 Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me. |
Forum Index : Electronics : DC rating of MOSFET Safe operating area
| Author | Message | ||||
LadyN Guru  Joined: 26/01/2019 Location: United StatesPosts: 408 |
Am I thinking right about DC rating of MOSFET Safe operating area? Background: I want to build a high side switch using a N-Ch MOSFET that will be the core component of my SSR. These SSRs will be on for hours at a time - atleast as long as the PV arrays are producing power. They will connect 150v DC from PV arrays to loads like laptop, phone chargers, fans that use universal motors etc. Eventually I will also be connecting drills, trimmers etc to them once I learn about bypassing inductive kicks correctly but we can ignore that detail. I need to use the MOSFET as a switch because none of the switches and contacts/sockets on these loads are DC rated. They are expected to be used on AC and rely on the sinusoid to quench arcs during turnoff. So before I turn off/disconnect these devices from a socket that provides DC, I have to kill power to the socket. Evnetually I plan to drive these loads using "biphase" DC using a full hbridge to emulate zero crossings but one step at a time. Application: Basic reading has indicated that I need to look for MOSFETs that are DC rated - those that mention pulsed switching in their SOA are not really well suited for my application? Only two MOSFETs that I have on the way mention a DC rating in their Safe operating area: 1. SPB20N60S5 : i. Drain current chart indicates I can expect it to allow 12A at around 100 degree C ii. At Vds = 200V, Id = 10A So IF I were to use the SPB20N60S5, for my 150v DC usecase, I should be able to power 150 * 10 loads (ensuring I keep the SPB20N60S5 at 100 degree C or less) 2. HY4008 (seems like the company that made it got acquired and the new company does not make them anymore?) i. Drain current chart indicates I can expect it to allow 90A at around 100 degree C ii. At Vds = 100V, Id = 5A So IF I were to use two HY4008 back-to-back, for my 150v DC usecase, I should be able to power 150 * 5 loads (ensuring I keep the HY4008 at 100 degree C or less) This does assume that connecting two HY4008 back-to-back, each of which are individually rated for Vds_max = 80v, CAN indeed support 150V: it's likely I am wrong in this assumption and if so, please correct me! Honestly, since the HY4008 is so marginal for this specific usecase, I wont be using it except to experiment with it, expecting it to be destroyed and will likely end up using the SPB20N60S5 instead, but I primarily made this post to think aloud and see what mistakes I am making in my thinking. So please point out my mistakes and things I am not taking into consideration. The heatsinks will be made out of steel disc brake rotors from cars |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Good project, I have quickly drawn a proposed circuit for a load switch using back to back mosfets, so others here can better visualize what you are talking about. Individual mosfets have to be rated for the voltage you will be using, suggest 200v min rating if using 150 vdc. HY4008 will blow up here. They are not being switched off and on as per an inverter, so the driver circuit can be extremely simplified. Here I have used 2 voltage generating opto-couplers, each outputs approx 8 volts at very low current, so two in series will output approx 16 volts, the zener limits this to 12, or you could go to 15; would make no difference. The gates in the mosfets will charge up over a few tens of microseconds; they do switch off faster as the drivers have active turn off. So the load as seen by the mosfets is a pulse period in microseconds where the load transition occurs. Power dissipated is minimal as determined by the (RDson * current) per device, you will not require very big heat sinks. Essentially they are operating at DC, all mosfets have a DC load rating dominated by RDSon * current and size of internal die connections. If you limit the current load to approx 80% device rating, shouldn't have too many problems.  Cheers Mike |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
This is definitely the best way to build a high power isolated switch, for either ac or dc. There are a whole family of photovoltaic isolators around, the TPL191 is just one member that produces a fully isolated floating dc voltage. The switching speed will be relatively slow, and its highly unlikely any dangerous transient voltages will be produced. It might still be a good precaution to fit a transient voltage suppressor between PS1 and PS2 for higher voltage operation. Cheers, Tony. |
||||
Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
Could this switch be used for isolating a string of solar panels? Also what about substituting a IGBT to do the switching? These are usually rated for high voltages and I have a few, rescued from dead inverters. Klaus |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
No reason you couldn't use IGBT's to isolate your PV array, leave out the zener diode, normally you have to also place anti-parallel diodes across each device. MOV protection will be required from possible static lightning induced voltages. Mike |
||||
| hotwater Senior Member  Joined: 29/08/2017 Location: United StatesPosts: 120 |
IGBT are preferred for higher voltage/lower current circuits. Regardless of the size there will be about a 2V forward voltage. That takes a lot of heat sink. So, avoid if less than 300V. FETs at 100C. I won't even talk about that. That is the reason for multiple FET and low on resistance. |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
Mike, first of all THANK YOU for the schematic. where would one place the MOV(s) and what should their rating(s) be? I wanted to understand this a bit more - why would not the 80V Vds rated HY4008 put in "series" be able to withstand 150V switched across them? My other concern, which is why I made the original post, is, the SOA I quote is computed than empirically generated but that we will see |
||||
| Warpspeed Guru Joined: 09/08/2007 Location: AustraliaPosts: 4406 |
Just check the data sheet for the specific IGBT. A great many have an inverse internal diode and function somewhat like a mosfet in the reverse direction. If its out of an inverter it will very likely be one of those. Some IGBTs are without any internal diode, and a reverse polarity can kill the IGBT. Even stranger, some rather rare IGBTs even have an internal diode in series with the device. Mosfets conduct in either direction when turned on, IGBTs do not. Cheers, Tony. |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
Place MOVs across Drain, Source pins, no MOV is going to stop damage caused by a direct strike, its low power static and induced voltages that I would be more worried about, get some with a breakdown voltage that will protect the mosfet. 80 + 80 individual rating wont give you 150 with back to back mosfets, one or other of the internal diodes is always conducting. Mike |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
What if I already have a 180V 1.5kW TVS diode across DS already? Place a MOV in parallel to that as well? Ah! You're right, and since the internal diode is but an integral part of the mosfet structure, a higher than spec'd DS would/could cause breakdown of the mosfet structure, damaging it. I should have thought about it. I did not think deeply enough. I think I should take a break from MOSFETs and start with something simpler, like a TL431 to get my thinking sharper. If I made any mistakes in this post, please do highlight. That's the only way for me to improve! |
||||
| Tinker Guru Joined: 07/11/2007 Location: AustraliaPosts: 1904 |
That little circuit of Solar Mike's had me intrigued so I went looking for suitable Mosfets. Not a lot on offer in the 200V low RDS on range so, how would that one suit: IRFP 4427PBF, 25mOhm, 65A TO247, $2.41ea. I prefer the TO247 package BTW. Those TLP191 Opto's cost twice as much each. My source, LCSC parts Klaus |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
Klaus, thank you for taking interest in this! Appreciate you looking for more suitable MOSFETs as that would definitely help. 1. I think you meant to type IRFP4227 (irfp4227pbf). The IRFP4427PBF is a dual low side driver. 2. For this application we need a MOSFET that is rated for DC vs. pulsed load. The SOA for the IRFP4227 shows an Id = 1A @ 200 Vds w/ 100uS pulsed load. Compare this to the SPB20N60S5 : i. Drain current chart indicates I can expect it to allow 12A at around 100 degree C ii. At Vds = 200V, Id = 10A So IF I were to use the SPB20N60S5, for my 150v DC usecase, I should be able to power 150 * 10 loads (ensuring I keep the SPB20N60S5 at 100 degree C or less) |
||||
| Solar Mike Guru Joined: 08/02/2015 Location: New ZealandPosts: 1138 |
I use the IRFP4668 in the mppt controller 200v, 130a 9mR, these would be ideal as a load switch. LadyN: you are not interpreting the SOA curve correctly, when the device is turned on it has virtually no DC voltage drop across it, so is operating in the DC part of the curve where max current is determined by the silicon and its internal wiring. As for 100 degrees C, they will not be getting anywhere near that. Cheers Mike |
||||
| LadyN Guru Joined: 26/01/2019 Location: United StatesPosts: 408 |
Mike, Thank you for pointing this out! I DID NOT think of the near 0 DC voltage drop but I was designing for the worst case situation by considering the stability of the system when there's a short circuit. These switches will be outside in the sun and our ambient temperatures can reach 50 deg C in the summer. My plan is to have an overcurrent feature for this switch that will either use a latch to gate the drive off OR crowbar a fuse to turn itself off. I don't know which one I'll go for or whether I should have both. This overcurrent detector should act in a few uS. Until I figure this out, I will need to use a MOSFET that can withstand a short circuit current until a ESP32 connected to an ACS712 can gate the drive off, which is going to be my intermediate overcurrent detector. Of course, since the processing is done through a uC that runs a RTOS, the response time will be in the mS range. Not very fast. So the MOSFET better handle the heat during that. Perhaps I can put a TL431 + ACS712 to gate the drive off instead and it would be faster than the ESP32? These are options I am thinking about. Maybe overspec'ing it out of fear of the unknowns. The reason why I did not like the 4427PBF is because the SOA does not mention a DC rating at all. However, both the IRFP4668 and the IRFP4468 do (infact, with a quick look, they seem identical!). I would rather, atleast for now, use a part that will withstand the worst case situation, if only, to shield me through my ignorant mistakes :D Thank you for recommending the IRFP4668. |
||||
 Print this page |
