Go Down

Topic: Switching ON/OFF High Current Load (Read 2050 times) previous topic - next topic

Paul__B

Not going to quote you but it does not make sense to me!

You have a 12 V source and two devices.  You need a single "switch" - three of the FETs you previously cited - to control each, so that is just two such switches.  Since you are switching inductors, each of them will require a diode across it as you previously described:

Of course, you mean 12 V where it says three and so for the other components.

What is correct in this diagram is that the "switch" - the three FETs - must be in the negative line to the inductor so that its source(s) are directly grounded in order to be controlled by an Arduino.  And the Arduino ground terminal itself does not take part in the path between the negative supply and the switching FETs, but connects separately to the FET source terminal.

MorganS

What you are missing is MOSFET drivers. The Arduino pins can only deliver 20-40mA to charge and discharge the MOSFET gates. This is important to switch quickly and not waste a lot of power during every PWM cycle.

A MOSFET driver can deliver tens of amps for fast nanosecond switching.
"The problem is in the code you didn't post."

nuggetchris

I'm putting together a new schematic.

Where does the MOSFET driver connect/play in? Is it connected inline on Arduino --> MOSFET pathway or directly to Arduino?

MorganS

"The problem is in the code you didn't post."

MorganS

Mouser has over 1000 MOSFET gate drivers in stock.

Most of them are intended to drive a pair of MOSFETs as in an H-bridge configuration. Only a few will be 'simple' 1-channel devices. Most will be surface-mount, which makes it more difficult for a hobbyist to buy one and test it. Filtering to through-hole and 1 channel gives me the TC4421 at the top of the list. That looks simple enough to plug into a breadboard. Do not forget to give it a decoupling capacitor. A 1uF ceramic is mandatory.
"The problem is in the code you didn't post."

nuggetchris

Quick question

Looking at the TC4421 it has 5 terminals as seems common with others.
INPUT, GND, VDD, GND, OUTPUT.

I am making the assumption I connect as follows inline:

INPUT - ?
GND - To Ground
VDD - From Arduino GPIO
GND - From Arduino GPIO
OUTPUT - To FET Bank

Please correct/confirm me.

If I am thinking right, it should be one gate driver per FET bank ("Relay").
Since there are two FET banks in circuit, there should be two drivers each connected to Arduino.

Paul__B

What you are missing is MOSFET drivers. The Arduino pins can only deliver 20-40mA to charge and discharge the MOSFET gates. This is important to switch quickly and not waste a lot of power during every PWM cycle.
My understanding of the OP is that PWM was never even a consideration.

Hi! For a project I am designing a system of very high-power electromagnets. Each electromagnet is fed a current of around 75A. To allow them to cool, I alternate between feeding power between one and the other around every 10 seconds or so.
Has something changed since?

nuggetchris

Has something changed since?

Nothing has changed. Generally 10s firing period on each.

My understanding of the OP is that PWM was never even a consideration.

I was under the impression the gate driver was necessary for operating a number of FETs from Arduino. If it is really for use of PWM then I don't know. I'd like constant current flow through the inductor during each firing cycle.

allanhurst

#38
Jul 17, 2018, 02:52 am Last Edit: Jul 17, 2018, 03:13 am by allanhurst
Be careful of MOSFET current specs. The quoted rating is generally a pulse rating.

And look at the package as well - a TO 220 may claim to handle lots, but I wouldn't trust it for more than 10A continuous. Look at the lead thickness - will that REALLY  handle 20A for long?

 2.5 sq mm house wiring cable is only rated at 20A and that's a sturdy bit of solid copper about 1.8 mm in diameter. Use some common sense.

 Use a bigger package - eg the 'nut and bolt' type modules . And heatsink adequately.


Or, as I previously suggested,a big relay may be more appropriate. Much less likely to emit expensive magic smoke.

Allan

nuggetchris

Be careful of MOSFET current specs. The quoted rating is generally a pulse rating.

Here is a data sheet from one which is 100A rated
https://www.mouser.com/datasheet/2/427/sqd100n02-3m5l-1064262.pdf

They appear to distinguish from pulse rating and continuous drain current, so I really shouldn't have reason to be alarmed if it claims to deliver a certain value continuously.

If you can find a quality relay from a dependable site other than ebay that can deliver on that scale post it. I would be happy to consider.

allanhurst

Quite.

That devices's 50A 'continuous' rating is for a pulse < 300uS long, 2% duty cycle.

Not on for 10 seconds.....


Magic smoke in your application!

Try eBay etc for suitable starter solenoids.

Allan

nuggetchris

#41
Jul 17, 2018, 04:23 am Last Edit: Jul 17, 2018, 04:24 am by nuggetchris
There really is no major problem with this setup.

a big relay may be more appropriate. Much less likely to emit expensive magic smoke.


Any thoughts on the use of mosfets Paul, or a good suggestion on mouser? I'm trying to figure this out soon.

MorganS

Sorry, I got diverted by the PWM part of the discussion.

It would probably work OK for infrequent switching with no driver.
"The problem is in the code you didn't post."

Paul__B

Any thoughts on the use of mosfets Paul, or a good suggestion on mouser? I'm trying to figure this out soon.
I'm not sure if I checked before, but if the first FETs nominated are rated for 35 A continuous DC, then three of them in parallel (and using the diode as well) should be reasonably conservative for a 75 A load.

MarkT

An alternative to paralleling MOSFETs is to get a suitably rated single device, for instance: http://uk.farnell.com/ixys-semiconductor/mmix1f520n075t2/mosfet-n-ch-75v-500a-smt-21/dp/2674812?st=iyxs  (although that requires surface mount).  Also http://uk.farnell.com/ixys-semiconductor/ixfn200n07/mosfet-n-sot-227b/dp/4905659?st=iyxs
Low on-resistance device or devices are mandatory for low voltage high current load. Paralleling is
probably a good option, but be
sure to use gate resistors and to keep their value low, a few tens of ohms might be good.  Without
gate resistors the weakest device takes all the current briefly, and RF oscillation is possible.  The main
advantage of a single device is simplicity (and the use of a package designed for high current wiring,
ie stud/screw terminal connection)

BTW you shouldn't be looking at the current rating, its never as relevant as the on-resistance and the
power rating.  You'd normally never run a MOSFET anything like its full current rating, and here you
have a 3V load so its really important not to, as the voltage drop at max current can be several volts!!

For this high current level a MOSFET driver chip is mandatory, you have to be able to push
lots of current into the gate to switch fast enough to prevent the MOSFET vaporizing.  Halfway through
the switching process the device will be dissipating massively so you want to keep that down to micoseconds with high power load. 

The best place to learn about MOSFET driver chips is in MOSFET driver chip datasheets.  For a simple
single low-side switch checkout the MIC4422 which can drive many amps of gate current.
[ I will NOT respond to personal messages, I WILL delete them, use the forum please ]

Go Up