Need Heavy Duty Transistors for Controlling Larger Motors with PWM?

[kolleamm]

I want to control the speed of a larger sized motor (roughly 8" x 4" x 4", 48v, 5.8A, 30A max ) using a transistor and PWM. The only problem is I can't find any transistors that have enough resistance to prevent the current from passing directly through.

I purchased some mofset transistors online and when the circuit is closed the resistance reads 20 Ohms, and when open 0.6 KOhms.
So I'm assuming a higher current would just pass right through it.

 

[DaveE]

I want to control the speed of a larger sized motor (roughly 8" x 4" x 4", 48v, 5.8A, 30A max ) using a transistor and PWM. The only problem is I can't find any transistors that have enough resistance to prevent the current from passing directly through.

I purchased some mofset transistors online and when the circuit is closed the resistance reads 20 Ohms, and when open 0.6 KOhms.
So I'm assuming a higher current would just pass right through it.

Transistors are good at turning off, all the way off, like MΩs. However, your circuit has to allow them to do this. So:

1) Maybe your circuit isn't as "closed" as you think.
2) Maybe you aren't making the measurement properly.
3) Maybe the resistance you are measuring isn't through the "off" transistor, but through other parts of your circuit.
4) Maybe the MOSFET gate is left floating and retains some previously applied voltage which will keep the device on.
5) Maybe that MOSFET has been damaged. This is less likely than people usually think, though.

You may need to select a different transistor, but not because you can't make it turn off. They all turn off when done right.

It's very hard to give useful advice though when none of the details are provided.

 

[hutchphd]

20 ohms? That is far too large. Do you mean milliohms?

 

[kolleamm]

Transistors are good at turning off, all the way off, like MΩs. However, your circuit has to allow them to do this. So:

1) Maybe your circuit isn't as "closed" as you think.
2) Maybe you aren't making the measurement properly.
3) Maybe the resistance you are measuring isn't through the "off" transistor, but through other parts of your circuit.
4) Maybe the MOSFET gate is left floating and retains some previously applied voltage which will keep the device on.
5) Maybe that MOSFET has been damaged. This is less likely than people usually think, though.

You may need to select a different transistor, but not because you can't make it turn off. They all turn off when done right.

It's very hard to give useful advice though when none of the details are provided.

Thanks for your advice.
I tried a similar test with a small dc motor and it did work, but still no luck with the larger one.

This is the transistor I bought :
https://www.sparkfun.com/products/10213

I have a 10KOhm resistor across the gate and source as well to remove any charge from the gate while it's off.
Also, the gate and source are connected to an Arduino (the gate being in the PWM channel and the source connected to the Arduino's ground).

 

[kolleamm]

20 ohms? That is far too large. Do you mean milliohms?

It wasn't milliohms

 

[hutchphd]

Yes well the "on" current spec with any appreciable current is ~20 mohms thank goodness. I point out that 30 A thru 20 ohms is kWatts of heat...do you understand? The reading of your ohmeter is not reliable because the response ids nonlinear. Live and learn...

 

[Baluncore]

I purchased some mofset transistors online and when the circuit is closed the resistance reads 20 Ohms, and when open 0.6 KOhms.

Get a copy of the datasheet.
http://www.farnell.com/datasheets/1774697.pdf

Can you post the circuit you used to test them.

 

[DaveE]

How to test a MOSFET with a normal DMM:

1) Turn your meter on to the "ohms" scale; "auto" range if it does that, otherwise start at the highest range for these tests except when you expect a low value, then use the lowest range that gives you a reading. If you choose "auto", you can ignore everything below about switching ranges.

2) With the test leads installed, but disconnected from the device, verify that the reading is really really big. If it doesn't say at least 1MΩ, go buy another DMM, the Chinese ones on eBay are cheap.

3) With a MOSFET out of the circuit, connect both DMM leads to the source terminal on your device (DUT=device under test). A perfect meter will read 0Ω. Yours probably won't. wiggle the leads a bit to get a feel for how well the meter connects to the DUT. Write down the lowest consistent reading you get. This value represents a 0Ω reading (you are really measuring the resistance of the test leads, connections, etc.). You will subtract this value from any low Ω reading you get later.

4) Not really a step - a comment: Get in the habit of doing these sanity (or calibration) checks with all of your measurements. DMMs, Oscilloscopes, Network Analyzers, whatever. You need to know if your instruments are telling the truth, or if you're using them wrong. Occasionally this will save you hours of confusion.

5) Connect the negative lead to the Source terminal, and the positive lead to the Gate terminal and measure the Drain Source to Gate resistance. Verify that it's >1MΩ. Reverse the polarity of the test leads and measure the Drain Source to Gate resistance. Verify that it's >1MΩ. It really should be >10MΩ, but don't worry for now if it's not.

6) Connect the negative lead to the Source terminal, and the positive lead to the Drain terminal (this is for N-Channel MOSFETs, you'll reverse the polarity for P-Channel devices). Connect the Gate terminal to the Source terminal (a short circuit, you can do this with a knife blade, paper clip, whatever is conductive) and measure the Drain to Source resistance. Verify that it's >1MΩ. Disconnect the short from Gate to the Source and verify that the Drain to Source resistance doesn't change much, at least for several seconds.

7) Set your DMM to the highest Ω range. The range is important; you want the DMM to put out the largest test voltage it can, and often the low Ω ranges use low test voltages. Connect the negative lead to the Source terminal, and the positive lead to the Gate terminal for at least 1 second. Now, carefully move the positive lead from the Gate terminal to the Drain terminal, without accidentally connecting the Gate to anything else (like the Source). Change the Ω range to a low setting and and measure the Drain to Source resistance. Verify that this is a low value (let's say <10Ω or so) for at least a few seconds. What you are doing in this step is charging up the Gate capacitance to the DMM test voltage and leaving the charge there while you measure the saturated channel resistance. Any accidental connection to the Gate may allow that charge to escape.

8) If your DMM has a diode test feature you could also test the body diode (anti-parallel to the Drain-Source channel) of the DUT. But it's unnecessary, so I won't describe it unless you ask me to.

- I can't guarantee this process works for all DMMs, but I can't recall a meter I ever used (out of many, many meters) that wouldn't do it properly.

- Be careful how you hold the device (or better yet, don't hold it) you don't want to mistake the resistance between your fingers for the DUT measurement. Heat sink tabs are connected to the transistor.

- This is a basic functional test to spot damaged devices, not a method to verify specifications. That's why the datasheet (THAT YOU NEED TO READ!) has test circuits and footnotes.

- This test is for discrete, single MOSFETs. If your MOSFET is part of an IC then you are measuring whatever circuit is in there, which may not act like a simple MOSFET.

edit: fixed some typos.

 

[kolleamm]

Hey guys, here is my circuit.

 

[Baluncore]

You have drawn an NPN BJT, with E,B,C, not an N-channel MOSFET with S,G,D.

Are you certain that the Arduino output is going to 5V ?, might it be only 3V ?
The 10k pull-down resistor makes it harder for the Arduino pin to go high. That resistor is not needed if you can quickly initialise the port to output and low.

A motor is inductive. You should connect a reversed power diode across the motor to catch the positive flyback voltage spike that could kill the MOSFET. That damage may have happened earlier and could explain your resistance readings.

 

[Baluncore]

Here is an example circuit with MOSFET and flyback diode.

 

[kolleamm]

You have drawn an NPN BJT, with E,B,C, not an N-channel MOSFET with S,G,D.

Are you certain that the Arduino output is going to 5V ?, might it be only 3V ?
The 10k pull-down resistor makes it harder for the Arduino pin to go high. That resistor is not needed if you can quickly initialise the port to output and low.

A motor is inductive. You should connect a reversed power diode across the motor to catch the positive flyback voltage spike that could kill the MOSFET. That damage may have happened earlier and could explain your resistance readings.

Ah interesting, okay that gives me a lot to work with.
My question is if the transistor could handle the initial current wouldn't it be able to handle the flyback voltage spike as well?
Also what specs should the reversed power diode have? I found a bunch of them on eBay but am having a hard time choosing the right one.

 

[Averagesupernova]

My question is if the transistor could handle the initial current wouldn't it be able to handle the flyback voltage spike as well?

Take a close look at what you asked. High current and high voltage are two different things. Each has a different mechanism for causing damage.

 

[DaveE]

The underlying concept to keep in mind is that you must always provide a path for the motor current to flow through. Also other inductive loads, like solenoids, transformers, etc. If you try to disconnect an inductor with current flowing in it it will make, theoretically, as much voltage as necessary in order to keep the current flowing. This is how the ignition system in older cars works, they make thousands of volts to generate a spark, which is probably not what you want in your circuit.

 

[Baluncore]

Also what specs should the reversed power diode have? I found a bunch of them on eBay but am having a hard time choosing the right one.

The flyback diode needs to be a fast recovery diode, rated to carry the motor current.

An experimenter can sometimes cheat and use another cheap MOSFET with a good body diode specification. That is the equivalent of a half bridge circuit, but without the high-side control.
Here is an example;

 

[DaveE]

I found a bunch of them on eBay

Unless you have to buy from them, it's time to ditch eBay, IMO. These parts aren't that expensive for a one time build. I mostly use digikey.com, you can get (nearly) exactly what you want and they have traceable quality (i.e. know supply chain and links to documentation).

There are many other distributors like digikey, https://www.newark.com/ is another good site, but I won't list all of them. 2nd tier are hobby sites like Sparkfun, Adafruit, Jameco, etc.

Bottom tier is eBay. They do have good stuff, but they also have questionable stuff.

 

[DaveE]

Also, I'm a bit concerned that you are using a 60V MOSFET in a circuit with an inductive load and a 48V power supply. Switching transients are very likely to exceed the transistor rating unless you are an excellent circuit designer (great layout, fast diodes, snubbers, etc.). It can be done with 60V, but it's not that expensive to buy a 100V (or 200V) MOSFET as insurance. If your motor controller works for 15 days and dies, this may be why.

 

[Frodo]

Curtis Instruments make a range of controllers including the 1223 model for my wife's heavy duty off road mobility scooter. The 1223/1233 range goes from 24 to 36V, with current limited from 45 to 110 Amps, and with a 1 hour rating of 30 to 40 Amps.

The output circuit is multiple power MOSFETs in parallel with a copper bus necked at each MOSFET to balance the currents. I could dig out the part number if required.

It is fully programmable, has soft start and stop and full four quadrant operation. There are comprehensive manuals on the web site.

RS Components is a good UK source of components and controllers. Their United Automation, DC Motor Controller is PWM, input voltage 6 - 24 V dc, current 40 A max. It is £33.

 

[kolleamm]

Thanks for all your input everyone, I will continue to work on the project with the knowledge you've given me and give an update soon.