Regulator that Protects the Input

Thread Starter

dreens

Joined Mar 7, 2023
2
Hi all, I'm in a situation where I have to provide power from a well-regulated clean source to something noisy, and I don't want the noise to get back to the clean power source. The well-regulated clean source will be 50 VDC, and the noisy things are compact brushless DC switching fans, four of them. Case constraints make it difficult to use any other cooling technique.

I know I can put the noisy things behind a CLC filter, but it'd be nice to do even better. I wish I could add a voltage regulator, but designed to sortof back-regulate and minimize disturbance on the clean power at the expense of the fans, which shouldn't much care how stable their voltage is. I plan to run the fans in series, with capacitors parallel to each, for a total voltage requirement of 48 VDC.

My suspicion from looking at the operating principle of a typical linear regulator is that they actually present a rather noisy and variable load on the supplying circuit. One alternative I could think of would be to use a current source directing current into the load in parallel with a zener reference. This would seem to minimize noise to the supplying circuit since the load draws constant current at constant voltage. I can find what seem like halfway decent current sources ( NSI45030T1G from on semi ) but they're all designed for keeping LED's evenly bright and the datasheets don't include things I care about like bandwidth or rejection ratio v frequency. This leads me to suspect I'm not on the right track.

Does anyone have any recommendations in this regard? I'd prefer options that keep my BOM smaller, so IC's are preferable to filter cascades. Cost is not a very significant factor, nor is energy efficiency.
 
Last edited:

Papabravo

Joined Feb 24, 2006
21,159
I recommend a combination of common mode and differential mode chokes. Your other schemes do not appear to have easy realizations. You might want to start by putting some numbers on your requirements so you will know what "done" means.
 

LowQCab

Joined Nov 6, 2012
4,029
The Sharp-Switching-Edges are probably the only real problem.
An "adequate" Common-Mode-Choke is probably the best answer.
.
.
.
 

Thread Starter

dreens

Joined Mar 7, 2023
2
Thanks for your thoughts on this guys. I’m having trouble deciding whether to use a differential mode or common mode choke. I don’t think I want to allow either kind of fast transitory. Maybe I should just use a pair of separate inductors?

As far as how to know when I’m done… Thanks for bringing this up. I did some measurements and found that the power supply I had been wanting to avoid making dirty, was already quite full of low frequency noise, especially at 240 Hz as it turned out. About 0.1 V ripple at that freq. I didn’t get it on the spectrum analyzer to check its noise components at all frequencies, but the gist is that it is nowhere as clean as I had thought, and I think a single choke or inductor should be sufficient.
 
Top