Teensy 4.1 - Recommend pump controller method?

Status
Not open for further replies.
Hi all,

I am looking for some advice on the best method to control a 12V, 4.5A Pressure Pump using a Teensy 4.1?

The system was previously run using an Arduino Mega and a 24V 5A IRF520 Mosfet Driver Module to control this motor. I am having a hard time find a finding a suitable Mosfet alternative for the Teensy? I'm also a bit worried I am going about this the wrong way (e.g. would a motor controller work better/use less power/etc)?

  1. Is there a better method than a Mosfet module to control the pump for a teensy (I am very new to circuits, so forgive my ignorance)? The current mosfet gets hot.
  2. I have seen that some people suggest the FQP30N06L, but read conflicting comments suggesting this would not work with 3.3V signal and get really hot under load? Would a driver chip be a method to fix this as a way to work with 3.3V?


Thanks for any advice! I'd also appreciate any guidance on topics I should read up on more:)
 
Could you give us a link to the specific "24V 5A IRF520 Mosfet Driver Module"? I tried a quick Google search, but it just turns up the actual IRF520 transistor, so I don't have any info tech about the module which I could use to give you an opinion about whether it would work with Teensy.

Regarding the FQP30N06L transistor (not a module, just the transistor) and nearly all large MOSFET transistors, the main issue is the "total gate charge" spec, which appears on page 2 of the datasheet. Many transistors also need more voltage, but this one says its maximum threshold is 2.5V, so the 3.3V from a Teensy pin could turn it on. But as you can see in figure 1 on page 3, it works much better if you drive the gate with 6 to 10 volts.

The main issue is gate is a large capacitor, so you need a resistor between the Teensy pin and the gate to limit the current pulse when Teensy tries to change the voltage. I would recommend 150 ohms as the lowest safe value.

150 ohms will limit the current to about 20mA. So at least during the beginning of the pulse, you'll be delivering 20mC of charge per second. The max gate charge is 20nC. So if 20mA were sustained, you'd charge the mosfet gate in about 1us. Of course it will slow down as it goes. You could go a lot of non-linear math, or just figure it'll take several microseconds for the mosfet to transition from fully off to fully on, or vise versa.

Whether several microseconds matters is a good question. During that time, before the transistor is fully on, much of the 24V will be across the transistor. Some will still be on the motor too, and what will happen with the current is a complicated question because the motor (probably) has substantial inductance. But a very worst case way of thinking is to assume the full 24V and 5A will be borne by the transistor during those microseconds.

Figure 9 in the datasheet tells how well this transistor can survive those conditions. The good news is it's pretty tough for short pulses, as you can see by the 10ms and 1ms curves. But if those conditions were to persist indefinitely, you can see in the DC curve the transistor can only survive about 3 amps while it has 24V between drain to source.

The other big issue is heat build up. During those microseconds, the transistor will dissipate a lot of power. If you only turn the motor on once per minute and leave it on, probably no big deal. But if you use PWM to rapidly switch the transistor on/off to control the speed or torque, then you'll suffer that power in the mosfet thousands of times per second. It can easily overheat. There are ways to estimate the temperature, but this message is already getting quite long.....
 
So regarding using the FQP30N06L (just the bare transistor, not a module with other parts), connecting it to a Teensy pin with only a 150 ohm resistor is kinda marginal, as you'll be running it along that lowest performance curve of figure 1, but it will probably work as long as you don't use PWM and don't turn the transistor on & off rapidly over and over again.

The other alternative is to use a special mosfet gate driver chip. This lets you turn the mosfet on/off very quickly, which minimizes the time it can dissipate power. But the downside is you need another chip, and that chip needs 10V to 15V power. A chip I have personally used with good results is TPS2814. It can control 2 mosfets, but you can use just 1 of its drivers and leave the other output unused. If you look at its datasheet, you'll see they make several versions with other features. Some have a 12V regulator built in, which can accept up to 40V input. Those might be more convenient, where you could run the chip from your 24V power without having to create a special 12V supply with even more parts.

One other really important thing to consider, if you haven't already, is the diode that's normally needed across the motor. Normally these need to be "fast rectifier" type diodes. See figure 32 in the TPS2814 datasheet (and assume L1 is the motor). The faster you cause the mosfet to go from fully on to fully off, the more you need this diode to protect the mosfet from voltage spikes caused by the "back EMF" energy stored in the motor's magnetic field.
 
Thanks Paul, that's really helpful! I appreciate the references to the important parts to check out in the datasheets as well!

The project is currently running of a 12V battery! The previous mosfet was a IRF520N on a Duinotech module.

I had put together a circuit based on what I've been reading in the forums so far:

pump_power.PNG

Does the fast rectifier diode need to be located at the motor, or can it be on the PCB? The wire from the mosfet to the pump is about a meter long. I didn't build the original system, so could one of these diodes be inbuilt in the pump, as the original mosfet doesn't look to have a diode?

I will have a read/google the TPS2814 gate driver. I could wire 12V from the battery to the gate driver if that is a good option to get around the 3.3V on the Teensy? Reading the specs of the TPS2814 notes a peak current of 2A. Would I need one that can handle 4.5A though because of the pump specs (12V, 4.5A Pressure Pump)? The driver gate MCP14A0451/2 has a High Peak Output Current of 4.5A (typical). Is this the value I need to meet? So much info in the datasheet...

Thanks for all the pointers and advise! So much to research!!
 
Last edited:
Reading the specs of the TPS2814 notes a peak current of 2A. Would I need one that can handle 4.5A though because of the pump specs (12V, 4.5A Pressure Pump)?

No. The 2A peak spec is determines how hard it can drive the mosfet's gate, turning the mosfet on/off. Remember, the mosfet needs a certain amount of charge on its gate. How fast you can move that many electrons in/out of the gate determines how quick the mosfet will go from on to off and vise verse. With Teensy and a 150 ohm resistor, your peak gate current would be 0.02A. So if you add that chip with strong output, the mosfet can turn on/off about 100 times faster than if the weak output of a Teensy pin controls the transistor.

Two things will cause the mosfet to heat up. While it stays on, the 4.5A current flows through the resistance of the mosfet, which is 0.0035 ohms if turned on with 10V. This power is I^2 * R = 4.5 * 4.5 * 0.0035 = 0.07 watts. If driven by only 3.3V, the resistance will be higher, so the steady on power will be higher. The other power is during the switching time. This could be as high as 100 watts, but only for a very short time. Using the driver chip will cut that time by a factor of about 100. Without the driver chip, if you only switch once per minute, probably not a big deal. If you use PWM which switches thousands of times per second, it very well could be enough power to burn the mosfet without that driver chip.

Clamping diode physically close to the transistor is usually good. The faster the mosfet goes from on to off, the more important that diode is. If you use the driver chip, you're switching 100X faster, so definitely use a good (fast) diode. The diode does not need a large current rating. It only conducts in the brief moment after the mosfet turns off. Nearly all diodes rated for 1A can handle 4.5A pulses. The "reverse recovery time" spec is what matters.
 
That really sells the driver chip. For a few extra dollars, it seems like a worthwhile investment. Thanks for clarifying. I'll do some reading, and try come up with a design then that uses a driver alongside a mosfet for the Teensy!

Really appreciate the support and advice! Thanks:)
 
This is the updated design using a FQP30N06L and a NCP81074A Single Channel 10A High Speed Low-Side MOSFET Driver based on the video from MicroType Engineering.

Posting in case I missed something obvious?

pump_power_2.PNG

So I need to learn what tuning resistors at the gate means in relation to switching speed? Will post as I learn more!
 
You are using a 10A driver and then limiting it to 40mA of gate current. Use something far lower than 300 ohms.
 
You are using a 10A driver and then limiting it to 40mA of gate current. Use something far lower than 300 ohms.

Thanks for the heads up! I did a few calculations with a lower resistor, but will need to look into the switching gate resistors so I choose a suitable option.

gate.PNG
 
This circuit works with a 5V VCC, but when I swap it out with a 7.4V or 12V, the mosfet gate driver doesn't seem to turn on the pump, and voltage is delivered to the gate? I'm so confused :confused:

You haven't indicated any Vcc in the diagram so its not at all clear what you mean.
 
Apologies, I should have said the 12V signal on the schematic (edited to clarity). The design is for a 12V battery to power a pump, but when I was testing the PCB I had it hooked up to a 5V power source. So when the 12V was set to 5V it seems to work, but when I plug in the 12V battery, the gate driver doesn't seem to work?
 
Well as the MOSFET is logic level I'd expect it to work at any of those voltages no problem.
I'd suggest careful checking of your circuit, especially the OUTL connection.

Note the 100nF cap on the output should be more like 1nF to 10nF and be a ceramic type mounted on the motor terminals directly,
else its doing very little to suppress RF.
 
Well as the MOSFET is logic level I'd expect it to work at any of those voltages no problem.
I'd suggest careful checking of your circuit, especially the OUTL connection.

Note the 100nF cap on the output should be more like 1nF to 10nF and be a ceramic type mounted on the motor terminals directly,
else its doing very little to suppress RF.

Thanks for the feedback! I was hoping the response would be that I'd made an obvious mistake or messed up the values, but if the design looks fine then there must be a fault elsewhere. I'll need to do some more investigating. Cheers all
 
Status
Not open for further replies.
Back
Top