Controlling Inrush Current in a Power Supply

[D from BC]

I got a case of: Thermistor Vs Something Else

+----(high side lim)--+-----+
| | |
\/\/\/170VDCpk 180uF Rload~85ohm
| | |
+-----(low side lim)--+-----+
|
Com

I'm not too familiar with inrush thermistors.
I know they get hot.. which can suck.
Don't hotter electronics die sooner...??

I looked at a thermistor app note...
http://www.epcos.com/web/generator/...ta__en.pdf;/PDF_InrushCurrentLimiting_AN2.pdf
groan...

The design process looks like a pita at first glance.
Bad enough I'm turned off by the heat emitted.

I'm more familiar with the math for transistor limiting
https://www.onsemi.com/pub/Collateral/AND8164-D.PDF
And...it's spiceable..

What's the popular circuit these days for controlling the inrush
current to bulk capacitors in power supplies?

D from BC

 

[Phil Allison]

"D from BC"

I got a case of: Thermistor Vs Something Else

+----(high side lim)--+-----+
| | |
\/\/\/170VDCpk 180uF Rload~85ohm
| | |
+-----(low side lim)--+-----+
|
Com

** 180 uF and 85 ohms is not a good combination.

The ripple voltage will be huge - about 50% or 80 volts p-p.

I = C dv/dt



......... Phil

 

[D from BC]

"D from BC"

** 180 uF and 85 ohms is not a good combination.

The ripple voltage will be huge - about 50% or 80 volts p-p.

I = C dv/dt



........ Phil

oops...maybe I oversimplified that schematic....


+----(high side lim)--+--------+
| | |
\/\/\/170VDCpk 180uF reg
| | |
+-----(low side lim)--+--------+
| |
Com1 Com2

reg = DC-DC convertor acting kinda like an 85ohm resistor.

But yeah...there's lots of ripple but the convertor doesn't drop out.
D from BC

 

[Phil Allison]

"D from BC"

oops...maybe I oversimplified that schematic....

reg = DC-DC convertor acting kinda like an 85ohm resistor.

But yeah...there's lots of ripple but the convertor doesn't drop out.

** Have some consideration for the poor, 180uF cap.

A 2 to 5 ohm NTC is all you need for inrush surge limiting - one rated for
about 4 to 5 amps max will be fine.




........ Phil

 

[D from BC]

oops...maybe I oversimplified that schematic....


+----(high side lim)--+--------+
| | |
\/\/\/170VDCpk 180uF reg
| | |
+-----(low side lim)--+--------+
| |
Com1 Com2

reg = DC-DC convertor acting kinda like an 85ohm resistor.

But yeah...there's lots of ripple but the convertor doesn't drop out.
D from BC

oh crap....took awhile to sink in..
Not 85ohms....there's a load of 260ohms across the cap.
Anyways...I don't think it'll affect the surge current issue too much.

Either way, there's still going to be a big current spike when power
is applied.
D from BC

 

[D from BC]

"D from BC"



** Have some consideration for the poor, 180uF cap.

A 2 to 5 ohm NTC is all you need for inrush surge limiting - one rated for
about 4 to 5 amps max will be fine.




....... Phil

Yeah...high ripple current---> self heating--->shorter lifespan
I goofed the load value and posted that prior to this post.


About the thermistor..
Wow.. only 2 to 5 ohm R @25C (for the original posted circuit)

I was thinking if I clicked on the power switch just at the mains peak
then the current spikes to:
170V/5ohm = 34 amps

However...I suspect that's about as good as it gets...

D from BC

 

[colin]

I got a case of: Thermistor Vs Something Else

+----(high side lim)--+-----+
| | |
\/\/\/170VDCpk 180uF Rload~85ohm
| | |
+-----(low side lim)--+-----+
|
Com

I'm not too familiar with inrush thermistors.
I know they get hot.. which can suck.
Don't hotter electronics die sooner...??

I looked at a thermistor app note...
http://www.epcos.com/web/generator/...ta__en.pdf;/PDF_InrushCurrentLimiting_AN2.pdf
groan...

The design process looks like a pita at first glance.
Bad enough I'm turned off by the heat emitted.

I'm more familiar with the math for transistor limiting
https://www.onsemi.com/pub/Collateral/AND8164-D.PDF
And...it's spiceable..

What's the popular circuit these days for controlling the inrush
current to bulk capacitors in power supplies?

D from BC

NTC thermistor makes a good inrush current limiter,
but it only limits it when its cold,
as it heats up it disipates less energy,
its best to insulate it so it doesnt wats emuch enrgy,
however it then takes a long time to cool down.

when its hot a short power interruption will give no protection against
inrush.

a smart way is to have a mosfet acros the NTC,
well actually we used relay back in the days ....

ofc if you have a mosfet there it could control the inrush all on its own
anyway.

Colin =^.^=

 

[Phil Allison]

"D from BC"

About the thermistor..
Wow.. only 2 to 5 ohm R @25C (for the original posted circuit)

I was thinking if I clicked on the power switch just at the mains peak
then the current spikes to:
170V/5ohm = 34 amps

** Most folk make the mistake of thinking inrush surges ought to be
completely eliminated.

Why ?????

A 250 watt incandescent lamp surge is up to 35 amps at switch on.

A 300 VA transformer surge is up to 100 amps at switch on.

A TV set or CRT monitor surges up to 50 - 100 amps at switch on, due to the
de-gauss circuit.


Each of the above are household items PLUS the inrush surge duration they
generate is ** waaaaay ** longer than that of a SMPS cap charging.

The most insidious example around now are CFLs ( ie compact fluoros) where
the inrush surge is anything up to 50 amps for a lamp that draws only 20
watts when running !!

Put 10 or 20 on the same domestic lighting circuit and see what happens to
the switch and cct breaker.



........ Phil

 

[D from BC]

NTC thermistor makes a good inrush current limiter,
but it only limits it when its cold,
as it heats up it disipates less energy,
its best to insulate it so it doesnt wats emuch enrgy,
however it then takes a long time to cool down.

when its hot a short power interruption will give no protection against
inrush.

a smart way is to have a mosfet acros the NTC,
well actually we used relay back in the days ....

ofc if you have a mosfet there it could control the inrush all on its own
anyway.

Colin =^.^=

It's certainly on the flip side to thermally insulate a part..
Usually I'm bolting semi's to heatsinks..

NTC across the mosfet??
Like:

+---------------------+-----+
| | |
\/\/\/170VDCpk 180uF Rreg
| | |
+-----NTC----+--------+-----+
| | |
+-----S[_D---+ Com2
| G
| |
|--?----|
Com1

I'm not sure what the control is like...
I'm guessing the NTC reaches max conduction and then the mosfet turns
on with a latch circuit.
This lets the NTC to cool down to R 25C.
A mains power disruption resets the latch...When power is back on..the
surge current is limited by the NTC.

Is that the idea?
D from BC

 

[Phil Allison]

"colin"

NTC thermistor makes a good inrush current limiter,
but it only limits it when its cold, as it heats up it disipates less
energy,
its best to insulate it so it doesnt wats emuch enrgy,
however it then takes a long time to cool down.

when its hot a short power interruption will give no protection against
inrush.

a smart way is to have a mosfet acros the NTC,
well actually we used relay back in the days ....

** Using a relay (or mosfet ?? ) gives no protection against inrush
surges if the interruption is briefer than the time to reset the device to
its open condition and AC supply interruptions can be as brief as only 1
cycle.

If the inrush surge current of some item is so * horrendous * that it must
never happen, then a missing cycle detector is essential, followed by
instant shut down and eventual re-start of the equipment.

See my other post re widespread inrush surge paranoia.


BTW

Use a damn spell checker - there are no less that six errors in the above
quote from you.




........ Phil

 

[D from BC]

"D from BC"


** Most folk make the mistake of thinking inrush surges ought to be
completely eliminated.

Why ?????

A 250 watt incandescent lamp surge is up to 35 amps at switch on.

A 300 VA transformer surge is up to 100 amps at switch on.

A TV set or CRT monitor surges up to 50 - 100 amps at switch on, due to the
de-gauss circuit.


Each of the above are household items PLUS the inrush surge duration they
generate is ** waaaaay ** longer than that of a SMPS cap charging.

The most insidious example around now are CFLs ( ie compact fluoros) where
the inrush surge is anything up to 50 amps for a lamp that draws only 20
watts when running !!

Put 10 or 20 on the same domestic lighting circuit and see what happens to
the switch and cct breaker.



....... Phil

I didn't see the trees in the forest

Thanks


D from BC

 

[colin]

NTC thermistor makes a good inrush current limiter,
but it only limits it when its cold,
as it heats up it disipates less energy,
its best to insulate it so it doesnt wats emuch enrgy,
however it then takes a long time to cool down.

when its hot a short power interruption will give no protection against
inrush.

a smart way is to have a mosfet acros the NTC,
well actually we used relay back in the days ....

ofc if you have a mosfet there it could control the inrush all on its own
anyway.

Colin =^.^=

It's certainly on the flip side to thermally insulate a part..
Usually I'm bolting semi's to heatsinks..

NTC across the mosfet??
Like:

+---------------------+-----+
| | |
\/\/\/170VDCpk 180uF Rreg
| | |
+-----NTC----+--------+-----+
| | |
+-----S[_D---+ Com2
| G
| |
|--?----|
Com1

I'm not sure what the control is like...
I'm guessing the NTC reaches max conduction and then the mosfet turns
on with a latch circuit.
This lets the NTC to cool down to R 25C.
A mains power disruption resets the latch...When power is back on..the
surge current is limited by the NTC.

Is that the idea?

yeah thats the idea, the mosfet just turns on as soon as the caps have fully
charged.
maybe just a short delay.
the mosfet will ofc see the full 170v.
the ntc will disipate whatever power it needs to reach a certain
temperature,
so bolting it to a heatsink will just make it disipate loads of power.

insulating it will not make it run much hotter.
you can get them poted in an insulating capsule.

the mosfet just makes it so it limits inrush during a short power
interuption,
and saves a little bit of power,
most equipment that claims to be inrush protected just has the NTC.

Colin =^.^=

 

[Eeyore]

insulating it will not make it run much hotter.

That depends ENTIRELY on the power dissipation and the thermal resistance of the insulator.

Please don't make unquantified silly claims like the above.

Graham

 

[MooseFET]

I got a case of: Thermistor Vs Something Else

+----(high side lim)--+-----+
| | |
\/\/\/170VDCpk 180uF Rload~85ohm
| | |
+-----(low side lim)--+-----+
|
Com

I'm not too familiar with inrush thermistors.
I know they get hot.. which can suck.
Don't hotter electronics die sooner...??

I looked at a thermistor app note...http://www.epcos.com/web/generator/Web/Sections/ProductCatalog/Nonlin...
groan...

The design process looks like a pita at first glance.
Bad enough I'm turned off by the heat emitted.

I'm more familiar with the math for transistor limitinghttps://www.onsemi.com/pub/Collateral/AND8164-D.PDF
And...it's spiceable..

What's the popular circuit these days for controlling the inrush
current to bulk capacitors in power supplies?

If you have the room, great big inductors are kind of nice for inrush
limiting. You put a low resistance in parallel with them so that
the resistance charges the capacitors before the inductor saturates.

I have seen a truely huge case where the capacitors were charged up in
3 steps with relays shorting out resistances. The relays were closed
as the voltage neared fully charged.

 

[colin]

That depends ENTIRELY on the power dissipation and the thermal resistance
of the insulator.

Please don't make unquantified silly claims like the above.

Graham

As a NTC heats up above some temperature its resistance drops
so it disipates less,
in other words it behaves like a temperature control device
when used as a series element.

so insulating it has its benefits in reduced power loss.

ofc there are limits to how low the resistance goes etc,
and the sharpness of the drop.

Its not a claim but a brief explanation of how the thing behaves.

If you want quantified claims the OP has to specify quautities such as
running current
and max inrush current, then you can look up a datasheet for a suitable
device.

Colin =^.^=

 

[D from BC]

As a NTC heats up above some temperature its resistance drops
so it disipates less,
in other words it behaves like a temperature control device
when used as a series element.

so insulating it has its benefits in reduced power loss.

ofc there are limits to how low the resistance goes etc,
and the sharpness of the drop.

Its not a claim but a brief explanation of how the thing behaves.

If you want quantified claims the OP has to specify quautities such as
running current
and max inrush current, then you can look up a datasheet for a suitable
device.

Colin =^.^=

At current, the NTC in my app is the hottest component. A little
annoying for a 90% eff convertor.

I find the thermal insulation option interesting to keep the heat off
other components (if worthy) or the reduce the temp. inside the box.

I think I get the idea...
I'm imagining that a thermal/power equilibrium is reached at a lower
resistance when the NTC has insulation.
In other words...
NTC open air power disp > NTC insulated power disp.
(I^2*R1) > (I^2*R2)
where
R1=NTC in open air
R2=NTC insulated

But...slower cool down as you mentioned before.
D from BC

 

[Fred_Bartoli]

D from BC a écrit :

At current, the NTC in my app is the hottest component. A little
annoying for a 90% eff convertor.

I find the thermal insulation option interesting to keep the heat off
other components (if worthy) or the reduce the temp. inside the box.

I think I get the idea...
I'm imagining that a thermal/power equilibrium is reached at a lower
resistance when the NTC has insulation.
In other words...
NTC open air power disp > NTC insulated power disp.
(I^2*R1) > (I^2*R2)
where
R1=NTC in open air
R2=NTC insulated

But...slower cool down as you mentioned before.
D from BC

Why not simply dropping the NTC and put there a true current limiter
(mosfet+bjt).
With a 68mR source resistor you'll get pretty low dissipation, 10A surge
limiting, and 'instantaneous recovery' since there's no memory.
With an additionnal cap, you can even make it a 'slow start surge
current limiter'. First connect -> no current 'surge'. Fast connect
after disconnect -> surge limited to 10A or whatever.

 

[D from BC]

If you have the room, great big inductors are kind of nice for inrush
limiting. You put a low resistance in parallel with them so that
the resistance charges the capacitors before the inductor saturates.

I have seen a truely huge case where the capacitors were charged up in
3 steps with relays shorting out resistances. The relays were closed
as the voltage neared fully charged.

Nah...
The inrush problem is in the front end of an offline smps.
I did alot of work to reduce the power inductors for the smps...
Not gonna ruin the design with a relatively massive inductor.
It's like a tumor.

I going to use a thermistor.

But I'm still interested in alternative ideas...


D from BC

 

[Eeyore]

Why not simply dropping the NTC and put there a true current limiter
(mosfet+bjt).

Because a thermistor in China costs tuppence ?

Graham

 

[D from BC]

On Mon, 06 Aug 2007 18:37:00 +0200, Fred_Bartoli

[snip]

Why not simply dropping the NTC and put there a true current limiter
(mosfet+bjt).
With a 68mR source resistor you'll get pretty low dissipation, 10A surge
limiting, and 'instantaneous recovery' since there's no memory.
With an additionnal cap, you can even make it a 'slow start surge
current limiter'. First connect -> no current 'surge'. Fast connect
after disconnect -> surge limited to 10A or whatever.

I'll probably do that when I find out that the NTC heat is a zit in
the design.
Otherwise, an active inrush circuit just bloats the design.

On the PCB..I had to move the NTC away from anything heat sensitive.

About the cct..
Oh yeah...bjt BE junction across a I sensing R (68mR)...yadda yadda
yadda...
My 1st choice for active inrush control..

However, I'm using a thermistor for now...


D from BC