Connect with us

Designing a lab power supply for tube experimenting

Discussion in 'Electronic Design' started by Peter Andersen, Oct 21, 2005.

Scroll to continue with content
  1. Hi

    I would very much like to have some ideas, schematics, tips on how to design
    a lab power supply for experimenting with tube designs. The specifications
    could be like 0-500V 0-100mA... (+/-?)

    I would like to build a good regulated power supply. Could i just design it
    like a normal low voltage PSU? like a opamp, voltagedivider, a BU508 as
    seriesregulator, and a couple of high voltage caps.??

    Please mail me any kind of schematics and give me some help and hints about
    this project...

  2. Jon Yaeger

    Jon Yaeger Guest

    Save yourself a lot of grief and buy a used Heath SP-17A power supply. I
    saw one for sale for $50 here:

    Or do the eBay thing.

  3. Please mail me any kind of schematics and give me some help and
    I Jon

    Yes, I have seen it a couple of times when I was browsing around the net,
    but it is fun to make and design these things, and I have all the parts I
    need for this kind of project.
    I just want to read som experiences from others.
  4. Peter Andersen wrote...
    One problem in designing such a power supply is the short-circuit
    current-limiting aspect of the design. BJT's have a horrible 2nd-
    breakdown effect that will greatly exacerbate your design problems.
    To get an idea of the problem, study the BU508's SOA curves. Ouch!
    For example, at 40 volts, you're allowed 120 watts of dissipation,
    but at 400V you're allowed only a wimpy 36 watts!!! These are 25C
    values, which have to be derated as the transistor's case heats up.
    I recommend that you use high-voltage MOSFETs as pass elements, and
    avoid the issue.
  5. Bret Ludwig

    Bret Ludwig Guest

    Jon Yaeger wrote:

    The Heathkit supply is nowhere near the quality of the Lambda and
    Kepco supplies that are available just as cheaply. However, with
    refurbishment, the Heathkit will do okay.

    Glass Audio ran an article by Bruce Rozenblit on a very nice bench
    supply you can build some time ago.

    I prefer using a homebuilt HT supply which consists simply of a off
    the shelf power transformer-mine is that of a Fisher receiver whose
    OPTs went into Fender guitar amps-with diodes and caps I can just plug
    into a Variac or into the wall for rated voltage, with nice recessed
    test ports. I have a (shortable)light bulb socket in series with the
    primary. I only use the HT winding, I use, and you should own, a
    variable DC low voltage supply. Mine is a commercial Asian made one,
    but older HP, Sorenson, Behlman Invar, Tek, or Lambda ones are hamfest
    constants, or you can build one. Joe Carr has a good book with nice
    plans. A Schauer or Schumacher garage sale battery charger can be
    salvaged for a power xfmr and case.People throw them out all the time
    when the leads are bad.
  6. Do you have a link, pdf, scan..? I would like to read it.

    Yes, I have done this for some time, but now I want to make a real HV power
    supply instead of the HV transformer/variac method. I have several HV
    transformers from 500-2KV.
    The real supply will of cause have a 6,3V output on the same box.
  7. John Fields

    John Fields Guest

  8. Cat got your tongue?

    Best regards,
    Spehro Pefhany
  9. Pooh Bear

    Pooh Bear Guest

    It puzzled me too at first

    Look for the up arrows under the word case ( although it may have moved now on
    account of the fixed/proportional spacing issue )

  10. Bret Ludwig

    Bret Ludwig Guest

    The article is the property of AudioXPress but you might email Bruce
    Rozenblit, who might be able to get you a copy.
  11. west

    west Guest

    Me too, Bret. Do you recall the year & month of that Glass Audio article?
  12. west

    west Guest

    Me too, Bret. Do you recall the year & month of that Glass Audio article?

  13. west

    west Guest

  14. Fred Bartoli

    Fred Bartoli Guest

    Plus it's plain wrong. The max power dissipation depends on the case
    temperature. If the junction was at 25°C there will be no power limit.
  15. It will be helpful if you can include some other supplies in the same

    6.3V AC (with tappings for 4v if you want to use really vintage
    valves). A couple of tappings at + and - 15% will be useful to check if
    your design is unduly sensitive to emission and mains supply variations.

    A negative low-current supply for grid bias. O - 10 v for signal valves
    or 0 - 30v for power valves. It may help to have several supplies
    available on potentiometers from one stabilised line, so you can test
    several stages at once. They may need to source or sink small currents,
    so a potentiometer is better than a stabiliser chip for these.

    A second high voltage low-current supply for screened grids or beam
    plates. Again it should be potentiometer derived because it may have to
    sink current. A hefty rotary switch and a string of high wattage
    resistors is the way to go. A two-bank switch with staggered resistor
    chains and a separate rotary potentiometer connected between the wipers
    will give coarse and fine control if you need it.

    It can either be connected across the main HT supply before the
    stabiliser or it can have a small supply of its own. It doesn't need to
    be particulary accurate, but it should be smoothed with a hefty
    electrolytic capacitor. If it uses a small dedicated supply, this can
    also be heavily smoothed and usefully brought out to another set of
    terminals for powering the earlier signal stages.

    It helps to have the HT switchable independently of the LT, that way you
    don't have to keep switching the heaters off and on while you make
    adjustments (and you don't get tempted to save time by leaving the HT on
    while you fiddle). Use a big obvious switch - label it clearly - and
    have a large, bright, red panel lamp to indicate when the HT is on.
    Always try to work with one hand in your back pocket, you will reduce
    the risk of a shock across the chest.

    *** All storage capacitors must discharge rapidly when the HT supply is
    cut - otherwise you will get some very nasty shocks.***

    See if you can get hold of wireless constructors magazines of the 1950s
    or 1960s, they often had articles on HT and LT PSUs. Another source
    would be radio amateur handbooks of that era.

    Get hold of the valve manufacturer's data books and read them - they
    will save you a lot of expensive trouble and anguish.
  16. Pooh Bear

    Pooh Bear Guest

    It depends which way you care to calculate it.

    Either way it's actually the die temp that ultimately matters. Case temp simply
    factors in the junction to case thermal resistance making the calculation easier
    for the novice. If an insulating washer is used then you need to go back to die
    temp and add the washer's resistance in the thermal circuit.

  17. Pooh Bear

    Pooh Bear Guest

    I forgot to say ----- because it's not very practical to measure case temp directly.
    Heatsink temp is the way it's done in practice.

  18. Jon Yaeger

    Jon Yaeger Guest

    Not for many power amps under some conditions. No problem for preamps or
    low wattage stuff.
  19. Fred Bartoli

    Fred Bartoli Guest

    You know this: maximum power dissipation is given at TJ=max and Tcase=25.
    You have to derate the power with increasing Tcase. TJ is already at max and
    stays there.

    How can this depends on the way you calculate it?

    Indeed, but that wasn't the point.
  20. John Stewart

    John Stewart Guest

    Does the PS need to be SS? There are many bullet proof tubed designs.

    And why go down to zero volts? That really complicates things & many practical
    regulated PS's are not that stable, zero to 100 volts. My own preference would
    be two supplies.
    That would be a zero to 150 SS & a 150 to 500 volt tubed version. That way one
    can do the biasing while the other handles the B+.

    My thoughts, anyway. John Stewart
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day