Variable Bench Switch Mode Power Supply

That'll be a 'no' then, I guess! What about less ambitious but still widely variable switch mode designs?

Scrim

By the time you get involved in designing at the 2KW power level, you should have enough sense to realize that it's a lot easier, and makes more sense, to address the specific requirement. At that power level, the power supply itself is a major portion of the design.

There's nothing to stop you hacking into the control circuitry of fixed-output designs, to produce output variations that are either voltage or current regulated. You will discover the limitations of the devices involved, when you do.

These limitations are sometimes addressed commercially through modularization - where for a specific application, modules with varying characteristics are selected and applied in series or parallel. For example, flyback converters give good voltage range compliance, forward converters give good current range control and linear variations can provide noise-critical performance.

Switchers are generally not good at regulating around zero voltage or zero current, unless they are configured for bipolar output voltage/current or can dissipate/store returned load energy.

RL

Thanks for that. I'm really after a variable supply though for general lab use where I need from milliamps to tens of amps. Usually an amp or two suffices but recently I wanted to test several Peltier devices and needed a supply of unknown high current capability at between 8 and 28 volts. Mainly I was wondering if any of the hobby magazines had published a switch mode bench supply design.

Scrim

That sounds more like 300W, with a voltage compliance that is within the range of conventional integrated circuitry, driving external power devices - cost shouldn't be an issue, if you stick to the immediate application requirement. The bulk source of isolated DC may already be available to you.

Something like an LT1339 or LT3740, running off a simple bulk 48Vdc source with a 12V control supply rail, should do it. LTC offers app notes, pre-assembled evaluation printed circuit boards and even free simulator software for these parts. They are intended to control external mosfets for up to ~20A of buck-regulated current. As synchronous switch controllers, they are capable of discharging load capacitance under transient conditions.

Such information is regularly presented in publications like EDN, EETimes or ECNews

RL

Thanks again. Lowering my power output expectations somewhat, here's an example of the type of thing I'd really like, and for cost reasons, to build myself:

I expect that would find a permanent place in my equipment cupboard.

Scrim

Nice. BTW i used some supplies very like that some 25 to 30 years ago, they were made HP back when they still made stuff. Single output, up to 40 V, up to 8 A, about 140 W max, GPIB.

HP Used to be one of my favorite companies. I particularly liked the how the few products from them I had included circuit diagrams and lots of info to help you calibrate and repair stuff. And encouragingly the manuals included errata rather than them pretending everything's perfect.You certainly paid for it though!

I think I'm going to try the idea of hacking an ostensively fixed voltage switch mode PSU. Breaking into the voltage control loop shouldn't be too hard if it's a fairly dumb supply, but I suspect I'll have to build my own constant current circuit externally. Would be well worth it if it worked though!!!

Alan

And you may be able to get the original schematics from HP for that old thing if you can find the model number.