I need the lowest rds available on mosfet

Lowest rds on mosfet available

Hi,
I have an application in which I must switch very high currents through a coil, at low voltage (<10V), with an n channel mosfet and I need the absolute lowest rdson commercially available, other criteria are irrelevant. So far I found IPT004N03 with 370 microohm, typical. Do you have any recommendation, is out there a lower rds on available? Thank you.

Re: Lowest rds on mosfet available

Terminal resistance matters, you'll probably prefer a parallel circuit anyway. In so far, choose the best available and economic transistors.

Re: Lowest rds on mosfet available

Hi,

I recommend to use some search engines at mosfet distributors or at mosfet manufacturers.
Distributor´s benefit: overviw across different manufacturers
Manufacturer benefit: latest news

Klaus

Re: Lowest rds on mosfet available

The lower the Rds(on), the higher the gate charge. You may want to look into the highest switching frequency you need to switch it.

Re: Lowest rds on mosfet available

The lower value of r_ds(on) is also not infrequently associated with the package of higher thermal dissipation capacity, which in other words means that it will be likely very expensive.

Re: Lowest rds on mosfet available

You should first quantify "very high" current and then
figure just how much voltage drop and switch Pdiss
this rolls up to, per ohm Rds.

Then, Digi-Key / Mouser / Allied / etc. will give you
parametric selection "choosers" that you can assign
acceptable BVdss, Ron value ranges to and sort as
you see fit.

You might like to look at the EPC GaN FETs if you
are able to handle their PCB attach; they have no
package or wire bonds to add to net Rds. Beware
however that EPC plays some games with current
handling specs (like, the advertised ratings are
narrow-duty-cycle and the FETs (their solder balls)
cannot possibly handle the rated current at the
rated temperature at DC - by a factor of 10 or more).

Beware also that the PCB may contribute more to
on resistance than the FET itself, when you are in
the sub-milliohm range (see copper sheet resistance
vs weight, and the realities of routing).

10V and 1mOhm allows 10kA which is a lot of juice.
100A and 1mOhm is 10W power dissipation in the
switch, which is not much for a fat FET. Not like
0.37mOhm is liable to be -that- much better, at
least in this ballpark. Quantify.

Re: Lowest rds on mosfet available

I must switch very high currents through a coil, at low voltage (<10V)....

Click to expand...

Do you need high frequency operation? What is the typical number for a very high current?

If the current is around 100A, the voltage drop will be 37mV and the power lost will be <4W. But there will be switching effects (if you want to switch it on and off at high frequency)

Need more info.

Re: Lowest rds on mosfet available

Hi,

10V and 1mOhm allows 10kA which is a lot of juice.

Click to expand...

I doubt any mosfet datasheet specifies 10kA at 1mOhms Rdson.
This gives a huge power dissipation of 100.000W.

Usually the datasheet mentiones the peak current rating.

Klaus

Re: Lowest rds on mosfet available

Many electromagnets run at 100-200A current (the coils are water cooled- the conductors themselves are copper water pipes wrapped with fiber glass tapes) but they are slowly getting out of fashion these days. Superconducting magnets use higher currents but they work on a closed turn- very few in fact. But they are tough to switch on and off fast- the total energy involved in the magnetic field is rather large.

I note the specified device (IPT004N03) is rated at max 300A and perhaps you can use it at 200A without much problem. Perhaps you can use several devices in parallel in case you need much higher currents.

Re: Lowest rds on mosfet available

Thank you all for answers. I plan to do some experiments with high energy magnetic fields. The current range is around 10-12ka and working mode is discontinuous, with pulses under few milliseconds and large pauses between them (seconds)
Of course I will have to use several mosfets in parallel and I count on approx. 1200a peak current for each transistor, according to datasheet of IPT004N03. I also plan to use pcb only for driving circuit, the actual drain and source connections I need to find a way to solder them directly to a copper bar, as I don't think there is any pcb capable to withstand such currents. My problem is that the entire circuit resistance is distributing the power only slightly over 50% in the actual coil, and the rest is loss (busbar connections, mosfet rdson, power supply esr). I cannot change the coil requirements nor power supply esr, the only room for improvement is in the mosfet rdson.

Re: Lowest rds on mosfet available

The current range is around 10-12ka and working mode is discontinuous...

Click to expand...

The real challenge will be to switch on/off a bank of mosfets with a very high inductive load (and not the lowest rdson). Please let us know how did your experiment went- just being curious.

Re: Lowest rds on mosfet available

It's a hobby project, so I do not have an exact time schedule, but sure, as soon as I have some news, I will.

Re: Lowest rds on mosfet available

Sounds like you've given a fair amount of considerations to the main challenges, especially interconnect resistance which will probably dominate over FET resistance (especially if you compare it to the coil's resistance). However you haven't mentioned what you plan to do with the energy in the coil after your pulse... if you're not careful, that energy will cause a huge flyback voltage at turn-off, which will surely destroy the FETs. Consider using a beefy flyback diode, or driving with a full bridge to steer the energy back towards the power supply...

Re: Lowest rds on mosfet available

The back emf is surely a concern. The basic idea was to use flyback diode but your idea with steering energy back to the power supply is very interesting.
The main connections are designed so far with 20x3mm rectangular copper bars and fortunately, this is not the main source of circuit resistance, as I can stack them if it is necessary.
Main problems are with transistors, they are a significant part of the total resistance and there also some mechanical issues, as those transistors do not have anything flexible in them. If solidly soldered directly on the bar with both connections, they may be damaged either due to thermal movement of the entire assembly (at the time of soldering) or due to mechanical shock which may be present when a pulse is engaged/disengaged. They also must be soldered in such a way so that the current will be distributed equally among several paralleled transistors.

Re: Lowest rds on mosfet available

The 100A spec limit and the SOA curve for the IPT004N03 is to prevent wirebond fusing rather than junction thermal limits.

Also consider a better figure of merit being the lowest R*Q product for both Qg and Qgd and Ron at Vgs used.

Then the same device ought be used to drive up to 50 parallel devices. Distributed switches will also lower the ESL layout inductance with shunted paths. Consider IPB09N03LS https://www.infineon.com/dgdl/Infineon-MOSFET_OptiMOS_revolutionizes_Oring-PB-v01_00-EN.pdf

https://www.infineon.com/dgdl/Infineon-MOSFET_OptiMOS_datasheet_explanation-AN-v01_00-EN.pdf

I recall the challenges of instrumenting a tube seam arc welder that used 4" thick solid copper wheels for inner/outer electrodes in the late 70's and the induction challenges to prevent measurement errors for currents over 10kA. But it worked. It was officially called a Zirconium shim Diffusion Bonder at Bristol Aerospace ( now Magellan) for nuclear steam tubing.