• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Bench Power Supply Questions ?

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.


  • HP 6448B DC Power Supply.jpg
    HP 6448B DC Power Supply.jpg
    17.3 KB · Views: 204
Last edited:
I've connected supplies in series before. Even different brands of supplies. You'll make sure the supplies are truly floating - i.e. their outputs are not referred to chassis ground. In addition, most if not all supplies will have a spec for the maximum voltage allowed between the negative output terminal and chassis ground. You need to ensure that this spec is not violated.

I don't know if all supplies like to be hooked in series. The HP and Sorensen brand supplies I have don't seem to mind...

I have also connected bench supplies in series. Sometimes you get away with it, sometimes you dont. Granted my most spectacular displays of burnt power supplies happened when I connected several in series then overloaded the combination.

When you series power supplies, and then draw enough current to activate current limiting the power supply with the lowest current limiting will fold back. It then eats the output of the other power supplies, backwards. If each power supply has an internal (or added externally) diode across its output that can eat all the current without failing, then you are OK.

The 6448B is an old style 60 Hz switcher. Its control loop is VERY SLOW, hundreds of milliseconds reaction time. The current limiting is equally slow. There is also a quart sized electrolytic cap directly across its output. Have you ever shorted its output while it is operating at 650 volts? It will blow a #18 wire in half, or blow an electrolytic cap right off a PC baord. Don't even think about fusing its output with a glass fuse. Yes, I have one, and yes, I learned all these things the hard way!

I think seriesing two of these is a good way to blow up one or both of them. Each one will output 1 Killowatt, and draw about 1200 watts at full tilt. Can you even power two of these. I have blown my 15 amp bench breaker with mine several times, under normal operation, no shorts, just a 525 watt tube amp. What are you building?????
Look for an old Kepco BHK 1000-0.2M They go from 0 to 1000 volts and 0 to 200 mA. There is also a 0 to 500 volt 400 mA model and a 0 to 2000 volt 100 mA model. They are solid state except for 8 pass tubes. The tubes are 8068 which are expensive, but 6CD6's will work with minor socket rewiring. The curent limiting is fast and accurate enough to keep things from exploding.

I used to have one of each. I sold the 500 and the 1000 volt version for $50 each and parted out the 2000 because no one wanted it. I have seen these things sell for $500+ on Ebay, but I got each of mine at hamfests for less than $50.

There are newer digital Kepcos with similar specs, but they are way out of my price range. BHK 1000-0.2MG is the numberfor the 1000 volt version.

Some builders of HV amps use Electrophoresis power supplies. These can sometimes be found on Ebay cheap. I got two different "tested" units that didn't work. Return shipping costs as much as the refund, so I parted them out.

I don't build high voltage amps anymore since my 845SE doesn't get used and the 833A never got built.

I still build big power sweep tube amps, and the 6448B is just about right for those. Mine makes 650 volts and 1.7 amps turned up all the way, and it has caused at least 3 big parts explosions. The fan is loud too. Not a big deal when you are building 500 watt amps. I have a Fluke 407D and a Knight KG766 for smaller stuff.

If I have 2 HP 6448B, can I hook them up in series to get 1200V ?

Each 6448B is 600V max.

You should find these important information in the operational manual. I believe to know that HP allows two identical units connect in series in master/slave configuration.

But no matter what I believe to know,refer to the manual !

1200 Volts are no joke, they are deadly serious!

Edit: Manual of the HP 6448:


Last edited:
I did not have the manual for my unit, so thanks for sharing. The manual talks about parallel operation, but makes no mention of series operation.

The schematic reveals what I suspected. There is 500 uF of capacitance directly across the output. When the supply is cranked up there is the energy in these caps (about 200 Joules) plus the output of the supply available at the output terminals. All this must be dissipated before the current limiting can activate. It makes parts dissapear.
Joined 2009
Paid Member
500 uF, 600 V. That packs a serious punch. My math says 90 joules. U = 0.5*C*V^2. Still. That's not to be taken lightly. That's more than enough energy to kill you or turn your circuit into a claymore mine.

Unless the project requires THAT much energy, I suggest some of the still deadly but somewhat more sane options listed in Post #5.

Personally, I use an HP 6209B (320 V, 0.1 A) supply.

Last edited:
My math says 90 joules. U = 0.5*C*V^2.

DUH...I forgot the 1/2 part......and at least on my unit V goes up to 650 and I have used it at that level. The biggest OOPS came with about 620 volts applied to Pete's big red board. There was a 10K 1W dropping resistor feeding a 47 uF 450 Volt cap. The cap was eating about 550 volts for half an hour or so when it decided to short. Applying 620 volts across a 10K 1W resistor caused an immediate flashover removing all evidence that the resistor ever existed, except for a black spot on the PCB. The lightning bolt that ate the resistor then took out the capacitor. No it didn't vent, it exploded. The black base and two wire stubs remained in the board. Paper confetti and shards of aluminum were scattered all over the room. All of this happened before my hand moved about 1 foot to the master kill button on my workbench.

Unless the project requires THAT much energy, I suggest some of the still deadly but somewhat more sane options listed in Post #5

I only use the big guy when it is needed. The same board with new caps and bigger tubes:

The first picture shows one of Petes boards with a single channel being tested. The big HP is set on 650 volts and one channel of the amp is drawing over 624 mA. The yellow meters read cathode current (volts across 10 ohms).

Second picture shows the output trace at 266 watts. Note the Sorensen QRD20-4. As you stated it's good for heaters, bias and solid state stuff, makes about 24 volts all the way up. Rated for 4 amps, mine goes to about 4.5.

Third picture shows the second channel now operational and wired in parallel with the first through a common OPT. Still idling here.

Fourth picture shows the 8903A reading "watts in to 8 ohms" (19.0 SPCL). 504 watts! Note the Knight KG-664 (I had the # wrong in the previous post). This was a $25 hamfest find and goes from 0 to 400 volts at 200 mA.


  • 266watts_A.jpg
    335.3 KB · Views: 108
  • 266watts_scope_A.jpg
    233.6 KB · Views: 99
  • IMG_1079_A.jpg
    291.7 KB · Views: 102
  • 504watts_A.jpg
    207.4 KB · Views: 101
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.