Fluke 407D HV Power Supply – Repair/Restore

Another HV power supply  I acquired. Was in reasonable shape/condition physically . I liked the 550V output with a regulated -250V bias  out which can do  5mA. The Heathkit 2717 is very low on bias supply current and if not careful, you can burn out the pot. Here we have wirewound pot for same. Also MAX output is @ 400V. This one can do 550V with precision voltage control to 1s and .1 volts.

For someone who is new to the whole game, these power supplies are used for vacuum tube electronics. So they have a plate voltage output, which is referred as the main HV ( from 0-550V in this instrument), a negative Grid Bias voltage – referred as Bias output – , -ve output ,  -0 to -250V in this instrument,  and two heater supplies for tubes, which are 6.3V each, which can be cascaded to 12.6 as 6.3 and 12 are the most common tube heater voltages.

This is the initial condition I brought this in.  The usual dust and mud and labels.

 

After extensive  cleanup, as typically done in other projects.

Now a quick overview of the internals. 3 X 807 is used as series pass aka main regulator tubes.

As you can see, loads of bumble bees, and all leaky. You can see the insulation resistance as almost zero on the TO-5

There are plastic caps as well ( film) which are good and dont need replacement.

Remember, check resistors too. I found many which were drifted.

PCB !!!

The rectifier board, which also hosts the  transistors, which are used to regulate filament supply to regulator  comparator/amplifier tubes.

Check the height of the transformer.

and so called Amplifier board, which is the main voltage regulator board. 5651 VR tube is shielded and at the far bottom. 2 X 12AX7 as comparators. One for 250V  bias out and other for main output. 6AW8 Pentode/Triode combo as error amplifier.

And fluke precision wire wound resistors. 

Burned out 807 cathode resistors, a clear sign that the unit was used for long and may be 807s are low on Gm.

Almost dead neons, NE2E.

Note: Different serial No# uses different type of neons, verify the type used in your exact unit with the service manual.

I checked them on my Tek 575 CT and @ 20V/div, its sustain voltage at 130V…

….instead of the ~60-80V specification. Here is how a new NE2E performs on the same CT with same settings. Sustain ~ 65V.

For details on 575, Check my 575 Restore Project

Massive transformer !!!

The transformer is almost 8 inch tall . I ve kept a 5651 VR tube next to it to show you a relative size of it.

Now to restore :-

::WARNING:: DANGER ::

THIS IS A DEADLY INSTRUMENT WITH UP TO 850-900V DC INSIDE, WHICH IS CAPABLE OF DELIVERING STRONG CURRENT.

DO NOT ATTEMPT TO REPAIR/OPEN/OPERATE/FOLLOW  UNLESS YOU REALLY KNOW WHAT YOU ARE GETTING IN TO. THERE ARE DEADLY VOLTAGES INVOLVED AND IT CAN  KILL YOU INSTANTLY

THERE ARE BETTER WAYS TO DIE IF YOU ARE UP TO THAT.

Even if you are familiar with electronics, be very careful when you work on something like this.

Use test  instruments with HV OR PROFESSIONAL PROBES.DO NOT USE CHEAP PROBES. Most multimeters can accept only 1000V from ground level, and this unit develops close to 900V inside.

A quick overview of the design.

Main transformer with multiple secondary windings. Main secondary have multiple taps  for different voltage ranges.

There are two 6.3V/5A wingdings for filaments, which is directly provided to the front panel.

Input voltage is switched according the the voltage range selected, else the tubes will end up dissipating way too much power on lower voltage ranges.

For example if we were to use a single large secondary winding which delivers one single voltage,  say of 800V and @ 100mA load  current, if the output is set at 100V, the tubes will have to dissipate (800V-100V)*100mA = 70W!!

the voltage range selector switch  feeds the rectifier, which is very well designed, with caps across diodes to avoid ringing, and RC snubber at the input to avoid switching transients.  This output as per schematic can reach up to 840V on 500V range, I measured 900Vs on mine. BE CAREFUL.

Main filter is 2X 450V caps, with resistor divider to load share. Now this goes to the main regulator (series pass) array of 3 X 807 tubes. Cathode of the 807s deliver the output, grid is controlled by voltage regulator part of the network. Output sampling resistors (precision) get switched to match the voltage range in use and is part of the main voltage selector switch. This is fed back to the voltage regulator comparator to sense the output.

 

And finally the meter circuit. Simple switching circuit to measure HV output, Bias out and HV current. Nothing special here. HV out is disabled when the meter is off.

Now the 807 screen grid bias circuit, simple half wave rectifier …

 

..and associated biasing, and frequency compensation. Watch out for the neon and resistors in the section.

 

Now to the filament  power supply regulator for the comparators- 12AX7s.  Main comparator tubes are provided with regulated DC  filament voltage for better stability. So there is a dedicated regulated  DC filament power supply  for V9 and V13, using transistors and a Zener. The only two 3 leged semi conductor devices in the unit.

Now to 250V Bias voltage  regulator.  Dedicated secondary winding and half wave rectifier, filter, snubber,  feeding to the  series pass tube which is a 6AQ5. Regulation reference is provided by a 5651 VR Tube with standard comparator and error amplifier design. 5651 VR Tube provides reference for the main HV output regulator too.  12AX7 is used as the comparator, and 6AW8 as the error amplifier. R47 is the long tail.

 

Moving  to main HV regulator. Tube lined up is same as -250V Bias regulator. 6AW8 as error amp, 12AX7 as comparator. Input reference comes from 5651 VR Tube and output voltage is sampled via precision resistor  divider to suit the voltage range in use. Additional pots on the output sampling line controls R63 and R63 provides 0-50 and 0-.5 vernier control (fine)  for the output.

 

And the main regulator circuit, which refers back to the 5651 VR Tube  for reference and samples output via precision divider. Also note  the extensive  high frequency compensation/ripple control is employed throughout the design.

 

Restoration

Basic checks identified several resistors drifted way too away. replaced all, including the cathode resistors on the 807s, which were burned out.

All bumblebees capacitors were replaced so do two neons and pilot lamp.

NOTE: please clean the PCB after soldering, flux residue can catch fire as the voltage level involved are too high.

Amplifier (regulator) board after recap and re-resistor.

Rectifier  board after recap and re-resistor.

Tube Testing : 807 Tubes were on the lower end, still okay for use, as they are not totally dead. 6AW8s were completely dead, one was short, so replaced them. 5651 Tube was precisely on the mark.

Electrolytic capacitors –  I tried reforming them and found many had bad leakage current. So decided to replace with Nichicon 450V/105°C.

The OIL capacitor, at the final output was all good, so didn’t replace it. Its leakage was way low and was all intact till 600Vs while testing with my IT11 Tester.

Finally power up with variac, monitoring all rails. Meters are attached to filter capacitors to measure the rectified output and ensure all are coming up properly.

And @ 40V input AC, here is the reading from all rails.

And finally after few minutes of warm up, full input and here it is back to life.

{{ I broke the meter cover while moving it around during cleaning 😦 }}

Inside view –

And full throttle…

I load tested it, checked ripple, regulation, and all are under reasonable value considering I dont have great 807s in there.

Here is the final result @ 550V DC

WARNING : DO NOT ATTEMPT TO OPEN/REPAIR UNLESS YOU REALLY KNOW WHAT YOU ARE GETTING INTO. ISOLATION TRANSFORMER AND STRICT ADHERENCE TO SAFETY PROCEDURES ARE A MUST DUE TO TH EXTREME VOLTAGES AND CURRENT CAPABILITY INVOLVED IN THIS INSTRUMENT. IT CAN KILL YOU IN A SECOND. 

 

BACK TO HOME PAGE FOR OTHER REPAIR/RESTORATION PROJECTS