"Sub Zero" Low Voltage Submini Tube Push/Pull Amplifier Project

[MetalGuy]

This is a $0.32 (single quantity) FET made by Fairchild: http://www.fairchildsemi.com/ds/FQ%2FFQP10N20C.pdf - if you can't afford this then I doubt you can even afford to build effects!

I was talkin about FETS with ON resistance les than 0.1 Ohm.

As far as that pesky "transformer" goes, it's all about finding an inexpensive but reliable core for us to wind the transformer on.

Believe me if you want this project to be popular try to avoid any transformer winding at home! Better look around for something that anyone can get online.

How much are you currently paying for a 30 W transformer?

http://www.tube-town.net/ttstore/product_info.php/info/p2358_Tube-Town-Powertransformer-EI-230---200---6-3-7-VA.html

http://www.tube-town.net/ttstore/product_info.php/info/p2192_Torodial-30VA-Preamp.html

So this leaves us with the question of how much power you guys are asking from the power supply.

Personally I would need a PS that can power up a 2 or 3 12AX7s tube pedal or a submini tube amp which is like 250-350V/5-10mA depending on the application. For applications that require more power I think I'll stick to the transformer.

I think I can beat that with a switcher... I'm pretty busy this week but my father, a designer of SMPS's for about 30 years, has also taken a vested interest in this project.

This is very good. I'll be looking forward for your project but once again don't make me wind transformers!

[kristoffereide]

Can you post a perf layout for both the supply and the amp for idiots like me?

[frequencycentral]

Can you post a perf layout for both the supply and the amp for idiots like me?

I haven't done a perf for this project. Though my Superfly perf, which includes the SMPS and amp on one board could easily be adapted to this design. I ended up building a Superfly instead of a Sub Zero because it's a more versatile design, capable of crunch as well as clean, and most of the major components are the same.  Having said that, if I can get a 300v SMPS running I'd be all over a Sub Zero build. If you want me to do an adaptation of the Superfly perf to be a Sub Zero perf I'll go ahead and do it though. 

[Ripthorn]

How much are you currently paying for a 30 W transformer?

http://www.tube-town.net/ttstore/product_info.php/info/p2358_Tube-Town-Powertransformer-EI-230---200---6-3-7-VA.html

http://www.tube-town.net/ttstore/product_info.php/info/p2192_Torodial-30VA-Preamp.html

.6A for a heater supply is only enough for two tubes, though that is a nice looking little transformer.  I wish I could find me a toroidal one like you posted for American primary voltages.

[MetalGuy]

.6A for a heater supply is only enough for two tubes, though that is a nice looking little transformer.  I wish I could find me a toroidal one like you posted for American primary voltages.

Email Dirk at TT and ask him about american voltages or universal primary. I'm sure he can arrange something.

[kristoffereide]

Can you post a perf layout for both the supply and the amp for idiots like me?

I haven't done a perf for this project. Though my Superfly perf, which includes the SMPS and amp on one board could easily be adapted to this design. I ended up building a Superfly instead of a Sub Zero because it's a more versatile design, capable of crunch as well as clean, and most of the major components are the same.  Having said that, if I can get a 300v SMPS running I'd be all over a Sub Zero build. If you want me to do an adaptation of the Superfly perf to be a Sub Zero perf I'll go ahead and do it though. 

Superfly huh... Which one is the best for just having a little practice amp at home? Could you give a link?

[markeebee]

Superfly huh... Which one is the best for just having a little practice amp at home? Could you give a link?

I asked the same question a little while ago.  Answer was......Superfly.

http://www.diystompboxes.com/smfforum/index.php?topic=78302.0


Also very worth looking at the original Firefly

http://www.ax84.com/index.php/oldprojects.html?project_id=firefly

[markeebee]

Calling Clifff, calling Cliff.....

....or anybody else that can help......


I'm trying to get my head around the whole SMPS v Linear thing.

One question:  am I right in thinking that if I use a switcher, there isn't a requirement for bigarse filter caps?  And, ergo, does that mean there won't be 300V (or whatever) floating around the circuit after it's shut down?

That would be a big benefit.  The HV troll lurking in the caps seems to be the thing that most puts people off building tube amps. 

[Cliff Schecht]

The physical cap sizes on the output of a switcher are a function of the amount of switching noise you can deal with at the output and how much ripple current the caps need to be able to handle. With a (relatively) high wattage switcher like the one we've been talking about, the caps need to be able to handle quite a bit of RMS current without heating up. This tends to makes it so that the caps are physically large. With a higher frequency switcher however, the noise is all well above audio frequencies and we can have a lot of switching noise without any noticeable negative side effects.

A useful trick that I used recently in a tube amp is to put electrolytic caps in series. While the effective capacitance is reduced in half with two caps, the voltage handling capabilities double (assuming both caps are the same value and voltage rating). When you do this though, it is smart to include some sort of resistance in parallel with the cap. This not only drains the capacitors quickly when the amp is turned off, it helps to evenly distribute the load amongst the capacitors in series. If you look up "series capacitor drain resistor", you'll find some formulas that will help you determine the necessary resistor value and wattage.

As far as the switcher caps go, Nichicon makes some excellent low-esr and low-impedance caps that are quite compact in size as well. I'm currently eying the E-PW or E-PM series (specifically, the 330uF 100V parts) for a relatively cheap yet high performing cap.

[davidallancole]

Hi Cliff,

Are you aware of any chips that use an external switch and have both voltage feedback and current sense for a simple boost supply?

[Cliff Schecht]

I was planning on using a variant of the old reliable Unitrode UC3842 for the SMPS. It uses an external switch and can do a few different topologies quite well. It is a current-mode controller, but any current-mode controller can be turned into a voltage mode device by "averaging" (or filtering) the output of the current sense circuit so that it is just a continuous signal. This isn't something that you would really want to do though as you lose your pulse-by-pulse regulation and introduce an unexpected pole into the current sense path. This pole essentially adds more delay to the switcher and kills the transient response of the amplifier (not desirable).

[davidallancole]

Thats what I was looking for.  Looks like a standard boost supply can be made with this with voltage feedback and current limiting.  Would the boost topology not be easier for the average diyer to build?

[Cliff Schecht]

I may switch topologies. It seems that nobody is interested in winding their own coupled inductor (the "transformer" in a flyback) and I've actually located a Coilcraft part that seems capable of providing 300 V at the current levels that would be used by your standard submini design. Here is the part: http://www.coilcraft.com/da2032.cfm#table

I'm not 100% sure this part will work, but it's definitely a candidate because people don't want to make their own transformers. I can get the actual switcher designed pretty quickly once we all nail down a set of specifications, but nobody has given me any solid information. As of now, I'm designing the switcher for 300 V at about 0.1 Amps (30 Watts). But if you guys want your negative grid bias voltage then you have to be willing to wind your own transformer, I highly doubt there is currently a usable commercial part. Aside from a lack of good specification, I've already got the power supply designed.

[frequencycentral]

I can get the actual switcher designed pretty quickly once we all nail down a set of specifications, but nobody has given me any solid information.

Cliff, apart from voltage and amperage, what else should we be specifying? I'm kinda assuming there are variables here of which I am unaware? Small physical size would obviously be great.........

[davidallancole]

It looks to me from the sub-mini designs I have seen that negative bias is not required.  If it ever is another switcher chip could give a negative rail.  Most of the sub-mini amps seem to run a B+ between 125-200VDC so an adjustable supply between those voltages would be fine.  It looks like the Nixie supply that the fellows have been using only supplies a few 10's of milliamps, so my guess is it would be fine if it supplied up to 50mA.

To summurize,

1.  Adjustable from 125-200VDC
2.  Aproximately 50mA maximum Io
3.  Input voltage of 9-18VDC
4.  Easy to build and source parts for
5.  Uses a common controller chip (I heard some people say they choose the 555 timer design since that chip has been around for ever and will continue to be)

Thats my 2 cents.

[Cliff Schecht]

Eliminating the negative grid bias rail makes my job a whole lot easier. I can now switch to a simpler boost topology and granted that I find some good enough magnetics on Mouser (my preferred supplier), it shouldn't be too long before I can have the switcher up and running. The specifications you guys just gave, specifically the current draw, is what I need to properly design the thing. I would like to overrate the output current capabilities by about double (300 V at 0.1 A is reasonable to me) so that even at "full load" the supply stays cool and runs without any problems. I'll even get a board going for people to make, at 100 kHz layout becomes much more importance and a well done, compact controller section can make all of the difference.

David: A smaller range on the input voltage section makes the design a bit less tedious and so I'm going to go with 12 V nominal +/- 3 V (so 9V to 15V). All of the parts I'm choosing will be from Mouser and since I'm a poor college student, the overall parts cost will be of high importance. As far as a separate SMPS supply for providing the negative grid bias, this is not a desirable thing to do unless all of the switchers inside the system are syncronized. Otherwise the two separate power rails can intermodulate and throw crap into your precious audio path that you wouldn't otherwise have. I'm going to scrap the need for the negative supply (for now) and design a simpler boost circuit that uses modern, inexpensive and easy to come by parts. If anybody wants that negative grid bias voltage then I can post a design for that as well, but you'll have to be willing to wind your own custom magnetics. This isn't necessarily hard, but winding 167 turns around a torroid (what I calculated earlier) isn't a trivial task either .

More to come!


Edit: Also, the adjustable output complicates things a bit. What may be better is to choose a few specific voltages that people are interested in and posting the required circuit components for these voltages. I'm pretty sure that changing the output voltage requires the compensator to be redesigned as well (I will check on this though). If this is a problem then the easiest thing to do is to place a voltage dropping resistor in series with the SMPS output to get the voltage levels you want. It's common practice anyways with just about any tube amp you look at.

[davidallancole]

I will take a look at the common circuits and pick out there voltages today Cliff.

[davidallancole]

I don't know if the rest will agree, but maybe 150VDC, 200VDC, and 250VDC supplies would be useful.  I myself would like a 175VDC supply.

Do switching supplies like steadier currents then not?  What I am wondering is if they would handle the dynamic current draws of a class AB amp?

[Cliff Schecht]

One of the major advantages of a current-mode switcher is that they have good "transient response". A switcher is usually designed to operate over a certain range of input and output voltages and currents. As long as the load (all the tubes and what have you) stays within the range designated in the specifications, then the SMPS will ALWAYS be ready to provide good, clean power. A current-mode controller does peak-by-peak monitoring of the switch voltage and will compensate for changes in the output load at the frequency of operation. This means that if we are operating at 100 kHz, the output voltage gets corrected for at 100,000 times per second.

Dynamic current draw is not a problem, but conversely there will be NO sag from the switcher. This will have to be implemented in another form, as we discuss in another thread (dying battery thread).

[davidallancole]

Sounds good to me.