5V DC to +/- 15V DC

[cncjerry]

I need a DC to DC converter made from discrete components.  I had been using a monolithic converter but I keep popping them and until I get the 15V opamp circuit under control, I would like to use cheap devices.  I only need 50 to 100ma out of the +/- 15V.  I looked around and see some high voltage circuits but not something in this range. 

Thanks

Jerry

[T3sl4co1l]

Why not buy one that's current limited?

Tim

[SiliconWizard]

You could use some kind of isolated DC/DC converter such as the Murata NDTD0515C: https://power.murata.com/en/ndtd0515c.html . It has short-circuit protection. About 15 bucks and it just works.
Don't know whether isolated could be a problem in your case or not, but those Murata are sturdy parts.

[Benta]

Use a Traco Power module for prototyping your opamp circuit. After that, you have time to spin your own.

[cncjerry]

I was trying to keep the cost low as I've blown those converters up a couple of times.  I figured if I could build one with a couple of transistors I could pop them in when needed.

Thanks

[Buriedcode]

At >50mA charge pumps are pretty much out so you're left with either a boost or push-pull with custom magnetics or an off-the-shelf module with built in current limiting as suggested.  What volume are we talking here?  If it's really that cost sensitive you could just go with a cheap(ish) converter and add your own current limiting. 

[Wolfgang]

Hi,

I made one myself to power a low noise amp from USB. Look here:

https://electronicprojectsforfun.wordpress.com/power-supplies/a-usb-powered-15v-supply/

Its quite some circuitry, but results are OK.

[Zero999]

I was trying to keep the cost low as I've blown those converters up a couple of times.  I figured if I could build one with a couple of transistors I could pop them in when needed.

Thanks

Getting a current limited DC:DC converter will not fix the problem. Figure out why the 5V to +/-15V converter is failing first.

[T3sl4co1l]

I was trying to keep the cost low as I've blown those converters up a couple of times.  I figured if I could build one with a couple of transistors I could pop them in when needed.

Thanks

Getting a current limited DC:DC converter will not fix the problem. Figure out why the 5V to +/-15V converter is failing first.

Um?  That's basically the cause and solution in two sentences.

FYI, if a DC-DC converter doesn't specifically say it's current limited or short protected, it will release the magic smoke if you try.  They're usually a two-transistor chopper configuration, no control, no regulation, no limiting at all.

Particularly common in the very small ones, that can't afford any room (or budget) for those functions.

Tim

[Zero999]

I was trying to keep the cost low as I've blown those converters up a couple of times.  I figured if I could build one with a couple of transistors I could pop them in when needed.

Thanks

Getting a current limited DC:DC converter will not fix the problem. Figure out why the 5V to +/-15V converter is failing first.

Um?  That's basically the cause and solution in two sentences.

FYI, if a DC-DC converter doesn't specifically say it's current limited or short protected, it will release the magic smoke if you try.  They're usually a two-transistor chopper configuration, no control, no regulation, no limiting at all.

Particularly common in the very small ones, that can't afford any room (or budget) for those functions.

Tim

Whoops. I meant to say figure out why the device the DC to DC converter is powering, is drawing too much current and causing it to fail. Replacing the DC:DC converter with a current limited one will not solve the problem. The device it's powering will just cause it to shut down or current limit and it still won't work. Fair enough, it offers the ability to do some debugging, but it will not solve the problem.

I'd advise building/buying a +/-15V current limited power supply supply. It's easy to make one with a 15-0-15V transformer, a bridge rectifier, some big capacitors and the LM78L15 and LM79L15. Note I suggest the L versions to limit the current to <200mA and they're cheap too. Now the circuit can be debugged and put back into the original system, once it's working.

[T3sl4co1l]

Whoops. I meant to say figure out why the device the DC to DC converter is powering, is drawing too much current and causing it to fail. Replacing the DC:DC converter with a current limited one will not solve the problem. The device it's powering will just cause it to shut down or current limit and it still won't work. Fair enough, it offers the ability to do some debugging, but it will not solve the problem.

Ah right, that makes more sense

I assumed the problem was clumsy probes and debugging, so the problem will be solved in due time (heh, hopefully?).  If this were a replacement sort of thing for example, yeah, something else is broke, yo.

Tim

[Ian.M]

An add-in 100mA current limiter for debugging could be quite simple - use the classic OPAMP + N-MOSFET constant current sink circuit for the -15V rail, and its P-MOSFET mirror as a constant current source for the +15V rail, All its going to need is a dual rail to rail OPAMP, the MOSFETs on a small heatsink, a couple of Zeners for Vgs clamping and a few passives.   

It will spend most of its time with the OPAMP outputs stuck at opposite rails as it tries to turn the MOSFETs on harder to increase the current, but when you hit the limit, after a very short surge while the OPAMP output unsticks from the rail and starts slewing hard towards the MOSFET Vgs threshold voltage it will cut the current back to its limit.   Foldback limiting could be added with an extra resistor from each OPAMP output to its control voltage (in+) input (from a resistor divider between the rail and ground) to cut back the commanded current once the output has unstuck from the rail, sized so it never takes the commanded current below say 10mA.

Even some 50mA polyfuses, + clamping diodes so you cant drag the + rail below ground or + rail above ground would probably save your day.

[TiN]

I used LTC LT3439 for isolated DC-DC for exactly those voltages, was very happy with low noise and easy design. Also uses off-shelf transformer, no need DIY stuff.
Crude prototype board, showing regulator (on the right). Just 9 parts in BOM:

I had been using a monolithic converter but I keep popping them

What makes you think that discrete regulator would be less poppable? 

[ebclr]

Just in case

https://www.ebay.com/itm/5V-6V-9V-12V-15V-24V-Positive-Negative-Dual-Output-power-supply-DC-DC-Step-up/183250767422?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

[DaJMasta]

There's also the option of wiring in a lab power supply and when it works, you have an easy way to readout your current requirements and pick an appropriate module.

[Ian.M]

Yep, and if power sequencing is an issue, but not excessively critical, simply run the lab PSU feeds through a DPST relay with its coil driven by the 5V rail, with schottky clamping diodes from the output rails to 0V so none can cross 0V due to slight differences in the relay contact open or close times.

[cncjerry]

Just in case

https://www.ebay.com/itm/5V-6V-9V-12V-15V-24V-Positive-Negative-Dual-Output-power-supply-DC-DC-Step-up/183250767422?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649

How did you know I bought a half dozen of these?

The +/- 15V is for a curve tracer I am redesigning based on one published in QST two years ago.  I have Parkinsons so I shake and probing has become difficult. The unit I've been using is  a Murata MEA1D0509DC.  They are 5.45 per and I need two for +/- 18v though +/- 15 works as well.  I have two Universal Micro units that are large and work fine but I don't want to blow these.  They are pricey and purchased for another project.  I've blown three of the Murata units already and to this day I don't now why, they are on a breadboard and they just stop working and I'm out 5 bucks.  I am using the Universal units now and I was thinking of building the curve tracer into a TM5000 unit which would solve the power problem.  But in the meantime, I was thinking of building my own converter if I could find a circuit.  That way if I pop a small signal transistor it's no big deal.

Anyway, thanks for the input.

Jerry