PDA

View Full Version : Question about soldering for pumps


nrbelk
11/01/2011, 08:52 AM
I have been using a lot of information from Disc1 on how to create (and wire) my DIY dosing pumps.

I am now in the soldering stage. I had set up the circuitry pretty much the same way that Disc1 shows in his schematic that he has shared (except I added an npn transistor that switches the mosfet transistor).

The problem I am having right now is that the current never seems to turn all the way off on the pump. It will rev up during its "on" cycle (I have a test program that turns them on for 5 seconds and then off for 5 seconds). But during its "off" cycle, the pump just slows down, but never shuts off completely.

I tried to do some troubleshooting and I found that when I make a connection using a multimeter from ground to the gate pin of the mosfet, it starts working as expected. Does this mean anything? How should I test this? Why would it not turn off all the way?

Thanks!

der_wille_zur_macht
11/01/2011, 09:14 AM
Can you perhaps link to a schematic of the circuit you are trying to build, and describe the intended functionality?

nrbelk
11/01/2011, 09:37 AM
I don't have a link to my schematic but I could try to explain it.

Arduino pin 9 -> resistor -> basepin (TIP31AG)
Arduino 5v -----------------> collector (TIP31AG)
emitter (TIP31AG) --> gatepin (IRF510)
Pump Negative --> drain (IRF510) -> diode -> 12v
Ground --> source(IRF510)

A line off the 12v goes to the pump and the diode is a feedback protection thing.

Does that help?

nrbelk
11/01/2011, 09:38 AM
Oh, and arduino ground and the ground from the 12v are connected together

nrbelk
11/01/2011, 10:05 AM
forgot to describe the functionality.

I am sending an on/off (high/low) signal from pin 9 on the arduino. This should allow the 5v at the transitor to trigger the gate pin on the mosfet. The mosfet will then complete the circuit on the pumps and allow current to continue flowing to ground.

nrbelk
11/01/2011, 11:45 AM
bump

kcress
11/01/2011, 01:28 PM
Here's the problem. FETs have such low drive needs because they are voltage controlled that the leakage current coming out the the processor pin is enough to partially turn ON the FET it's hooked to. Because of this you must ALWAYS include a pull-down resistor from the FET's gate to ground. (N-FET) The resistor provides a path for the processor's leakage current to dissipate. Use a 47k ohm or anything between 10k and a 100k.

nrbelk
11/01/2011, 01:58 PM
Thanks so much for that information!!! That makes sense with what I saw happening when I was trying to find the problem.

So let me explain what I need to do back to you to see if I understand it right. To fix this, I should just be able to connect a 46k ohm resistor to ground and connect the other end of the resistor to the MOSFET's gate pin?

Thanks!

nrbelk
11/01/2011, 04:24 PM
Well, I hope that I understand that alright, I'm off to buy more resistors!

Is it just me or is the DIY forum more popular now these days? I never seen threads move down as fast as they are lol.

nrbelk
11/02/2011, 08:15 AM
so I soldered in a pulldown resistor. While it appeared to make everything more stable (I would get the same results each time rather than the fluctuating results), I still have the issue with the pump running.

I tested the resistance between the source and drain pin while the circuit was running. It would go down to really low resistance at the right time and the pump would run at normal speed.

But during the time that the pump should be off, the resistance never registered as out of limit, but rather around 52k (if I'm reading my multimeter right). Does this mean my mosfet is bad?

DustinB
11/02/2011, 09:02 AM
Try using a resistor with a lower resistance, something along the lines of a 10k or so.

nrbelk
11/02/2011, 09:15 AM
The pulldown resistor should have a lower resistance? Swap out the 47k ohm for a 10k?

nrbelk
11/02/2011, 10:00 AM
I just read that without a pulldown resistor a mosfet can self destruct easily. So easily that I have destroyed my MOSFET?

Here is the link that says this:

http://www.cadvision.com/blanchas/hexfet/nch-sw.htm

nrbelk
11/02/2011, 10:12 AM
Here is some more information I found:
<quote>
If you use the Atx power supply, you will not need the 5V regulator. Instead, connect the positive fan lead directly to the power supply's 5V line. The negative side of the capacitor is connected to Gnd and the positive side is connected to 5V. I suggest you use this one. Also, all Gnds should be connected together. I have condensed everything down into this for you:

Arduino IO Pin -> Current Limiting Resistor -> Mosfet Gate Pin

Mosfet Gate Pin -> Pull Down Resistor -> Mosfet Source Pin -> Gnd -> Atx Gnd -> Arduino Gnd -> Negative Side Of Capacitor

Mosfet Drain Pin -> Negative Fan Lead

Positive Fan Lead -> Atx 5V -> Positive Side Of Capacitor
</quote>

This is somewhat like what I'm trying to do. Although I'm not using any transistors.

My question now is if I need to connect my gate pin to the source pin with the pulldown resistor like what I highlighted in red shows?

nrbelk
11/02/2011, 04:22 PM
bump

kcress
11/02/2011, 07:21 PM
Yes, not having a pull down can cause the FET to sit in some intermediate ON-ness which can dissipate a ton of heat in the FET. Too much and yes, a damaged FET is the result.

Here's the wiring you need:

I/O pin to a 1K resistor to the gate(This is only to protect the processor pin from screw-ups.)

Go with a 10k from the gate to GND.(Overkill but what the heck. Normally 47k is more than needed but I haven't messed with Arduinos much.)

Hook your supply to the fan+.

Hook the fan- to the FET drain.

Hook a diode from the drain to the source on the FET. It needs to have the BAR on the drain and the other end on the source. This is to prevent inductive kicks from blowing your FET, a likely possibility in your case.