Help with Mosfet

[BobZilla]

Hi Guys,
I am wanting to play around with cap dumping using mosfets and I have actually ordered these specific models.(Link to data sheet below)

LINK:
http://www.infineon.com/dgdl/IPx50R1...39d43facbe02fb


My basic idea is to use a smaller NPN to switch the gate on these guys which will control the cap dump through the drain/source. I will be putting the switching on the neg side to the charge batt.

I thought I understood what I was looking at but now I am a a bit unsure of the specs for driving the gate on these and I am hoping someone with more experience can look at the sheet and help me understand. At what voltage and current would I want to hit the gate with to have it fully open the D to S?

I initially really liked the looks of these mosfets because of the very low resistance across the DS. In addition they are rated for high voltage and 60+ amps so I think they should do well.

I am not experienced with these devices so if anyone could help me figure out the proper way to switch them I would appreciate the help.

Thanks,
BobZilla

[Joster]

Hey Bobzilla,

I am by no means an expert but I am working on the same type of cap dump.

You have here an n-channel MOSFET with a min/max Vgs of -20/20V. To the best of my knowledge 20V at the gate will turn it all the way on. I'm not too sure if 0V or the -20V turns it all the way off. I would think 0V would do it.

As far a current goes there is hardly any current drawn from a MOSFET. There is no physical connection between the gate and source/drain. All you need is the 20V potential across Vgs and you are good.

One thing to keep in mind though is that you need a resister (about 100K should be good) across Vgs. You can't just say, (hypothetically) hook a 20V battery across Vgs and dump the cap. I don't really know why it is needed but I couldn't get it to work without one.

What are you using for a timing circuit? I tried building one using 555 timers but ended up using an arduino.

Anyway....hope that is somewhat of a help!

-Joe

[BobZilla]

Thanks for the advise Joe.

I have chosen the arduino as well ;-)

I had never used one before but it seemed like a good place to start. So far I have just been writing small loops to control LED's while I wait for my mosfets in the mail. I plan to work with small voltages and practice switching a light bulb at first and then on to a cap dumper once I'm comfortable with it.

I have also been thinking of other possibilities with the arduino. I think SCR's could be used instead of mosfets if you code another channel for a pnp to break the flow and keep the SCR from latching. It's all about setting up a clever timing system and with the micro controller it should all be possible.

Perhaps you and I can exchange some ideas for using this platform once I get my feet wet.

[Joster]

Absolutley! What are you using for a storage Cap?

[BobZilla]

I have ordered a 60k uf 75 volt cap. This will be a bigger system than I have played with so far.

I have made various systems with bunches of flash caps in parallel and some other 1500uf that I got dirt cheap but now I am wanting to build bigger. I have not used this method for switching before but I have had a lot of success with my small systems with various switching methods so I'm confident I can move up to bigger and better now.


At mouser it is part #75-36DX603G075DJ2A


I have been playing with my solid state builds for the last year but I'm starting to shift back to the mechanical wheel now. SS is fun but I think more can be done with a mechanical setup.

I did a test circuit with the arduino tonight using a small npn tranny and just had it switching a light bulb off and on. I was powering it with a 9v battery and I tell you it did not last long (few hours). I did a current check from the battery and it was pulling 50-60ma just to run the board and switch the transistor. I'm talking about the switch circuit, not the bulb load which I was driving from a different source. I will be working on adding a generator coil to my mechanical setup and see if I can pull enough off that to run the switching circuit.

[Martin B]

Bob,
Have a look at US Pat 6677730.
https://www.google.com/patents/US6677730

I've tried the cap-discharge side of that circuit long ago using a 555 instead of the SG3524.
I was dumping a 22,000uf cap into a 12v LAB at about 24v.
I had 2*10 turn pots on the 555 and was able to adjust the dump time to almost nothing so the cap didn't discharge completely.

With a very short dump time, the battery was "ticking" just like when using the "spanker" (2A12) and all the wiring was jumping around.
The mosfets were getting hot only when the dump time was very short. I was using IRFP250Ns.
No proper scope so I can't tell you how long that dump was...

Martin B

[John_Bedini]

Just a tip of advise if your working on this cap dump charger. You need the resistor 10K to pull the gate off or the fet will stay on. The SG is ideal you just need to understand how to use it. It can be made to run at 10 to 100 cycles but remember that is divided by two. Pin 10 of the SG can be made to regulate the battery. The best data sheet is the TI somewhere I have the original RCA data sheet that tells you much more then what is now on the net. You should run this in inverted mode you just need to switch the fets around. Many things can be done with SG 3524 that most people do not understand, but then again I'm an Analog Engineer and have been so for 40 plus years. SCR will work and you need nothing to turn them off if you run inverted mode with them, you can not run them above 30 cycles or they will not turn off, think about what I said here.
John B

One more thing I want to bring up is that if your going to run the SG at high voltage use a resistor zener diode combination. also zeners can be used on the gates on the fets for over voltage.

sg2524.pdf

Hey Bobzilla,

I am by no means an expert but I am working on the same type of cap dump.

You have here an n-channel MOSFET with a min/max Vgs of -20/20V. To the best of my knowledge 20V at the gate will turn it all the way on. I'm not too sure if 0V or the -20V turns it all the way off. I would think 0V would do it.

As far a current goes there is hardly any current drawn from a MOSFET. There is no physical connection between the gate and source/drain. All you need is the 20V potential across Vgs and you are good.

One thing to keep in mind though is that you need a resister (about 100K should be good) across Vgs. You can't just say, (hypothetically) hook a 20V battery across Vgs and dump the cap. I don't really know why it is needed but I couldn't get it to work without one.

What are you using for a timing circuit? I tried building one using 555 timers but ended up using an arduino.

Anyway....hope that is somewhat of a help!

-Joe

[BobZilla]

Thank you Mr. Bedini. You have given me much to think about.

I am trying to use an "Arduino" for switch control. I'm not sure if you are familiar with these but they are programmable micro controllers for hobbyist. Through software you can set timing for different channels to flip on or off etc. very precisely. I am trying to emulate the principals you have laid out but I do not have the level of experience with components required to build it as you would.

I look at it as a way for someone like myself to cheat a bit. I know you have a profound understanding of circuits and you know the best way to do things.

Again I want to say thank you for everything you have shared with us all and it's really nice to see you back on the forums! ---Bob

[BobZilla]

*EDIT*
I wanted to make clear that I really appreciate your uploading the data sheet and that I am not dismissing in any way what you have said. This is very valuable information and I am going to try and understand.

[nte]

Thank you Mr. Bedini. You have given me much to think about.

I am trying to use an "Arduino" for switch control. I'm not sure if you are familiar with these but they are programmable micro controllers for hobbyist. Through software you can set timing for different channels to flip on or off etc. very precisely. I am trying to emulate the principals you have laid out but I do not have the level of experience with components required to build it as you would.

I look at it as a way for someone like myself to cheat a bit. I know you have a profound understanding of circuits and you know the best way to do things.

Again I want to say thank you for everything you have shared with us all and it's really nice to see you back on the forums! ---Bob

You may be able to use a Logic Level MOSFET with something like a 100ohm resistor from the arduino pin to the gate and a 10k resistor from the gate to the ground. The 10k resistor prevents the gate from floating like john said. If you don't use it your cap can dump while doing things like programming the arduino.

Another option is to connect a 10k resistor from vcc (probably needs to be 10+ V) to the gate and connect an optocoupler between the gate and the ground. With this arrangement the cap will dump when the arduino pin is LOW, although I may be wrong. See this DLE-TEST25 : Regenerative Loading with the J. Bedini's unipolar motor piloted by the APMC v1.0b

[BobZilla]

Hey thanks for the link NTE that is pretty cool.

I think I will make a thread to document this attempt at a dumper once I get under way with it. So far It is only an idea, waiting on parts. I do understand the need for the resistor at the gate to pull it down. I was able to switch a bulb with the arduino that was driven by a 12 volt source and used a very small NPN transistor. This is the approach I have thought I would take but until I start actually bench testing who knows. I am thinking if I can already switch a 12 volt bulb than I can replace the bulb with a high power fet and then dump the cap from it. Basically a small transistor to switch a larger one for the heavy load that the cap presents, well more than one but that is the idea in a nutshell. Today I am going to try and drive the arduino from a rectified generator coil, as I said before just the board and the little npn pulls 50-60 ma so I would love to drive that from the machine rather than a supply battery. Who knows I may just smoke the darn thing but no pain, no gain huh.

Depending on what I can pull from a generator coil I may be able to switch the mosfet with that too, if not perhaps I can pull that from the primary. I will just have to try different things to see what may be possible.

Thanks for the advice everyone. I will post some progress as this develops. ----Bob

[BobZilla]

For anyone who is also trying to understand MOSFET operation I have found a very nice paper dealing with parallel configurations and matching issues. This paper also helps me understand the general operation.

http://www.irf.com/technical-info/appnotes/para.pdf

---Bob

[nte]

Hey thanks for the link NTE that is pretty cool.

I think I will make a thread to document this attempt at a dumper once I get under way with it. So far It is only an idea, waiting on parts. I do understand the need for the resistor at the gate to pull it down. I was able to switch a bulb with the arduino that was driven by a 12 volt source and used a very small NPN transistor. This is the approach I have thought I would take but until I start actually bench testing who knows. I am thinking if I can already switch a 12 volt bulb than I can replace the bulb with a high power fet and then dump the cap from it. Basically a small transistor to switch a larger one for the heavy load that the cap presents, well more than one but that is the idea in a nutshell. Today I am going to try and drive the arduino from a rectified generator coil, as I said before just the board and the little npn pulls 50-60 ma so I would love to drive that from the machine rather than a supply battery. Who knows I may just smoke the darn thing but no pain, no gain huh.

Depending on what I can pull from a generator coil I may be able to switch the mosfet with that too, if not perhaps I can pull that from the primary. I will just have to try different things to see what may be possible.

Thanks for the advice everyone. I will post some progress as this develops. ----Bob

To better understand MOSFETs operation I recently did this experiment:
I used a MOSFET, a 10K pot, an amp meter and an optional load.

The pot connects to VCC and Ground, the wiper to the gate of the MOSFET.

The amp meter connects to VCC and the MOSFET Drain. (load optional)

MOSFET Source connects to ground.

Start adjusting the pot until the current starts to flow, Note the voltage at the gate when the current barely starts to flow. This will be around 3-4volts but varies with the MOSFET.

When the MOSFET is not fully open it acts like a variable resistor, expect the MOSFET to get burning hot in seconds if you have no load in the circuit. If VCC is 12v and your load drops 10v the mosfet will need to dissipate (2v x current flow) watts.


A couple years ago tried pulsing a SSG style coil with an arduino and a NPN, the arduino kept freezing up and would need to be reset. Adding an optocoupler solved this problem for me.

[BobZilla]

Great advice, thanks!

I have been reading up quite a bit on both mosfets and bjt's. I am quickly realizing that there is much more to understand than I thought previously.

I expect that I will be tinkering for quite awhile before I can really get to the business of making a circuit behave properly. I will probably burn up quite a bit of components in the process but it's all part of the learning process.

BTW:
I did try driving the arduino from a rectified generator coil and a regulator. It worked but at least in my setup it was not really practical. I had to increase my wheel load too a point to get the speed required and in doing so I think it would be better to just pull from the primary directly. It was neat to tinker with but not practical. Maybe if I dialed in the coil length, core size and wire size to produce exactly the right voltage and current it may have an advantage but for my initial testing with a coil I had laying around it was not gaining me anything to drive it this way.

[Joster]

Just a tip of advise if your working on this cap dump charger. You need the resistor 10K to pull the gate off or the fet will stay on. The SG is ideal you just need to understand how to use it. It can be made to run at 10 to 100 cycles but remember that is divided by two. Pin 10 of the SG can be made to regulate the battery. The best data sheet is the TI somewhere I have the original RCA data sheet that tells you much more then what is now on the net. You should run this in inverted mode you just need to switch the fets around. Many things can be done with SG 3524 that most people do not understand, but then again I'm an Analog Engineer and have been so for 40 plus years. SCR will work and you need nothing to turn them off if you run inverted mode with them, you can not run them above 30 cycles or they will not turn off, think about what I said here.
John B

One more thing I want to bring up is that if your going to run the SG at high voltage use a resistor zener diode combination. also zeners can be used on the gates on the fets for over voltage.

[ATTACH]2626[/ATTACH]

Hi John,

You are a huge inspiration for me and thank you so much for all the awesome info you have provided.

Regarding the Comparator MOSFET Cap dump, I am able to get the MOSFET to switch on and off ok when I test it to just turn a light on and off but as soon as I hook it up to my caps I can't seem to get the MOSFET to fully turn off even with 0V at the base.

Also, is it imperative to use photo flash caps? Right now I have 4 6800uF Photoflash caps that seem to be working well. Just wondering.