Alternative to the IRF3708
DavidGreg
Posts: 38
Hi-
Per recommendations from this forum, I've been using the IRF3708 in various projects requiring switching large loads. My typical application is just on/off control, but I occasionally have need for using PWM to control power delivered to the load.
I was browsing digi-key today and I ran across a different MOSFET that I would like some advice on. The data sheet is here: http://www.nxp.com/documents/data_sheet/PSMNR90-30BL.pdf
The RdsOn is less than half of the IRF3708, and it appears to switch on "fully" some where between 3 and 3.5V, which is nice. However, the gate charge is quite a bit higher than the IRF3708 - does this mean that in a switching application it would be inferior to the IRF3708? I usually use 100 ohm resistors in-line with the gate supply from the propeller pin in order to get high currents into the gate.
-David
PS the 3708 datasheet is here: http://www.irf.com/product-info/datasheets/data/irf3708pbf.pdf
Per recommendations from this forum, I've been using the IRF3708 in various projects requiring switching large loads. My typical application is just on/off control, but I occasionally have need for using PWM to control power delivered to the load.
I was browsing digi-key today and I ran across a different MOSFET that I would like some advice on. The data sheet is here: http://www.nxp.com/documents/data_sheet/PSMNR90-30BL.pdf
The RdsOn is less than half of the IRF3708, and it appears to switch on "fully" some where between 3 and 3.5V, which is nice. However, the gate charge is quite a bit higher than the IRF3708 - does this mean that in a switching application it would be inferior to the IRF3708? I usually use 100 ohm resistors in-line with the gate supply from the propeller pin in order to get high currents into the gate.
-David
PS the 3708 datasheet is here: http://www.irf.com/product-info/datasheets/data/irf3708pbf.pdf
Comments
This is a newer part than the IRF3708 so advances in low gate voltage operation should be better.
I wish they had it in the TO-220 package though. The TO-220 can dissipate a bit more heat in free air with no heat sink.
It's a tad expensive at $2.98us compared to $2.04us.
Hi falf; The PSMNR90 appears to be quite a bit slower than the IRF3708.
MOSFETs aren't exactly rated in frequency, (to many variables).
However. it looks like the PSMNR90 can switch in 10mS or so driven by a Prop. Don't quote me here as I haven't tried it yet
The IRF3708 can easily switch at faster than 1mS.
Of course, both can switch much faster if a high current gate driver is used. the IRF3708 still has the edge here.
One clear advantage of the PSMNR90 is the max gate voltage is +-20V, the IRF3708 is only +-12V.
Duane J
The PSMNR90 appears to be about 2mOhm. So the power dissipation would be about 0.128W as opposed to 1.9W for the IRF3708
It would seem that you could go to 22A at 1W free air.
Duane J
As it stands the PSMNR90 has about 100nC on the gate, and if the plateau is typically at 2.9V as the graph implies it'll take about 25us to charge up, 3.5us to discharge. I'd suggest this limits the the PWM to 300Hz for reasonable switching losses - rather poor. Drive that gate from 10V and a MIC4422 with 5 ohm gate resistor and it'll switch both ways in less than 300ns I suspect.
Calculating switching time: First realise that the gate charge isn't like a simple capacitor - its best to look for the gate charge/voltage graph in the datasheet. Expect the bulk of the charge (when handling high drain currents) to be in the plateau there. If say the plateau is at 2.9V as here, you really want to drive with 6V or so - then the switch on and off are matched - however from 3.3V this means very slow turn on - there's only 0.4V to push current through the gate resistor. Switch off we have 2.9V available, a lot faster.
So the missing figures on the datasheet are the min/max plateau voltages, we just get the rather useless threshold voltage min/max... I suggest a good strategy is to assume the plateau variation is the same as the threshold variation, and the average is as shown in the aforemention gate charge/voltage graph.
So if we've got a plateau at 5V in the graph, a nominal gate drive of 10V and a threshold quoted as 2..4V, then we assume the plateau is between 4 and 6V. For this the 10V drive is good - symmetrical switch on and off. To calculate the switching time you work out the current through the gate resistor and determine how long that current needs to flow to supply or drain the charge. Away from the plateau you can assume an exponential RC style decaying current if you want to model more accurately - the slopes in the graph show the capacitances are somewhat different before and after the plateau - this is not really a capacitor!
For driving from 3.3V we ideally want an average plateau at 1.65V, and that would imply a threshold in the 0.4 to 0.8V range I believe. No high power MOSFETs go that low - the thin gate oxide is just too vulnerable to the high switching transients.