Parallel MOSFETs

Discussion in 'The Projects Forum' started by keyturbocars, Nov 22, 2010.

  1. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    Moderator, Please delete this message. Sorry that I had a double post, but my browser crashed when I tried to submit the first time, and I ended up making 2 posts. Don't see how I can delete this duplicate post. Thanks.

    Hello, This is my first post here. I was actually searching on Google about parallel MOSFETs and some posts on this site came up. I am not very strong in my electronics understanding, so I thought I would see what some electronics experts think.

    I have a wind turbine (10 ft diameter blade rotor) that is rated for 2kW and I've seen peak output up to 3.5kW on a couple occasions. The wind turbine charges a 48V battery bank. Peak amperages I've seen are around 65A from the wind turbine during very strong wind gusts. The controller has a rectifier to convert the 3 phase wild AC turbine output to DC and then a control board that dumps wind turbine power to big heating resistors when the battery is fully charged. In addition, when the wind turbine RPM is too high, the control dumps the turbine output to the resistors to slow down the turbine. Heating resistors are 4kW total. The controller I am testing uses MOSFETs to dump the power to the heating resistors. I'll attach some pictures.

    There are 3 MOSFETs attached to a heat sink. They are made by IXYS and their part number is: IXTQ82N25P . Here's the spec sheet on it.

    http://ixdev.ixys.com/DataSheet/99121.pdf

    It lists a maximum amperage of 82A.

    I've got 3 questions for you electronics experts...

    1) Does the 3 of these MOSFETs mean that they can control a maximum of 3 x 82A?

    Just curious what the power capacity of these 3 MOSFETs really is in the real world.

    2) Could 2 more identical MOSFETs be solderd to the board and increase the current switching capacity further, or would the rest of the circuit board not be able to drive the 2 additional MOSFETs?

    That might be a dumb question, but I am not familiar enough with electronics to know. The 3 MOSFETs might be more than adequate as is as long as I keep them cool enough.

    The 3 MOSFETs that are on there now appear to be screwed to the heat sink dry. No thermal paste. I think that was intentional, because the heat sink is actually used as a conductor for the dump circuit. The heat sink screws to the board over a large tinned copper trace. I've tested continuity and the heat sink is used as a conductor. I think that's why they didn't put thermal paste on there. I'm concerned that the MOSFETs will get too hot this way, but I don't want to add thermal paste and cause other problems. I thought if I added 2 more identical MOSFETs, I might increase the reliability of the system. There are spots on the board and the heat sink is drilled/tapped for 2 additional MOSFETs. Just don't know if the driver circuit could handle 2 more. This same circuit board is used for a larger wind turbine where they apparently use 5 MOSFETs. I don't know if the only difference between my board and the one in other controller is 2 less MOSFETs.

    Not sure if I should even try to add 2 more MOSFETs or just try to keep the existing 3 MOSFETs as cool as possible with some better fans.

    I'll attach some pictures.

    Thanks for any advice,

    Edward
     
  2. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    International Rectifier's App Note on Parallel MOSFETs

    It can be done with matched MOSFETs, depending on the application and supporting circuitry.

    The rating is more than just the Max current of the MOSFETs (2kW is about 24V at 82 Amps). The power handling of 65A should be within the design of the circuit and heatsink.

    To get better efficiency out of it, I'd suggest increasing storage capacity so the amount of power dumped as heat through the resistors is minimized.
     
  3. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,677
    900
    Real world DC capacity is often less than the advertised capacity (I_{D25}). That value is usually determined by the junction. When case and leads are considered, it can be considerably lower. In this example, the lead limitation reduces it to 71A, see p.2 I_{D(RMS)}. Also, if you look at figure 12, you will find a graph showing the safe operating area for various V_{DS} and DC current.

    You can often parallel additional mosfets to increase current capacity, but you need to consider the whole design. Is the problem that the current exceeds the mosfet's capacity or is the problem heat conduction. More mosfets will not reduce the amount of heat produced, unless they are each operating at a lower RDS(on) as a result. Adding more mosfets to the same heat sink (for example) won't change the heat sink's capacity. You would also need to consider the gate drive capabilities. Can it handle 5 mosfets? Finally, the mosfets may have gate drive resistors. You would need to include the same value resistor (as a first approximation) for each additional mosfet you might add.

    It is common to use the mosfet tab as a high-current conductor. In some designs, it is even soldered to the heat sink/board to get better heat conduction. If it does not have paste and is apparently a conductor, do not add paste.

    You answered your own question. Sorry to be duplicative.

    John
     
  4. John P

    AAC Fanatic!

    Oct 14, 2008
    1,632
    224
    I don't think that's true. Adding more identical MOSFETS would reduce the array's resistance, to a total of (0.035/n) ohms. That means that the lost power would go down in inverse proportion (I^2R loss). It wouldn't change the heat sink's capacity exactly, but for a given current it would indeed be cooler. And for a d.c. situation, I can't see how the driver could fail to control 5 transistors instead of 3, though if there's PWM control, there'd be a reason to check what the driver can do.

    My concern would be with the generator and mechanical parts. Can they run at a higher power level than they're designed for? I thought wind power systems were designed to protect themselves by turning away from the wind if it gets too strong.
     
  5. wayneh

    Expert

    Sep 9, 2010
    12,101
    3,033
    Ditto. The only question in my mind is whether the 2 extra MOSFETs can be driven properly. Time spent at less than full ON will cause rapid heating at those currents, so you want a nice clean fast switching to "on". They're like capacitors, so driving them at high frequency would especially be a question. If they're switching at turbine (ie. low) frequencies or lower, there's not much concern. If they're set up like a furnace with simple on/off control over seconds or minutes, no problem. But it would make sense to put them under PWM control, and that's where a concern arises.

    The problem is in-rush current, which the driver must supply briefly to charge the gates. Gate resistors are used to limit in-rush, to protect the MOSFET but this also protects the driver. Increasing the total in-rush by 5/3 shouldn't be the straw that breaks the camel's back, but it is a question mark. You could add one and see how it goes, or invest more time to research the drive circuit. Closely comparing the higher capacity circuit to your plan might alleviate all your worries, if it is indeed exactly the same except for the MOSFET count. That seems likely.
     
  6. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,677
    900
    Sorry, I wasn't fully awake. Ohm had a rough night.

    John
     
  7. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    Thanks for all the replies. Maybe I should just leave the 3 MOSFETs and concentrate on making sure they stay cool. Not sure how I can help them stay cooler if there is no thermal grease between the MOSFETs and the heat sink. Here are some pictures of my wind turbine system:

    http://www.facebook.com/album.php?aid=31445&id=100000264867099&l=6c791ef013

    The pictures show a different controller that is relay based instead of MOSFET based like I'm testing right now. I thought the MOSFET based controller might be more reliable, but I'm still testing and have yet to decide which controller to use.

    Thanks again for the feedback.
     
  8. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,677
    900
    Great pictures. They remind me of western Nebraska (Scottsbluff). Good ol' farm ingenuity. John
     
  9. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    Thanks John. We live in a canyon in Eastern Washington.

    Is there a way I could go back to my very first post of this thread and delete my first paragraph that says "Moderator please delete this post....". Or can a moderator do that for me?

    On this cirucit board, this same board was used for a 2kW wind turbine (3 MOSFETs) and a 3kW wind turbine (5 MOSFETs). In my controller, they just left out the 2 extra MOSFETs, but there are places for them on the board and on the heat sink. I'm thinking that it's very possible that the gate driver circuit is set up for the 5 MOSFETs even on my board.

    Would there be any harm for a circuit board to be designed for 5 MOSFETs but then was only run with 3 MOSFETs?

    In other words, I'm wondering if there is any harm in having "too powerful" of a gate driver?

    If not, then I suspect they just kept the rest of the board the same and just left out the 2 MOSFETs for money savings on the 2kW controller. This controller was manufactured in China and it's extremely difficult to get any detailed information on the circuit, so I can't just ask the company for a schematic or inquire about the 2kW and 3kW board differences.
     
  10. bertus

    Administrator

    Apr 5, 2008
    15,646
    2,345
    Hello,

    You can edit the first post yourself.
    There is a edit button on the bottom right in each of your own posts.
    This forum has no timeout on editing, like others sometimes do.

    Bertus
     
  11. wayneh

    Expert

    Sep 9, 2010
    12,101
    3,033
    I agree. Just hard to prove. Not so hard to validate, though, by testing.
    Nope. The drive circuit supplies primarily a voltage, with almost no current except when charging the gates. If it can handle enough charge for 5 gates, 3 would be no issue.
    You could only over-voltage a MOSFET gate, but the driver circuit won't do that. The MOSFET will just take whatever current it can (thru a gate resistor) until it's charged, just like a cap.
    This was likely driven more by wanting to differentiate the two products for two-part pricing (standard marketing tactic). The COST difference is trivial; the MOSFETs would have maybe cost $1 to the manufacturer.
     
  12. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    Hello Bertus,

    For some reason, that Edit button disappeared. I just went back to try again after logging in, and it's not there anymore. It was there when I first made the post, but disappeared later.

    Thanks,

    Edward

     
  13. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    WayneH,

    Appreciate the comments. You mentioned that it would be possible to determine if the gate driver is sufficient for 5 MOSFETs by testing. Here is where my electronics weakness comes to the surface...

    What would be the best way to test this?

    I only have a couple digital multimeters around here, and not much experience to know how I could test the existing circuit to see if the driver can handle adding 2 more MOSFETs. I really like the idea of adding 2 more strictly from the standpoint of reliability and safety. If a MOSFET ever failed, then I'd have a bigger factor of safety having redundant MOSFETs in parallel. Also, the fact that the MOSFETs don't have any thermal grease concerns me, and by adding 2 more MOSFETs it seems to me that I'd have more surface area touching the heat sink and hopefully better overall heat transfer to the heat sink.

    I used this controller before on another wind turbine for several months, and I became very familiar with how it works. When wind speed increase to around 30mph, then the controller begins to try to slow the wind turbine blades by loading it down with the 4kW heating elements. This is where the MOSFETs dump the wind turbine power to the heating elements. This is the electromagnetic braking that is supposed to keep turbine RPM safe. It starts off by pulsing the heating elements. It's not a high frequency pulse. It's around 1 pulse per second. The reason I know is because I can hear what's happening. There are 2 fans that blow over the heating resistors and whenever power is dumped to the heating resistors, then the fans come on. If the wind blows harder, then the controller will cause the dump to go continuous. The 2 fans then run continuously. Finally, if the battery voltage gets to a set point (fully charged), then the controller will dump all the power to the heating elements again on a continuous basis to protect the batteries. This doesn't happen very often, because I have a different diversion controller setup for heating hot water for the house. I divert power off the battery bank into the hot water heater. I have it set up so the hot water heating keeps the battery voltage under control and prevents the main turbine controller from just heating air with the 4kW resistors. Heating the air is not helpful in the summer time. Hot water is something my large family can use all year round.

    So, to clarify:

    1) If wind is slow, then power trickles into battery bank.
    2) If wind is moderate, then MOSFETs pulsed (~ once per sec) to load down & slow turbine with 4kW heater.
    3) If wind is strong, then MOSFETs dump continuously to try to slow wind turbine.
    4) If batteries fully charged, then MOSFETs dump all power to heating elements to prevent batteries from overcharging.
     
  14. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    I know that my last post was long winded and my question probably got lost in all my wordiness.

    WayneH mentioned that it would not be hard to validate whether my existing gate driver circuit was sufficient to drive 5 MOSFETs instead of 3.

    What would be the best way for me (not an electronics guru) to test my existing circuit to find out if it's safe to drive 2 more MOSFETs (total of 5)?

    Thanks!
     
  15. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    Can you get a closeup photo of both sides of the board?

    The traces running to Q5 and Q7 (the missing MOSFETS) may have more missing components in the path.
     
  16. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    Here are some close up pictures of the circuit board on the top and bottom. Two pictures of the top side on either side of the MOSFET heat sink, and one picture of the bottom.

    I only see 2 empty spots (a resistor and capacitor) that are right under the long IC chip. I can see that all 5 traces going to the gate pin on the MOSFETs have a resistor in place. After seeing this, my guesstimation is that it'd be a plug and play process to install 2 more MOSFETs, but I'd appreciate more experienced feedback.

    Also, I was looking at the underside of the board at the traces for the dump on the MOSFETs, and I wonder if the dump pins on the 5 MOSFETs could handle all the current, and if I could use thermal paste on the MOSFETs where they attach to the heat sink.

    Thanks.
     
  17. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,677
    900
    Are R26, R28, R30, R31, and R34 the same value? If so, the board manufacturer has done you a favor by including appropriate gate resistors for 5 mosfets. I would just go for it and add the two additional mosfets. Of course, that advice comes with a money back warranty. :D

    John
     
  18. thatoneguy

    AAC Fanatic!

    Feb 19, 2009
    6,357
    718
    It looks good to go. Order identical MOSFETs, or better yet, 5 new ones so they are from the same lot, and replace all in one batch. Then you have a few spares.

    Adding thermal grease shouldn't be an issue, just make sure there isn't already a thermal film on the back of the MOSFET that is providing electrical insulation from the heatsink before adding grease. The insulating film (if present) is usually a good heat conductor anyway.

    Lastly: Do this at your own risk! The dump resistors may not be rated for the full current, that could be the only other difference between this and the next higher capacity model.
     
  19. keyturbocars

    Thread Starter New Member

    Nov 21, 2010
    15
    0
    Yep, all 5 resistors have the same number printed on them. I think I'll go order the MOSFETs. Found them at Mouser.

    Thanks for the advice guys!
     
  20. jpanhalt

    AAC Fanatic!

    Jan 18, 2008
    5,677
    900
    Be sure to send a picture.

    How much more expensive is it to incorporate variable pitch/constant speed to the blades?

    John
     
Loading...