Maker Pro
Maker Pro

Using MOSFET for wireless Electricity transmission

Mongrel Shark

Jun 6, 2012
260
Joined
Jun 6, 2012
Messages
260
So I have stepped up from Joule thieves to Wireless transmission.

Circuit with NPN transistor: Bifilar pancake, one end to base (via resistor if running from much more than 2v) one end to collector. Pos to centre tap, emitter to neg. Thats it. Super simple. If you need to see a drawing, Just like a joule thief with no appliance/load. Do it with speaker wire 2n2222 and flat AAA and it will run down to .3v (or less, I keep disconnecting them, never run one into the ground) and give about an inch of range...

That was a bit boring, so I took 30 meters of cat 5 and a 3055..... 10 cheap Chinese 3055's later and i'm pretty sure the +200v on collector and -300v on base is killing them... Ok that was a bit obvious. Get a good protection diode, and keep amp draw below 500ma, and they last for a while though. I get about 50cm of range at around 3-5 watts. I'm pertty sure I can improve this. I need a switching device that can handle the volts though. Amps is not such an issue. Unless I can double the range, there will be no great need to draw more than half an amp. Higher freq would be good though. I currently have 11-25khz, depending on transistor variation (did I mention they are cheap knock offs?). If I can get a mhz or more I would be very happy.

Which leads me to MOSFET's, Not because I think its the best part, just because I have 2... I have this one, which we will call 4.5v due to gate threshold and this one Which tests with gate threshold of 3.15 On a single heatsink from a plasma I stripped.

I've done my best to google info. watched and read some tutorials etc... Looks like the main problem I will be facing is the negative pulse at the gate. Can I use a protection, reverse bias diode? Like with a regular NPN? Both transistors have gate-source BDV of +/- 30v So I figure the .200-300 I'm getting will kill them in short order. I'm also wondering about gate resistor. I know its not normal practice, but I can see pos up to 30-40v on the base before it goes negitive. with my 3055, this would be pushing the limit for the MOSFET's.. I can upload scope shots if anyone wants to see.

I was also wondering if I can use the two in parallel, and if it would be of benefit. While they both have drain tabs that would normally contact the heatsink, they have been stuck on with insulating thermal tape. So no common drain. They appear to have different roles on the circuit board too. which seems to be some kind of HV driver.


In testing them (with my atlas dca55 semi tester) I also noticed that if you just connect to drain and source, they both test as a diode, with the source being anode, drain cathode and frv of .52v. Is that normal, or are they damaged?
 

Harald Kapp

Moderator
Moderator
Nov 17, 2011
13,700
Joined
Nov 17, 2011
Messages
13,700
In testing them (with my atlas dca55 semi tester) I also noticed that if you just connect to drain and source, they both test as a diode, with the source being anode, drain cathode and frv of .52v. Is that normal, or are they damaged?
That is normal. It is the body diode.

Sorry, no answers to the other questions
 

Mongrel Shark

Jun 6, 2012
260
Joined
Jun 6, 2012
Messages
260
Thanks for reading and confirming they work Harald. I found some more on the same board. Just looking up data sheets now. Also blew my last 2 3055's. Pretty keen to keep experimenting (i haven't even tried octo-filar yet!). So I'm probably going to try soon and find out the smoky way...

If I use a small resistor on gate, and an apropiate protection diode, it might work. Worth risking a more common K2995 toshiba I think...

I have 3 more Plasmas in the pile of stuff I need to strip. So i guess there will be more MOSFET's...

I'd still love an answer if anyone can help though :)
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
if you're getting spikes at the gate, you can use a pair of back to back zener diodes and connect them between gate and source. The reverse breakdown of the zeners should be greater than the gate voltages you expect, and less than the rated maximum for the device.

If the spikes have significant energy, a gate resistor would be advisable.

Yes, you can place a reverse biased diode across drain and source. The body diode is generally very poor in performance and unless specified otherwise, can't be relied on to save the device.
 

Mongrel Shark

Jun 6, 2012
260
Joined
Jun 6, 2012
Messages
260
I don't know if I have any Zenners? What kind of devices would I find them in? (I could google that, but advice here is more valued. I'll google later. I have crt's, plasmas, dvds, vcrs, vacumes, fridges and an air con in the pile. Was council collection last week :)))

By back to back. Do you mean cathode to cathode or anode to anode (if there is a difference please point it out), so one has to break down for current to flow? If so I would assume I need one with reverse breakdown higher than gate threshold, but lower than gate-source breakdown, with the appropriate rating to handle the spike? ie 300v or more.

Yes, you can place a reverse biased diode across drain and source. The body diode is generally very poor in performance and unless specified otherwise, can't be relied on to save the device.
So if I have no zenner's this is ok? I can see the advantage to zenners, I could keep some of the neg spike... Protection diodes tend to flatten it. If they don't, the transistor tends to blow, indicating the diode wasn't doing its job.
I think I'm getting a similar thing with the 3055's. Yes 200v on the collector is pushing it, but it's on for 1us and off for 10-40us. I think the -300v on the base is whats really killing them. even with 1n4007 in place. On other similar systems, I have tried putting the protection diode from neg terminal to coil, bypassing resistor and other stuff on base, protects the transistor well, but flatens the neg spike, which I think is important in this application. As in reverse/negitive power JT's like the Dr Jones one. (thats a real brain melter. I'm determined to show whats going on there. going to need a lot of diodes and amp meters). While I am prepared to play around with flat neg spike, I'd love to be able to keep it... Turns 200v into 500v. ( I put one into each hand the other day, not recommended! Hit me up to the sholders, I think I wasnt far of zzzzzzThunk...)


while i'm rambling, If anyone has any thoughts on what taking -300v from a 12v SLA neg and putting it into the pos will do, I'm all ears. It's not really charging from itself. That would violate the laws of thermodynamics. I'm describing it as un-flattening atm, but that's not really accurate either... How do I explain this to people?

[edit] re significant energy. Its induction spike. It has no amps until it goes through an diode or spark gap like device.. Put 300v induction spike through a 0.7 diode, and you get 0.7v, The rest turns into amps, sort of. I don;t know how to explain it. Watts law applies though. Minus losses.
 
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Back to back means cathode to cathode OR anode to anode -- it makes no difference.

Can you show me a circuit diagram of what you're doing, it will make things a lot easier.
 

Mongrel Shark

Jun 6, 2012
260
Joined
Jun 6, 2012
Messages
260
attachment.php
For a better example of the coil, see here Or here

Ok there's my horrible drawing. I left resistors out, because they are not always required, And I don't know how to draw a water resistor like the one I am using. Depending on voltage and coil resistance and maybe induction/magnetic amplification. I seem to get more voltage drop on one side than explained by resistance, and a voltage increase on the other side, even from flat DC..

For 12v I am running around 1.5-2kohms on the base, and the biggest protection diode I can find from emitter to base. I'm probably going to move the cathode of the diode to the coil resistor junction, Seems to work well with Joule thieves. Which this is as well, if you put a spark gap or diode (again depends on voltages) from collector to emitter in the normal JT style. This will reduce Wifi rang though. I don't want a load on transmitter. I want as high a voltage differential as possible


So to change the NPN transistor to a N chanel MOSFET, I have tried moving the base to gate, collector to drain and emitter to source. It conducts an amp or more (from 6v with 2x k2995 and 1n4007 protection) BEFORE I connect the gate.... When I connect the gate, it draws a bit over twice that, without getting too hot right away. Looks promising for the current! Except the only way I can get it to oscillate, is to tap the neg lead on the source, or the pos lead on the center tap. It's the on-off inductive field that causes the wifi to work... Someone on youtube (I have a few videos including 12v cat5 and 6&12 speaker wire using water resistor in some, Water resistor is really cool to but that's a whole other thread..) Suggested I need to have a 100k-1mohm resistor from gate to ground. to make it start in the "open circuit" position. I'm about to try this...

Does that help?
 

Attachments

  • Wireless transmission bad drawing..png
    Wireless transmission bad drawing..png
    39.8 KB · Views: 1,067

CDRIVE

Hauling 10' pipe on a Trek Shift3
May 8, 2012
4,960
Joined
May 8, 2012
Messages
4,960
I don't see a power source... and you're right,... it's no Rembrandt.

Chris
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
A resistor is drawn as a squiggly line :)

I'm also not sure as to how this is supposed to generate an AC voltage in the coil. I presume it's DC that you supply it with?
 

Mongrel Shark

Jun 6, 2012
260
Joined
Jun 6, 2012
Messages
260
Joule Thieves and simple bifillar ossilators for dummies.doc

Yes. dc to the center tap. Just like a joule thief. It is a joule thief, if you add the resistor to the base. So with a transistor 2n2222a in place, and a flat AAA of say 1.2v it works like this: Positive charges the center of coil to the base enough for it to close c-e junction, as soon as that happens there is an easier path of flow, from c to e, as c should have a similar charge to b if the coils are wound evenly (*1) so the base goes off again, Causing the Flux to collapse. Causing a voltage increase. As the collector terminal is + at this point, the other end of the coil, at the base, must be negative. Depending on coil ratio (*2) this can be much greater than the the positive spike. Magnetic amplification plays a big part here. As does the combined Capacitance and induction at the same time.

Remember, the turns are opposing and side by side the whole way So you have oppositely charged conductors separated by a thin dielectric. Ie its a capacitor whenever in transition state. The only time its not a capacitor, is the time the base is conducting to the emitter before the c-e junction closes. once the c-e junction closes, the base turns negative, causing it to switch of, causing the Flux to collapse, into the capacitance of the coil(*3*4). If you have a spark gap, or more commonly today a semiconductor like a 3v led or a diode, the voltage will charge to this point, then discharge the capacitance causing flux to generate again. helping kick of the next cycle.




*1 Remember it is bifillar, with oposing directional windings, so if windings are not equal in number of turns magnetic amplification takes place, In this way it is posssible to start a silicon transistor from as low as 0.3v and germanium well under 0.1v, by harnesing the amplification of voltage, at a cost of amps)

*2 Again, depending on turn ratio of opposing winds you can amplify volts or amps as required. At a cost of the other. Watts law must apply in addition to resistance losses. If you have high volts at one end of the coil, you will have high amps at the other. in an opposite voltage. This gets really complicated, and i'm not aware of any formula to work it out.


*3 crerated by the highly oposing charges seperated by a thin insulator. According to Collumbs law the opposite charged particle's in this case electrons to attract to each other through the insulater, causing electron saturation fo the conductor, and posibly the surfaces of the insulator.

*4Sometimes there is a back-charge into the battery at this point, and this is how the reverse (Dr Jones) joule thief works. This often causes false power readings if your using a multimeter you don't see the back-spike, so it can appear as if you are drawing more amps than you really are. This is why there are so many false over-unity claims with this type of oscillating coil. Rating the efficiency of theses circuits can only be done with a joule count. Or possibly a rough measure buy metering power from appliance over time. ie how long is the led at x amount of brightness from a capacitor of known charge, Personally I will be doing some run cap to charge cap test soon. Results will be on my YouTube, which is a non profit open source project.


Here is a simulation I made for anyone that has Java up to date of a Joule thief running an LED, I have some scope graphs at the bottom, if you mouse over a scope, it will turn that part of the circuit blue. The transistor is represented in the middle in white, collector on top, base on bottom. left scope has collector coil connection on top and resistor - coil connection on bottom, the right graph shows the volts, amps, and power consumed by battery. as its a power supply, it's doing the opposite of consuming power, so it's a negative figure, I don't know how to reverse this, sorry. Its an online circuit simulator. so you can change anything you like with a right click. resistor value, Led battery scops etc. just right click the bit you want to change. It's fully interactive. If you remove the led it will show you what I see on my scope with the wifi system OR it will crash the aplet... Mine keeps crashing... Turning the sim speed way down my help. it is crunching some wicked maths...

Here is a screenshot for those that don't have Java
attachment.php


This is a screenshot I got just before it crashed on me...
attachment.php




Note the nearly 4kv on on the ends of the coil, shown on left scope graphs.... I didn't think this would work, but there it is, in the sim...



Now I have just done one of the best bifillar self blocking oscillator Joule thief explanations you will find, which took me a few hours BTW... Could someone please help me find a transistor and help me drive it.... I think I worked out why the MOSFET isn't working... My explanation solved my problem for me... The Mosfet dosn't conduct amps from gate (base) to source (emitter) does it? if not that explains everything. The back to back Zenners might get it going though. think I found some on the board the mosfets came from. been too busy typing all that ^^^ to check them out....


(phew)
 

Attachments

  • JT sim ss.JPG
    JT sim ss.JPG
    50.7 KB · Views: 622
  • SWES sim ss.JPG
    SWES sim ss.JPG
    47.5 KB · Views: 729
Last edited:

Mongrel Shark

Jun 6, 2012
260
Joined
Jun 6, 2012
Messages
260
P.s. My channel Can be found Here:http://www.youtube.com/user/MongrelShark?feature=mhee I hope its ok to link to it. As I said. everything on there is open source and nonprofit I don't even allow ads to show (let me know if you see any) It's a community education project, and I will get a website up when I find the time. If anyone has a (reasonable) problem with me linking here let me or one of the mods know and the link will be taken down.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
OK, so it's a blocking oscillator.

Yep, they're pretty hard on the transistor.

Here is a page showing some alternate forms of this oscillator. One has a zener diode between the base and emitter which protects against excess negative voltage

If your collector voltage is rising too high, you can connect a zener between the collector and the base to turn the transistor on if this happens. It should have a voltage just under the Vceo for the transistor.
 
Top