That's also a good approach, depending on the circumstance. This is a little fart, about 2x5cm split into line and isolated sections, so, a bit tight for squeezing in a relay to short out the big ass-resistor.(*)
(*) Thanks for the xkcd explanation!
Cheers, James Arthur
My limiter attempt was an electronic variation, of sorts. I modified the classic two-BJT limiter produce a voltage foldback current limiting based on the pass element's Vce.
Rather than Carver's attempt to protect their transformer, my aim here was to protect the limiter's *pass* transistor. A straight 1A series limiter, for example, could've seen 170V @ 1A. So I made the current limit foldback based on the pass transistor's Vce. You can make i(c) fall to zero, for example, for Vce >10V, reducing peak dissipation from 170W to 10W.
I guess I could've added a timer, too. But it was turning into a hairball, an inelegant, unnecessarily complicated approach.
Simple is better.
Cheers, James Arthur
The OP should consider and test what happens if there is a short-circuit downstream, in case the resistor consistutes a fire hazard in its intended location within the product. Often a mild overload causes more heating than a dead-short which opens the resistor more rapidly and safely. Whether such a mild overload is likely depends on the circuit.
A "fusible" ww type may be safer (at least in the opinon of the safety testing lab and perhaps in reality) in the event of an overload, and may make it unnecessary to also fit a fuse in some circumstances.
Phil makes a good point about WW types and reliability - film types fail easily in surge testing. Adding the WW resistor will likely also make it more difficult to damage the rectifier during surge testing, and may make the whole SMPS more reliable than without it.
Clearly none of those measures solve the problem.
NT
A capacitor ahead of a big capacitor no doubt will reduce the initial charge current, but should the inductance be ahead of the bridge or after the bridge but before the capacitor ?
In the tube rectifier days, the limited cathode peak current forced the use of choke input (LC) or pi (CLC) arrangement after the rectifier. In regular operation, the inductance also improved the power factor, which is a issue these days.
The problem with the DC side inductances is the need for an air gap.
here's my "electronic limiter" idea; use a mosfet and cap to limit charging current thru the center tap of a bridge doubler configuration, switch off MOSFET when filter caps are charged to go back to full wave. No pass transistors or resistors dissipating power in the load's current path
(I haven't tested this!!!)
Bifiliar wound, run the ground return current thru it too?
A fusible resistor is what one would spec in this service.
Cheers, James Arthur
They don't solve the instant problem, that's correct.
Cheers, James Arthur
Riffing on Sphero & Phil's input, putting the resistor after the bridge would let you short it with a MOSFET (instead of a relay) after an appropriate time delay.
I just cracked open a compact fluorescent and an LED bulb. One uses no inrush limiting, the other has a 5-ohm fusible resistor.
Cheers, James Arthur
Cheers, James Arthur
Standard cermet surface-mount resistors are not certified for that use, but they seem pretty good. I took a few regular low-ohm 1206's and connected them across the AC line; they made a little spark and a pip! sound and opened up.
There are parts spec'd for this.
On the other hand, we've found official surface-mount fuses to be unreliable.
-- John Larkin Highland Technology, Inc lunatic fringe electronics
And I bet both are festooned with CE/UL/VDE approval text.
You're only charging 15 uF, so the current must be small. Just a series resistor should work.
-- John Larkin Highland Technology, Inc lunatic fringe electronics
Sometimes whatever the cheapo ones use don't fuse - I came home one afternoon several years ago to find a CFL I had left on sputtering and smoke pouring out of the base.
Once everything had cooled down I opened it up to find the "fusible" resistor in place - blackened, scorched and melted but apparently still conducting just fine.
Never cheapo again, and all LEDs now...
That's how mainframes solved the problems in the '60s - '80s. There were huge inrush limiting resistors in the plenum right before the fans and above the power supplies. It all worked great until they started controlling power sequencing with software. ;-) (The problem was actually hardware but...).
Some PSUs use a resistor shorted by either a relay or SCR when the cap has had time to charge. You have to use a flameproof or fusible resistor in case the relay fails to activate.
Jon
A resistor to limit the inrush current, and a relay to short the resistor after a second or two.
Exactly, that's what I spec'd- a fusible 'cement' resistor, similar to these:
--sp
-- Best regards, Spehro Pefhany
Close -- don't time it, though: let it "decide". When the voltage reaches, say, 90%, /then/ short it out.
Flip side: what if it never reaches 90%? Say the load is shorted, or the resistor burned out (or both, more likely!). You still want a timer, but this one shuts the whole thing down if it never reaches 90% and turns on.
You'd probably get away with doing this, with a pair of relays or TRIACs, an optoisolator, and some auxiliary power for the logic, and the logic would be a comparator (LM311?) and some R's and C's. This could maybe be simplified and implemented with higher voltage logic (a couple transistors?) so you don't need the auxiliary supply.
Tim
-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website: http://seventransistorlabs.com
** That is very uninformative.
** Massive non sequitur.
As I already posted, the peak current surge is huge - in the order of 200 amps if there is no limiting.
** Billions of CFLs cannot all be wrong..... Phil
Well I had a reasonably powerful microcontroller to synthesize the waveforms for the inverter and auto-diddle the output frequency (it was a controller for an ammo making operation), but did not know what
90% was- measuring bus DC voltage for feed forward but not the incoming AC voltage. Timing with load removed was simple (implicit assumption the AC voltage is more-or-less constant, of course)Not sure what you mean here, but if something is damaged (relay doesn't close or load is shorted) any kind of (safe) shutdown is fine, it's got to come back to the factory.
I could see that working and being useful in some cases.
--sp
-- Best regards, Spehro Pefhany
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.