Forums

MOSFET switching an inductive, high-frequency load

Started by Don Gilmore February 14, 2014
Guys:

I'm an ME working on a project where I must switch voltage to a small coil =
on and off at a fairly high frequency (around 25 kHz) with a MOSFET transis=
tor.  The DC resistance of the coil is very low (about 1 ohm), but the impe=
dance is much higher because of the frequency (about 100 ohms).  I'm switch=
ing 24 Vdc.

Should I specify the max. pulsed current of the transistor based on the DC =
current (24/1 =3D 24A), or the max. continuous current based on the current=
 pull with impedance (24/100 =3D 240mA)? The "pulsed current" ratings appea=
r to be only about twice the "continuous current" ratings.

Thanks for any help you can give.

Don
"Don Gilmore"

I'm an ME working on a project where I must switch voltage to a small coil 
on and off at a fairly high frequency (around 25 kHz) with a MOSFET 
transistor.  The DC resistance of the coil is very low (about 1 ohm), but 
the impedance is much higher because of the frequency (about 100 ohms).  I'm 
switching 24 Vdc.


** It is crucial to know if the coil has a core material - like iron or 
ferrite.


Should I specify the max. pulsed current of the transistor based on the DC 
current (24/1 = 24A), or the max. continuous current based on the current 
pull with impedance (24/100 = 240mA)? The "pulsed current" ratings appear to 
be only about twice the "continuous current" ratings.

Thanks for any help you can give.


** Describe the coil properly and say what the on-off duty cycle is.

...    Phil




On Friday, February 14, 2014 10:24:49 PM UTC-6, Phil Allison wrote:
> "Don Gilmore" > > > > I'm an ME working on a project where I must switch voltage to a small coil > > on and off at a fairly high frequency (around 25 kHz) with a MOSFET > > transistor. The DC resistance of the coil is very low (about 1 ohm), but > > the impedance is much higher because of the frequency (about 100 ohms). I'm > > switching 24 Vdc. > > > > > > ** It is crucial to know if the coil has a core material - like iron or > > ferrite. > > > > > > Should I specify the max. pulsed current of the transistor based on the DC > > current (24/1 = 24A), or the max. continuous current based on the current > > pull with impedance (24/100 = 240mA)? The "pulsed current" ratings appear to > > be only about twice the "continuous current" ratings. > > > > Thanks for any help you can give. > > > > > > ** Describe the coil properly and say what the on-off duty cycle is. > > > > ... Phil
The core is ferrite and the duty cycle is 50% square wave. I have added a snubber cap to get rid of the spikes. Don
In article <1cf776ca-f14d-4568-9a49-b8e292a63f34@googlegroups.com>, 
eromlignod@outlook.com says...
> > Guys: > > I'm an ME working on a project where I must switch voltage to a small coil on and off at a fairly high frequency (around 25 kHz) with a MOSFET transistor. The DC resistance of the coil is very low (about 1 ohm), but the impedance is much higher because of the frequency (about 100 ohms). I'm switching 24 Vdc. > > Should I specify the max. pulsed current of the transistor based on the DC current (24/1 = 24A), or the max. continuous current based on the current pull with impedance (24/100 = 240mA)? The "pulsed current" ratings appear to be only about twice the "continuous current" ratings. > > Thanks for any help you can give. > > Don
Be careful when switching that load. I don't see a schematic here to look at but I can only assume you re doing the conventional approach of pulling down one side of the coil that is being supplied from the other end? When the switch is off(open), the drain will see ~ 2 times your PS voltage. In your case this would be 50+ volts, so steps much be taken to select a MOSFET that can handle that. Of course, this also assumes you are using a cap across the coil? If not, then your resulting voltage when the coil collapses will most likely destroy the MOSFET unless you plan on using a catch diode for that? Personally, I would use a MOSFET that can handle 24 amps with proper mounting continuously.. Just to be on the safe side. P.S. A current limit switch maybe advisable here. This can be done using a low R value power resistor on the SOURCE of the MOSFET as a reference back to a comparator to switch (latch) the circuit in the off state, it can also be used to create the oscillation if you wish. Seems funny because just last night I plugged in a test circuit in LTspice based exactly on that same idea for a plasma generator :) Jamie
On Fri, 14 Feb 2014 17:58:29 -0800, Don Gilmore wrote:

> Guys: > > I'm an ME working on a project where I must switch voltage to a small > coil on and off at a fairly high frequency (around 25 kHz) with a MOSFET > transistor. The DC resistance of the coil is very low (about 1 ohm), > but the impedance is much higher because of the frequency (about 100 > ohms). I'm switching 24 Vdc. > > Should I specify the max. pulsed current of the transistor based on the > DC current (24/1 = 24A), or the max. continuous current based on the > current pull with impedance (24/100 = 240mA)? The "pulsed current" > ratings appear to be only about twice the "continuous current" ratings.
Post a schematic someplace (Google Docs seems to work for that sort of thing these days). Then give us a link. Calling out the impedance of the coil when the excitation is a square wave is pretty dodgy. What's the coil inductance? You don't mention what happens when you turn the transistor off. Your biggest problem could be the voltage spike from inductive kickback. We can't tell you anything unless you post a schematic. You mention in another response that the duty cycle is 50%. You could push things a bit and calculate the average I^2R losses in the FET, but you'd be leaving out any calculation of the switching losses. You're probably safer going with a FET that's rated for the peak current applied to it. The peak current that goes through the FET depends on a lot more than the coil's DC resistance or its inductance. We can help you with that if you post a schematic. Did I mention that you should post a schematic? -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
"Don Gilmore"
> > The core is ferrite and the duty cycle is 50% square wave.
** What one earth are you up to ?? 24VDC switched with a 50% duty cycle has an average value of 12 VDC - no matter what the frequency is. .... Phil

"Don Gilmore" <eromlignod@outlook.com> wrote in message 
news:1e5d5bf0-75f8-488f-a7b1-dbc96cc06121@googlegroups.com...
> On Friday, February 14, 2014 10:24:49 PM UTC-6, Phil Allison wrote: >> "Don Gilmore" >> >> >> >> I'm an ME working on a project where I must switch voltage to a small >> coil >> >> on and off at a fairly high frequency (around 25 kHz) with a MOSFET >> >> transistor. The DC resistance of the coil is very low (about 1 ohm), but >> >> the impedance is much higher because of the frequency (about 100 ohms). >> I'm >> >> switching 24 Vdc. >> >> >> >> >> >> ** It is crucial to know if the coil has a core material - like iron or >> >> ferrite. >> >> >> >> >> >> Should I specify the max. pulsed current of the transistor based on the >> DC >> >> current (24/1 = 24A), or the max. continuous current based on the current >> >> pull with impedance (24/100 = 240mA)? The "pulsed current" ratings appear >> to >> >> be only about twice the "continuous current" ratings. >> >> >> >> Thanks for any help you can give. >> >> >> >> >> >> ** Describe the coil properly and say what the on-off duty cycle is. >> >> >> >> ... Phil > > The core is ferrite and the duty cycle is 50% square wave. I have added a > snubber cap to get rid of the spikes.
In most situations the current in the inductor rises linearly with time for the duration the MOSFET is on - if its on too long, the core saturates and any hint of inductance vanishes, when that happens the resistance becomes only the DC resistance of the winding wire. The math for all the aspects of inductors is pretty much a field in itself - a common pitfall is gathering formulas from various sources and end up mixing standards (SI, CGS, MKS etc)

"Maynard A. Philbrook Jr." <jamie_ka1lpa@charter.net> wrote in message 
news:MPG.2d695567d5a6f32f9897aa@news.eternal-september.org...
> In article <1cf776ca-f14d-4568-9a49-b8e292a63f34@googlegroups.com>, > eromlignod@outlook.com says... >> >> Guys: >> >> I'm an ME working on a project where I must switch voltage to a small >> coil on and off at a fairly high frequency (around 25 kHz) with a MOSFET >> transistor. The DC resistance of the coil is very low (about 1 ohm), but >> the impedance is much higher because of the frequency (about 100 ohms). >> I'm switching 24 Vdc. >> >> Should I specify the max. pulsed current of the transistor based on the >> DC current (24/1 = 24A), or the max. continuous current based on the >> current pull with impedance (24/100 = 240mA)? The "pulsed current" >> ratings appear to be only about twice the "continuous current" ratings. >> >> Thanks for any help you can give. >> >> Don > > Be careful when switching that load. I don't see a schematic > here to look at but I can only assume you re doing the conventional > approach of pulling down one side of the coil that is being supplied > from the other end? > > When the switch is off(open), the drain will see ~ 2 times your > PS voltage.
The flyback voltage can be many times the PS voltage! It depends a lot on the purpose of switching the inductor, how you go about controlling that. Maybe an MOV - if it can't be damped, I'd prototype with the 600V MOSFET from an old SMPSU and measure the actual voltage, use a diode/capacitor peak detector and run it a few minutes to fully charge the cap, then you can select a suitably rated permanent MOSFET.
On Friday, February 14, 2014 8:58:29 PM UTC-5, Don Gilmore wrote:
> Guys: >=20 >=20 >=20 > I'm an ME working on a project where I must switch voltage to a small coi=
l on and off at a fairly high frequency (around 25 kHz) with a MOSFET trans= istor. The DC resistance of the coil is very low (about 1 ohm), but the im= pedance is much higher because of the frequency (about 100 ohms). I'm swit= ching 24 Vdc. Hmm guessing at numbers... the coil has an inductance of ~600uH (100 ohms a= t 25 kHz) Then the L/R time is 600uS. But if you are swtiching at 25 kHz = (40 uS period) and are only on for 20 us... well you are just never going t= o get that much current through the coil. Perhaps that is you intent? (I = could go ahead and estiamte your max current... L*di/dt =3D V =3D~24(20/60= 0) ~1 amp.) =20 I'm assuming that the coil either has 24 volts across it or is shorted to g= round.. not an open circiut. That may be totally wrong...=20 As Tim said, a circiut would help.=20 George H.
>=20 > Should I specify the max. pulsed current of the transistor based on the D=
C current (24/1 =3D 24A), or the max. continuous current based on the curre= nt pull with impedance (24/100 =3D 240mA)? The "pulsed current" ratings app= ear to be only about twice the "continuous current" ratings.
>=20 > Thanks for any help you can give. >=20 >=20 > Don