Ok, so, fun with heatbeds and Mosfets… following @Thomas_Sanladerer 's recent video and having a new Migbot Prusa i3 incoming I decided to try switching with a 'fet over a SSR, which my other 2 Migbots run… I keep a stock of IRLB3034’s which are high amp N channel MOSFETs, however the resistance is a little on the high side, so I put 2 in parallel to negate this. Wired it up following the diagram on the video, using a 15k ½ watt resistor in place of the 10k (shouldn’t matter too much, as far as my understanding goes) and constant heating, then realised the 'fets that were recommended are P Channel, so I then switched the polarity of the circuit, which should have worked, but the same thing happened, constant heating, 'fet gets warm. I’ve since ordered in a few IRF3708’s wired it up again, but still the same thing happens. I feel like I have the dumb at this point. Board is a MKS Base v 1.2 powering a MK2B heater (these are going to be replaced at some point for a Silicone one as I need fast heating), the only thing I can think of at this point is to put the 100 Ohm Resistor in line to see if that helps, and check the switching voltage (I’ve assumed it’s 5v as the MKS is just an all-in-one RAMPs solution), the 15K is resistor is too much (as I stated above it shouldn’t be), or I have a total dumb and not wired it correctly, I’m right in saying that wiring the negative side on the source pin to the PSU and signal from the board is fine? I can’t think of any other place that would go to ground on the printer.
What you seem to be describing is a MOSFET that is not fully switched ‘on’.
When a MOSFET is fully on, the drain-to-source resistance should be tiny. The datasheet for the IRLB3034 says that when fully on, the resistance is 1.7 milli-Ohms.
A MOSFET that is only partly on will exhibit much higher drain-to-source resistance than the full-on resistance, and thus dissipate a significant amount of power internally (it gets very hot).
The ‘on’ voltage is 4.5 Volts at the gate (according to the datasheet), but that voltage is more likely to be the point where the device begins to switch on. In fact it may not be fully on until 5 volts or higher.
If my guess is correct, you can probably just change to another MOSFET type with a lower ‘on’ gate voltage, say around 3.5 V, so that switching with a 5 V signal will ensure that it is fully on and thus it will dissipate almost no power internally.
@Paul_Gross aye, although I think the 4.5v switching should be fine, the MKS is a RAMPS all in one after all. I build unregulated MOSFET Vape Mods for a living, the same behavior is exhibited when the resistor fails, it’s a very similar circuit, hence the comment above about the resistor not being a high enough wattage. Basically what I’m saying is if the resistor fails between the gate and source, you get constant on.
The 15k resistor is a pulldown, keeping the gate from floating when the mcu powers up. It also helps a little with noise. The value of this resistor is not as important as you think, and the current flowing through it is very small.
If you are using an n channel mosfet, the source should be connected to GND.
@Stephanie_A - Aye, I realise the Ohm value isn’t an issue, hence using the 15k over the 10K that Tom recommends in his video it was more the watt rating I’m interested in. Basically, if that isn’t enough it’s going to act like it’s not there as the resistor won’t hold, and the MOSFET will be permanently open. I say this as I’ve had a set of resistors that have held then failed on the circuit I use on a daily basis, I moved from 1/4 to 1/2 watt and they hold without an issue. I could be totally wrong here, and am going on my experience of the N channels, but I’m using the IRF3708, recommended from Tom’s video, which is P channel. However, basically you’re saying the 1/2 watt should be fine?
My pulldown resistors are 1/24th watt. Ohms law dictates how much power the resistor will dissipate. Assuming your gate voltage is 5v, the equation would be p=(5*5) /15000 quite insignificant.
However, since you are using a p channel, the circuit completely changes. See http://www.electronics-tutorials.ws/transistor/tran_7.html
P channel turns on when Vgs is negative. Also, the source pin is attached to the positive voltage. A pullup resistor connects the gate to the source.
A p channel operates in the opposite of an n channel. When Vgs is zero, the mosfet is off. In this case, the gate would need to be 12v. When Vgs is negative, the mosfet turns on. So if you apply gnd or 0v, the potential is -12v.
Now, since the arduino runs on 5v,the max voltage to the gate is 5v. The source is 12v, so Vgs would be -7v,and the mosfet would still be on.
Ok, I think I got that one, admittedly my knowledge is limited, so, the IRF3708 isn’t a good 'fet to use? RDS(on) is 4.5v Vgs, so I’d have thought it would have held, it’s also one that Tom recommended in his video. So I’d have thought it would have worked, unless I’ve missed something. Just to clarify, with a P 'fet… Resistor bridges the gate and source, positive of the controller to gate, negative of the board and PSU to source, positive of the bed heater to the PSU, negative of the bed heater to drain?
What Tom suggests is an n-channel fet. The irf3708 is n-channel.
Since it is an n-channel the source should be connected to negative, gate connected to the i/o pin of the mcu that is designated for it, and negative side of the heatbed to drain.
I would recommend leaving p-channel MOSFETs alone unless you really understand what you are doing. Even some experienced people seem to make mistakes with the p-channel devices, because high-side switching is almost always a more complex problem than low-side, which is exactly what @Stephanie_A is suggesting.
Also, p-channel MOSFETs have a slightly higher on-resistance than n-channel, so they will get hotter faster, which is generally not desirable.
Stephanie’s suggestion is good. It is the easiest and most common solution, and should give the best results for you.
Again - remember to make sure you choose a MOSFET that is fully switched on with 5 V at the gate, otherwise it can get very hot while conducting in the partly-on state. There are many low-cost n-channel MOSFETs that switch with gate voltages around 3 V that you could use.
@Stephanie_A , @Paul_Gross Think I may have derped somewhere along the lines, I do have the IRF3708, somewhere I must have had some crossed wires when researching and got it into my head that the IRF3708 was P channel.
So, that just leaves my wiring and I think I know what I’ve done. I had a wire coming from the MCU and another wire to the PSU on the negative side to the Source pin. If I understand Stephanie correctly, it’s connect the source pin wire to the PSU,The gate to the positive of the MCU (the MKS Base has a heat bed header) and the Drain to the negative if the heat bed, I should be golden.
Also, if I hadn’t derped on the N Channel and the wiring, the IRLB3034s in parallel would probably have worked… facedesk
That sounds about right, but you’re also confused on the gate connection, you call it “positive of the mcu”, what you’re actually connecting it to is the i/o pin. The mcu should be configured to toggle this pin to turn on the heater. The source is connected to the negative of the psu, and your 15k resistor from the negative/source line to the gate. The 100ohm resistor on the gate is optional, it prevents ringing and protects the gate and mcu pin. If the wire to the gate is long, I would add it (it needs to be as close to the gate as possible). It’s not critical though.
Make sure your wires are sufficient gauge too. The gate wire can be small considering it has very little current, but the other 2 must carry the full current of the heater, for a heated bed this can be anywhere from 10-15 amps.
@Stephanie_A The reason I called it that is basically due to the fact that there is a Positive and negative connection block on the MKS base, basically to power the heat bed off the board. I realise my terminology isn’t totally correct, just what worked in my head so I know what I’m doing, but thanks for the help, I’ll give that a shot tomorrow morning when I’m at my printers and apologies for having a dumb with the P and N channels >_<
Ok, small update, even wired as above, took a multimeter to the bed out on the MKS board. it’s shunting out 12v… so those MOSFETS aren’t any good. As it is, I’ve got a DC-DC SSR in for the new printer now, but have a Fusebox CoreXY build in the works, so I’ll use them for that…