Inrush current Protection in transformers

Thread Starter

SVS

Joined Aug 16, 2012
89
Hi everyone,

Is this circuit a good solution to limit inrush current in a transformer ??

It is done using a MOSFET and a resistance that is connected in parallel, the MOSFET is switched ON after a delay so that the current will be flowing through the resistor and the inrush current to the transformer would be avoided.
I have attached the circuit .

Please throw in your views about this solution.

thank you!
 

Attachments

#12

Joined Nov 30, 2010
18,224
A few things need to be changed. Your transistors are backwards polarity and the duration of the square wave needs to be named.
 

MaxHeadRoom

Joined Jul 18, 2013
28,617
The transformer winding inductance inherently limits inrush current, so you don't need the resistor/MOSFET.
+1, And I have installed some pretty hefty transformers with large reservoir caps and never had a problem, I believe a problem is often caused by under fusing, or wrong fuse type (delay/Slo-Blo etc)?
Max.
 

AnalogKid

Joined Aug 1, 2013
10,986
Usually inrush current limiting is applied to a capacitor-input power supply or power supply front end. In tis way the dischargd filter capacitors do not appear as a momentary dead short across the input. In your circuit the H bridge transistors are trying to appear as dead shorts on every cycle to improve efficiency. Your circuit can be used todelay the output, but it will have no effect on the shape of the transformer current pulses.

ak
 

KMoffett

Joined Dec 19, 2007
2,918
The transformer winding inductance inherently limits inrush current, so you don't need the resistor/MOSFET.
Not all transformers. We purchased several 120V/20A auto transformers. They intermittently (about 50% of the time) tripped 30A breakers at turn-on. Even with no load. Though they were imports, a guy at STACO (the real VARIAC) said that was normal. :(

Ken
 

t_n_k

Joined Mar 6, 2009
5,455
I think the OP is addressing the magnetizing inrush current issue. Presumably they want to protect the pwm transistor bridge. A more careful study of the transient behaviour may preclude the need for this protection anyway.
 

AnalogKid

Joined Aug 1, 2013
10,986
Not all transformers. We purchased several 120V/20A auto transformers. They intermittently (about 50% of the time) tripped 30A breakers at turn-on. Even with no load. Though they were imports, a guy at STACO (the real VARIAC) said that was normal.
Yup. Same thing can happen with normal little offline transformers for linear supplies. At turn-off the transformer core is left with a "magnetic bias" (someone else's words). If the input AC phase angle at power-on is just right, the current spike is significant. Hence slo-blo fuses.

ak
 
Last edited:

Thread Starter

SVS

Joined Aug 16, 2012
89
Usually inrush current limiting is applied to a capacitor-input power supply or power supply front end. In tis way the dischargd filter capacitors do not appear as a momentary dead short across the input. In your circuit the H bridge transistors are trying to appear as dead shorts on every cycle to improve efficiency. Your circuit can be used todelay the output, but it will have no effect on the shape of the transformer current pulses.

ak
Sorry . the circuit I posted was just a rough one.
now i have attached the proper schematic.
To stabilize my DC power supply , I have 250uF*4 electrolyte capacitor at the input.
so , these capacitors deliver a pretty high starting current during Turn ON.
That s the reason for using a MOSFET / resistance to control this inrsuh current.
 

Attachments

Thread Starter

SVS

Joined Aug 16, 2012
89
A few things need to be changed. Your transistors are backwards polarity and the duration of the square wave needs to be named.
Please check the actual circuit which I have newly attached ..
The old one was just a rough sketch--
 

Thread Starter

SVS

Joined Aug 16, 2012
89
The transformer winding inductance inherently limits inrush current, so you don't need the resistor/MOSFET.
I have a ferrite core transformer where my winding resistance is pretty low and the inductance value is on the higher side.
So, i have to protect my transistors and the transformer primary side from this short high current spike.
 

Thread Starter

SVS

Joined Aug 16, 2012
89
You presumably have control of the PWM drive. Just perform a proper soft start (search for "soft start" on the web) and your problem is solved.
not really. the sad part is i can just control the frequency of my PWM signal and not the dead time.
But i make sure that the MOSFETS are driven using ZVS technique.
thank you!!
 

Alec_t

Joined Sep 17, 2013
14,280
i have to protect my transistors and the transformer primary side from this short high current spike.
1) That current spike is through the reservoir caps C5-8, not through the transistors/transformer.
2) M5 has no gate-drive source to give a Vgs ~10V.
3) What is the purpose of inductors L3,L4 in series with the transformer windings L1,L2?
4) The higher the inductance of your transformer winding the slower will be the rate of increase of the current in it.
i make sure that the MOSFETS are driven using ZVS technique.
How do you do that? I see nothing in your schematic for sensing zero-voltage across the FETs.
 
Last edited:

alfacliff

Joined Dec 13, 2013
2,458
the simpliest one I have seen is a relay (properly sized) and a resistor, also properly sized. the relay is connected with its contacts shorting the resistor when it drops out, when you turn on the system, the relay is pulled in, putting the resistor in series with the incoming power. after a short delay, the relay is dropped out, shorting the resistor. that is the way square d, allen bradley, and others did it to protect the rectifiers in large drives.
 

Thread Starter

SVS

Joined Aug 16, 2012
89
1) That current spike is through the reservoir caps C5-8, not through the transistors/transformer.
2) M5 has no gate-drive source to give a Vgs ~10V.
3) What is the purpose of inductors L3,L4 in series with the transformer windings L1,L2?
4) The higher the inductance of your transformer winding the slower will be the rate of increase of the current in it.
How do you do that? I see nothing in your schematic for sensing zero-voltage across the FETs.
1. To protect the current spike from the reservoir capacitors , I need the MOSFET M5 and resistance R5 in parallel so that during the start ,
The current flows through the resistor R12 and the transformer .
After 0.2 s , the the MOSFET M5 switches ON which has a low Rds ON.
2. M5 is driven using a photo isolated opto coupler , but LT spice doesn t have a Model. So, i just added a normal opto coupler.
And the MOSFET turns ON at Vgs = 5V.(datasheet)
Why should Vgs be 10V??

3. L3 and L4 are the leakage inductance .

4. I drive this circuit inbetween 30- 50 KHz and I have a reasonable dead time in between my PWM signals . but still ZVS sensing / switch will be my next steps.

But right now my worry is the current spike during TURN ON from the capacitor bank , which has to be minimized...
 

Thread Starter

SVS

Joined Aug 16, 2012
89
the simpliest one I have seen is a relay (properly sized) and a resistor, also properly sized. the relay is connected with its contacts shorting the resistor when it drops out, when you turn on the system, the relay is pulled in, putting the resistor in series with the incoming power. after a short delay, the relay is dropped out, shorting the resistor. that is the way square d, allen bradley, and others did it to protect the rectifiers in large drives.
That s just a time delay relay ...

And in my circuit ,
MOSFET (M5) in parallel with a resistor(R12) works on the same principle ..
 

alfacliff

Joined Dec 13, 2013
2,458
but relay contacts probably have less on resistance. this means less heat and loss. also, any power line spikes will not kill the relay. the reason the relay is used is that a whole cabinet of 6000 mfd600 volt electrolytics in paraless take a bit of time to charge, and also a bit of current, would take a very large and expencive mosfet. also, there is almost no circuitry involved that would have to be powered up before turning on the drive.
 

Thread Starter

SVS

Joined Aug 16, 2012
89
but relay contacts probably have less on resistance. this means less heat and loss. also, any power line spikes will not kill the relay. the reason the relay is used is that a whole cabinet of 6000 mfd600 volt electrolytics in paraless take a bit of time to charge, and also a bit of current, would take a very large and expencive mosfet. also, there is almost no circuitry involved that would have to be powered up before turning on the drive.
Not really.. A NMOS power mosfet with low Rds would do the job and it is very cheap compared to a Solid State relay. A SSD relay for 220 V DC, rated for 5 A would cost me 50 € here.

But this circuit with MOSFET, optoisolator and resistance , i end up spending not more than 5€ max...
 
Top