Variable DC power supply

[hesamadman]

Can anyone suggest a good variable DC power supply for a breadboard? Does anyone use these on their breadboard? What do you use for heater power source?

[Willabe]

Look at these;

http://el34world.com/Forum/index.php?topic=18354.msg187427#msg187427

http://el34world.com/Forum/index.php?topic=15445.msg148808#msg148808

[eleventeen]

The comment I would make is, what are your plans for your breadboard? If you plan to build 4 x 6550 amps, you need a mother power supply. And regulating it is very expensive, because the pass elements need to be high voltage and high power. You may need 200 ma at 550 volts if you want full power out of such a tube complement. 


A breadboard P/S that *only* has to power a few preamp tubes and a PI and NO power tubes could probably get by with 25 mils capacity (probably not actually using more than 10 mils) and 250-300 volts.


And by the time you have a tube power supply suitable for a preamp, say 300 volts, 25-50 mils, there's hardly any need to regulate it because the current draw is so light.


So all I'm saying is, if you want to build up and use a breadboard, don't let not having a 200 ma power supply stop you.

[sluckey]

A 350-0-350VAC @ 200ma and a variac would probably handle most any of your B+ needs. Use a separate filament transformer.

[2deaf]

Can anyone suggest a good variable DC power supply for a breadboard? Does anyone use these on their breadboard? What do you use for heater power source?

I have an experimenting chassis with a Stancor PC8408 PT.  I bought that transformer brand new some number of decades ago and used it for various amps until I put it in this chassis for good (I think).  It is rated at 6.3V @ 2.5A, so I soak up extra heater current when necessary with one or two 6550's that my 300PS shredded also decades ago.  I have a switchable load on the power supply, but I wanted more versatility so I built an outboard VVR in a box.  So now I have a chassis and two outboard modules.  If I were to start from scratch, I would do what Sluckey suggested. 

[PRR]

This seems to do what your dual-pot design does (split voltage over two MOSFETs), but with just a one-gang pot?

[DummyLoad]

A 350-0-350VAC @ 200ma and a variac would probably handle most any of your B+ needs. Use a separate filament transformer.

+1 

a 6A 6.3V and 6A 5V transformers are handy to have around.

http://www.hammondmfg.com/167.htm    167Q5 and 167Q6

or simply use the PT you're going to build with: USUALLY you should have a good plan of what you want to use for the output stage, so the PT and OT are known, the preamp(s) and driver(s) are the variables then.

these are another option:  https://edcorusa.com/tbpwr-hv-1 

two models available to choose from depending on your mains voltage.

a 300V @ 300mA PT should cover most things except for 500V+ power amps.

most variacs can be wired to supply 15% over mains voltage, e.g. 120V mains variac can run up to 140V.

--pete

[2deaf]

This seems to do what your dual-pot design does (split voltage over two MOSFETs), but with just a one-gang pot?

What if there is some current flowing through the gate?

[PRR]

> What if there is some current flowing through the gate?

Then the MOSFET is blown, zapped, kaput, deceased.

MOSFET Gate is hyper-thin glass. The only way it can conduct is from damage, when an arc sputters electrode metal through a hole in the glass. Un-blown, a MOSFET Gate is one of the lowest-current "wires" in electronics, much lower than reverse diodes (solid or vacuum) or even most PCB-stuff.

[2deaf]

MOSFET Gate is hyper-thin glass. The only way it can conduct is from damage, when an arc sputters electrode metal through a hole in the glass. Un-blown, a MOSFET Gate is one of the lowest-current "wires" in electronics, much lower than reverse diodes (solid or vacuum) or even most PCB-stuff.

Allow me to rephrase my question:  What if there is some current flowing through pin one of at least two different flavors of ST Microelectronics MOSFETs?

The current is there, but it may be going through the internal protection circuit or some other path as opposed to going through the gate itself.  All of my ST's do this and they are not blown, zapped, kaput, nor deceased -- I recognize those characteristics in a MOSFET having rendered a couple of them as such.

The voltage drop is no big deal and you have a good design there.  I used a similar design with a rotary switch a couple of months ago when I was trying to determine what bias to use with different voltages. 

[PRR]

There shouldn't be Gate current, not enough to measure with bench equipment.

Are you sure?

With 220K-470K resistor networks, a 10Meg VTVM/DMM's loading will cause voltage shift, only when poking at it.

Use test-leads, don't touch the meter while testing. Measure the Gate-Source voltage. It should be +1V to +5V. It must not be outside +20V to -0.5V, or the protection diode will conduct, the MOSFET not working much at all.

[jjasilli]

I have a Heathkit PS-4.  It can putout regulated 400 VDC @ up to 150mA, so it's under-powered for big bottles.  Also puts out Bias voltage and 6.3VAC @ 4A.  The old Fluke 410 can put out I think 600VDC @ substantial amperage.  The Peter Bench website (I think) has plans for a home-brew PS.


I've used my Heathkit only for trouble shooting, not design.  The issue I see with any bench PS is that ea type of guitar amp has its own type of PS which, statistically speaking, is never regulated.  However, a powerful, regulated DC heater supply sounds nice.  It might keep heater noise issues out of the design phase, even if the final amp uses an AC heater supply.  Quieter heaters might be helpful on an open, unshielded breadboard.

[2deaf]

Are you sure?

No.

With 220K-470K resistor networks, a 10Meg VTVM/DMM's loading will cause voltage shift

I am sure about that.

[DummyLoad]

The Peter Bench website (I think) has plans for a home-brew PS.

do you mean pete millett?

http://www.pmillett.com/file_downloads/HV_BENCH_PS_SCH.pdf

--pete

[jjasilli]

Yes, Pete Millet. Good catch!

[eleventeen]

If you look at that Pete Millet schematic, you will see that the supply is built around an old Tek 'scope transformer. See the three HVAC windings in series?  I have sometimes found some of those "high voltage" windings on Tek scope trannies to be low current. It's not uncommon for one of those windings to be a 50 ma special winding for just three tubes. I have not looked at the tranny in question although I DO have 2 of them sitting on my porch from taking apart a Tek scope PS (separate box from the scope) but I can assure you that a grand total of ONE of those windings has any sort of ampacity that 5 qty 807 tubes (basically 100 ma per tube) can handle. The pass section of this supply is way oversize for the tranny.


Secondly, it uses 5 qty 807 (you could certainly delete 1-2-3) My Fluke 407 with 3 qty 807 can make 555 volts @ 300 mils. It's a brute. 807's are good in the sense that they are cheap and plentiful.


jjasilli makes an excellent point that you will in 99% of cases NOT find a regulated supply in an actual amp.


If I wanted to build such a thing, I would mimic the Heathkit PS-4 and that would give you a place to use your cruddy old 6L6 tubes.


Heath schematics: http://www.vintage-radio.info/heathkit/

[PRR]

Didn't Weber have a hi-current PT with about a dozen voltages? And a factory screw-up so they sold it cheap?

Weber VST has "improved" their site so bad that I can't find anything now.

{later} Argh. Found the fugger.

WPTGP Transformer $36.75
http://www.tedweber.com/media/import/wptgpsch.JPG

You can get some DC values from 56V and 100V to 966V(!!), including 290V, 350V, 390V, 420V, 440V, and 574V. These choices, plus some 1K 10W resistors for lower-power steady-current jobs, will cover a lot of cases. US/Can users may also run the 240V primary on 120V for half of all the above (28V, 50V, 145V...287V), but you lose the heater supplies. All at nominal 150mA.

     BTW: I am 97% sure this was a CT+bias winding, but someone got one (or more) winding wrong. Take Red-Red as the ends. Red/Wht-Red/Wht are 70V lower taps. Red-Yel is the (intended!) CT. Blue(Red) is the (intended!) bias-tap. All the colors make sense. But one side of Red/Yel is 300V, the other side is 40+210= 250V. If you ass-ume it was supposed to be 250V/40+210=250V each side of Red/Yel CT, you can get 445V DC (safe for 450V caps) and 350V (tweedier)- great plan! But someone snuck another 50V in. I'm sure when Ted tested the shipment and found the error, he pencil-checked ALL possible variations and determined it simply CAN'T be run CT. It will run "single winding" (with any series string already wired-in) into a full wave bridge and do DC just fine.
    Look at W025130EU-- almost exactly what the WPTGP may have been before someone mis-counted. So 35% off the price, in return for some fancy figuring.
    I'd be wary of 690V @ 150mA, because the intended CT scheme runs each side half the time, the FWB alternative runs all windings both half-cycles. But who wants 966V DC?? If you take half the windings, under 500V DC, then the total heat at 150mA is about as-designed.

[shooter]

Weber VST has "improved" their site so bad that I can't find anything

  yup!
I have used the GP tranny a couple times and is handy

I found this one for BB, just haven't started yet

[DummyLoad]

just doodling... 500V @ 500mA variable PS w/ 6-150 neg V. bias supply. based on PRR and other's VVR stuff. i quit adding crap (whistles & bells) when the bulk of the project hit near on to 400 bux with a case.


$hitty sim says it should work. heatsinks indicated are overrated. using a 100W 3U power amp chassis from china. 1/2 the price of a bud, hammond, it's a par-metals case.


http://www.ebay.com/itm/HiFi-DIY-Audio-amp-chassis-table-top-enclosure-20-16123N-/371499127500?hash=item567f12d2cc:m:m3ewThzrKGErogXuX4DApVg


please feel free to comment; add; delete; flame; etc..


--pete

[2deaf]

That looks really cool to me.

I don't know much about VVR and I couldn't seem to learn squat from my minimal online search abilities.  So I bought some of those MOSFET big boys last July and tried them out for myself.  I had already trashed an IRF820 in a source-follower arrangement when I inadvertently hooked the the gate-to-source zeners in series instead of butt-to-butt.  So I ordered in some big MOSFETs that had built-in zener protection to compensate for my inadequacy.  Anyways, I don't see internal zeners in the IXFA10N80P and I wonder if it wouldn't be prudent to put them externally for positive gate-to-source voltage as other folks have done.   

[DummyLoad]

That looks really cool to me.

... zener protection to compensate for my inadequacy.  Anyways, I don't see internal zeners in the IXFA10N80P and I wonder if it wouldn't be prudent to put them externally for positive gate-to-source voltage as other folks have done.

thanks!

IMO, it's a bare minimum to get a breadboard up without a variac and the hassle of builing the PS onto the breadboard. i was going to add VT rectifier (GZ34) with switching from SSR to VTR, but it's just too dang busy as it is now. KISS prevailed.

it would be wise to add the zener protection diodes. i selected the MOSFETs in this design because the SIM SW i'm running had SPICE models for them. was going to use the IRF820 for the B+ supply, but wanted beefier parts for the B+ to keep the devices well within SOA with ample heat-sinking in case of dead shorts before the fuse pops.

i smoked some IRF820 MOSFETs as well experimenting with a transformer loaded cascode for a type 4AB reverb tank driver. damn thing was picky about the bypass cap and the gate stop resistor: it would fly off into oscillation super-heat the OT, fry the IRF820 and 2N4393 within seconds. fixed by eliminating the bypass cap and used 22K gate stop. 

--pete

[PRR]

> PS w/ 6-150 neg V. bias supply

Since the negative bias supply is floating on the transformer and not used for anything else....

You could design the bias supply as a "positive variable supply" then ground the positive output to get a negative output. Advantage is that now you use a N-type MOSFET which can be the same type as the ones bought for the main output (overkill but WTF). Disadvantage is debugging is a bit more mind-bending.

> wise to add the zener protection diodes

Agree.

Builders should note that the Gate-stopper resistors need to be _AT_ the Gate, within an inch or less. That Gate is hyper sensitive. Hang any wire on it, it will find a way to oscillate in the cell-fone bands (well, maybe only high police-radio bands).

[DummyLoad]

> PS w/ 6-150 neg V. bias supply
Since the negative bias supply is floating on the transformer and not used for anything else....

have seen that before - toft audio designs mixer power supply i repaired for a friend is +/-18volt. PS is 2 positive power supplies strapped in series. junction of PS1(-) and PS2(+) is virtual ground.

the big FETs are $3.91 ea. the bias supply FET is $1.66 - no brainer there: negative polarity design wins. 

> wise to add the zener protection diode
Agree.

added zener diode protection with updates to schematic attached. also attached is a mostly complete BOM.

OUCH! 482 bux! save moolah: use old RM PC case & use salvaged test equipment PT & choke... 

--pete

[HotBluePlates]

If you look at that Pete Millet schematic, you will see that the supply is built around an old Tek 'scope transformer. See the three HVAC windings in series?  I have sometimes found some of those "high voltage" windings on Tek scope trannies to be low current. ...

You have to check out Millett's site for complete details. And according to data on his Tek transformer page the 3 windings used in series are rated for 0.5-0.8A each. So it turns out the 5x 807's are appropriate.

[eleventeen]

Yeah, I have:


Tektronix 555 oscilloscope 120-128:

234V mains: 1-4 (link 2, 3)
E/S screen: 27

Winding Voltage Current
9-10 208V 160mA
20-21 204V 50mA
11-12 126V 250mA
13-14 126V 250mATektronix 555 oscilloscope 120-127:

234V mains: 1-4 (link 2, 3)
E/S screen: 21

Winding Voltage Current
5-6 6.3V 0.35A
10-11 6.3V 10A
12-13 6.3V 6A
14-15 148V 160mA

There's no question these are heavy transformers and all the windings are there for a serious bench supply.

But: If you wanted to make such a supply capable of as much as the tranny can put out for its ~~20 lbs, you'd want to avoid using the 50 ma winding in your series string. Oh, and by the way, you noticed there are no heater windings?

For that, you need the "little brother" 17 lb transformer.

Such a bench supply could do very well keeping your bench from floating away.

I *have* messed around with some other Tek scope trannies. The ones I have came from the separate box type of supply so Tek did not have to worry about compactness. They had this whole box to work with. I *have* had other trannies that would be better than the two I have (point being, they are not all created equal for this direction of salvage.)

Underneath the connection board (which you see in the pix of the underside of Millett's supply) with the turret connections which ordinarily goes under mounting surface for the tranny, you can see some of the wires coming from the windings. You just have to be careful because *some* of the windings are intended to just run a couple of tubes and they have "not much" current capacity. That was my point. You'll see big thick wire for the 6A 6.3 connections, then 24 ga for another one.