Voltage Multipliers (for our tube projects...)

[Gus]

The linked two transformer supply is missing bleeder/balance resistors, also you need to be mindful of the first and second VA ratings of the transformer.  Fred N. used a supply like that for a project that has been on the web for some time IIRC.

Did you google the Royer mxl2001 mod yet?

[Darkness, Darkness]

Did you google the Royer mxl2001 mod yet?

Tkanks Gus. I had a few difficulty to get the schematic but it is all the more interesting as I would like to try these on of these days tube mod for mics 
If someone is intesrested (you should!) here is the article PDF link http://www.diyfactory.com/projects/royerproject/royermod_2.pdf

This power supply, proving 95VDC is interesting for the subminis. For my current tube project (Marshall preamp clone), I still prefer the Weber transformer idea, because it provides almost out of the box higher voltage, closer to the voltage of the Marshall preamp

[frequencycentral]

Ok, question time:

I'm running the nixie SMPS that I posted on page 1 of this thread. I've got it hooked up to a small amp, basically a Firefly using a 6112 and a 7327. The two tubes' heaters are in series, so that's 300ma current draw. I'm using a 12 volt regulated wallwart that's rated at 500ma, but I know by experimenting that in reality it gives out more like 700ma. The more voltage you ask of the SMPS, the more current it draws. I've been 'soak testing' it all afternoon, I've left it on and keep coming back to it to check it's not on fire (!). You can tweak the trim pot just back from where it starts to hum because it's asking for more ma than the wallwart can give. Periodically, I have to tweak it back, as half an hour later it will be humming when it wasn't before. I started out at 206 volts, now I'm down to 186 volts. So what's happening here? It seems to me that the longer I run it for the lower the voltage it will supply, though it seems to have bottomed out at 186 volts. I realise that I could use a higher rated wallwart and teh problem would go away, I just want to understand what's going on.

Anyone?

[Mick Bailey]

Unless you have a heat dissipation problem the output should remain constant. My own amps don't drop voltage at all, but I had to both heatsink and fan cool the switching transistor as well as run the PSU off a 4A supply. If your transistor is not overheating, check that the wallwart power supply is not suffering from thermal overload.

[frequencycentral]

I have the IRF740 screwed to a big aluminium heatsink 3" x 4" with vents - far too big for the build I'm planning. I'll try another (more thermally conductive?) heatsink. What sort of heatsinks do you use Mick, size and material? Also, my SMPS is still on breadboard, so is it possible that the capacitance of the breadboard itself is causing issues. I'll also try a higher rated bench supply, just to check that it's not thermally overloading my wallwart. The storage cap I'm using is the 2.2uf specifed in the nixie SMPS schematic, I notice you use 47uf. While I don't think this is the source of the problem, I should use a bigger cap to smooth things out I guess.

[brett]

Hi

I have the IRF740 screwed to a big aluminium heatsink

What type of heat transfer compound/washer are you using?  MOSFETs are unique in that they conduct less when hot (unlike BJTs that suffer thermal runaway). 

Uou probably know this, but I'll state it for the new folk: it's often the thermal resistance at just one junction that causes the problems.  Heat transfer is only as good as the weakest link. (Thermal resistances are just like series electrical resistances - one high resistance ruins conduction through all of the low resistances.)
cheers

[frequencycentral]

Brett, I'm a noob regarding heat dissipation. I just screwed the IFR740 to a big aluminium heatsink. 

What do I nned to do to get this right? Transfer compound? Washer? Not quite sue about what to do re . "it's often the thermal resistance at just one junction that causes the problems" - any enlightenment greatly appreciated!

[Ripthorn]

Hey Rick, your solder is not high thermal resistance (as we know) but if heat is anything like electricity, when you match up your mosfet to your heatsink, you could have an impedance mismatch, which minimizes heat transfer.  You could get something like Arctic Silver thermal compound which is really good stuff (I have both the compound and the epoxy and both work great).  I used it for mounting my regulators in my submini amp to the chassis and it works much smoother now (and the chassis only gets mildly warm as opposed to the heatsinks which got blazing hot).  Hope that helps.

[JOHNO]

Really interesting reading here. Ill be picking up the parts for the Nixie psu tomorrow. Ive been trying to make a small tube amp for a while now but with only 48v dc not had much luck. About connecting two transformers like FC posted on pg4, is that safe? Ive got two identical transformers 240v primaries and the secondaries are multitapped for 12,15,17.5,20,24,27.5,30 volts (30VA) 1amp. Is it possible to connect them together to get higher voltage?

[brett]

Hi

but if heat is anything like electricity

Yep, heat transfer is very much like electricity transfer (and good electricity conductors, such as metal, are usually good heat conductors).  Your problem might be that air is a really bad conductor (of heat and electricity).  That's why your computer has a fan - to overcome the poor conduction of air by adding lots of convection.  In your case, you're not putting a fan in your device, so you need to excluse as much air as possible from between your MOSFET and the heatsink.  A thermally-conductive grease or thermally conductive plastic/rubber washer does the trick. 
cheers

PS the maths are easy.  Thermal resistances (degrees Celcius (C) or Kelvin(K)/W) can be treated like series resistances (volts/amp).  If your MOSFET has 1C/W resistance(from the junction to the metal backing) and the heatsink is 10C/W (fairly big) and the junction 20C/W (fairly bad), then the total resistance is 31 C/W.  Generally, you want the juction to be less than 100 C hotter than the surrounding air (Si melts at 150C).  So, in this case, the MOSFET and heatsink can safely dissipate 100C/31 C/W = 3.3W.  But if you install a better junction (5 C/W instead of 20 C/W), the total resistance is 1+5+10 = 16 C/W, the total dissipation can be 100 C/16 C/W = 6 W (amost twice as much).  Because junctions are often poor conductors (high thermal resistance), it's usually best to improve it before you go replacing a smallish heatsink (10 to 20 C/w) with a monster (2 C/W).  Also, adequate ventiation of the box (slots/openings etc) really helps keep the air cool, allowing for smaller heatsinks and poorer junctions.

[JOHNO]

Heres a pcb for the nixie psu. "NOT VERIFIED" please cast your eyes over it.

[JOHNO]

Crap that didnt work let me try again. NOT VERIFIED
dam it didnt work again

[Ripthorn]

Really interesting reading here. Ill be picking up the parts for the Nixie psu tomorrow. Ive been trying to make a small tube amp for a while now but with only 48v dc not had much luck. About connecting two transformers like FC posted on pg4, is that safe? Ive got two identical transformers 240v primaries and the secondaries are multitapped for 12,15,17.5,20,24,27.5,30 volts (30VA) 1amp. Is it possible to connect them together to get higher voltage?

It is possible to get more than 240VAC on your HT line, but I would highly warn against it.  Say for example your first transformer is pulling 10mA at your 24V tap.  You connect that tap to your 12V secondary of the other transformer.  In theory, that should give you 480VAC on the other side of your second transformer.  Sounds good, right?  The problem is that you can very easily blow the second transformer doing this because you are using a higher-than-rated secondary voltage, which is really stressing the core and windings on both sides of the transformer.  Not only that, but you are also transferring twice the power, which means you may very easily get the first transformer in on the blowing up action.  So, it is possible, but I don't recommend doing it.

Actually, with a multi-tap transformer like you just said, it should be safe as long as you don't exceed the power rating.  My above explanation is primarily for single-secondary transformers.  So in your case, yes you can, just be very careful about how much power is getting through each winding.  You never want ti to exceed 30W, period.

[JOHNO]

I uploaded the PCB in the gallery. search for "nixie psu". I made some of the pads pretty large so as to suit a wide range of caps and inductors. ITS NOT VERIFIED. So you guys should check it closely for mistakes before you build on it. Hope its useful.

[JOHNO]

Opps just realized that the layout is reversed. Hmmm or is it? You guys will figure it out.

[Mick Bailey]

I use solid copper heatsinks bent from 16 gauge sheet with thermal compound used for PC processors. without a heatsink the MOSFET gets hot, voltage drops and it eventually dies.

The test for your heatsink is to run the PSU until you see the voltage drop, switch it off then feel if the transistor is hot. Should run just warm to the touch if all is well.

[JOHNO]

OK the PCB is verified. Getting 105 to 303 volts dc for the Nixie psu. Running it at 12 volts 500ma. The only thing that has me worried at the moment is that the circuit kind of whistles or buzzes. Anyone else experienced this? Only have it connected to the multimeter at the moment.

[frequencycentral]

OK, I found a small heatsink on an old PC graphics card, I've scraped all the thermal compound onto the back of my IRF740 and attached it to the heatsink, and will do extended testing today,

OK the PCB is verified. Getting 105 to 303 volts dc for the Nixie psu. Running it at 12 volts 500ma. The only thing that has me worried at the moment is that the circuit kind of whistles or buzzes. Anyone else experienced this? Only have it connected to the multimeter at the moment.

It whistles and buzzes? Weird. And it's not powering anything? Mine hums if I ask too much of it, but that's through the speaker.

[earthtonesaudio]

http://www.cosmicrays.org/muon-%^&*roft-walton-calculation.php
http://www.diystompboxes.com/smfforum/index.php?topic=73905.0
http://www.diystompboxes.com/smfforum/index.php?topic=73808.0
http://www.diystompboxes.com/smfforum/index.php?topic=70582.0
http://dustbunny.physics.indiana.edu/~paul/cwbase/
http://www.techlib.com/files/voltmult.pdf

I just realized that the first and most important link in this post was screwed up by the forum's "bad word censor" software.
The part that says "%^&*roft" should be spelled "c o c k roft" (no spaces) as in the Nobel Prize winner.  Pssh.   

[amptramp]

There were a number of tubes made for car radios which operated off 12 V filament and plate supplies.  One of the more interesting was the 12K5 which was a space-charge grid tetrode (control grid operated at +12 V, input on the screen).  The 12K5 was matched to an 800 ohm output impedance (so 600 ohm transformers would work) and has a 9000 micromho transconductance and a 40 mW output power.

There was a whole series of these tubes (about two dozen different types) with a maximum plate voltage rating of 30 V although a few were rated at 16 volts.  The car radios all had germanium output transistor power stages, so no vibrator or transformer was needed.  Many functions were represented including dual triodes, pentodes, diode-pentodes and diode-triodes.

This allows you to use tubes with only a 12 V supply for both filament and plate.