SELV smps eval kit for control algorithm design

[nimish]

Does a low voltage (<48Vac) dev platform exist that allows for safe development of PFC and other control algorithms without worry of isolation breaches exist? I want to explore some ideas but I also don't really want to deal with isolated probes and other safety reqs for handling energized components. The idea would be that the SMPS would be built as if it was dealing with mains but actually just takes input from a stepped down AC source.

[Faringdon]

You'd have to just buy a big mains transformer and use that as your input.
Youre better off just using mains and a diff probe though...wear safety specs in case of explosion.
Be careful not to touch stuff.
But yes you can use any  PFC chip with a 48VAC input...just change the dividers etc.
But you really want to be with the real mains.....partly because the feedback loop depends on the mains level...the multiplier does a vin^2 calc at least.

Safety regs are AYK,  not applicable to you  if you are just alone in a lab...do as you please.
I recomend proper ESD shoes...so you  dont get electrocuted and dont esd damage stuff.....dont touch 2 places at once on the pcb.......to be honest, i have had strong Lhand to Rhand 240vac mains.....it made my pecs vibrate!...but was painless and did no harm......but yes, be very very careful...specially with HVDC!!!

I am not sure, but i wonder if most deaths from 240vac are from  the fall from the surprise and hitting their head on the floor....but HVDC is bad for you cuzz theres no zero cross. My worst shock was off a charged mains cap...but it only hurt in the spot where i had  bashed a ligament when falling off a bike....not sure if woudl have still have hurt if i hadnt already the injury

If you are doing feedback loop stuff, you want to be with the real mains......as it will be in your application...with the same output caps , inductor value, and all.

[NiHaoMike]

I had the idea of a "educational" version of a grid tie inverter that connected via an AC wall wart, in order to make it much safer to play with.

[nimish]

You'd have to just buy a big mains transformer and use that as your input.
Youre better off just using mains and a diff probe though...wear safety specs in case of explosion.
Be careful not to touch stuff.
But yes you can use any  PFC chip with a 48VAC input...just change the dividers etc.
But you really want to be with the real mains.....partly because the feedback loop depends on the mains level...the multiplier does a vin^2 calc at least.

Safety regs are AYK,  not applicable to you  if you are just alone in a lab...do as you please.
I recomend proper ESD shoes...so you  dont get electrocuted and dont esd damage stuff.....dont touch 2 places at once on the pcb.......to be honest, i have had strong Lhand to Rhand 240vac mains.....it made my pecs vibrate!...but was painless and did no harm......but yes, be very very careful...specially with HVDC!!!

I am not sure, but i wonder if most deaths from 240vac are from  the fall from the surprise and hitting their head on the floor....but HVDC is bad for you cuzz theres no zero cross. My worst shock was off a charged mains cap...but it only hurt in the spot where i had  bashed a ligament when falling off a bike....not sure if woudl have still have hurt if i hadnt already the injury

If you are doing feedback loop stuff, you want to be with the real mains......as it will be in your application...with the same output caps , inductor value, and all.

I'm ok with the minor changes due to voltage, but it's less about regulations and more about me not wanting to electrocute myself. 400V is nothing to sneeze at esp since it's DC as you said!

I might have to build it myself with a 10:1 stepdown transformer and a 24V scaled down version

I had the idea of a "educational" version of a grid tie inverter that connected via an AC wall wart, in order to make it much safer to play with.

Baffles me why there isn't one already. Some of the videos of big companies selling transistors in SMPS have some truly cowboy shit going on near dangerously high voltages. One guy just picked up an inductor coil barehanded without checking for it bring energized...scary

[Faringdon]

Yes, or you coudl set it up in a perspex enclosure, and have a contactor on the door, such that the mains doesnt even get to it till the door is closed and its not possible to actually touch it.
Because even with a mains transformer...youre going to have exposed live mains somewhere?
Admittedly not HVDC

[nimish]

Yes, or you coudl set it up in a perspex enclosure, and have a contactor on the door, such that the mains doesnt even get to it till the door is closed and its not possible to actually touch it.
Because even with a mains transformer...youre going to have exposed live mains somewhere?
Admittedly not HVDC

Of course, and the interlock system would, I'd think, be basic practice in a commercial lab...having a little bit of live mains is something we have to deal with but having a huge wirewound toroid inductor just out there on a desk seems like inviting trouble.

I did find: https://www.microchip.com/en-us/development-tool/dv330101 which fits the bill. $500 though :O

[schmitt trigger]

A good isolation transformer is your best friend when designing and troubleshooting AC mains-level power supplies. 

[T3sl4co1l]

The thought has occurred to me, also in context of say hi-fi audio, where there's no shortage of enormous (often toroidal) transformers with terrible, terrible capacity -- because they love to use them with even more enormous capacitors, on the assumption that that's better for bass capacity, or the reduced ripple helps, or something.  So the power factor is abominable, like 0.2 or something.

You know, audiophile stuff, it doesn't need explanation, it just is...

So, there might even be a small but useful market for such things.  It'll be a while before I have time to dedicate to play with such a circuit, though.

If nothing else, you can take an off-the-shelf eval kit, and scale all the power-facing components appropriately.  L down, R down, C up.  If the inductors have taps/windings, adjust ratios accordingly.  Pick appropriate transistors and diodes.  Everything done to ratio, the controller doesn't even know anything has changed (think in terms of invariants like this: it's a very useful approach in general!).  Assuming the controller doesn't need high voltage itself, of course (so, avoid the self-starting types, I guess).

You won't be able to run much higher current than the board was designed for, of course -- PCB tracks aren't scaling here, obviously!  But you'll get roughly the same dynamics (e.g. switching loop stray inductance giving same peak voltage, if dI/dt is the same and Coss isn't much higher), so that might be handy too.

Tim

[nimish]

A good isolation transformer is your best friend when designing and troubleshooting AC mains-level power supplies.

Isolation transformer doesn't do jack against a 400V DC Bus capacitor

The thought has occurred to me, also in context of say hi-fi audio, where there's no shortage of enormous (often toroidal) transformers with terrible, terrible capacity -- because they love to use them with even more enormous capacitors, on the assumption that that's better for bass capacity, or the reduced ripple helps, or something.  So the power factor is abominable, like 0.2 or something.

You know, audiophile stuff, it doesn't need explanation, it just is...

So, there might even be a small but useful market for such things.  It'll be a while before I have time to dedicate to play with such a circuit, though.

If nothing else, you can take an off-the-shelf eval kit, and scale all the power-facing components appropriately.  L down, R down, C up.  If the inductors have taps/windings, adjust ratios accordingly.  Pick appropriate transistors and diodes.  Everything done to ratio, the controller doesn't even know anything has changed (think in terms of invariants like this: it's a very useful approach in general!).  Assuming the controller doesn't need high voltage itself, of course (so, avoid the self-starting types, I guess).

You won't be able to run much higher current than the board was designed for, of course -- PCB tracks aren't scaling here, obviously!  But you'll get roughly the same dynamics (e.g. switching loop stray inductance giving same peak voltage, if dI/dt is the same and Coss isn't much higher), so that might be handy too.

Tim

This is likely what I'll have to do. Do you have any resources on scale invariant SMPS designs? Everything I have seen is very tied to the absolute values of the components.

[T3sl4co1l]

Hmm, maybe that would make a good article...

In the mean time: the design equations are your symmetries.  Inductance scales inversely with Fsw and Ipp (current ripple), and proportionally with V_L (Vin, Vout-Vin, whatever the topology applies to L).  And likewise for resistors, and inversely for capacitors.  A controller doesn't care what it's driving or sensing, as long as it sees the voltages it expects; for example, you can use a regular boost controller (e.g. UC3843) for an inverting Cuk converter, by wiring a unity gain inverter using an outboard opamp, then you get the same normalized 2.5V feedback signal, no problem.

Tim

[nimish]

Hmm, maybe that would make a good article...

In the mean time: the design equations are your symmetries.  Inductance scales inversely with Fsw and Ipp (current ripple), and proportionally with V_L (Vin, Vout-Vin, whatever the topology applies to L).  And likewise for resistors, and inversely for capacitors.  A controller doesn't care what it's driving or sensing, as long as it sees the voltages it expects; for example, you can use a regular boost controller (e.g. UC3843) for an inverting Cuk converter, by wiring a unity gain inverter using an outboard opamp, then you get the same normalized 2.5V feedback signal, no problem.

Tim

Ah ok, makes sense. I'd read those articles. A lot of the intro/intermediate stuff just elides the design equations or just focuses on implementation quirks. Not much for the hobbyist aside from scanning papers and theses

I'm looking at implementing a wacky multilevel PFC + active EMI filtering and active power decoupling, so no off the shelf controllers unfortunately. Have to implement control algorithms by hand, so I want something that I can safely handle for debugging

The

[T3sl4co1l]

Interesting; link us some reports if you are able

Tim