# Using MOSFETs to control resistive loads

#### durg

Joined Oct 23, 2019
2
Hello everybody!

I want to dump energy from some solar panels into heating by using two heating elements (R1 and R2). To maximize the power output from the solar panels I plan to use a microcontroller to change between single heater, series and parallel configurations.

Now, I tried to come up with a circuit using switches (see the attached image file below).

Using relays, I think this would work nicely. However, I was thinking it could be a smart move to design it using MOSFETs instead of relays for higher reliability. Now, I am not very familiar with using MOSFETs as switches in such a configuration. (I have only used MOSFETs for low-side switching of a single load using N-channel MOSFETs).

The maximum current (when both heaters are connected in parallel) would be around 8-10A at 220VDC. I will have unregulated 36V (from a single PV) and regulated 12VDC available.

I guess S1 should be okay using a N-channel MOSFET?

Any tips for choosing and configuring the MOSFETs S2 and S3 or pointers in the right direction (reading material, example circuits, etc.) would be greatly appreciated. I suppose it would make sense to use a P-channel MOSFET for S3, what about S2?

#### shortbus

Joined Sep 30, 2009
8,591
To maximize the power output from the solar panels I plan to use a microcontroller to change between single heater, series and parallel configurations.
Can you explain how dumping power increases the power output?

#### SLK001

Joined Nov 29, 2011
1,548
I would stick to relays.

If you DO want to use MOSFETS, first figure their individual power dissipation capabilities. Find the rds-on for your chosen fet, then calculate the power dissipation. At 400V, rds-on is probably around 0.2 ohms, so Pd is around 20 Watts per device. That means you'll have to cool your devices. Also, the Vgs for full rds-on will be quite high (for digital control), probably around 20+ volts. Given all this, yeah, I'd stick with relays.

#### crutschow

Joined Mar 14, 2008
27,232
use a P-channel MOSFET for S3, what about S2?
Yes, it's easiest to use P-channels for both.
But since N-channels have lower on-resistance, you might consider using those with a high-side driver with built-in bias generator.

But I agree with SLK001, for those voltages and currents, using relays is likely the best choice, since I would assume the switching will not be occurring that frequently.

#### durg

Joined Oct 23, 2019
2
Can you explain how dumping power increases the power output?
Thanks for replying! Sorry, I was a bit unclear. All the energy from the solar panels are going to be stored by heating oil in an insulated tank. Which can later be utilized for heating water, cooking and maybe other things.

I'm a newbie, but as I understand it there is a optimal point (maximal power point) on the IV curve where the power is maximal. And I want to adjust the load (depending on watt/m^2) so that we are close to that point.

I hope that makes sense.

Also: Thanks SLK001 and crutschow for the suggestions. I think I will try to use relays to begin with and see how frequently it will switch.

#### shortbus

Joined Sep 30, 2009
8,591
Thanks for replying! Sorry, I was a bit unclear. All the energy from the solar panels are going to be stored by heating oil in an insulated tank. Which can later be utilized for heating water, cooking and maybe other things.
Most people over the years have dumped the resistor heat directly into the water tank not in to some type of oil. What happens if you get a leak in the oil to water heat exchanger? You would contaminate your potable water if that happened.

You might be thinking wrong on your parallel and series of the heater too, I don't know but will give some information to you. The heaters are basically a resistor so they will follow the resistance rules. https://www.kitronik.co.uk/blog/how-to-calculate-resistors-in-series-and-parallel/ So parallel will give you less heat into the liquid.

Last edited:

#### crutschow

Joined Mar 14, 2008
27,232
I want to adjust the load (depending on watt/m^2) so that we are close to that point.
For that you could use a microcontroller to monitor the output power for each setting and select the one that gives the most power.
You do this periodically (say once every few minutes) to see if the optimum power point has changed.

But I'm not sure how close you will get to the optimum with only four settings.