Switching constant current source for electrolysis bath

[ricko_uk]

Hi,
I am designing a 2A constant current source for an electrolysis bath. Because of limited space and no cooling available, I need to use a switching regulator. So I came up with the attached schematic.

The most important characteristic is that the current through the plates has to be very clean DC. Especially free from any switching noise from the regulator.

UPDATE
I have changed the schematic to the attached one. It is now using a synchronous FET SMPS with a resistance of 75mR. That should keep the SMPS IC relatively cool. But just to be sure, how do I calculate the temperature of that SMPS from the FET resistance, current and frequency?

Thank you

ORIGINAL

The most important characteristic is that the current through the plates has to be very clean DC, free from any switching noise from the regulator.

I have chosen the inductor relatively large at 100uH and was also wondering whether to add a second low-pass LC filter after C1. I guess it can clean up the voltage further but would slow the regulator response?

An alternative I also thought of, is to do add the LC as just mentioned but then insert a current sense resistor but between C1 and the LC filter to implement a high-side current sensing for regulator's feedback pin. Would this be a better idea? And if so would it allow me to decrease the inductor value so to use a smaller one?

Thank you

[John B]

That regulator will still need a heatsink at 2A. It has a darlington output. I would look for something with FET switches. You could use a lowside controller, ones even used for LEDs and at 2A there's plenty of ones that will have internal switches.

[moffy]

If you want to clean up some of the high frequency switching noise, a 10uH inductor and 0.1uF ceramic cap following your LC might help without effecting loop dynamics too much.

[ricko_uk]

Thank you John and moffy,

@John,
good point about the Darlington output. I can see various SMPS with FET of 75mR so that should be low enough to keep it relatively cool? How do I calculate the regulator losses from the FET resistance?

By low side LED controller, do you means simething like this? https://www.diodes.com/assets/Datasheets/AL8843Q.pdf

What would the advantage be compared to my solution (if I do use a FET output switch in my schematic)?

@moffy
So there are a no advantages to make the current sensing high side?

Thank you

[John B]

Thank you John and moffy,

@John,
good point about the Darlington output. I can see various SMPS with FET of 75mR so that should be low enough to keep it relatively cool? How do I calculate the regulator losses from the FET resistance?

By low side LED controller, do you means simething like this? https://www.diodes.com/assets/Datasheets/AL8843Q.pdf

What would the advantage be compared to my solution (if I do use a FET output switch in my schematic)?

Something like that chip, yes. Though most I have used will have low side current sensing too.

Low side regulation would be more common because they can use N-channel MOSFETs which will have lower losses compared to a similar high side P-FET. Also the current sensing circuitry is built into the IC, saving you from incorporating a separate opamp and dealing with potential instabilities.

Another option is something like a TL494 controller which gives you more options in slowing down the control loop. You'd combine it with a gate driver and discrete N channel MOSFET. It's an old part, but I've found plenty of uses for it, especially if you have a load that doesn't change rapidly - as I would assume in the case of electrolysis?

[NiHaoMike]

A current mode controller would be ideal for your application.

[ricko_uk]

Thank you both.

I updated the OP with the new schematic following your feedback and more research I did. Also attached here.

I didn't mention that I still need the external amplifier to feed the current to a micro for monitoring. So using the LED driver as suggested by John although nice might not provide that much more advantage.

Any thoughts on this new configuration?

Thank you

[NiHaoMike]

Use 35V caps for the input and output, then you can use larger values for the same cost and size which will give you a higher ripple current rating.

[moffy]

@moffy
So there are a no advantages to make the current sensing high side?

Thank you

It is 6 of one and half a dozen of the other. Generally low side current detection is easier (you don't have to deal with a common mode voltage) but it depends on what you are using, sometimes high side might make sense.

[Circlotron]

The most important characteristic is that the current through the plates has to be very clean DC. Especially free from any switching noise from the regulator.

Why is low noise necessary? i would have thought it would be similar to charging a battery, where noise and voltage ripple are, up to a point, not an issue.