Hewlett Packard / Agilent 6632B Upgrade

HP / Agilent 6632B front pan­el with upgrad­ed bind­ing posts

Upgrading a Hewlett Packard / Agilent 6632B System DC Power Supply with front pan­el bind­ing posts and cor­rect­ing an orig­i­nal design error.

My local mak­er­space recent­ly had a dona­tion of a HP / Agilent 6632A and a 6633A System DC Power Supply. I was not famil­iar with this series of sup­plies, so when I returned home I start­ed doing some research on them. What I found was very inter­est­ing.
The 663X series are “two quad­rant” pow­er sup­plies, mean­ing that they can source cur­rent and also sink cur­rent at one volt­age polar­i­ty.
After a quick search on Ebay, I was now the proud own­er of a sup­pos­ed­ly non-work­ing 6632B pow­er sup­ply, which is the new­er ver­sion of the 6632A.
The auc­tion descrip­tion stat­ed “When attempt­ing to con­nect a device to the pow­er sup­ply, the pow­er sup­ply boots into safe mode. I was unable to get a pow­er out­put from the device.
The HP / Agilent 6632B has a rat­ing of 20 volts at 5 amps for a 100 watt out­put, and also has out­put cur­rent mea­sure­ment capa­bil­i­ty in the microam­pere range when mea­sur­ing less than +/- 20 mA. It has a cur­rent sink­ing capa­bil­i­ty up to the max­i­mum rat­ed out­put cur­rent of 5 amps.

HP / Agilent 6632B back pan­el with out­put connections

The unit arrived in great shape, with only a few scuffs and scratch­es from nor­mal use. After a quick cleanup and some basic safe­ty checks, includ­ing the fuse, trans­former, leaky capac­i­tors, and visu­al dam­age to com­po­nents I was about ready to pow­er it up.
In prepa­ra­tion for attach­ing a mul­ti­me­ter to the out­put ter­mi­nals on the back of the unit, I noticed that the screws were very loose includ­ing the ones for the sense ter­mi­nal jumpers.
After tight­en­ing the screws, and con­nect­ing the mul­ti­me­ter the unit pow­ered up with no errors. After try­ing var­i­ous loads across the oper­a­tional range of the sup­ply, the unit oper­at­ed per­fect­ly with almost no issues. The only prob­lem that I could find was a loose volt­age adjust­ment shaft on the front pan­el rotary encoder.
My guess is that the loose screws on the sense con­nec­tions, were the source of the Ebay sell­ers issues with the unit.

HP / Agilent 6632B front pan­el inside before modification

To access the rotary encoder the front pan­el con­trol board must be removed. Gaining access is not very hard, after remov­ing pow­er from the unit, first remove the knob and then the T10 torx screw on each side of the front pan­el. These are hid­den under a thin plas­tic cov­er that is attached with adhe­sive. The cov­er can then be removed by detach­ing the rib­bon cable and four pow­er con­nec­tors on the switch. Document the cable posi­tions before remov­ing. The front pan­el cir­cuit board is attached with one T10 torx screw, 10 slid­ing plas­tic fin­gers, and one tab near the screw. Gently bend the cir­cuit board fin­ger to clear the tab, and slide the cir­cuit board towards the switch. Once the plas­tic fin­gers are cleared the board should eas­i­ly lift straight up.

HP / Agilent 6632B front pan­el loose rotary encoder

On my unit the four small met­al fin­gers that hold the rotary encoder togeth­er were bent out caus­ing the loose shaft. A quick re-crimp­ing with nee­dle nose pli­ers solved this prob­lem for me.

HP / Agilent 6632B front pan­el option cutouts

You may have noticed the strange look­ing cutouts on the back side of the front pan­el. This was for a fac­to­ry fit­ted option #020 front pan­el ter­mi­nal. This option pro­vid­ed a pair of bind­ing posts for the pos­i­tive and neg­a­tive out­puts in par­al­lel with the back pan­el con­nec­tors.
Unfortunately this was just an exten­sion cable from the back of the unit with no volt­age sense con­nec­tion at the front pan­el ter­mi­nals. Under load this can cause a volt­age drop at the front ter­mi­nals, along with the inabil­i­ty of using remote sens­ing with­out con­nect­ing to the rear pan­el sense terminals.

I want­ed to add front pan­el out­put and sense ter­mi­nals, along with a chas­sis ground ter­mi­nal on my unit.
In a pro­duc­tion envi­ron­ment which these units were designed for this would not be an issue, as these sup­plies would most like­ly be rack mount­ed and attached to a sys­tem wiring har­ness for use.

3D print­ed ter­mi­nal adap­tor for the 663X series front panel

I was able to find a 3D print­ed adapter on Thingiverse.com for HP 66xxA 65xxA series pow­er sup­plies by NearFarMedia. This insert also works for the 663xB pow­er sup­plies. They show how a HP 6644A sup­ply was upgrad­ed to uti­lize front pan­el bind­ing posts.
This was very sim­i­lar to what I want­ed to accom­plish on my sup­ply except for the front pan­el but­ton which I left out.

Test of 3D print­ed insert with ter­mi­nals on 0.75″ stan­dard centers

I ordered the rec­om­mend­ed Sato Parts #T‑45 bind­ing posts on Ebay, and was hap­py with the qual­i­ty of the bind­ing posts, even though they were a bit pri­cy $$.
I was able to find pic­tures and schemat­ics for the orig­i­nal option 020 cir­cuit board, and craft­ed my own enhanced ver­sion of it.

HP / Agilent 663xB series front ter­mi­nal inter­face board top view
HP / Agilent 663xB series front ter­mi­nal inter­face board angled view

The inter­face board con­sists of a high cur­rent fer­rite core and a pair of film capac­i­tors for each pos­i­tive and neg­a­tive terminal.

HP / Agilent 663xB series front ter­mi­nal inter­face schematic
HP / Agilent 6632B ter­mi­nal inter­face board installed

On the 663xB series of pow­er sup­plied there is anoth­er option for a relay board, that was not pop­u­lat­ed on my unit. The con­nec­tor foot­print for this sits just in front of the option 020 con­nec­tor on the A1 board. Unlike the 4 pin option 020 con­nec­tor, this con­nec­tor foot­print J320 also includes the out­put sense con­nec­tion and the neg­a­tive trace shield connection.

HP / Agilent 6632B A1 board J320 relay board con­nec­tor footprint
HP / Agilent 6632B A1 board with new J320 con­nec­tor mod­i­fied and installed

I decid­ed to dis­able my back pan­el con­nec­tor by remov­ing 4 jumpers R456 thru R459, this also removes the rear ter­mi­nal capac­i­tors from the J320 con­nec­tor. There were a cou­ple of noto­ri­ous RIFA capac­i­tors that I replaced with Y2 safe­ty capac­i­tors, even though they are not used in my con­fig­u­ra­tion.
I used a Molex #26–61-4090 ver­ti­cal 9 pin head­er which required some mod­i­fi­ca­tion to clear an exist­ing resis­tor and jumper. I also removed jumpers R451 and R454, which con­nect the sense cir­cuit to the rear pan­el ter­mi­nal. No need for the extra noise on the sense lines if I won’t be using that ter­mi­nal. The cable side con­nec­tor hous­ing used was a Molex #09–93-0900 along with #08–65-0115 crimp con­nec­tors for 18 AWG wire.

HP / Agilent 6632B Noctua 60mm fan test and front ter­mi­nal wiring

A pair of 18 AWG shield­ed three con­duc­tor cables were used from J320 to the front pan­el ter­mi­nal inter­face board. I con­nect­ed the red and black wires of cable #1 (red) to pins 1 & 2 of J320 (+OUT_P) and the white wire to pin 6 (+SENSE), no con­nec­tion was made to the shield. On cable #2 (black) the red and black wires were con­nect­ed to J320 pins 3 & 4 (-OUT_P), with the white wire to pin 7 (-SENSE), and the shield wire con­nect­ed to pin 9 on J320 with no con­nec­tion on the ter­mi­nal inter­face board end of the cable shield. I col­or cod­ed the two cable ends with red and black heat shrink for easy iden­ti­fi­ca­tion. By con­nect­ing two 18 AWG con­duc­tors in par­al­lel, I end­ed up with the equiv­a­lent of a 15 AWG con­duc­tor for the pos­i­tive and neg­a­tive con­nec­tions, which is more than enough for the 5 amp. rat­ing of this supply. 

HP / Agilent 6632B fin­ished front pan­el bind­ing post connections

Before fin­ish­ing up the project there were sev­er­al small things to com­plete. I need­ed to fin­ish replac­ing the remain­ing prob­lem­at­ic RIFA safe­ty capac­i­tors locat­ed on the pow­er input RFI board. These capac­i­tors are always con­nect­ed across pow­er and ground when the pow­er cord is plugged-in, even if the pow­er switch is off. The orig­i­nal RIFA capac­i­tors were a met­al­ized paper type that were sus­cep­ti­ble to humid­i­ty, which caused crack­ing of the insu­la­tion over time. After being stored for many years these capac­i­tors are noto­ri­ous for explod­ing or catch­ing fire short­ly after hav­ing pow­er applied.

HP / Agilent 6632B A1 board and RIFA safe­ty capac­i­tors from pow­er RFI board

My RIFA capac­i­tors looked to be in excel­lent shape, but it’s only a mat­ter of time before they fail, so they must go.

Example of bad RIFA safe­ty capac­i­tors from my Tektronix 2430A

The oth­er issue that need­ed to be addressed, was an orig­i­nal design defect on the A1 board. When the engi­neer drew the orig­i­nal schemat­ics for the 663xB series, they left out a junc­tion dot where two fil­ter capac­i­tors on the +/- 15 volt reg­u­la­tor inputs crossed ground.
This error was ini­tial­ly dis­cov­ered by Dr. Frank on the EEVblog.

HP / Agilent 6632B A1 board miss­ing ground junc­tion. Connect to U304 pin 2 Gnd
HP / Agilent 6632B A1 board jumper from capac­i­tors C496 & C497 to ground

The fix is a sim­ple jumper from where C496 and C497 con­nect to a near­by ground on U304 (+ 15 volt reg­u­la­tor) pin 2.
The orig­i­nal cir­cuit works OK, until one of the tan­ta­lum capac­i­tors starts to become leaky, plac­ing a high per­cent­age of the total +/- unreg­u­lat­ed sup­ply busses (52 volts) across one capac­i­tor, exceed­ing it’s rat­ed volt­age of 35 volts. Typically tan­ta­lum capac­i­tors in this type of cir­cuit fail with a break­down short cir­cuit or explo­sion when their volt­age rat­ing is exceed­ed.

On a side note, I did try replac­ing the noisy high-air­flow heatsink fan with a much qui­eter Noctura 60mm fan. Unfortunately the Noctura did­n’t pro­duce enough air­flow to cool the heatsink even at full speed. It was worth a try though.

Below are some extra images of inter­nals and exter­nals of the 6632A:
(click the images to enlarge)

EagleCAD V7.7 Terminal inter­face schemat­ic and board files Rev. 1.1

Leave a Reply

Your email address will not be published. Required fields are marked *