[TCML] Driving a flyback off of mains gate drive woes.

Carl Noggle cn at q.com
Sun Oct 9 08:32:09 MDT 2011


Hi, Scott,

I forgot to mention that the capacitor from the upper input lead (to the 
center of the diodes) should be AC rated for high current, such as a 
motor-start or a motor-run cap.  These are available from Grainger, 
etc.  Polypropylene caps may work, too--you will have to check the ratings.

---Carl





> Hi Carl,
>      Interesting, so I don't need a transformer, that certainly makes 
> things easier since I already need 12V for the frequency generator.  
> There is one thing though, If I use a rectification configuration like 
> is used for the PVM400 driver, to get 240V out of 120 using two filter 
> caps, tie center to ground, that floats the ground of the MOSFET 
> source pin below ground by 120V, so will a driver on a separate power 
> supply grounded to a true ground still work?  I really don't know how 
> they work or hook up, does it reference the MOSFET ground somehow so 
> regardless of its voltage it is OK, or will it burn out using that 
> configuration?  Thanks.
>
> Scott Bogard
>
> On 10/8/2011 11:44 AM, Carl Noggle wrote:
>> Hi---
>>
>> Since you are using an N-channel MOSFET, the source is grounded and 
>> most drivers will work fine.  No need for a xfmr--that will slow the 
>> transition times and increase the dissipation in your device.  Just 
>> make sure the driver will handle your gate capacitance and will 
>> output enough voltage to turn the FET on completely.  Most drivers 
>> will do that.  You will need a separate low voltage supply for the 
>> driver.
>>
>> The advantage of a half- or full-bridge is that the FETs are 
>> self-protecting against voltage spikes by their body diodes.
>>
>> ---Carl
>>
>>
>>
>>
>>> Hi Mike,
>>>     Hence my issue, for the moment I want a true flyback driver, not 
>>> a half bridge.  My issue is I can't find a MOSFET driver that will 
>>> work Vcc on that high of a voltage, and from what I know of MOSFETs 
>>> to run a driver off of a different voltage source, biases them on a 
>>> different voltage, and they are stuck on or off no matter the 5 volt 
>>> swing.  From what I can see (did some perusing at work) several 
>>> solid state Tesla coils use a gate transformer, so it is biased to 
>>> whatever the MOSFET source voltage is, but I don't know where to get 
>>> one of these, or how to make one, or if it will work on PWM over a 
>>> range of duty cycles...  I would imagine it would be easy, and I 
>>> actually have an oscilloscope now to make sure it is working 
>>> (finally) but I'm hesitant to just throw stuff together, I don't 
>>> want to blow out the ARduino I'm using as a frequency generator, as 
>>> it would take time and money to replace (project is kind of on a 
>>> personal deadline, I'd rather get the correct part numbers.)  If it 
>>> makes a difference I am settled for the moment on IRFP250 MOSFETS, 
>>> but plan on upgrading to IRF450s in the future.  So I guess my 
>>> question is how to make a gate drive transformer, and what MOSFET 
>>> driver would be suitable for driving said transformer (assuming it 
>>> is done that way) and how to wire up (external components) a MOSFET 
>>> driver, since I've never done that before.  Thanks again guys.
>>>
>>> Scott Bogard.
>>>
>>> On 10/7/2011 5:37 PM, Michael Twieg wrote:
>>>> That schematic isn't even for a flyback supply.  It's a half bridge 
>>>> driver,
>>>> driven by a hi/low side bootstrapping driver.  For a simple flyback 
>>>> you will
>>>> need some kind of simple low side gate drive IC.  The IXDD414 is 
>>>> one option,
>>>> but it's overkill (also it's obsolete and very hard to find, unless 
>>>> you go
>>>> for the newer versions).  The requirements of the gate driver will 
>>>> depend on
>>>> the FET you're driving and the rise/fall times you want.
>>>>
>>>> -Mike
>>>>
>>>> On Thu, Oct 6, 2011 at 10:26 PM, Scott Bogard<sdbogard at gmail.com>  
>>>> wrote:
>>>>
>>>>> Hi Dave,
>>>>>      Yeah I saw that schematic, it kind of gave me the idea, 
>>>>> switchable from
>>>>> 120 to 240 volt operation, super high power capacity, but there are
>>>>> limitations with that chip I'd like to circumvent.  Namely that it 
>>>>> is stuck
>>>>> on 50% duty cycle (overall flyback duty cycle is independently 
>>>>> controlled by
>>>>> a second oscillator at a low frequency) and that you have to use 
>>>>> the built
>>>>> in frequency generator (I have a frequency generator, that I can 
>>>>> program
>>>>> frequency routines into.)  Basically I need some kind of plane 
>>>>> Jane MOSFET
>>>>> driver that I can mount to the high voltage rails like that chip, 
>>>>> but that
>>>>> accepts an external input 5V logic, I don't think such a thing 
>>>>> exists, so
>>>>> there must be some way to build one...  Can I make a BJT totem (I 
>>>>> think that
>>>>> is what it is called) fed by like a 3k resistor or will that pop 
>>>>> under that
>>>>> voltage?  Thanks in advance for your help guys, it is so much 
>>>>> appreciated by
>>>>> this solid state newbie!
>>>>>
>>>>> Scott Bogard.
>>>>>
>>>>>
>>>>> On 10/6/2011 8:20 PM, David Speck wrote:
>>>>>
>>>>>> Scott,
>>>>>>
>>>>>> Below is a link to a full schematic for a flyback driver off the
>>>>>> Information Unlimited website.  No personal experience to know 
>>>>>> how well it
>>>>>> works, but it seems to be a close match to the sort of thing you 
>>>>>> are looking
>>>>>> for.
>>>>>>
>>>>>> http://www.amazing1.com/**download/PVM500BASICSCHEMATIC.**pdf<http://www.amazing1.com/download/PVM500BASICSCHEMATIC.pdf> 
>>>>>>
>>>>>>
>>>>>> Dave
>>>>>>
>>>>>> On 10/6/2011 6:40 PM, Scott Bogard wrote:
>>>>>>
>>>>>>> Greetings all,
>>>>>>>      Back up to my old tricks again.  After procuring a 
>>>>>>> schematic for a
>>>>>>> PVM 400 flyback driver, and noticing it runs off of mains 
>>>>>>> voltages, I am
>>>>>>> attempting to create my own version, with programmable 
>>>>>>> adjustable duty cycle
>>>>>>> and frequency.  What I don't seem to know how to do, is drive 
>>>>>>> the gates,
>>>>>>> what I have is an ATmega chip producing a nice 5 volt square 
>>>>>>> wave, of
>>>>>>> variable duty cycle and frequency (3-300 kHz) and some good 
>>>>>>> flybacks.
>>>>>>>   Obviously I cannot just plug that 5 volt signal into a bunch 
>>>>>>> of MOSFET
>>>>>>> gates, it would likely not have the guts to hold them open, even 
>>>>>>> if it was
>>>>>>> the correct ground, which it is not.  So basically what hardware 
>>>>>>> is needed,
>>>>>>> a gate drive transformer?  MOSFET driver?  Could somebody please 
>>>>>>> point me to
>>>>>>> some straightforward literature about biasing MOSFETS, and 
>>>>>>> possibly provide
>>>>>>> part numbers, not just reference "generic BJT."  I know it must 
>>>>>>> be possible,
>>>>>>> people have run solid state TCs off of MOSFETs for a long time, 
>>>>>>> so what am I
>>>>>>> missing...  Thanks in advance, I'm looking specifically to run 
>>>>>>> it off of
>>>>>>> about 320V DC (+-160V) low side drive (not half or full bridge) 
>>>>>>> if that
>>>>>>> makes a difference.
>>>>>>>
>>>>>>> Scott Bogard.
>>>>>>>
>>>>>> ______________________________**_________________
>>>>>> Tesla mailing list
>>>>>> Tesla at pupman.com
>>>>>> http://www.pupman.com/mailman/**listinfo/tesla<http://www.pupman.com/mailman/listinfo/tesla> 
>>>>>>
>>>>>>
>>>>>>   ______________________________**_________________
>>>>> Tesla mailing list
>>>>> Tesla at pupman.com
>>>>> http://www.pupman.com/mailman/**listinfo/tesla<http://www.pupman.com/mailman/listinfo/tesla> 
>>>>>
>>>>>
>>>> _______________________________________________
>>>> Tesla mailing list
>>>> Tesla at pupman.com
>>>> http://www.pupman.com/mailman/listinfo/tesla
>>>>
>>> _______________________________________________
>>> Tesla mailing list
>>> Tesla at pupman.com
>>> http://www.pupman.com/mailman/listinfo/tesla
>>
>> _______________________________________________
>> Tesla mailing list
>> Tesla at pupman.com
>> http://www.pupman.com/mailman/listinfo/tesla
>>
> _______________________________________________
> Tesla mailing list
> Tesla at pupman.com
> http://www.pupman.com/mailman/listinfo/tesla



More information about the Tesla mailing list