Need opinion RE HP/Agilent 6623A power supply

[member_xyz]

I am looking for a older and  less expensive alternative to newer Agilent E3631A triple output power supply.
Does anyone own or has experience with older HP 6623A triple output power supply to comment on ease/difficulty of use in operating the unit via front panel keypad, overall accuracy and reliability?

Thanks

[free_electron]

i have 6624's ( 4 channel version )

drawbacks :
- no 'dial' you need to type in the voltages on the keypad . this makes 'playing' hard ...
- heavy as hell
- noisy as hell ( fan noise )
- if you fry the supply kiss it goodbye. the hybrids are unobtainium ...

advantages ( if you can call em that )
- bulletproof design
- rock solid output
- overdesigned and built like a tank.

the 662x series is NOT a lab power supply. it is not intended for that use. it is intended to sit in an instrumentation rack under software control and perform automated testing. it will happily sit in a corner and do this for 30 years 24/7  without even blinking.

in terms of accuracy they are on par with the 3361.
in terms of reliability : they are more reliable than the 3361. the 3361 is a bench instrument , the 662x is a rack instrument different breed
in terms of manual control : the 662x stinks... need to go up 100 millivolt : start punching keys ... and if you miss the dot you may get settings you don;t like .... the rotary encoder is much more forgiving than the keypad ...

now , the 662x have serial port in the back... take a pic or whatever flavored cpu , hook up 2 rotary encoders ( one for voltage , one for current ) and simply send in the numbers ...
if i got a moment later today , i'll pop the cover off one of my 6624's. they are really built like a tank. the circuit boards are massive. it's amazing how much they crammmed in that chassis.. the transofrmer is like half the size of the chassis and custom built for hp. its got like 20 secondaries. ( each channel has 2 or 3 power windings so they can switch dynamically , and like 4 or 5 'assisting voltages' that are all isolated. there is a truckload of cables coming from that transformer.

[member_xyz]

Thanks free,

This is exactly the type of information I am looking for.

If you get a chance please pop the cover off one of your units and post some photos of the internals.

[casinada]

I Have an HP6627 Agilent Flavor and an HP6624. As Free_Electron mentioned they are built like tanks and they are designed to be mounted on racks. There are no relays inside like other designs and they don't have front Connectors so you have to bring the cables to the front somehow. They Have a big Fan in the back that sucks air from 4 heat sinks and make noise. On both of my units one of the outputs was bad. The power supplies wouldn't work unless you disable the self diagnostic (internal jumper or switch depending on the vintage). The good thing is that Agilent still makes them so there are parts available for them. In the end I had to replace the Custom 6805, a couple of fuses and a capacitor. Each Processor cost me $20 from Agilent. The Capacitor would have cost me $50 from Agilent but i Replaced it with a Nippon Chemicon for less than $10 (via china). The Little Fuse 7A Fuses cost about $1 a piece. The design must be sound as it hasn't changed since its creation. The Transformer is a beauty.
Pros as mentioned: Reliable, Built like a tank, stable, repairable, GPIB remote control. Keypad entry.
Cons: Rear panel connections, Big, Heavy, Noisy
It definitely has its place in industry and labs were the unit will be connected for long periods of time and/or controlled remotely.
I can post pictures if requested
Thank you

[Dago]

I have a couple 6632A power supplies and I made a blog post about them with some pictures and info if you are interested: http://www.dgkelectronics.com/hp-6632a-power-supplies/

[member_xyz]

Thanks guys for valuable and informative feedback.

Dago, you have an interesting USB/GPIB interface adapter. Have you considered including/adding rotary encoders  as free electron has suggested ? You may be able to expand it further by adding a display  and turning it into a stand alone remote control unit, as well as USB interface to a PC?
Since 6623 has voltage sensing terminals, it should not be an issue to mount/place it away from the bench and run longer output leads and remote terminals?

[free_electron]

6632 is a different animal from the 6623 ! the 6632 is a single channel the 6623 is a modular machine that has 3 channels. there is a 6622, 6623 and a 6624. the mix is in the chjannels (2 3 and 4 ) and you can have different 'options' in terms of max voltage and max current.

I'll snap some pictures tonight. weekend was too busy.

[member_xyz]

Thanks free,

Yes 6623A with triple output is what I am looking at.

[casinada]

Some Pictures....

[casinada]

The First picture is the GPIB controller board, the second one is the massive transformer, the third and four pictures are the power supply boards. Each of the capacitors cost about $50 from Agilent direct. The transformer alone probably cost $$$$$$

[free_electron]

you got the SMD version of the GPIB controlle.r mine are older and the controller is thru-hole. You also have the 'pcb' version meaning there are no hybrid modules. there is a pcb coming up along the heatsink going to discrete transistors. mine have hybrids. those are unobtainium when yu fry them ( although they are very hard to kill )

[casinada]

The Agilent 6627A has the New GPIB board. The 6624A has the old type GPIB board with Jumpers (first picture). The second picture is the Hybrid Chip, a $46 part from Agilent. The third picture is the Leaky $45 capacitor and its replacement. I really wanted to get the original part I tried to solder the replacement capacitor to the original base but I was unsuccessful. I guess I'm not a good alchemist.

[Dago]

Thanks guys for valuable and informative feedback.

Dago, you have an interesting USB/GPIB interface adapter. Have you considered including/adding rotary encoders  as free electron has suggested ? You may be able to expand it further by adding a display  and turning it into a stand alone remote control unit, as well as USB interface to a PC?

The board has a header with GPIO that can be used for rotary encoders (and this was the reason why I put the header in there in the first place ) and the board has an UART header that could be connected to an auxiliary board which has a display. Still need to work on the software.

[member_xyz]

Thanks guys for valuable and informative feedback.

Dago, you have an interesting USB/GPIB interface adapter. Have you considered including/adding rotary encoders  as free electron has suggested ? You may be able to expand it further by adding a display  and turning it into a stand alone remote control unit, as well as USB interface to a PC?

The board has a header with GPIO that can be used for rotary encoders (and this was the reason why I put the header in there in the first place ) and the board has an UART header that could be connected to an auxiliary board which has a display. Still need to work on the software.

Please keep us posted on the progress of GPIO board.
I am now looking for one possibly two 662XA power supplies with multiple outputs. Damn things are very heavy and shipping cost is usually higher than the units themselves.

[Ordinaryman1971]

I have a question about HP 6623A, I have one that reports error: No model number.
I can't toggle between channels, the first channel outputs something but it's about half of the set voltage.
I was trying to switch channels but it still reports the same and acts the same regardless of which channel I connect to the first connector.
I have another one "for parts" unit but the GPIB board is older, thru hole when the original pcb is smd.
When I've replaced the smd board with the thru hole one, the power supply's LCD screen stops working.
That's the extend of testing I've done so far.

[casinada]

It has been a while since I repaired them. If I remember correctly If one of the output boards fail the self test the power supply will fail to operate. You can bypass the self test by changing a jumper on the GPIB board. The rest of the outputs will work. I think you can also unplug the control connector from the GPIB board to the defective output board and the power supply will also start.
Good luck. Service manuals are available and probably describe the troubleshooting procedure better than me. 

[free_electron]

you can reset the model number through a GPIB command. if no model number is provided the firmware doesnt know the range of the installed modules.

looks like the eeprom (93c46) is fried.

the GPIB controllers are NOT interchangeable ! ( the smd version is NOT the same as the thru hole version ! ) especially on the display part.

[ScribblesOnNapkins]

casinada: Thanks for posting those pictures. I have an HP 6627A and I noticed something.

Your photos show U338 as a group of discrete components. In mine and all the copies of the manual it's a single large hybrid.

1. Can you please post more photos of just that sub assembly.
2. Does anyone have a copy of the manual where that section is described? In the copy I have from the following URL it's just described as a sub assembly.
http://literature.cdn.keysight.com/litweb/pdf/5989-6871EN.pdf?id=1268102

PS: Don't you love how they hinged the top cover and used a minimal number of screws to hold each board in.

[SV1EOR]

I have a Hp6627a on my bench in order to service one of the output boards with problem in U338 Hybrid as the trouble scouting charge tells me.
Has anyone found the circuit or a part list for it in Hp manuals?
I have the parts on the small hybrid boards, but I'm not sure ...

[rfclown]

I have a few 6627,6624. The pros:
1. Amazing for the eBay price. 4 channal clean linear HP GPIB supply for $300 bucks? (just looked on eBay and this was on top, including shipping)
2. Solid. I've used them for about 30 years (not the same one) with no failures/issues
Cons:
1. Loud
2. Display only shows one channel at a time (if I could change only one thing, it would be this)
3. connections come out the back. You can make something to help accomodate that (see picture).
4. heavy

I used these at Motorola. At one time I had a setup where I needed many supplies, and I used three of these (12 channels). Used LabVIEW/GPIB to sequence proper power up and down and ensure no desctructive fault conditions (might shut down one supply if another one current limited). When I left Moto after 22 years and started contracting, I bought equipment that I was familiar with, which included these DC supplies. I never had any issue with them. I couldn't believe how cheap I could get these things on eBay. I later got some single channel 35V and 60V units (really heavy).

I still have one in my bench stack, but today I find that it not the first DC supply that I turn on at my bench. Why?
1. Loud
Most of the time I'm doing something which my small no fan 18V/2A linear supply can handle. No fan, so I can hear the transformer hum. With the 6627A when I turn it on it sounds like a jet starting up.

[Bud]

These multichannel supplies have a big con that you cant enable/disable outputs at once together, you have to cycle through the channels individually. Sometimes this is a problem.