Buck boost converter confusion / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

Buck boost converter confusion

1 2 3 4 next

Move Thread

LAN_403

Inducktion

Wed Apr 03 2019, 11:09PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Hi all

Trying to understand this application circuit from the LTC1625 datasheet.

NK4yck

I understand they're trying to do synchronous rectification with the P channel fet to increase efficiency, but my question is...

Would it still work with higher output voltages? Say, 30 volts max?

I know right off the bat the gate drive IC would become very unhappy with 30 volts input; obviously, one would disconnect it and feed it straight 12 volts or something as an example.

But, otherwise, would the P channel fet still turn on? P channel mosfets have always confused me and to this day I still don't quite understand their operation.

Appreciate the help.

DerAlbi

Wed Apr 03 2019, 11:26PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

P-Fets are actually like N-Fets, just the opposite. If you know how to turn on a light bulb with an N-Fet, then built the same circuit with the P-Fet, and connect the supply voltage in reverse

Everything will work.
Regarding your question, i think it can be solved at system level without talking about the circuit:
At 12V output voltage, loosing 0.6V over a diode is, right there, a 5% systematic efficiency loss. So synch. rectification makes sense to get back some efficiency.
At 30V output voltage, this does not matter so much anymore (0.6V = 2% efficiency hit)

So, i would say, dont care if this works at 30V because it does not make sense to use a mosfet then anyway.

But you are correct, if you would leave the P-Fet in, it would not really work in that exact circuit. Actually, the LTC1693 dies first, since it only can live below 13.2V. So this limits your output voltage here.

Inducktion

Thu Apr 04 2019, 12:12AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Is it worth synchronous rectification even with 10 amps output?

With 10 amps, over a 0.6 v drop schottky, gives about 6 watts of power dissipated.

With 10 amps through a 0.02 ohm PFET, that gives about 2 watts of power dissipated, plus the added complexity of the circuit...

There's another application circuit without synchronous rectification.

J24vwa

Should I just stick with this version, non-sync and not deal with the gate driver nonsense?

DerAlbi

Thu Apr 04 2019, 12:33AM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

You can scale that to any current you want, it will always remain the same argument. The efficiency impact is proportional to the voltage drop in relation to the overall output voltage. No matter if you draw 1A or 30A. In your example, if you assume 30V, 10A, then you deliver 300W. Do you really care about 4W savings somewhere? Even 2W on a component will most likely need some form of cooling (at least consideration). Once you are there, you can also cool the 6W

I cant really give more advice without knowing your constraints. I would try to reduce complexity. Its hard enough. So yeah, go for a diode for a start.

2Spoons

Thu Apr 04 2019, 05:06AM

Registered Member #2939 Joined: Fri Jun 25 2010, 04:25AM
Location:
Posts: 615

To complement what DerAlbi is saying, synchronous rectification might be worthwhile if you absolutely have to have max efficiency - maybe because you are running on a battery or other limited power source, or maybe that extra 4W is something you just can't get rid of easily.
Horses for courses, as usual.

Inducktion

Thu Apr 04 2019, 05:19AM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

Gotcha. I'll keep it simpler. Gonna do some low side drivers for the "low" side fets (the fets I have are higher gate capacitance than the datasheet says the SMPS chip can handle on its own), and just use a low V-drop Schottky.

Which leads me to two more questions;

1. Why do really low v-drop schottkys have absolutely abysmal reverse leakage current? There were some I was looking at, that when "hot" would leak several hundred milliamps of current!

2. To "add" constant current to this, would it realistically be as simple as just adding a differential amp and a comparator, something like this?

W7hG8i

The potentiometer sets the current "limit" by comparing the amplified voltage to some reference. Is there anything I should be careful of when doing this?

Again thank ya'll for your insight.

DerAlbi

Thu Apr 04 2019, 07:14AM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Modifying the feedback path is not a fun thing to do. The amount of gain you bring in with a comparator will likely lead to oscillation- If you really want to do this, only use the OpAmp, so amplify the shunt voltage to the nominal FB-voltage at your target current. But you should 2 things to help to stabilize the regulation loop:
a) add a resistor between the output of the OpAmp and the FB-Pin. Then connect a capacitance between the output voltage and the FB-Pin. If there is a steep output voltage change, the regulation can work without phase delay from the OpAmp.
b) Your OpAmp adds gain to the feedback loop. Its bandwidth will require limiting . so add a capacitor parallel to the feedback resistor from Out to the inverting input.

----
In all honesty, such modification is BS

Just use a constant current driver, if you need a constant current. LED-Drivers are the way to go.
You still have not really communicated your requirements. But if i only go for the keywords ("buck boot, current regulation) then i would recommend the LT3755 for example. The easiest circuit would be the sepic-configuration depicted at page 17 of the data sheet. Not good for high power however.

johnf

Thu Apr 04 2019, 08:46AM

Registered Member #230 Joined: Tue Feb 21 2006, 08:01PM
Location: Gracefield lower Hutt
Posts: 284

Induktion
you need to look up PID control (proportional integral differential)
this is how gain and phase shift are managed in real world systems mechanical to electronic it doesnt matter.
As for synchronous rectification as others have said it rapidly losses favour as the output voltage goes over 12 volts and would not be used at 48 volts output

Inducktion

Thu Apr 04 2019, 04:14PM

Registered Member #3637 Joined: Fri Jan 21 2011, 11:07PM
Location: Buffalo, NY
Posts: 1068

I see. Well, my requirements:
I'm trying to design an actually decent buck boost module that one could tack on a 12v supply and get good regulation, current capability and reliability. Constant current is a good thing to have which is why I'm trying to add it here. 0 to 30 volts with 10 amps current capability is what I'm shooting for

I know Chinese modules exist that do this sort of thing but I wanted to give designing one myself a try. It's relaxing to layout pcbs and actually try to make something new.

Patrick

Thu Apr 04 2019, 06:12PM

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

Inducktion wrote ...

I know Chinese modules exist that do this sort of thing but I wanted to give designing one myself a try. It's relaxing to layout pcbs and actually try to make something new.

Yes it can be.

Beware of PID loops which are dependent on microprocessors. They can work but if theres an isolation boundary, you can get a 6 volt oscillation on a 12 volt regulated supply thats stable and wont go away... but only magically appears with special load conditions you cant expect. so PID loops can be out-smarted. Thats why they're rarely used. Theres a lot of industry pressure to use PIDs to eliminate the whole Vector Network Analyzer phase of development.

Control loops and SMPSs are a lot of fun for me. but no small matter either. I recommend a feedback loop without isolation and no wall sockets for you. I have to admit a 450 volt 330uf cap bursting with sparks and steam is fun though. (if you still have a house.)

1 2 3 4 next

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