[Berni]

Mine has a lot of audible transformer hum too and yes the fan is loud.

I have not been able to fix any of that. The hum is likely coming from the transformer being such a flat design, giving it a lot of leaking magnetic field around it and that then vibrates the top and bottom case covers. As for the fan i found that it can't be run slower than it already does. When the output is enabled with no load at all the heatsink still gets pretty hot due to the bias current of the output stage.

As for the encoder i have no idea what the part number is, but from a glance it does look like a standard part. Try measuring the distance between the pins and the height of the shaft and comparing it to the similar looking parts on Digikey.

[Micke]

I bought encoders at good price from Ebay, a seller in Germany.
This encoder is common in a lot of other HP/Agilent gear, so getting 10pcs is nice 
http://www.ebay.com/itm/10x-Inkremental-Encoder-12fach-ALPS-EC16-ohne-Rastung-/302025367910?hash=item46521d6166:g:3KIAAOSwrx5UYhWb

[artag]

Attached is a binary dump of my 66332A firmware version A.01.03 if you are curious enough to try loading it on your 6632B. I also have a couple of 6632B and haven't gotten around to trying that myself yet.

That firmware contains this list of power supplies. Presumably it supports them all, though I don't know how it decides which one to be. Maybe a factory calibration option, or a link.

HP6611C HP66312J9 HP66312J8 HP66311J1 HP66311B HP66308A
HP66332A HP66312A HP66311A

I have only the 6632B. It has EPROM AM27C2048 5080-2573 A.01.04

There are 2 other programmable devices that might be customised though - a xilinx config rom 1736e marked 5080-2527 and a PAL 126V8 marked 5080-2528. Are the same parts fitted to the 66332A ?

[AndyC_772]

It's all in the service manual:

The dc power supply's GPIB address and model number as well as other constants which are required to program
and calibrate the supply are stored in a EEPROM on the A2 Interface board. The Interface board also contains
references and other components that will affect the alignment of the supply. If the Interface board is replaced, the
supply must be reinitialized and calibrated. To initialize the power supply:
a. Enable the Calibration mode
b. Simultaneously depress the "0" and "9” keys.
c. Using the Up/Down arrows select the appropriate model number
d. Press "Enter"
The dc power supply will go through the turn-on self test sequence. It is now re-initialized and must be calibrated.
See Appendix A of the User’s Guide for the calibration procedure.

[Berni]

It's all in the service manual:

The dc power supply's GPIB address and model number as well as other constants which are required to program
and calibrate the supply are stored in a EEPROM on the A2 Interface board. The Interface board also contains
references and other components that will affect the alignment of the supply. If the Interface board is replaced, the
supply must be reinitialized and calibrated. To initialize the power supply:
a. Enable the Calibration mode
b. Simultaneously depress the "0" and "9” keys.
c. Using the Up/Down arrows select the appropriate model number
d. Press "Enter"
The dc power supply will go through the turn-on self test sequence. It is now re-initialized and must be calibrated.
See Appendix A of the User’s Guide for the calibration procedure.

Oh neat, so the service manual tells you how to hack it in to a higher up model.

[BFX]

Front terminals changed to much nicer ones 
http://www.tme.eu/en/katalog/#search=BS-244MN&cleanParameters=1

Old one here:)
https://www.eevblog.com/forum/testgear/agilent-66332a-dynamic-dc-source-20v5a-repair-and-mods/msg951238/#msg951238

[Berni]

Front terminals changed to much nicer ones 
http://www.tme.eu/en/katalog/#search=BS-244MN&cleanParameters=1

Old one here:)
https://www.eevblog.com/forum/testgear/agilent-66332a-dynamic-dc-source-20v5a-repair-and-mods/msg951238/#msg951238

Those do look rather nice, very close to the stock ones that originally come on HP gear.

I did want to add sense terminals to the front panel of mine....i should buy some.