Solid state soft start for psu

[Analog_Fan]

YouTube's algo suddenly came with video from people repairing amps.

I'm trying to prevent some inrush current problem.
My first drawing was resistor, capacitors only.
The second was similar to a youtube's amp repair video, using resistor, cap, npn.

But i don't wanna use relay's and the constant current.

I tried some solution found on stack exchange, but none of them work.

I need to apply this to both rails, Falstad show my circuit worked on the negative rail, but not positive rail whilst being the same accept PNP.
although they should, i even got the same results in LTSPICE.
No rising "edge", pretty much the mosfet turns on instantly.

I wanna place this after bridge rectifier.
The added image is supposed to work solution posted on stack.


any one got a solution?

 

[JohnRoberts]

Back in the 1970s when I designed my 4x250W audio amp I used the 3 position power switch from the old WE power supply chassis that I built the amp into. In the 1st position there was a resistor added in series with the mains feed to the power transformer. In the power switch 2nd position the series resistor was shorted out and the loudspeaker relay was energized. This prevented turn on surge current and loudspeaker thumps/clicks.

JR

 

[Analog_Fan]

Back in the 1970s when I designed my 4x250W audio amp I used the 3 position power switch from the old WE power supply chassis that I built the amp into. In the 1st position there was a resistor added in series with the mains feed to the power transformer. In the power switch 2nd position the series resistor was shorted out and the loudspeaker relay was energized. This prevented turn on surge current and loudspeaker thumps/clicks.

JRView attachment 129962

That's indeed a neat solution.
: )

it seems there is also a possibility to use a triac or so, but not look in to it any further and also i don't have this part at hand.

The problem is that the solution (from above circuit) on the positive rail can't be used on the negative rail's mosfet gate.
There is almost a difference of 40 volts, so the negative rail turns on much faster or very much unequal to the positive rail if you use a N channel on the negative rail. a LM393 looking at the positive rails capacitors coupled to a CD4013 (as latch) would be the best solution, if a zener can be used in a PSU that's starting up for the LM393?
to be sure the MOSFEF is fully open.

Ltspice and Falstad shows a massive surge if you use single 22000uF capacitor, 1+ Kilo amp during a very brief moment.

I really don't want to use a relay for a speaker set that's turned on a big part of the year.

 

[ccaudle]

I wanna place this after bridge rectifier.
The added image is supposed to work solution

If that circuit is supposed to represent the circuit you build the voltage sources are not correct. There is not a constant DC voltage after the rectifiers.

 

[Analog_Fan]

If that circuit is supposed to represent the circuit you build the voltage sources are not correct. There is not a constant DC voltage after the rectifiers.

I'm no regular user of Ltspice, but Falstad, but some people here "insisted" i should learn to use Ltspice.
That's the voltage source from the component list, not good?

The coupled inductors as transformer, have 220v on the input but only 16V comes out and i don't know how to change it. i don't know the inductance of my transformer, it only states the output voltage and VA.
the capacitors where calculated by hand, 470uF / 1000000, etc, but somehow Ltspice suggested the right values when selecting real capacitors from the list, so i did.

 

[ccaudle]

That's the voltage source from the component list, not good?

If I understand correctly that the circuit is intended to be placed between the diodes and the capacitors to decrease the magnitude of the current into the capacitors.
The voltage at that point is a pulsing rectified waveform. If you want to accurately represent that then include the diodes in the simulation and use an AC voltage source at the input to the diodes, just like the real circuit.

 

[Analog_Fan]

If I understand correctly that the circuit is intended to be placed between the diodes and the capacitors to decrease the magnitude of the current into the capacitors.
The voltage at that point is a pulsing rectified waveform. If you want to accurately represent that then include the diodes in the simulation and use an AC voltage source at the input to the diodes, just like the real circuit.

This is what i made 2 days ago.


It's the positive rail only.
i tested with 0.25 second duration as wel, 3, 5 & 10 seconds.
Figuring out what R1 & C1 should be.

 

[MidnightArrakis]

any one got a solution?

On the AC-side of things, there are these devices:

https://www.digikey.com/en/products/filter/inrush-current-limiters-icl/151?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=PMax_DK+Supplier_Focus Suppliers&utm_term=&utm_content=&utm_id=go_cmp-19976712210_adg-_ad-__dev-c_ext-_prd-_sig-EAIaIQobChMIk8_G6v_BhgMViFFHAR2kWwUAEAAYAiAAEgI_qfD_BwE&gad_source=5&gclid=EAIaIQobChMIk8_G6v_BhgMViFFHAR2kWwUAEAAYAiAAEgI_qfD_BwE

-- INRUSH CURRENT LIMITER --


I just picked a 10-Amp device at random, but these devices go up to a 40-Amp of continuous current rating.

/

 

[ccaudle]

This is what i made 2 days ago.

I suspect that V2 should not be in the circuit. That is an independent voltage source, so unless you have a separate power supply from somewhere, I assume that is supposed to represent the behavior of the transformer circuit, which is already in the simulation.

Are those 0.022F capacitors? I.e. 22,000 uF? Your circuit is on the wrong side to prevent inrush current problems, all the current will be going into charging up those large capacitors. The circuit does nothing very useful on the output side of the capacitors.

 

[Analog_Fan]

I suspect that V2 should not be in the circuit. That is an independent voltage source, so unless you have a separate power supply from somewhere, I assume that is supposed to represent the behavior of the transformer circuit, which is already in the simulation.

Are those 0.022F capacitors? I.e. 22,000 uF? Your circuit is on the wrong side to prevent inrush current problems, all the current will be going into charging up those large capacitors. The circuit does nothing very useful on the output side of the capacitors.

Thnx, V2 is just 8 ohm load (emulating a speaker) controlled by frequency (audio) too see if this has an effect on the charge of the caps. I would like to try to have voltage rails not moving at all on a kick drum.

It's just a test, those 22000uF, i might as wel choose for 24 x 47 / 68 / 100 uF or something like that.

I have found some German transformer companies, I'm thinking about having dedicated transformers made, if their are willing to do this, than i would have made something like having 2 x 15 V bi-polar @ 3 Amps (low / mid range & tweeter) and 1 x 15 V bi-polar @ 500 mA. this 3rd circuit to turn on the other 2 circuits.

 

[Analog_Fan]

On the AC-side of things, there are these devices:

https://www.digikey.com/en/products/filter/inrush-current-limiters-icl/151?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=PMax_DK+Supplier_Focus Suppliers&utm_term=&utm_content=&utm_id=go_cmp-19976712210_adg-_ad-__dev-c_ext-_prd-_sig-EAIaIQobChMIk8_G6v_BhgMViFFHAR2kWwUAEAAYAiAAEgI_qfD_BwE&gad_source=5&gclid=EAIaIQobChMIk8_G6v_BhgMViFFHAR2kWwUAEAAYAiAAEgI_qfD_BwE

-- INRUSH CURRENT LIMITER --
View attachment 129992

I just picked a 10-Amp device at random, but these devices go up to a 40-Amp of continuous current rating.

/

Yeah, i was thinking about these as well, but i don't have them at hand now.

I saw some EMI filter circuits on Tube, that have these.
I'm going to use the above circuit, where the "EMI" filter is around the bridge rectifier as figured in the DartZeel amp.

 

[Analog_Fan]

On the AC-side of things, there are these devices:

https://www.digikey.com/en/products/filter/inrush-current-limiters-icl/151?utm_adgroup=&utm_source=google&utm_medium=cpc&utm_campaign=PMax_DK+Supplier_Focus Suppliers&utm_term=&utm_content=&utm_id=go_cmp-19976712210_adg-_ad-__dev-c_ext-_prd-_sig-EAIaIQobChMIk8_G6v_BhgMViFFHAR2kWwUAEAAYAiAAEgI_qfD_BwE&gad_source=5&gclid=EAIaIQobChMIk8_G6v_BhgMViFFHAR2kWwUAEAAYAiAAEgI_qfD_BwE

-- INRUSH CURRENT LIMITER --
View attachment 129992

I just picked a 10-Amp device at random, but these devices go up to a 40-Amp of continuous current rating.

/

"Please check your internet connection or disable your ad-blocker."
I can't see digikey, it thinks I'm some kind of robot.


I just went to shop around the corner.

His supplier can supply NTC, but witch.
10K Ohm is like 0.5 Watt.
10 Ohm is like 4/5 Amp.
--- or small value's can handle big loads.

What do you suggest?

My voltage is 19.6 Volt (rect) and a aimed consume of 1.25 amp if continuous loaded.

 

[MidnightArrakis]

"Please check your internet connection or disable your ad-blocker."
I can't see digikey, it thinks I'm some kind of robot.


I just went to shop around the corner.

His supplier can supply NTC, but witch.
10K Ohm is like 0.5 Watt.
10 Ohm is like 4/5 Amp.
--- or small value's can handle big loads.

What do you suggest?

My voltage is 19.6 Volt (rect) and a aimed consume of 1.25 amp if continuous loaded.

[Please check your internet connection or disable your ad-blocker / Location: Universe] -- It's possible that since you aren't from Planet Earth, that you are being seen as an "Alien" and Digi-Key is protecting itself from being overtaken by an "alien invasion"!!! Perhaps by getting a VPN and having your Internet connection setup as going through a U.S. server might resolve this issue with you being blocked. Which part of the "Universe" are you from???

[10K Ohm is like 0.5 Watt. / 10 Ohm is like 4/5 Amp] -- As I had mentioned in my original post, the "Inrush Current Limiters" are for use on the -- AC-mains -- side of things, meaning.....120VAC / 240VAC.....and not the secondary side of a transformer. The device I showed in my original post was for a -- 10-Amp -- device, which I had selected at random simply to show you what these devices looked like, not as THE device for you to use.

Try and see if these MOUSER -- 1.3-Amp -- devices are "seeable" for you:

https://www.mouser.com/ProductDetail/EPCOS-TDK/B57153S0330M051?qs=Uha62PnlbFCIYN2KNURR7w==

https://www.mouser.com/ProductDetail/EPCOS-TDK/B57153S0330M000?qs=AKDv8POSxR3UfQGA5zAfvg==

https://www.mouser.com/ProductDetail/EPCOS-TDK/B57153S0330M000?qs=AKDv8POSxR3UfQGA5zAfvg==

/

 

[Analog_Fan]

[Please check your internet connection or disable your ad-blocker / Location: Universe] -- It's possible that since you aren't from Planet Earth, that you are being seen as an "Alien" and Digi-Key is protecting itself from being overtaken by an "alien invasion"!!! Perhaps by getting a VPN and having your Internet connection setup as going through a U.S. server might resolve this issue with you being blocked. Which part of the "Universe" are you from???

If would own a company like Digikey, Mouser or anything else, i would not use ANY code or cookies from other "companies" to assure my customers privacy, now you're being forced by companies like DIGIKEY to allow cookies from 3rd parties, unknown parties, that's why i avoid online shopping, but people apparently can't read very well themselfs and let other do that on their behalf, climate change is such subject.
A few day's ago a reporter asked a climate protester how much CO² our country emits and the protester responded, i don't know you can look that up on Google!

Some of these 3rd parties like Hotjar seem to claim being able to sense if your happy with a website.
Their code follows your fingers, mouse, the position you scrolled to, so they know what image or text you're reading, seeing and for how much time.

The internet has become much of a waste of time, obsolete after 30 years.

[10K Ohm is like 0.5 Watt. / 10 Ohm is like 4/5 Amp] -- As I had mentioned in my original post, the "Inrush Current Limiters" are for use on the -- AC-mains -- side of things, meaning.....120VAC / 240VAC.....and not the secondary side of a transformer. The device I showed in my original post was for a -- 10-Amp -- device, which I had selected at random simply to show you what these devices looked like, not as THE device for you to use.

Try and see if these MOUSER -- 1.3-Amp -- devices are "seeable" for you:

https://www.mouser.com/ProductDetail/EPCOS-TDK/B57153S0330M051?qs=Uha62PnlbFCIYN2KNURR7w==

https://www.mouser.com/ProductDetail/EPCOS-TDK/B57153S0330M000?qs=AKDv8POSxR3UfQGA5zAfvg==

https://www.mouser.com/ProductDetail/EPCOS-TDK/B57153S0330M000?qs=AKDv8POSxR3UfQGA5zAfvg==

/

I was just reading this site, it seams you use a small value, like 5/10 ohm.

https://www.ntcsensors.com/How_to_use_NTC_thermistor_to_limit_surge_current_/

It also suggest to use a relay to bypass the power consumption of the NTC to safe on power.
Now i'm playing again with transistors again to open a MOSFET.
But some how a capacitor delay with transistor doesn't seem to work in Ltspice, Fallstad.

 

[Matador]

With the initial conditions at 0V, if whatever is pulling up the gate of the MOSFET is "off", then the gate is floating, and will follow whatever charge is in the channel. Also, once the NPN turns on, it will rapidly turn on the MOSFET once the threshold voltage is crossed, which only delays the inrush until 'later'.

You essentially want to ramp the gate voltage up slowly: so the pull up circuit has to RC charge the gate up slowly which causes the output to rise slowly. Even just a simple RC at the gate will accomplish this. Then you can use complementary devices for positive and negative.

 

[vintageampfixer]

Back in the 80s, when we were prototyping or production testing the big PA amplifiers at Audix, we used to soft start them via a variac; "switch on" was therefore winding the variac's control knob gently from 0% to 100%. We'd usually have an ammeter in circuit too, so the main current could be monitored. IIRC, the largest I ran up was 10kW RMS output power, so fairly substantial. The factory was in a small-ish village, so we also had the in-line "resistor" of the overhead cables which meant we rarely got the full 240v ... unless there was a power cut and we were running the generator. FWIW, the variac method is also a good way to slowly power on very old equipment if it's been standing for ages as it allows the eletrolytics to re-polarise ... hopefully without them going pop!

I saw a YouTube video just the other day where the chap had a set of high wattage tungsten halogen light bulbs in series with his supply to the "item under test". If things were good, the lamps would glow briefly in response to inrush current and if thing weren't the lamps'd just illuminate at full power (without whatever it was blowing itself to pieces, with a bit of luck )

The practical problem I've found with slow start circuits when the load is somewhat unknown, is that they're either an irritation because they switch on too slowly or they go too fast and blow things up anyway. Obviously they can work well for known-quantity situations.

 

[CJ]

Marantz 2270
delay can be controlled with C807

not what you want, just food for thought

 

[Analog_Fan]

With the initial conditions at 0V, if whatever is pulling up the gate of the MOSFET is "off", then the gate is floating, and will follow whatever charge is in the channel. Also, once the NPN turns on, it will rapidly turn on the MOSFET once the threshold voltage is crossed, which only delays the inrush until 'later'.

You essentially want to ramp the gate voltage up slowly: so the pull up circuit has to RC charge the gate up slowly which causes the output to rise slowly. Even just a simple RC at the gate will accomplish this. Then you can use complementary devices for positive and negative.

On the Node R1 to C1 it clearly shows what you would expect, a voltage ramping up, however Q1 is instantly turned on while still ramping up when enabling the simulation or resetting, in Ltspice as Falstad!

Maybe i should try this on a breadboard.
The transistors are there to pass the full supply voltage rather than including the voltage drop of R1, not having the mosfet fully open.

 

[Matador]

Right..because your RC is on the base of the BJT, not on the gate of the MOSFET. That's what I was saying: once the RC passes 0.6V on the base, the MOSFET will instantly switch full on, which just moves the inrush to a few hundred milliseconds later than it would with no circuit at all.

The RC needs to be at the gate of the MOSFET, since that's the thing you want to "ramp on".

 

[Analog_Fan]

...

probably best to use CD40106

This circuit gets activated somewhat at half the supply voltage.
You could time it as long as you want to.

Now figuring out how this works for the negative rail.
pin 14 (VDD) of the cd40106, now goes to GND, pin 7 (VSS) to -19.6V and a logic high = GND.
to activate P - channel mosfet.