Potentially A Simple Multi kW Bench PSU?

[Boscoe]

Hi all,

I recently made a low power 1mA/1mV accurate bench PSU for debugging and characterising low power (mainly battery powered) electronics. You can see it on my website at http://gswdh.co.uk/gallery.html#psu_project. The electronic design and code are here:

https://github.com/geowal19/PSU-Hardware
https://github.com/geowal19/PSU-Software

It was great fun creating this device and it has proved mighty useful over the past 6 months but (on some jobs) I'm finding myself running out of power. I began thinking of complimenting it with a high power version that doesn't need the high accuracy.

While digging through Digikey I found this interesting power module. It can buck/boost from 9 to 60V in and 0 to 60V out at 50A! The very crucial parameter here being it can output 0V which is very important when building a laboratory power supply in case of short circuit faults. It is a little pricey though.

https://www.digikey.co.uk/product-detail/en/recom-power/RBBA3000-50/945-3463-ND/10413764

However, I think overall a power supply could be built quite simply using one of those telecommunication rectifiers from eBay (easily get 1.5kW at 48V for 100€) and, with a little bit of external control, you could get some CV/CC working with this module. Please see the attached concept.

There's a risk that the constant current control wouldn't work due to potentially a large propagation delay through the power module making the control loop a bit useless. I would love some feedback on this idea, I might implement it.

[coromonadalix]

The rbba3000  is an interesting device, BUT  since its not an insulated converter   it give me some frills,  i need insulation ... for a 3kw  dc-dc  it will be massive

The H60SB0A050NRDC    too

[Boscoe]

Would the mains to 48V converter not offer the isolation you need?

[coromonadalix]

yep   but   you will need an 3kw transformer  ...   if you push the dc-dc  converter to the max, an toroidal x-former will reduce the footprint needed


Edit : it was my 2 cents comment   had failures in the past with non insulated dc-dc converters,  when some of them failed they simply busted the DUT components at their output

The H60SB0A050NRDC     with mcu control could be a great project.

[mawyatt]

Nice work on the "Lab Supply"

What do you need for voltage and current for the higher power supply?

Agree, for lab use you want a fully isolated output and the ability to have 0 volts output.

Maybe a good SMPS like the Mean-Well types with a post linear output regulator for CV and CC?

One idea that I've not tired is to "float" the isolated MW output negative return from the output negative return with a pair of back to back high current diodes. This would allow a small negative voltage relative to the output return to allow "zero volts" at the output, and also handle to output currents. The CV and CC control loops would be referenced to the return outputs and not the MW return. Could be issues and allow Murphy to step in, so maybe some good prior simulations and testing required.

Also, some of the linear Op-Amp type power devices might prove useful, like the LM3866, LM1875, and APEX devices for example.

Anyway, sounds like a fun project, please keep us posted on your progress.

Best, 

[T3sl4co1l]

Huh, why not drive the pins directly?  They're not error amp inputs, and you have direct voltage and current control.  Just pipe the DACs straight in.  It's already a "bench" supply, just add source.

For the ratings, it's not a bad price.  I can't do better myself, not without like, making a modest production run from China; and the result would probably be less compact, or less general (e.g. buck mode only).

Tim

[Boscoe]

It describes 'hiccup' behaviour on over current events. I believe this is turning the output off entirely and reattempting enabling the output every now and again to see if the fault has cleared. This is not CC behaviour. Please correct me if I'm wrong, it would make life a lot easier.

[nctnico]

yep   but   you will need an 3kw transformer  ...   if you push the dc-dc  converter to the max, an toroidal x-former will reduce the footprint needed

No. Just use a AC/DC converter. For example surplus from telecom equipment.

[T3sl4co1l]

It describes 'hiccup' behaviour on over current events. I believe this is turning the output off entirely and reattempting enabling the output every now and again to see if the fault has cleared. This is not CC behaviour. Please correct me if I'm wrong, it would make life a lot easier.

It... well, it doesn't really "describe" it?  It's not clear if the "overload" condition is meant the same as the current limit condition.  Surely it only goes into hiccup at low output voltages or something?  It would be rather bizarre to say that the output current can be "trimmed" if the means of that adjustment is as gross as a hysteresis control or worse.

The datasheet really doesn't discuss it at all; you'll have to find some more documentation hopefully (check mfg website), or get one and see.

Tim

[james_s]

yep   but   you will need an 3kw transformer  ...   if you push the dc-dc  converter to the max, an toroidal x-former will reduce the footprint needed


Edit : it was my 2 cents comment   had failures in the past with non insulated dc-dc converters,  when some of them failed they simply busted the DUT components at their output

The H60SB0A050NRDC     with mcu control could be a great project.

What are you talking about? The telecom "rectifier" he's talking about is a complete switchmode power supply, it provides the isolation, there is no need for a separate transformer.

[coromonadalix]

@James_s

You did not see the module specs, the module does not provide insulation between the input and the output, it operate like a buck boost

It will work fine, but in my case if i want isolation   it's not totally ideal

[voltsandjolts]

A telecom rectifier is a PSU in itself, typically spitting out a fixed 48VDC at lots of amps, and isolated from mains input.

[james_s]

@James_s

You did not see the module specs, the module does not provide insulation between the input and the output, it operate like a buck boost

It will work fine, but in my case if i want isolation   it's not totally ideal

No, I did see the specs, I still don't know what you're going on about. He wants to power the module from a telecom rectifier, those are isolated switchmode power supplies, the isolation is provided by the telecom rectifier, no additional isolation is needed. What part of this do you not understand?