max1771 9-16v in and upto 250v out

[Ice-9]

I 've been researching into powering 12ax7 and minitubes lately. Needing to get a high voltage for the plate of the valve was my main task. I started with %^&*croft walton ladders, but since moved on to researching SMPS with a 555 timer chip.  After plenty googling i found there are chips designed specifically for this type of psu.

I came across this little circuit which uses a maxim MAX1771 chip to switch a FET. Its capable of getting 250v from a 9v source with a very small component count. The articles i found were to power nixie tubes.
I was just wonderering if anyone else has had any fun playing with this type of PSU circuit.

Here is a schematic i have drawn up from the articles i have read.

[frequencycentral]

Looks interesting Mick, but why not just go with something like this one: http://www.ledsales.com.au/kits/nixie_supply.pdf

The parts count is similar, and 555 are cheap and easy to get.


EDIT: We talked about this already didn't we...!

[Ice-9]

Yeah Rick i had seen that in a thread here. They are very similar circuits, i was thinking that maybe the maxim one might be more efficient using a chip designed for that use rather than the 555.

[frequencycentral]

I bet that Maxim chip is expensive! Though it would seem to have a regulator built in that the 555 circuit uses the voltage divider and BC547 to do. Other than that , you're right - they are very similar. I wonder if the Maxim does have any advantage though.

Also, can you post the schematic a bit bigger please?!

[Ice-9]

I think the chips cost about £3-4 . Here is the full article with the schematic in it.

http://www.desmith.net/NMdS/Electronics/NixiePSU.html

[frequencycentral]

Zvex used a negistor based PS for the Nano, 5 components apparently:

i needed a switching power supply to provide my nano head with 230VDC from the 12VDC input (i decided to use DC so it could run off a gel cell for busking in the subway, or deep-woods woodshedding using your car lighter).  i looked around for switching circuits and everything was far too big... nothing would fit in the "two postage stamp" area i had reserved for the power supply, so i decided to build one from scratch using the negistor as the oscillator (never used in a commercial product before, just a laboratory curiosity).  i had a lot of experience with negistors and recognized the pulse-width v frequency possibilities.  the oscillator was three components... a negistor (2N2222A), one resistor (adjustable to set the operating frequency), and one ceramic oscillator cap.  to this, add a blocking cap and diode to direct the pulses, and we have 5 components.

negistors can provide a "pure" sawtooth when left nearly unloaded, but with loading, they change their waveshape significantly, almost to a pulse.  i needed to square up and strengthen this pulse, so i loaded the negistor with a high-gain transistor in clipping mode... excess loading and maximum gain are easy if you leave out unneeded resistors.  8^)  one transistor and one resistor.  now we're up to 7 components.

by changing the bias on the buffer transistor following the negistor using feedback directly from the output at the HV capacitor, i could change the pulse width and regulate the voltage...  so as the output voltage climbed, the pulses were pinched off to limit the time spent charging the coil.  so i created a little feedback/protection circuit using 4 high-voltage zeners and a resistor to feed d.c. bias back into the buffer/loading transistor following the negistor.  4 zeners, one resistor, and one trimmer to control how much bias was making it back for regulation, to set the output voltage.  6 components here for a total of 13.

another thing i had to prepare for was when the battery input voltage began to fall, say from a peak of 15 VDC to a low point of 9VDC (as you will often see in gel cells).  over this range the negistor oscillator's frequency would naturally drop (because it's a VCO) which would increase the "on" time spent charging the coil, boosting the output voltage.  because this happened automatically, zero extra components.

of course, there were two complimentary transistors needed to provide high-current/high speed drive to a power mosfet feeding a coil, resulting in 4 more components, and then a high-speed rectifier to take the power pulses off the coil and store them in the power caps, for a total of 5 components. now we're up to 18.

between these two types of regulation, the feedback network and the negistor VCO, i could control the output voltage within 5% with a 50% change in input voltage over a range of 0mA to 17mA output draw on the high voltage side.  i did all of this with 18 components, all discrete, in the space of 2 postage stamps, using zero data sheets.

i could have used chips and done a more practical mathematical modern approach to the thing, by reading datasheets and trusting modern methods, but i would have been forced to create a daughter board and even then it might not have fit and would have upped my costs per unit significantly.  i figure there's always a solution to doing things more simply and smaller out there, more often than not, in forgotten/obsolete/"lab curiosity" and maverick approaches.  often, data sheets slow me down and distract me from simpler possibilities.  8^)

Do-able? Replicatable?


EDIT: OK, maybe 18 then!

[The Tone God]

Using a switching supply with audio circuits, and in particular circuits with random current demands like amplifiers, is really difficult. In developing the switching supply for the TubeUlent we tried out pretty much every commercial solution out there including controller ICs. I still have a box full of test parts and circuits. Every thing had some kind of issue be it noise, supply response, custom parts, consistency, etc. I decided that it would take less effort to design my own supply from scratch then to try to use some oddball part(s) with a bunch of hacks to solve various cascading problems. The result is something that I have full control over and can tailor to fit my needs of the product.

In short this is a long and difficult road ahead especially if you are not well versed in power supply technologies.

Andrew

[frequencycentral]

Using a switching supply with audio circuits, and in particular circuits with random current demands like amplifiers, is really difficult. In developing the switching supply for the TubeUlent we tried out pretty much every commercial solution out there including controller ICs. I still have a box full of test parts and circuits. Every thing had some kind of issue be it noise, supply response, custom parts, consistency, etc. I decided that it would take less effort to design my own supply from scratch then to try to use some oddball part(s) with a bunch of hacks to solve various cascading problems. The result is something that I have full control over and can tailor to fit my needs of the product.

In short this is a long and difficult road ahead especially if you are not well versed in power supply technologies.

Andrew

Well I'm pleased with the PS in my humble  "Murder One"  offering, but that's just 70 volts, as that's all the 5672 tube I used could handle.

Any hints or pointers Andrew? What do you think of this one for an amp? http://www.ledsales.com.au/kits/nixie_supply.pdf

[The Tone God]

Part of the problem with a switching supply in the DIY community is that it is not like building a pedal where you drop in a set of common parts and the circuit should work. In an effect some substitutions will work and in most cases the worst case scenario is that the circuit will not function. In a switching supply parts matter, physical layout matters, circuit design matters, what you are hooking the supply to matters, what you are feeding into the switching supply matters. It is a big monster on to itself. If testing transistors for gains and biasing them to work in a circuit seems like alot then a HV switching supply should not be attempted. Your better off grabbing a transformer and using a time tested power supply circuit. Many problems will be solved with this.

Some of you are running into a common problem with beginners. You are thinking too much about voltage. Voltage is not was should be focused on. Current is the big thing in this application. Sit down and come up with a power supply design criteria. What input and output voltage, current needs, supply regulation and response, and many other factors will decide the parts that go around the clock. Caps, switching element, inductors, etc. will be dictated by the design criteria of power supply. These are big things that NO fancy controller will solve for you.

Andrew

Edit: Me no type english good

[Minion]

Also with these types of switching suplies if the switching frequency is in the audio frequency range the switching noise will probably leak into the audio signal ......

Useing 2 small Back to back 12v transformers it a good way of getting high voltage and 12v Filament voltage .....

[The Tone God]

Also with these types of switching suplies if the switching frequency is in the audio frequency range the switching noise will probably leak into the audio signal ......

And not just the core frequency but the harmonics resulting from this frequency, and the inductor's discharge characteristics, and the switching element's response time, and...

Andrew

[Ice-9]

The noise generated and picked up by the audio circuit is one of the problems that concerened me as the inductors do emit a lot of RFI, you can use a shielded inductor but the efficiency will drop dramatically. I have not done much in the way of psu design before but its strange how you can get quite hooked into psu design. (always considered it a boring, secondary neccessity). I've always avoided circuits with inductors as they have been a mystery to me in the past, so this is also all about learning.

Please remember these circuits can output a Very High voltage and are dangerous so be careful.