1. Home
  2. Electronics Design
  3. Inductor saturation during smps startup

Inductor saturation during smps startup

I was looking at using an LT1170 in a Positive-to-Negative Buck-Boost Converter, as per page 14 of this data sheet

formatting link

Thing is, when I put this into LTSpice (using Linear Techology's own model of their part), it shows that the current through the inductor rises to 10 amps, during start-up, settling to something more reasonable later. I don't want to have to use a 10 amp inductor, and the part they suggest isn't 10 amps anyway. I can limit the current by including a clamp circuit on the Vc connection, but it increases the component count, and would probably need to include a trimpot because the relationship between current limit and clamp voltage isn't specified.

Does it matter if the inductor saturates during start-up?

Note that in the spice schematic below, I've included the switching on and off of the 2 amp load at regular intervals to look at the transient response, and regulation (the latter not being good at all, though good enough in my application).

Sylvia.

Version 4 SHEET 1 1704 788 WIRE 96 -64 64 -64 WIRE 256 -64 176 -64 WIRE 368 -64 256 -64 WIRE 528 -64 368 -64 WIRE 368 -48 368 -64 WIRE 528 -48 528 -64 WIRE 64 16 64 -64 WIRE 64 16 -144 16 WIRE 64 32 64 16 WIRE -144 96 -144 16 WIRE -144 96 -160 96 WIRE 256 112 256 -64 WIRE 256 112 192 112 WIRE -160 128 -160 96 WIRE 256 176 192 176 WIRE 320 176 256 176 WIRE 368 176 320 176 WIRE 368 224 368 176 WIRE 480 224 448 224 WIRE 544 224 480 224 WIRE 640 224 608 224 WIRE 752 224 720 224 WIRE 816 224 752 224 WIRE 896 224 896 128 WIRE 896 224 816 224 WIRE 992 224 944 224 WIRE 224 240 192 240 WIRE 240 240 224 240 WIRE 320 240 320 176 WIRE 480 256 480 224 WIRE 816 256 816 224 WIRE 752 272 752 224 WIRE 896 272 896 224 WIRE 992 272 992 224 WIRE 1136 272 1040 272 WIRE 1040 288 1040 272 WIRE 240 352 240 320 WIRE 944 352 944 224 WIRE 944 352 896 352 WIRE 1040 352 1040 336 WIRE 1072 352 1040 352 WIRE 1136 352 1072 352 WIRE -160 416 -160 192 WIRE 16 416 -160 416 WIRE 64 416 64 320 WIRE 64 416 16 416 WIRE 240 416 64 416 WIRE 320 416 320 320 WIRE 320 416 240 416 WIRE 368 416 320 416 WIRE 480 416 480 320 WIRE 480 416 368 416 WIRE 560 416 480 416 WIRE 752 416 752 336 WIRE 752 416 624 416 WIRE 816 416 816 336 WIRE 816 416 752 416 WIRE 992 416 992 352 WIRE 992 416 816 416 WIRE 368 464 368 416 WIRE 304 512 304 496 WIRE 368 560 368 544 FLAG 368 16 0 FLAG -208 16 0 FLAG 528 32 0 FLAG 368 560 0 FLAG 16 416 NOMGROUND FLAG 256 176 FB FLAG 1072 352 0 FLAG 224 240 VC FLAG 896 128 0 SYMBOL diode -208 32 R270 WINDOW 0 32 32 VTop 2 WINDOW 3 0 32 VBottom 2 SYMATTR InstName D1 SYMATTR Value 1N4148 SYMBOL res 192 -80 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName 470 SYMATTR Value 470 SYMBOL cap -176 128 R0 SYMATTR InstName C1 SYMATTR Value 1E-6 SYMBOL cap 352 -48 R0 SYMATTR InstName C2 SYMATTR Value 100E-6 SYMBOL res 224 224 R0 SYMATTR InstName R2 SYMATTR Value 5000 SYMBOL cap 224 352 R0 SYMATTR InstName C3 SYMATTR Value 0.1E-6 SYMBOL res 304 224 R0 SYMATTR InstName R3 SYMATTR Value 1200 SYMBOL cap 464 256 R0 SYMATTR InstName C4 SYMATTR Value 2.2E-7 SYMBOL res 352 208 M90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R4 SYMATTR Value 10000 SYMBOL diode 608 208 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName D2 SYMATTR Value 1N914 SYMBOL schottky 624 400 R90 WINDOW 0 0 32 VBottom 2 WINDOW 3 32 32 VTop 2 SYMATTR InstName D3 SYMATTR Value MBR745 SYMATTR Description Diode SYMATTR Type diode SYMBOL res 736 208 R90 WINDOW 0 0 56 VBottom 2 WINDOW 3 32 56 VTop 2 SYMATTR InstName R5 SYMATTR Value 47 SYMBOL cap 736 272 R0 SYMATTR InstName C5 SYMATTR Value 1000E-6 SYMBOL res 800 240 R0 SYMATTR InstName R6 SYMATTR Value 470 SYMBOL res 880 256 R0 SYMATTR InstName R1 SYMATTR Value 6 SYMBOL voltage 528 -64 R0 WINDOW 123 0 0 Left 2 WINDOW 39 24 124 Left 2 SYMATTR SpiceLine Rser=0.5 SYMATTR InstName V1 SYMATTR Value 16 SYMBOL ind 352 448 R0 SYMATTR InstName L1

SYMATTR SpiceLine Ipk=7 Rser=0.013 Rpar=0 Cpar=0 mfg="Coiltronics" pn="CTX50-7-52" SYMBOL sw 992 368 R180 WINDOW 3 24 -12 Left 2 SYMATTR Value "" SYMATTR InstName SW SYMATTR SpiceModel MYSWITCH SYMBOL voltage 1136 256 R0 WINDOW 123 0 0 Left 2 WINDOW 39 0 0 Left 2 SYMATTR InstName V2 SYMATTR Value PULSE(0 2 20e-3 1e-6 1e-6 10E-3 20E-3 10) SYMBOL PowerProducts\\LT1170 64 176 R0 SYMATTR InstName U1 TEXT -232 464 Left 2 !.tran 100E-3 TEXT 688 528 Left 2 !.model MySwitch SW(Ron=.1 Roff=1Meg Vt=1)

Reply to
Sylvia Else
Loading thread data ...

What is the total resistance, thus what is the current? I wouldn't assume s pice will tell you unless you put _all_ the parameters in manually. An indu ctor won't mind overcurrent pulse during start up, they're slow thermally. The chips are rated 2.5-11A, so once you know your current you'll know whet her you're in safe territory. Usually soft start &/or current limiting take care of it.

snip spice, I've not looked at it.

NT

Reply to
tabbypurr

Bucks used as inverters are vulnerable to that problem, and to dragging down the input rail on start-up. We had a thread on undervoltage lockouts for the LM2549 a month or so back, and that's the fix.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

Soft-start works too, and some chips allow UV and SS to be combined.

Sylvia's sim crashes, some part missing, CTX50 something.

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

Agreed, unless somebody connects it to some lab supply with a low current limit _and_ a slow turn-on. (That's one way I test SMPSes these days.) Anything will work if you connect it to a car battery. ;)

If you use the buck as a buck, it's really just charging the output and reservoir caps, and it doesn't matter much if the inductor saturates doing that, assuming decent internal current limiting on the switch.

Didn't try running it.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Well, don't do that! The real nasty is when a switching wart drives a box full of switchers, which can devolve to a foldback supply driving a negative resistance. Bad load line. Or expensive oscillator.

(That's one way I test SMPSes these

--

John Larkin         Highland Technology, Inc 

lunatic fringe electronics
Reply to
John Larkin

Reminds me of the Bob Newhart classic:

My stuff gets used by grad students a fair amount, so eventually somebody will lose the wall wart and power it off some cheesy bench supply. In the QL01 (1 uA FS, shot noise limited above 10 nA in 1 MHz), I'm using an UVLO circuit on the inverting supply, and running the pilot LED off the negative supply, so it doesn't come on unless all the supplies are OK.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Sylvia,

May be i missed something but when i see the current in the storage inductor (I(L1) 50uH) I see it flows undirectional --> bad thing !

So for me it will saturate not only at startup but chances are that it will after few m(b)illions of cycles (let say few seconds or minutes) ... (No complete de-magnetization of the core).

With my experience of that sort of weird micropower things it will works on conditions ... or will overheat after hours of rated current.

Why don't you use a traditional flyback converter ? Many of my designs based on flyback topologies are working nicely and quietly (modern EMC requirements)

OTH Habib.

Reply to
habib

You could do dim for prime power available, bright for everything OK. I think some new laws discourage warts from having LEDs, to save energy!

I'm designing a new box now and hope to keep the number of power rails/pours down to a reasonable number... like 16 maybe. I usually have a schematic sheet or two that is POWER SUPPLIES but in this one we may have a bulk supply sheet and then a lot of local regulators scattered on various sheets, closer to their regional loads. That concept would extend to the physical layout.

It's going to be maybe 32 sheets of schematic, with analog, three FPGAs, Tiny Logic, ECL, LVDS, discretes, all that jumbled up. In a thing like that, the power supplies can become a nightmare. Air flow and cooling increase the fun.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

Nah, that's your standard buck connected as an inverter, with a bit of decoration. Basically you take a normal buck, ground the output, and let it pump its own ground pin negative. Works great if you allow for a few quirks, such as the inrush.

You want the current to be both unidirectional and continuous--it's basically a duty cycle integrator, with the inductor and output cap forming a lowpass filter.

The dynamics are a bit more complicated than a vanilla buck converter, because the chip sees an input voltage of V1-Vnomground, which varies all over the place during startup.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Fun.

I often have 4 or 5 even on a purely analogue board, more if you count bootstrap pours. Super quiet +/- for the front end, medium quiet +/- for the output amp, maybe +5 for some comparators and stuff. If I'm using dpots or something like that, they need their own separate quiet 5V.

The present one one is just an uber-bootstrapped nanoamp TIA, no series-connected photodiodes or noise cancelling or anything.

Sure makes it easy for the layout guy.

What is it?

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net
Reply to
Phil Hobbs

Well, sort of a pulse generator.

--

John Larkin         Highland Technology, Inc 
picosecond timing   precision measurement  

jlarkin att highlandtechnology dott com 
http://www.highlandtechnology.com
Reply to
John Larkin

I sort of guessed that. ;)

Cheers

Phil Hobbs

Reply to
Phil Hobbs

That's how I draw schematics, as well. Global regulators go on the regulators page but point-of-load regulators go next to the part they're supplying. BGAs have a separate block for power, so the POL regulator goes next to it.

That's where all the fun is.

Reply to
krw

Strange - I'm pretty sure it's only using standard stuff from LTSpice. I was running a year old version of LTSpice, but I've just downloaded the most recent version, and the simulation still works.

Sylvial

Reply to
Sylvia Else

On considering the various comments, and allowing for the fact that I'll probably have two of these running off the same mains powered SPMS, I think I'll implement the current limit clamp, arrange for the two to start in sequence, not concurrently, and not load either of them until they're both stable.

Sylvia.

Reply to
Sylvia Else

  • I would think that would create a magnetic bias in the core, enough to alter operational integrity..
Reply to
Robert Baer

Works fine for me, too.

Reply to
John S

The data sheet for that inductor (CTX50-7-52) says it is rated for slightly more than 10A. I don't think you have a problem.

Reply to
John S

Ok John i missed to explain on what I'm being hurt on that app note. The current waveform in the storage inductor smells funny ! In all DC/DC converters the current in an inductor increase or decrease smoothly (a ramp traditionally) although voltage (dV(t)/dt) across an inductor could be high. I repeat a square waveform current in the inductor reveals something very weird. Another thing i always been uncomfortable with LTC chips design is that ... there are no theory justification :-( "Do something ... and after do somehow anything ... R1, R2, comme-ci

I spent a long long time on DC/DC theory and learn the most valuable knowledge back in old days when Unitrode was a chips maker on power area.

LT has released a remarkable tools as LTSpice but for chips ... mmhhh i can't just used them when clients are requiring DDR (Detailed Design Review) from me. As an example please take a look at the LT3575 and please explain me how it works in details. We spent a long time in 2009 (Rolls Royce Civil Nuclear) to figure out how it works ... Finally a Spice simulation with hidden models is nothing less than a verification and nothing more or less.

Indeed.

Habib.

Reply to
habib

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.