The MAX1771 SMPS Development Thread

Started by frequencycentral, October 18, 2009, 02:38:29 PM

Previous topic - Next topic

frequencycentral

I found the circuit to be very unstable on breadboard, a hovering hand above it would cause noise, and touching the MOSFET made the voltage drop considerably. One I moved it onto perf it really settled down, it doesn't appear to mind the proximity of the hovering hand, or even touching the safe parts of the circuit. I'm using to to run a 12AX7 / 12AU7 SE amp into a Hammond 1750A. I'm getting a steady 240v out of it under load. The amp itself is on breadboard. The only real artefacts I've had have been solved by addressing the gain structure and power filtering of the amp itself.

TBH, before I built it onto perf I expected it would be pretty unstable, and had thought about making the unused pads into a ground plane. It really does seem very stable though, I've had the amp running for many hour continuously, just to soak test it. I'm a happy moo cow!
http://www.frequencycentral.co.uk/

Questo è il fiore del partigiano morto per la libertà!

zambo

So this design will deliver a solid 50ma at 240 volts under a load? can it deliver more than 50ma?
I wonder what happens if I .......

bacanador

Switching power supplies are a mistery to me, I studied them a little and was totally turned off by it, and specially because the pcb layout requirements. I wanted a HV power supply and I stumbled across the Nixie power supply, I downloaded the eagle files, and I was about to send it out to be made, but can't find the specified inductor, I picked a different one, but I am going to have to change the PCB footprint.

But now I am curious about the 555 design, didn't even occur to me that the 555 could be used for something like this. And it worked on a perf board, was this luck or this circuit not as sensitive as the other one?

Would it be possible to get up to 300V of either circuit?



bacanador

@frecuencycentral

Sorry I misread, I thought the MAX1771 didn't work on a perfboard. You made it work, but your perfboard work is really impresive. Might not work if I do it.

Cliff Schecht

I really dig this design now and may go a similar route with my future endeavors. I'm convinced that the current-mode boost topology I chose for my first few designs was both unnecessarily complicated and an antiquated design. A modern voltage-mode controller with voltage feedforward is just as safe/stable as a current-mode controller but uses less parts.

What I really want to start working on is a dynamic sag control that would allow you to simply flip a switch/turn a pot to increase/decrease the transient response (may take more trickery to get a tube rec/cap type response though). You could even use an envelope follower to add truly dynamic control of the power supply. I know others have done this, but I don't think it's been done on the scale of an entire tube amplifier before (at least intentionally!). I know Andrew did a sag emulator dealy once, but that's about the extent of what I've seen and I think that was for pedals, not tube gear.

zambo

mostly greek to me but are you talking about controlling the voltage from a knob? That would rule. would it change bias etc though?
I wonder what happens if I .......

Cliff Schecht

Well, no that's not what I'm talking about. But there is a knob to adjust the voltage as necessary. Some resistors and a pot could be used to fine tune the B+ or power scale the amp. This doesn't affect fixed bias circuits because it doesn't generate a fixed bias (this requires a separate circuit).

What I'm talking about is something that allows you to adjust the response time of the power supply. A tube rectifier/capacitor combo doesn't react instantly to changes in power needs, there is a delay. This causes the output voltage to sag for a bit, adding some distortion and compression to the signal. Most people don't like this, but we as guitar players do. What I want to do is accurately emulate this in a DC-DC converter and actually make the amount of sag adjustable. I have a few ideas for either switching in sag or dialing it in, but this is all just speculative still (as is my mini HV SMPS :D).

zambo

ok now i get it. I was not aware of that to be honest and i always wondered why a tube rectifier was considered to better or worse than solid state.....Thanks for the clue!
I wonder what happens if I .......

Pigyboy

Help Rick, Cliff, Nick, somebody,
Hey guys I have the SMPS on  pcb and followed Nicks layout comments. I have the proper voltages but am getting lots of ringing. Any ideas?
Thanks
And you'll have to admit, I'll be rich as shit
I'll just sit and grin, the money will roll right in....
                                                            - FANG

Cliff Schecht

Bad ringing on the output? This shouldn't be a problem unless you have excess parasitics somewhere (long, skinny traces for the power routes for example). What inductor did you use? If the Q is too high at your frequency (well frequencies since this isn't a fixed PWM SMPS), you could be resonating the crap out of the inductor (a small series resistance can drop the Q). I doubt this is the problem though. It could also be a bad solder joint. I can guess all day long but posting some scope shots of the ringing would really help out.

Also remember that tacking on a scope lead adds a lot of extra parasitics and could cause ringing that isn't really there.

frequencycentral

Not so technical as Cliff's reply, but when I was doing mine there was an optimum 'sweet spot' of the trim pot where I got clean output, not neccesarily the highest voltage but the best compromise between decent voltage and no noise.
http://www.frequencycentral.co.uk/

Questo è il fiore del partigiano morto per la libertà!

Pigyboy

#51
Ok, that helps. I was looking for a clean signal all the time. How sweet is the sweet spot? And you are getting it from the pictured layout? I talked to Nick de Smith and he suggested these
http://uk.farnell.com/murata/bnx002-01/filter-dc-power-10a/dp/9526943?Ntt=bnx002
Little rcl filters. I don't know if I have enough room to use them but you might try them. He also pointed out how important it is use a two sided board to separate the ground plane, mmmmmkay,
Back to the drawing board. Thanks, that does really help.
And you'll have to admit, I'll be rich as shit
I'll just sit and grin, the money will roll right in....
                                                            - FANG

frequencycentral

Well it's pretty 'sweet'. I slightly tweaked the schematic to work with a 1k trimmer instead of 10k, so I used different resistor ratios in the voltage divider to get the trimmer to work over the most useful range.

Something else to bear in mind is that the ringing you are hearing may not be from the SMPS, but from the (tube) audio circuit itself. If I try real hard I can get either or both of motorboating (low frequency) and squealing (high frequency) from a high voltage tube circuit. So you may want to look at the topology of your audio, ie rein back the gain between stages by looking at increasing cathode resistor values and/or grid stopper values, even by removing cathode bypass caps. You might also want to seperate the B+ feeding the different triodes into B1, B2 etc in the classic tube amp way of a voltage drop resistor and electro cap to ground. This should help a lot. Especially if you have in-phase triodes (1 and 3, or 2 and 4) sharing the same B+.

Getting HV tube audio to work well with SMPS is a big juggling act, you change one thing then you have to compensate somewhere else. There are many possible combinations that won't work well, it's a case of making small changes and noting down what differences there are until you blindly stumble upon (if you're like me  :icon_biggrin: ) the specific set of compromises that equates to the the optimal formula. Suspect everything - don't blame the SMPS until you're satified that it's not the audio.
http://www.frequencycentral.co.uk/

Questo è il fiore del partigiano morto per la libertà!

Cliff Schecht

This is really true when the SMPS is designed as a variable one. The requisite frequency compensation is setup and optimized for one DC level and so shifting around the feedback resistors will have a direct effect on stability. The best results will be achieved by overdesigning the power supply for your needs (keeps the heat down) and not making it uber-adjustable. Knowing what your needs are and designing to these specifications will almost always lead to better results, especially with touchy SMPS's. Perhaps I'll eventually get around to designing something for the forum..

Pigyboy

Thanks for the help everyone. I will go back to basics and start again.

Hey Rick, I like the tattoo joke!
I will let you all know how I am coming along.
Cheers,
Chris
And you'll have to admit, I'll be rich as shit
I'll just sit and grin, the money will roll right in....
                                                            - FANG

zambo

Is anyone making these for sale? I would love to just buy one with an instruction sheet. Just wondering. High voltage is intimidating already. Making this seems a little more intimidating.
I wonder what happens if I .......

defaced

There are some nixie power supplies on eBay that are similar to this circuit.  I haven't used them yet, but their price seems very reasonable. 
-Mike

zambo

If your refering to the 1363, 1364 from taylor electronics yeah ive used them. They are cool but it seems like unless you have them loaded up just right they whine like a mother. This design also is higher voltage and delivers more ma from 9 volts.Thanks though, always nice to get pointed in the right direction..
I wonder what happens if I .......

cfcubed

#58
First post, to a bit older thread & not really about a stompbox but somewhat relevant I think.  Since I got something out of this thread thought I'd post some info from some tests of mine related to tube hybrid headphone amps.

Not a real designer, more of an assembler,  I did the (mostly reference) implementation of the heater SMPS in Cavalli Audio's CTH.  SMPS heater was the only option to meet our stated goals for the CTH & its working well.  
It got me thinking why not investigate boost/HV SMPS for B+.   Could not find a lot of info around about this so figured I'd play around with it.

Anyway attached is a pic of something I've got mostly working, actively loaded grounded cathode / IRF510 SE w/all power from a 12.6v 1.5A DC supply.
I say mostly because as zambo relates, the taylor 1363/4 is not appropriate for the task (high freq whine).  Almost passable @ 200v only if heavily loading it (Rk / Rak @ 300R) so its switching freq rose to the upper limit of hearing & shielded the unit (use of a shielded inductor would have helped).  E.g. my teenage son can barely perceive the HV freq in headphone output with no signal source.  Play something through it & its pretty good.  BTW made some changes after this pic.



Well I decided to order the built ogi lumen implementation of Nick's MAX1771 project to see if that's up to this task.... But @ $65 USD its a bit pricey.

defaced

QuoteI say mostly because as zambo relates, the taylor 1363/4 is not appropriate for the task (high freq whine).  Almost passable @ 200v only if heavily loading it (Rk / Rak @ 300R) so its switching freq rose to the upper limit of hearing & shielded the unit (use of a shielded inductor would have helped).  E.g. my teenage son can barely perceive the HV freq in headphone output with no signal source.
Hmm.  That kinda sucks, I bought two of these finally hoping to do tube preamps with them, but completely forgot I recommended these and to check back on the thread.  So... and I'm pretty sure I know the answer, but the whine, is that coming from the SMPS unit itself or is it hash in the DC power or both?  I'm pretty sure the answer will be both but I figured I'd ask. 
-Mike