Inrush current-limiting

I've used the big black disk PTCs and, for some reason, they always work.

Or close the switch at the line zero cross.

---------------------------

All they use and need is a 1W, 10ohm resistor in the incoming AC. Use a WW type for reliability if you like.

A 10 ohm NTC would also do, but cost more.

.... Phil

---------------------

The OP made the error first.

..... Phil

what's the max current? I=Q/t= CV/10ms, 200mA? depletion FET?

George H.

It's a painful problem. Ideally input C.L. is built into the smps IC. Otherwise soft-start may handle it. But add an input current sensor, LTC6101, etc., and feed its signal to the FB pin via a threshold-setting resistor and diode.

Thanks Phil.

Correct--I meant 'NTC.'

Cheers, James Arthur

So did I.

But that's too late. This is the old line-->fuse-->rectifier-->reservoir cap., which then feeds the SMPS.

I designed a series-pass voltage-controlled current-limiter to go between the bridge and the filter cap. (The voltage-control aspect is to limit pass-transistor dissipation, rather than just current.) Three BJT's, but the whole thing is messy compared to just one part.

.-----. Line >---[fuse]---|~ +|---[limiter]--+---> to smps. | | | Neut >------------|~ -|---. --- 15uF '-----' | --- === | ===

An NTC is probably more reasonable, or even a series resistor (as Phil suggested).

Steady draw is 20W, or about 170mA @ 120VAC. A 5-ohm resistor would dissipate about an eighth of a watt, and, with other component resistances, limit worst-case inrush to

------------------

Nothing you do with the SMPS will help. The high inrush surge is purely due to connecting the filter capacitor to the AC supply at or near peak voltage. In theory, a huge peak surge can briefly flow - limited only by the impedance of the AC supply and the ESR of the cap.

Testing I did some time ago showed that a 47uF electro switched after a bridge rectifier created a surge big enough to regularly trip a 8amp thermal/magnetic breaker.

Interestingly, switching on about 5 or 6 22watt compact fluoros does the same thing.

.... Phil

If you can switch off the load virtually completely- what I did one time was to use a big ass-resistor (cement wirewound kind*) and time the charge so that the voltage across the resistor had dropped to virtually nothing, then shorted it with a relay contact. The power loss in the relay coil was less than the power loss in a typical NTC.

--sp

• formatting link

Well, yeah, kinda. The front end capacitor current can be limited by adding some AC -side inductance, and with a bridge containing SCRs instead of simple diodes. That'll let you ramp up the capacitor charge slowly. The old Carver Corporation 'Magnetic Field Amplifier' used this kind of limiting and an undersized AC transformer, with a control scheme that never fed the transformer any input that would saturate the core.

Precisely what I did ~1977 at start-up OmniComp (bought out by GenRad ~1987). ...Jim Thompson

----------------

At switch on, the triac firing point ramps up from zero crossings to about 65% of the peak voltage of the AC supply over about 2 seconds. Of course this eliminates any transformer of filter electro inrush surges.

In use, the firing point constantly changes to keep the internal DC rails steady as possible. It's a simple pre-regulator scheme that permits the use of a MUCH smaller AC transformer than otherwise.

However, if on bench test you hold the amp at full power with both channels running for a couple of minutes - smoke soon oozes out of the transformer. Ouch !!!

.... Phil

When overloaded, the device transitions to high resistance mode almost instantly.

If 10ohms cannot be left in circuit, bridge it with a relay.

.... Phil

When overloaded, the device transitions to high resistance mode almost instantly.

If 10ohms cannot be left in circuit, bridge it with a relay.

.... Phil

I'm not seeing how that would have any effect on the BR-C input.

NT

I think Win's thinking of the more general case of inrush into a loaded SMPS.

Undervoltage lockout and soft-start at the SMPS controller mitigate that component of the problem.

Cheers, James Arthur

Well that's one more way to skin this cat, isn't it?

In this case a relay would be larger than the board. In other situations bridging with a relay would work well--I think I had some old CRT monitors that did that. Hi-Fi sets used to do that too...now that you remind me.

Thanks Phil.

James Arthur James Arthur