Voltage drop when load (12V DC Compressor) is placed into circuit

Other Hardware General Electronics
1

Hello all,

I'm working on a school project involving controlling a 12V compressor's ON/OFF state with an Arduino UNO and a "switch" MOSFET. I am using a 12V car battery to power the compressor. An identical circuit to the one I have can be seen here.

When I send a HIGH signal to the MOSFET to close the switch while there is NO compressor in the circuit, I read ~12.5 volts across the terminals where the compressor would normally be attached. However, when I place the compressor into the circuit, and read the voltage across the same two terminals, I read ~5.5 volts. I'm not sure what is causing this voltage drop.

If anyone has any idea, I would appreciate the help.

Thank you!

Compressor
MOSFET Sparkfun Page
MOSFET Datasheet

2

How have you things connected up? (Post a picture of MOSFET and connections - it says more than 1000 words)

If you have 5.5V across the MOSFET and the compressor running, something will get VERY hot and go up in smoke in seconds.

// Per.

3

Zapro:
How have you things connected up? (Post a picture of MOSFET and connections - it says more than 1000 words)

The circuit that I linked in the original post is how I have wired my circuit. Unfortunately, I cannot post an actual picture of my circuit right now as I'm not on campus.

Zapro:
If you have 5.5V across the MOSFET and the compressor running, something will get VERY hot and go up in smoke in seconds.

Why do you say this? The MOSFET is rated for up to 60V, so I'm not understanding why 5.5V would damage it. The MOSFET does not generate a lot of heat during the 8 second intervals when the compressor is on. From what I felt it didn't anyways.

4

Did you measure battery voltage with the compressor on.
Perhaps the battery is bad, and the voltage collapses when the compressor starts.

That mosfet should have ~0.35volt across when that compressor is on (and draws 12Amp).
It has to dissipate ~4watt at that current. A small heatsink is needed.

Did you use a diode across the compressor, with the same current capability as the compressor.
Leo..

5

You need to measure the battery voltage under load to see what it is doing. As for your question about "why is 5.5v across the MOSFET a problem when it is rated for 60v" you are comparing apples and fish - the 60v is the maximum voltage the MOSFET can handle without shorting (breakdown). The 5.5v across it is a different issue - if you are dropping 5.5v across the MOSFET and pulling 2amps for example, that is 11 watts being dropped in the MOSFET - without a decent heatsink, it will rapidly overheat. Typically a small compressor will probably pull 2-5 amps. You need to measure voltages and figure out just what voltage is being dropped where in the circuit - starting with what is the battery doing under load.

EDIT - in looking at the compressor data sheet, I see it is rated for max 12A so if your MOSFET was dropping 5.5. volts at even 10A, you are looking at 55 watts ... the MOSFET would get very hot very fast.

6

The N-ch mosfet is a low side switch. In laymans terms, it puts ground to the circuit. If you tried this without the load (compressor) and connected the 12v to the drain and put 5v on the gate, it's going into meltdown. You just made the mosfet the load, and with a car battery putting out 100A or more, .....

When you hooked up the meter from drain to source, you SHOULD be reading 12v with no voltage to the gate. As voltage rises on the gate, the channel starts to conduct more and more, and voltage differential between drain and source drops while current increases.

7

Ok, I think I misread the OP.
If you are reading 12V without the load, and 5v across the load with the mosfet conducting, I would lean towards the battery being dead. The low side (ground) of the compressor should be at ground potential when the mosfet is conducting (provided the mosfet is fully on), which you can check by testing between the low side of the compressor and battery ground with the circuit on. You should get under a volt, and when the circuit is off, you should get Vbat (12v )
If so, charge the battery, or perhaps the battery is shot.

8

Wawa:
Did you measure battery voltage with the compressor on.
Perhaps the battery is bad, and the voltage collapses when the compressor starts.

Yes, I measured the voltage while the compressor was on. I should've specified that in the original post. When the compressor was on, I read a voltage of 5.5V across the compressor leads.

I also directly hooked the compressor up to the battery without going through the mosfet and other circuitry. It worked perfectly then.

gpsmikey:
You need to measure the battery voltage under load to see what it is doing. As for your question about "why is 5.5v across the MOSFET a problem when it is rated for 60v" you are comparing apples and fish - the 60v is the maximum voltage the MOSFET can handle without shorting (breakdown). The 5.5v across it is a different issue - if you are dropping 5.5v across the MOSFET and pulling 2amps for example, that is 11 watts being dropped in the MOSFET - without a decent heatsink, it will rapidly overheat. Typically a small compressor will probably pull 2-5 amps. You need to measure voltages and figure out just what voltage is being dropped where in the circuit - starting with what is the battery doing under load.

EDIT - in looking at the compressor data sheet, I see it is rated for max 12A so if your MOSFET was dropping 5.5. volts at even 10A, you are looking at 55 watts ... the MOSFET would get very hot very fast.

The mosfet was not heating up to any noticeably amount while the compressor was in the circuit and running, or when the compressor was out of the circuit. As for where the voltage is being dropped, I have no idea. Like I said, I read the correct 12V without the compressor in the circuit, but read 5.5V when it's in the circuit. I'll report back tomorrow after I've probed around a bit more.

tinman13kup:
When you hooked up the meter from drain to source, you SHOULD be reading 12v with no voltage to the gate. As voltage rises on the gate, the channel starts to conduct more and more, and voltage differential between drain and source drops while current increases.

With no voltage applied to the gate, I was also reading 0V between drain and source. While the 5V was being applied to the gate, I was reading 12V between drain and source. What I was reading is opposite to what you just said.

tinman13kup:
Ok, I think I misread the OP.
If you are reading 12V without the load, and 5v across the load with the mosfet conducting, I would lean towards the battery being dead. The low side (ground) of the compressor should be at ground potential when the mosfet is conducting (provided the mosfet is fully on), which you can check by testing between the low side of the compressor and battery ground with the circuit on. You should get under a volt, and when the circuit is off, you should get Vbat (12v )
If so, charge the battery, or perhaps the battery is shot.

I'm not sure that the battery is the problem because when I attached the compressor up to the battery directly, the compressor ran perfectly. I will charge it up just in case.

EDIT: Again, I read the voltage across where the load SHOULD have been and I read the correct 12V. The compressor was completely removed from the circuit, so I just had two open leads. This is probably not the correct way to test, but nothing melted. :slight_smile:

9

Hi,
Have you got the gnd of the battery connected to the gnd of the arduino?

Tom...... :slight_smile:

10

It sounds like you have not got it wired up like that thing you posted. It sounds like you have wired it up as a source follower.

11

It must be wired thus:

Where the coil represents the compressor motor. Note the diode across the motor - essential
to have this with an inductive load otherwise you'll have massive inductive spikes flying about.

Note the high value resistor between gate and source - without this you risk the MOSFET
half-turning on if the Arduino is powered down or reseting. If the MOSFET half turns on it
will dissipate loads and burn out very rapidly.

You should ideally use a better MOSFET, with an on-resistance of 10 milliohms or less
for a high current load. Note that the current rating of a MOSFET is not a useful parameter
as it represents the maximum current the device can handle with infinite heatsinking and
at maximum thermal stress. A 100A MOSFET is about right for a 10A to 20A load, but the
on-resistance is the parameter you actually work from.