MOSFET N - channel; half AC driving

[tester43]

Hi,

I am driving N-Channel mosfet normally from micro-controller using optoisolator. All ok. MOSFET is conducting half AC voltage from transformer.
Then I realised that I need to cut off all potential from Load- that leads me to options: buy P-channel mosfet or build high side N-Channel driver.
I wake up today with this question:
What would happen with MOSFET if I would put Gate voltage for a very short time (i.e. 1ms) at the level of 12V. Does it continue working with low resistance while Source voltage is rising (like Triacs do)... or does it overheat and stop due to low gate voltage when compared to Source?

ps.
It seems that due to use of AC I can't use traditionnal Bootstrap capacitor and diode. Capacitor would get charged to more or less 30-40V, then AC would fall to zero leading to immediate destruction (max Vgs = 20V).

ps2.
LOAD is resistance + thermocuple is series. To read TC voltage i need to shut off all other voltage from this area.

See the drawing for basic schematic

[tester43]

the longer I think about it - the longer it does not make sense.
I would need to have a separate voltage source higher by 10-12 volts than 24V peak following the AC all the time to avoid overvoltage on gate vs source.
I even tested that using lm7812 with gnd tied some voltage gives me expected outcome - lm7812 output is gnd voltage + 12;

somebody has a better idea?

[David Hess]

That is going to be a real mess.  You have a couple of options:

1. Use a TRIAC and optocoupler for driving TRIACs.  They will work fine at 24VAC and this is the same thing as a TRIAC based solid state relay.

2. Use back to back MOSFETs as a power switch.  They can be driven by a transformer or photo voltaic optocoupler.  This is the same thing as a power MOSFET based solid state relay which can work with AC or DC.

3. Same as 1 and 2 above but use an SCR or single MOSFET inside of a bridge rectifier as shown in the example below.

4. Is this suppose to be a battery powered application like a thermostat?  In that case, a latching relay or something like what is shown below is appropriate to minimize power drain.

[Audioguru]

The way it is now, the body diode in the Mosfet will conduct on the half-cycle you want it to be turned off.

[tester43]

but there is no other half of ac - it's rectified(?)

[Zero999]

MOSFETs are non-latching. If you need it to latch, until the current is interrupted, then you need an SCR.

What's the purpose of the circuit? There's probably an easier/better way to do this.

[tester43]

it's my hobby project: hakko t12 tip soldering station using correct power transformer (100-120VA)

[Zero999]

it's my hobby project: hakko t12 tip soldering station using correct power transformer (100-120VA)

Then why not  use a TRIAC and isolated TRIAC driver opto-coupler IC, such as the  MOC3022?
http://www.farnell.com/datasheets/97984.pdf

Power the microcontroller circuit from a rectified, smooth DC voltage and the element directly from the AC. To cut down on noise, zero crossing can be used. I was going to suggest an opto-coupler with built-in zero crossing, but the ones I could find aren't compatible with such low voltages.

[tester43]

first version of my project is exactly how you described it. I like mosfets better it's not producing heat - I found the one with _really_ low Rds_on.

second reason: I was reading JBC and Hakko schematics - they do not use triacs but mosfets - do not know exactly why yet

[Zero999]

first version of my project is exactly how you described it. I like mosfets better it's not producing heat - I found the one with _really_ low Rds_on.

second reason: I was reading JBC and Hakko schematics - they do not use triacs but mosfets - do not know exactly why yet

If the MOSFET is connected to a bridge rectifier, then it doesn't make much difference to the losses. If silicon diodes are used, it will be slightly worse than a TRIAC, because the voltage drops in the bridge, will be slightly higher, than the on stage voltage drop, of a TRIAC. Using Schottky diodes for the diode bridge, will make it slightly less lossy, than a TRIAC, but not much.

Two back-to-back MOSFETs and a photovoltaic MOSFET driver will give lower losses, than a TRIAC, (see link below) but only at low frequencies (below 100Hz or so) and low voltages <70VAC or so.
https://www.vishay.com/docs/83469/vom1271t.pdf

[tester43]

serious question: do you have any concept why commercial manufacturers do their soldering station power circuit using mosfets?

[Zero999]

serious question: do you have any concept why commercial manufacturers do their soldering station power circuit using mosfets?

I don't know why some commercial soldering iron stations use MOSFETs. There are probably various reasons, such as availability, being easier to control with a microcontroller, without an opto-isolater and being able to switch on and off, without latching, but I doubt efficiency is one of them. Look at the data sheet for a typical bridge rectifier, taking note of the full load voltage drop, now compare it with a TRIAC.

TRIAC
https://media.digikey.com/pdf/Data%20Sheets/NXP%20PDFs/BT138%20Series.pdf
On state voltage, 1.65V, at 15A.

Bridge rectifier
https://docs-emea.rs-online.com/webdocs/0fbb/0900766b80fbb6ae.pdf
1V per diode, at 20A, a little less at 15A, perhaps 0.9V, but there are two diodes in series, so that's a total of 1.8V!

Note, that I'm not even comparing like with like: on the data sheets linked above, bridge rectifier is rated to 35A and the TRIAC rated to 12ARMS. If I compared a bridge rectifier and TRIAC, of the same current rating, the bridge rectifier will look even worse.

The losses in the bridge rectifier will dominate the MOSFET losses and will be higher than the TRIAC losses, for the same load.

Also note, that it makes no difference, whether the AC is rectified and the element operated with unfiltered DC, or the element is in series with the bridge, with the MOSFET directly on the output: the losses due to the rectifier will be the same.

If you want low losses, then use back-to-back MOSFETs and a photovoltaic opto-coupler, similar to the one I linked to above.

If you want an easy life, run the element from unfiltered DC and the MCU from a regulated supply, consisting of a diode connected to the bridge rectifier, a  smoothing capacitor and the LM7805. Use a MOSFET to switch the element, but don't worry about using one with a really low on resistance, so long as it's good enough not to require a big heat sink.

[David Hess]

Thyristors and bipolar transistors are more economical than MOSFETs because they have smaller die sizes for a given power but the difference decreases at lower voltages.

Just for economy sake, I would use a TRIAC to switch AC to the heater on the low voltage side; this avoids rectification in the high current heater circuit and does not require 2 MOSFETs in series doubling cost to block AC.

The TRIAC can be driven directly rather than through an optocoupler.  Synchronization with zero crossings is still a good idea to minimize noise and this sort of thing used to be common in old temperature controlled soldering stations and power supplies with thyristor based preregulation which points to an examples to search for.