I have long wanted to offer a higher-capacity version of my Small Wart power supply, but I was discouraged by the high cost of the Magnetek FPT12-200 ($10.76 from Mouser). So I asked a Taiwanese factory to clone it, and they have done so. This pic shows the board that I did using a sample that they sent:

http://www.smallbearelec.com/Projects/SmWart/Board1.jpg

By happy coincidence, the transformer  fits neatly on a "medium" pad-per-hole board, with enough room left for a simple filter and regulator. I will offer a kit, of course, and the transformer will be available as a separate component for about five bucks. Two questions for the engineers, please:

--I used a heat sink (and silicone grease) Mouser #532-577002B00. Is this adequate at full load?

--To get the regulator IC to go flush to the board (and bar it from rocking), I took a small file to the widened areas of its pins where they enter the epoxy package. Is this acceptable?

I hope you are all well and making lots of good noise!

SD

well, the regulator is designed to be flat on a board or against a chassis to dissipate heat, so teh leads are designed to be bent into a board or left floating above the board to allow more space for airflow.  if the regulator isn't being asked to deliver much current (compared to its capacity) then it shouldn't matter at all that the unit is modified to go flush into the circuit board.  on the other hand, if the regulator is running at full capacity, it might be that it could use that extra airflow.  yet, with the heatsink you've chosen, i don't think you'll have a problem.  naturally most of the time the unit won't be operating at full power anyway.

I wonder how much heat is going to come from the transformer? I notice small transformers being a weak point in some commercial gear.

Some of the heat answer is going to depend on what regulator is used. An LM317 produces much more heat than an LT1129. http://www.linear.com/prod/datasheet.html?datasheet=243

The temp is also going to depend on the input and output voltages of the circuit.

I've used that same heatsink quite a number of times but I don't think it's really enough if you want to pull more than maybe 2.5 watts out of the regulator.

The temp that the transformer runs at under load is going to be BIG influence on how much heat sink area you need. If the transformer and regulator are in an enclosed box, you won't be able to pull much before the temp gets too elevated. Without knowing more about the transformer all I can say is box it up and test it in a worst scenario along with a temperature sensor for analysis.

Another thing I like about the LT1129 is that it has a really good thermal shutdown if it gets too hot.

Has that transformer been hi-pot tested and if so, to what voltage? It looks like the bobbin is split judging by the picture, but I'm wondering how that transformer would do for high(er) voltage tube pedals using 1/2 of the primary for supply and the other 1/2 of the primary as a 1:1 isolated output? Hmmm... maybe the 200 in the part number means 200mA??? So much for that idea if that's the case.

I can't find any listings for a "Magnetek FPT12-200". I'm guessing that's a 12V 200mA unit? If so, that heatsink will probably be fine with an LM317 sourcing 200mA.

Take care,
-Peter

Peter--

Thank you for your comments!

Quote from: Peter SnowbergSome of the heat answer is going to depend on what regulator is used. An LM317 produces much more heat than an LT1129. http://www.linear.com/prod/datasheet.html?datasheet=243 The temp is also going to depend on the input and output voltages of the circuit.

>>I didn't know about the LT1129 and will check it out.

I've used that same heatsink quite a number of times but I don't think it's really enough if you want to pull more than maybe 2.5 watts out of the regulator.

>>9 volts @ .25 amp (I have actually drawn that much from the board) would be 2.25 watts.

The temp that the transformer runs at under load is going to be BIG influence on how much heat sink area you need. If the transformer and regulator are in an enclosed box, you won't be able to pull much before the temp gets too elevated. Without knowing more about the transformer all I can say is box it up and test it in a worst scenario along with a temperature sensor for analysis.

>>That's the next step.

Another thing I like about the LT1129 is that it has a really good thermal shutdown if it gets too hot.

Has that transformer been hi-pot tested and if so, to what voltage?

>>Yes, to 2KV @ 60 Hz for one minute according to the spec.

It looks like the bobbin is split judging by the picture, but I'm wondering how that transformer would do for high(er) voltage tube pedals using 1/2 of the primary for supply and the other 1/2 of the primary as a 1:1 isolated output? Hmmm... maybe the 200 in the part number means 200mA??? So much for that idea if that's the case.

I can't find any listings for a "Magnetek FPT12-200". I'm guessing that's a 12V 200mA unit? If so, that heatsink will probably be fine with an LM317 sourcing 200mA.

>>Sorry, it's FP12-200, and, yes, 12 volts/200 ma. Will be available in about two months.

Take care,
-Peter

Thanks Steve. Very cool!

Two of these units back to back might be an excellent way to construct a high voltage tube pedal around a 12AX7.  :twisted:

With the current being so low... this is looking great. I don't think thermal issues should be much of a problem unless you live in a desert in summer. My answer was starting out based on pulling max current from an LM317 which is not the case here.

I'm looking forward to picking up a couple to test out.

Will there be any kind of certifications on these transformers?

Thanks!
-Peter

Steve,
I didn't follow why you were concerned about it "rocking"... What would vibrate the chip to "rock" it? The longer lead lengths provide some stress-relief under thermal cycling conditions, which will be present.

My only concern about you and your kit-builders filing them is that filing puts a bit of a strain on the lead frame deep into the molding, and might make problems that show up later in early IC failures when already stressed bonding wires inside the package are thermal-cycled, on top of being anchored to the PCB very close to the package. FWIW.

Peter's comments about heat from the tranny are very good, and the newer regulators do tend to be more stable (IF adequately bypassed).

Quote from: MR COFFEESteve,
I didn't follow why you were concerned about it "rocking"... What would vibrate the chip to "rock" it? The longer lead lengths provide some stress-relief under thermal cycling conditions, which will be present.

>>I didn't want the "tab" of the regulator to be able to hit the transformer frame under any circumstances.

My only concern about you and your kit-builders filing them is that filing puts a bit of a strain on the lead frame deep into the molding, and might make problems that show up later in early IC failures when already stressed bonding wires inside the package are thermal-cycled, on top of being anchored to the PCB very close to the package. FWIW.

>>I figured there might be some problem. OK--back to the drawing board. I was hoping to use an available perfboard, but maybe the ticket is to get a PC board made with room for a slightly larger, better-located heat sink. Having said that, d'you like the idea for the power supply?

Peter's comments about heat from the tranny are very good, and the newer regulators do tend to be more stable (IF adequately bypassed).

Quote from: Peter SnowbergSome of the heat answer is going to depend on what regulator is used. An LM317 produces much more heat than an LT1129. http://www.linear.com/prod/datasheet.html?datasheet=243


Another thing I like about the LT1129 is that it has a really good thermal shutdown if it gets too hot.

Peter, I can tell you have far more experience than me in this area, but (there's always a but)

the LM317 requires quite a few more parts, but has twice the current capacity of the lt1129 and 20db better
ripple-rejection --when the adjust-pin is properly bypassed (from the data-sheets).  The capacity isn't a problem in this application, but when I go to the trouble of building a power-supply I want it as clean as possible. The protection circuits (including thermal)in the LM317 seem to work very well --from experience

This chip (lm1129) appears designed for small-size, battery operation and low parts-count on tight circuit-boards --none of which are as important to me as performance. YMMV :-)

-Eric

Thanks for the comments and info Erik.

To be blunt, I've never used an LT1129 in audio. I've used a ton of LT1129-5 and LT1129-3.3 regulators for digital circuits which were battery powered, so I will yield to your experience. I like their low consumption and low dropout and I have yet to kill one. I can't say that about LM317s.

Take care,
-Peter

Say Steve,
Any chance of getting a clone of the transformer used in voodoo labs pedal power 2?
I fixed a pedal power 1 for a friend and the transformer it used had the mfg. name and model on it so I googled it and found their website. Of course, like the idiot I am, I didn't write that info down for fiture reference.
Brian

Quote from: Peter Snowberg
To be blunt, I've never used an LT1129 in audio. I've used a ton of LT1129-5 and LT1129-3.3 regulators for digital circuits which were battery powered, so I will yield to your experience.

Haha...I've built 3 power supplies (lm317) They are very clean, though.

QuoteI like their low consumption and low dropout and I have yet to kill one. I can't say that about LM317s.

I will yield to your experience in destruction
The low dropout could be useful -- everything's a compromise --isn't it?
It  would be nice if the data-sheet on the LT1129 had a bit more depth --though EVERYTHING looks good on paper .

I'll get a couple and see.

-Eric

Quote from: BDuguaySay Steve,
Any chance of getting a clone of the transformer used in voodoo labs pedal power 2?
I fixed a pedal power 1 for a friend and the transformer it used had the mfg. name and model on it so I googled it and found their website. Of course, like the idiot I am, I didn't write that info down for fiture reference.
Brian

Unlikely, because minimum order for these things is 500-1,000 pieces. But I wiould like to know the name of the mfr. for future ref. if you happen to hit it again.

SD

Fair enough.
I'll see what I can do.