Tweet
While working on the SWR SM900 last week, and setting up the remaining power amp modules from a boneyard SM400 to work in the SM900 chassis, I was powering up that amp module from an external bi-polar supply. The HP 6227B I often use only goes up to +/- 25VDC in tracking bi-polar mode. I brought out a Systron-Donner DL-40-1 dual supply that would go out to +/- 40V in bi-polar mode. While that was just enough to get past the point where the SWR Amp module’s output still shows significant DC offset (which usually doesn’t get back down below 200mV until the AC mains is up around 90-95VAC), I decided to ‘stack’ the outputs of my HP supply to that of the SD supply. Disconnected the jumper on the HP supply so the Master & Slave supplies would be independent. I’m pretty sure I had switched it to Independent mode. When I connected the Master supply on top of the SD Plus supply, I now had up to 65VDC. Connecting the Slave supply to the SD Minus supply, I got a spark, and didn’t complete the connection. The SD supply was still working fine, but now, my Slave supply on the HP didn’t work at all. Sigh……….
I did get the task completed with using the SD supply, then installed the module into the SM900 chassis and got that project completed, less the confusion with the fan powering.
So, Friday I began pulling this HP 6227B power supply apart. My original service manual isn’t here….it’s in my storage locker library, so I was able to download a copy of it of the internet, and printed out pages to paste up a fold-out of the schematic, then the troubleshooting guide, other useful pages to help the troubleshooting procedure. I’ve never had this supply apart, so after finding I’d have to remove the control boards that are mated to the heat sink assembly & rear panel to which they’re attached, I’d have to release two pairs of screws at the rear of the side frames. #6-32 Pozi-Drive Flat Head screws, with cone-shaped lock washers, and have been installed since 1977. My initial attempts were fruitless….those weren’t going to budge without my stripping out the heads. So, I seeped machine oil into the screw cavity, brought out a small pin punch and hammer, and hammered on the four screws, followed by holding a Pozi-Drive #1 bit into the head and hammered on it to hopefully reform the head pattern some, as the metal was already getting buggered. With all my weight pressed down on the screws, first driving them tighter, then trying again to loosen them, they finally gave way and came undone. What a relief. I wasn’t looking forward to having to grind those screw heads off.
When I first opened up all of the panels on this supply, I found on the bottom side two fuses, one of which was open. It couldn’t be that simple, could it? Replaced it with a fresh 3A Fast 3AG fuse, and upon powering up, the meter slammed hard off scale, so I had the pass xstrs source coming thru, with no control of them…perhaps shorted pass xstrs.
HP 6227B Sch-1.pdf
HP 6227B Sch-2.pdf
HP 6227B Sch-3.pdf
HP 6227B Sch-4.pdf
Then, I found how they build this sucker. The two side-panel PCB’s each contain the bulk of the regulator circuits, with the back end containing the two sets of pass xstrs. It took some time to navigate the schematic, and then the PCB that had failed. HP had a 12-pin TO-5 IC that contained two NPN-differential amps with an NPN tied to the common emitters of the diff pair. I hoped like heck I didn’t kill one of those. It measured just like the Master channel, which was still working, so quite nice to have one working board for comparison to check with. I finally found what appeared to be the fault…a driver xstr Q6 NPN, ahead of a TO-66 Driver stage that handled the 2nd half of the pass xstrs. The part looked like old GE plastic TO-92 size parts, while the base is on the right side, collector in the middle and emitter on the left side, like most European pinouts. HP had their own part number for it, and looked to be a screened part from the manufacture’s generic parts. The Pass Xstrs were all fine, as were the rest of the semi’s and resistors.
After sizing up the power supply range it and it’s partners in crime were running off, and looking to see what I had in my stores and data sheets, the two closest parts being considered were an MPS A-06 and a 2N4401. The 4401 has better current linearity out past 100mA, though only a rated 40V Vceo, while that of the MPS A-06 drops off quickly past 100mA, while it carried an 80V Vceo. After setting up the curve tracer, I found the 2N4401 was more than sufficient for the task, and, after checking the layout to find which way the base lead was bent, it was traditional TO-5 pattern, and soldered it in.
Then, re-assembled the supply, secured the assembly back into place, and cautiously went to power it up, first with the fuse removed. I had supply voltage following the adjustment control, so put the fuse in, and now had a working supply again. Set it back up for Tracking, and the repaired Slave is again tracking the Master as it had been doing for as long as I’ve owned it. The large computer grade electrolytics have date codes of 7714 on them, so this was probably built sometime in the late 1970’s.
I got to hand it to HP, having such a wonder build factor in their gear....certainly of this era, and really well-written service manuals. A true fine piece of gear, that's been working reliably for over 40 years, and built for service!
I did get the task completed with using the SD supply, then installed the module into the SM900 chassis and got that project completed, less the confusion with the fan powering.
So, Friday I began pulling this HP 6227B power supply apart. My original service manual isn’t here….it’s in my storage locker library, so I was able to download a copy of it of the internet, and printed out pages to paste up a fold-out of the schematic, then the troubleshooting guide, other useful pages to help the troubleshooting procedure. I’ve never had this supply apart, so after finding I’d have to remove the control boards that are mated to the heat sink assembly & rear panel to which they’re attached, I’d have to release two pairs of screws at the rear of the side frames. #6-32 Pozi-Drive Flat Head screws, with cone-shaped lock washers, and have been installed since 1977. My initial attempts were fruitless….those weren’t going to budge without my stripping out the heads. So, I seeped machine oil into the screw cavity, brought out a small pin punch and hammer, and hammered on the four screws, followed by holding a Pozi-Drive #1 bit into the head and hammered on it to hopefully reform the head pattern some, as the metal was already getting buggered. With all my weight pressed down on the screws, first driving them tighter, then trying again to loosen them, they finally gave way and came undone. What a relief. I wasn’t looking forward to having to grind those screw heads off.
When I first opened up all of the panels on this supply, I found on the bottom side two fuses, one of which was open. It couldn’t be that simple, could it? Replaced it with a fresh 3A Fast 3AG fuse, and upon powering up, the meter slammed hard off scale, so I had the pass xstrs source coming thru, with no control of them…perhaps shorted pass xstrs.
HP 6227B Sch-1.pdf
HP 6227B Sch-2.pdf
HP 6227B Sch-3.pdf
HP 6227B Sch-4.pdf
Then, I found how they build this sucker. The two side-panel PCB’s each contain the bulk of the regulator circuits, with the back end containing the two sets of pass xstrs. It took some time to navigate the schematic, and then the PCB that had failed. HP had a 12-pin TO-5 IC that contained two NPN-differential amps with an NPN tied to the common emitters of the diff pair. I hoped like heck I didn’t kill one of those. It measured just like the Master channel, which was still working, so quite nice to have one working board for comparison to check with. I finally found what appeared to be the fault…a driver xstr Q6 NPN, ahead of a TO-66 Driver stage that handled the 2nd half of the pass xstrs. The part looked like old GE plastic TO-92 size parts, while the base is on the right side, collector in the middle and emitter on the left side, like most European pinouts. HP had their own part number for it, and looked to be a screened part from the manufacture’s generic parts. The Pass Xstrs were all fine, as were the rest of the semi’s and resistors.
After sizing up the power supply range it and it’s partners in crime were running off, and looking to see what I had in my stores and data sheets, the two closest parts being considered were an MPS A-06 and a 2N4401. The 4401 has better current linearity out past 100mA, though only a rated 40V Vceo, while that of the MPS A-06 drops off quickly past 100mA, while it carried an 80V Vceo. After setting up the curve tracer, I found the 2N4401 was more than sufficient for the task, and, after checking the layout to find which way the base lead was bent, it was traditional TO-5 pattern, and soldered it in.
Then, re-assembled the supply, secured the assembly back into place, and cautiously went to power it up, first with the fuse removed. I had supply voltage following the adjustment control, so put the fuse in, and now had a working supply again. Set it back up for Tracking, and the repaired Slave is again tracking the Master as it had been doing for as long as I’ve owned it. The large computer grade electrolytics have date codes of 7714 on them, so this was probably built sometime in the late 1970’s.
I got to hand it to HP, having such a wonder build factor in their gear....certainly of this era, and really well-written service manuals. A true fine piece of gear, that's been working reliably for over 40 years, and built for service!