Motor Controller Power Supply

[dknguyen]

So the last bit for me to solve on my motor controller is this pickle...
I would like it to run off of 6V-42V input (preferably 6V-50.4V). This is 2 cell LiPo batteries all the way up to 10 cell Lipo (12 cell A123) or 12cell Lipo(14cell A123).

Anyways, the catch is that I also need a 5V/18V supply for the logic and gate drive circuits. The 5V isn't too hard to do. Just a buck converter which barely meets the dropout requirements. But the 18V is quite a bit harder.

Now, it seems no SEPIC, buck-boost, or flyback can do that and meet the input voltage rang. It is looking like I have to get a buck converter to step it down to power all the logic and then split off some of the output to a boost converter to bring it up to 18V.

Now...I'm not sure how I should go about this. Because I could power the logic and 18V boost converter straight off of the buck supply but that might makes some noise for the logic and create bad frequency interactiosn between two cascaded switching converters (what might happen here exactly? I'm not sure).

Or I could add filtering to isolate the two loads from each other after the buck converter.

I don't think using a LDO to isolate the logic from the boost converter will work because the switching frequencies are too high and can't be supressed by any linear regulator. Plus it also means the buck will be operating in dropout on a loaded 2 cell Lipo battery (not necessarily a problem since I searched high and low and found a PFET 100% duty cycle capable buck converter) and I might be able to squeeze the buck and LDO dropout into the little headroom I have above 5V and the battery input.

 

[dknguyen]

FIlters are too much hassle and uncertainty. It looks like I'm going to have to force myself to work with some MSOP-8 packages. Then I can easily get some boost converters with a synch-pin and use it to sync it to the buck converter. Problem solved.

 

[Hero999]

What about using a flyback transformer topology?

It's the same as the usual flyback except you take the power from a separate secondary coil wound on the outside of the inductor.

A 1:1 transformer will probably be suitable to both step-up 6V to 18V and reduce 50V down to 18V. It's probably better if it's wound bifilar style as it will geive the better coupling.

What current are you taking from the 18V rail?

I take it, you only 20mA or so for some op-amps?

If it's an off the shelf solution you're looking for a pulse transformer might even do for such a low current.

 

[dknguyen]

30mA from the 18V rail. I'm trying to avoid transformers since four of them need to go on a quadrotor helicopter.

I'll look into a flyback transformer though. If it would let me buck-boost with just one coil. It'd be nice to have a 18V supply that can run straight off the battery rather than sitting behind the buck converter, and SEPIC converters need 2 coils. It's also convenient that boost ICs (like the synchable one I am going to use can also be used for flyback...and SEPIC).

EDIT: Looks promising. THe main part being the duty cycle would be from 25%-75%, whereas the boost following the buck would always be running at 75% duty cycle which is a bit too close to the limits for my liking.

 

[dknguyen]

I found a tapped flyback transformer that will allow me to generate 18V and 6V simultaenously and I can just follow the 6V with a 5V regulator. SO I could replace everything with just one converter, but the transformer is rated for 18V-55V. I could still input 6V and it would still do it's job right? (Similar to the triac question asked recently by someone else)

https://www.coilcraft.com/pdfs/a0009-a.pdf

Could it be that the only reason the 18V is specified is because it is the limiting factor for the output rating of 3.3V@3A? (ie. the input coil can only take 3.3V/18V * 3A = 550mA?)

 

[dknguyen]

It seems I might have to go with the transformer method with a split transformer to produce all output voltages because

I don't understand this app note:
https://www.electro-tech-online.com/custompdfs/2009/01/AN-1082.pdf
about synchronizing a boost converter's clock to the output of a buck converter, but it also mentions that the input to the buck converter has to be in continuous mode and that the input voltage shouldn't vary by more than 10%- something which I can't guarantee.

 

[Ubergeek63]

It seems I might have to go with the transformer method with a split transformer to produce all output voltages because

I don't understand this app note:
C:\Documents and Settings\Dan Nguyen\Desktop\AN-1082.pdf
about synchronizing a boost converter's clock to the output of a buck converter, but it also mentions that the input to the buck converter has to be in continuous mode and that the input voltage shouldn't vary by more than 10%- something which I can't guarantee.

upload or link please. I have seen the technique used in PFCs and can explain it but am not sure if that is what they are talking about. synchronizing a boost before a buck ensures that the boost supplies as much of the power as possible instead of the hold up cap.

Dan

 

[dknguyen]

lol that's weird why is the link to my HD?

https://www.electro-tech-online.com/custompdfs/2009/01/AN-1082-1.pdf

I think I kind of understand it...now, but still not enough. THe node at the diode cathode goes HI because of the transistor turning on, and it has to be continuous so the schottky diode is always conducting to pull the node close to LO when the transistor is off. I need it to be able to work in discontinuous mode though and I still don't know why the 2/3 of the resistors are there.

 

[Ubergeek63]

lol that is strictly an EMI app. you do not need to worry about it. in PFC apps if you turn on the buck switch when you turn off the boost switch, you ensure the energy is going right to the output instead of being stored in the HV cap. It allows you to downsize it a bit.

Dan