EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: Darkwing on April 18, 2018, 11:28:35 pm
-
Hi folks! :)
I'm thinking of a project where I want to use n-channel MOSFETs in back-to-back configuration to switch (or even PWM) big AC loads on mains. (What I mean with back-to-back: see here (https://forum.allaboutcircuits.com/blog/controlling-an-ac-load-with-a-mosfet.518/) in answer 1.)
The ballpark: I'm expecting loads up tp 13 A on 230 VAC (Europe), so 3.000 W we have.
Looking for a pretty low RDSon of < 35 m?, this should get me a power loss of not more than 6 W per MOSFET — so hopefully not a big thermal problem.
The calc:
RDSon = 0.035 ?
Udrop = R · I
Udrop = 0.035 ? · 13 A
Udrop = 0.455 V
P = Udrop · I
P = 0.455 V · 13 A
P = 5.9 W
And this is even for DC btw.
Now here's the tricky part that makes me scratch my head:
When I take a look around (https://www.infineon.com/cms/en/tools/solution-finder/product-finder/mosfet-finder/?filterValues=~(BREAKDOWN_VOLTAGE_FILTER~(~%27240~%27250~%27300~%27400~%27500~%27600~%27650~%27700~%27800~%27900)~MAX_DRAIN_CURRENT_FILTER~%2720~MAX_ROS_FILTER~%2730)&visibleColumnIds=name,opn,productStatusInfo,orderOnline,522,5546d4624f205c9a014f3c76a4f51d29,681_nom,c352cb0d3c2944d88b366705fdc7bf26,641,473_max,451_max,559_max,559_nom,547_nom,547_max,633,618,551) to select a "big guy" like Infineon's IPW65R019C7 (https://www.infineon.com/dgdl/Infineon-IPW65R019C7-DS-v02_01-en.pdf?fileId=db3a30433e78ea82013e791a9f140093) for example, having a rating of 650 V and 75 A continuous drain current, I then stumble across the "safe operation diagram" and begin to wonder:
(https://img1.picload.org/image/doglgcpr/eev-mosfet-safeoperation-450.png)
So does this really mean, that I could only draw 100 mA at 230 VAC @ 50 Hz (= 20 ms "pulse") using such a MOSFET? :o
Can anyone point me in the right direction to find appropriate MOSFETs for my needs (maybe even with logic level gate [I dare to ask!])? Or maybe am I completly on the wrong track with reading the above diagram? I can't believe this ... 100 mA!? :-[
Thanks @all! 8)
PS: Please don't sidetrack with Relais, SSRs, IGBTs or Triacs, I want to discuss MOSFETs here. Thanks! :)
-
You will not have 230V across the FET while 13A are flowing through it.
-
That safe operating zone means with 230v between source and drain, if you slightly turn on the mosfet in it's linear region, IE slightly turned on to a resistance of 2300 Ohms and the voltage between drain and source is 230v, that is as far as you can safely go. Now, I hope you are turning these mosfets almost completely on when switching. IE, on resistance between drain and source will be 0.035 Ohm, not 2300 Ohm, and when the mosfet is on at 0.035 Ohm, what will be the voltage between your drain and source. I don't think it will be 230v, right? That is unless you are drawing thousands of amps of current...
Now, when turning off the mosfet, the resistance between source and drain should be greater than 2300ohm. It should be in the hundreds of megaohms region unless you have a faulty mosfet. Now, with that resistance and 230v between source and drain, will the mosfet be conducting more than, or anywhere near 100ma?
I think you are pretty safe concerning that safe operating area unless you are making a linear amplifier where the mosfet will roast if you try to conduct a linear 100ma continuous with a 229v drop across drain and source.
-
The SOA is a matrix of simultaneous measurements. You are reading it as if the MOSFET is blocking 230V when off that it can only conduct 100mA when on. If you run linearly and have the FET dropping 230V at 100mA that's 23W. It can pulse that for 10ms.
In your scenario you will be running a Vds closer to 0.5V when the MOSFET is fully on. I'm reading that SOA at 0.5V as about 30A DC (implicitly limited by RdsON).
-
Mains is usually quiet, but it can be a dirty, violent place.
You may want to consider:
1.. Using much larger transistors (>= 1000V), with a MOV:
a. in parallel with the switch (ensures transients do not blow the FETs; transient is passed to the load), or
b. in parallel with the source and load (ensures transients are absorbed, brute force)
2. Using a TRIAC, which will automatically turn on during a transient (passing it to the load), but which drops 1-2V.
#2 is the more traditional approach, both being cheaper, easier and more reliable. The power loss is not usually a problem; a small heatsink will suffice. Modules are commercially available (any AC-only solid state relay).
Tim
-
Aha! Thanks so much! ^-^
I knew that something felt wrong, when reading this diagram — thanks for pointing it out! I even calculated my voltage drop, but I never considered this to be VDS. ^^ Of course it is at on time.
Sometimes these SOA diagrams have an additional labeling saying "limited by RDSon"; now it is clear what this means.
(https://img1.picload.org/image/dogldrri/eev-mosfet-soardson.png) (For illustration purposes only.)
So looking back to my original example with the IPW65R019C7 (https://www.infineon.com/dgdl/Infineon-IPW65R019C7-DS-v02_01-en.pdf?fileId=db3a30433e78ea82013e791a9f140093): it has a rise time of 27 ns, so a much quicker "pulse" than what is measured in the SOA diagram. It can easily handle the 325 Vp (the sine wave peak voltage of AC mains) when it turns on. Also my VDS of 0.455 V at fully on state is not even displayed in the diagram, so unproblematic is it. I consider this to fall in the area of "limited by RDSon", so again, no problem here.
(https://img3.picload.org/image/doglddgw/eev-mosfet-onstate.png)
I would say that it is safe to say that it is safe. ;)
Thanks for helping me out on this! :-+