Driving a Buck-Boost Converter With A Switching Supply / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Driving a Buck-Boost Converter With A Switching Supply

previous 1 2

Move Thread

LAN_403

hen918

Thu Dec 06 2018, 01:35AM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

A Mazzilli should be fine with a SMPS (in general) so long as there is adequate capacitance on the output of the supply so the current limiting kicks in. Unfortunately if the supply starts to current limit, the ZVS ceases to oscillate, and "hangs-up" resulting in dead MOSFETs. As with any cascaded SMPS, the operating frequencies should be far away from each other, and any harmonics, although it's easier to just give it a go and see whether it works.

GeordieBoy

Mon Dec 10 2018, 01:08PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

The potential problem with driving one DC-DC converter from the output of another SMPSU is this...

DC-DC converters with a fixed load typically appear as a constant power load to whatever is driving them. This means that they will draw more current at low input voltage, and draw less current at high input voltage. Therefore they have a negative incremental resistance at their input port. If this DC-DC converter is used as the load for an up-stream switching power supply, the negative resistance has the potential to destabilise it's control loop.

It isn't a very common problem but can occur if the up-stream switched-mode power supply has a relatively small output capacitance and it's closed-loop stability is marginal to start with. Typical solutions if it does turn out to be an issue include adding smoothing capacitance between the two converters, and adding a small fixed resistive load to counteract the negative resistance of the downstream DC-DC converter's input.

-Richie,

Sulaiman

Mon Dec 10 2018, 03:09PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

"Therefore they have a negative incremental resistance at their input port."

Clearly true but I'd never considered this aspect.

I've reverse-engineered many industrial electronics power trains and I can't recall ever seeing a resistor in line used for anything other than for reducing power dissipation in linear regulators.
I guess that most designers just throw more capacitance at the problem ?

GeordieBoy

Mon Dec 10 2018, 03:59PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

I think the damping resistor normally goes across the supply rather than in series.

Adding more capacitance usually helps but you have to be careful where it moves the double-pole of the first converter's output filter with respect to the poles and zeros of it's closed loop compensator. It's a long time since I used to work on this stuff, so I don't remember the details now. I think the problem is less common these days due to current-mode control being more popular now.

There is another similar stability issue with DC-DC converters that is also little known, called Middlebrook criteria:

This explains how just adding an input-side filter for EMI suppression can result in unstable operation when combined with the negative resistance of the converter's input port.

-Richie,

Patrick

Mon Dec 10 2018, 06:21PM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

GeordieBoy wrote ...

.......with respect to the poles and zeros of it's closed loop compensator. It's a long time since I used to work on this stuff, so I don't remember the details now. I think the problem is less common these days due to current-mode control being more popular now.

yes, and CMC doesnt work well as it approaches 50% duty cycle. so most regulators have thier feedback taking into account a roll over to conventional VMC above 50% DC. You really need to model this with bode plots or make sure the load is controlled better than the power supply.

hen918

Mon Dec 10 2018, 10:13PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

Patrick wrote ...

This depends entirely on the topology of the SMPS, some topologies work best with VMC, some CMC, some will use a feed-forward voltage mode design, some will use a combination.
The OP isn't trying to design an SMPS, just use one, and the easiest way to see if it'll work is to try it.

Patrick

Tue Dec 11 2018, 07:11AM

Registered Member #2431 Joined: Tue Oct 13 2009, 09:47PM
Location: Chico, CA. USA
Posts: 5639

Active loads ( potentially negative resistance) may not allow a simple "try it and see" method to work. it may work one day and not another, with little rhyme or reason.

though i do agree. try it. if it works go with it. but dont be surprised if it gets fugly.

GeordieBoy

Tue Dec 11 2018, 12:42PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

If it was my project, I would lash it all together and do some load step tests to check for any possible stability issues with dummy loads.

Then deal with any stability issues if they are apparent, before using the arrangement to power anything more sensitive/expensive.

-Richie,

RickR

Thu Dec 13 2018, 02:07PM

Registered Member #93 Joined: Thu Feb 09 2006, 04:11PM
Location: Houston, Texas
Posts: 64

So, the votes seem to favor hooking it up and seeing what happens. That seems to be a reasonable approach. I'll try things out over the Christmas/New Year's break.

Now the question is - How would instabilities present themselves? Would the output voltage start wandering? Same for the output current? If I start adding inter-supply capacitance for stability should I start in 100s to 1000s of microfarads range, or is that something I just need to fool around with as well, since the actual values might vary by load?

Thanks again for the helpful responses and suggestions.

- Rick

previous 1 2

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob