Free Energy | searching for free energy and discussing free energy
Solid States Devices => Resonance Circuits and Systems => Topic started by: madddann on May 16, 2018, 10:22:58 PM
-
Hello!
I have a circuit that works in simulation, but not on the bench - at least not as expected.
My goal Is to make it work right, hopefully with your help.
This is the circuit simulated (with the desired voltage and current waveforms): https://postimg.cc/image/mpvl3w7md/
On my bench, the signal generator is an arduino nano completely isolated from the main circuit (powered by it's own battery) and is set to 25% ON duty cycle, the optocoupler is an LTV817 driving the transistor E13007-2. The coil is the size of a small contactor coil (36 Ohms) on i'ts laminated core about 80mm x 12mm x 9mm. The capacitor is 220 nF poly. For powering the transistor + tank circuit i use a 12V automotive starter battery.
This circuit (on my bench) generates the waveforms that look more like this: https://postimg.cc/image/lau0fawad/
After the initial 25% ON time pulse, the voltage drops to negative (the cap fills up with the backspike from the coil) but when the voltage climbs back up to about +21V, the waveform is cut flat for some time and then small oscillations continue until the ON pulse again.
Where the waveform is cut, it looks like something is coducting the current - maybe the transistor is turning on again for a brief time, but how is this possible what is causung this? ... negative resistance of the transistor?
Thx to anyone willing to help!
Dann
-
Hi Dann,
Have you checked the switching waveform with the scope whether it resembles the waveform coming from the signal generator?
I mean how well the opto coupler transfers the rectangular pulse to the base-emitter for correct switching.
It would help to see the bench scope shot on the base and the collector too, versus the common negative ground (if you have a 2 channel scope). Maybe you can make a photo from the display (when upload, maintain pixel size not higher than 900 horizontally, otherwise we have scroll back and forth to read the text lines) 8)
I see you use a 1 kOhm resistor between the base and emitter in your 2nd schema: does it make a difference? (it may help in a quicker switch-off of the transistor) Maybe you could change its value between 470 to 1000 Ohm?
Gyula
-
HI Gyula!
Thanks for posting and reminding me of what I should have done in the first place.
At first I was using a N-channel mosfet with the second scheme from previous post (with the resistor), but I thought that the mosfet's body diode was causing the cut-off wave (the simulations that I did were indicating that), this is why i switched to NPN transistor E13007-2. Now I'm not sure anymore what is causing what...
-----------------------------------------------------------------------------------------------------------------------------------------------------------------
The resistor in the circuit does make a difference, but I think the optocoupler may not be capable to deliver enough current to switch the npn transistor as it should be switched. Perhaps an alternate way of driving the transistor should be implemented - am I on the right track here?
Here are the scopeshots that I did today: without resistor, with 1k resistor and with 470 Ohm resistor, all with voltage + coil current at the LC tank and collector emitter (VCE) + base emitter (VBE) voltages at the npn transistor (E13007-2).
See scopeshots in this gallery - pictures from 1 to 7: https://postimg.cc/gallery/1mu9z7pi/ (https://postimg.cc/gallery/1mu9z7pi/)
Read the text below each picture when you open it, to see what is what.
Picture 1.2 is just a detailed view of picture 1.1 (the top and bottom of the waveforms in 1.2 are cut off)
Anyone that knows what is actually happening here please post your thoughts.
The goal is to keep the oscillations at max, so anyone that may know how to do that is welcome to present their circuit here.
THX!
Dann
-
Hi Dann,
Thanks for the scope shots. Yes, the 1 kOhm resistor is a bit better instead of a 470 Ohm between the base-emitter from the base-emitter driving signal point of view but maybe the overal driving is still not ok.
If you agree, please do the following: connect the collector of the opto output directly to the base of the transistor where the upper 1 kOhm resistor is also connected, this latter is coming from the +12V directly as before. And connect the emitter output of the opto to the emitter of the switching transistor i.e. to the negative rail. You may wish to check operation with or without another 1 kOhm connected also across the base-emitter.
This way the opto output will quasi short circuit the base-emitter of the transistor whenever the opto input receives an input pulse from the Arduino. In this new opto drive connection the opto will invert the input pulse as shown in the bottom right Figure of this data sheet, page 5: http://www.mouser.com/ds/2/239/LTV-817-827-847-647.pdf but I do not think this control signal phase inversion matters at all in this setup. The RL resistor shown there will be your upper 1 kOhm, ok?
By connecting the opto like that, the control pulse between the base-emitter of the transistor would resemble the most to the signal coming from the Arduino if that is a decent rectangular pulse. 8)
If the control pulse is okay, then we can continue with the collector side.
Gyula
-
Can you post a picture of your physical circuit ?
-
Hi Gyula, thanks for the suggestions, will try that today or tomorrow and report back.
Hi Forest, my circuit is just the arduino on a protoboard with the opto (both powered from 4x1.5V + small dc-dc converter to get 5V). The trasistor is also on the protoboard. The tank circuit, the current sensing resistor (0.1ohm) and the 4.7 ohm resistor are connected with clipleads to the starter battery and transistor.
this is all experimental setup right now, just to see how things behave.
Maybe you got confused by the Figuera schematics in my gallery - those were old experiments from the Figuera thread that didn't show much results, but I still learned some things from them.
I removed those pics now, so no more confusion.
Thx for your interest.
Dann
-
Hi Gyula and everyone else.
Sorry for the very long delay, but life just happens the way it does.
I tested the suggested circuit, but was unable to make the transistor to switch solidly, at best it was switching just faintly when tested with an automotive 10w bulb (the bulb had an lazy effect to turn on and off). Switching the tank circuit did not yeld any results at all.
The waveform on transistor base-emitter was a square wave, but the ON state was just about 30mV difference from the OFF state. I tried some resistors from 1.8k to 470 Ohm, I did not try a lower value because I think the opto can't handle the current.
Another resistor on the base-emitter did not help at all.
This transistor that I'm using (E13007-2) needs from 0.4A to 2A base current (from PDF), I think the opto is unable to drive it directly.
I also did a simple reverse polarity (reverse current) test on the transistor's collector-emitter with a 12v battery and 10w automotive bulb.
I connected + to emitter and - through the bulb to the collector, and guess what, the bulb lit almost fully up (the base was not connected to anything).
I think that this is exactly what happens here in this scopeshot - where the oscillating wave is cut off: https://postimg.cc/image/gmexmps2t/
Now I have the following question - what kind of semiconductor component would be able to do the switching in this circuit without cutting off the oscillation (when current reverses)?
Do IGBTs without a body diode exsist?
Anyone with an idea or possible solution is welcome to post it here.
Now I have some time (till Sunday) to test things out with what I have at hand.
Thanks Gyula and all!
Dann
-
Hi Dann,
The transistor is very likely defective, it should not conduct in the reverse voltage test. Also, the 30mV difference between the on and off states indicates that.
The reason I suggested to connect the opto output across the base and emitter was that in that case the base current would come directly via the the 1 kOhm resistor from the +12V rail and not through the opto with its 50 mA max output current. The opto would shunt the 0.7-0.8 V forward voltage of the base-emitter to as low as 0.1-0.2 V or so (to its own saturation voltage) and yes, less than 470 Ohm should not be used without abusing the opto, in this respect the opto is a limiting factor if a good power transistor you use for switching needs a minimum of some hundred mA base current.
Yes, there are IGBT types without the body diode, here is one of the manufacturers types from random search:
https://tinyurl.com/yd9uj85f (https://tinyurl.com/yd9uj85f)
(notice switching speed data for them)
and see and check digikey for instance
https://tinyurl.com/y7a5roga (https://tinyurl.com/y7a5roga) (notice I did not check here all whether they have body diode or not)
for instance Microsemi's APT25GR120:
https://tinyurl.com/y9grabqw (https://tinyurl.com/y9grabqw)
To drive an IGBT, you could follow the same latest opto circuit you used, where its output is connected across the gate-emitter of the IDBT and the gate would get the +12V from the 12V rail via also a few kOhm resistor.
Or you may wish to use a dedicated MOSFET gate driver to drive the IGBT, it may not be an overkill for your relatively low frequency tests.
Gyula
-
Hi Gyula!
Well, I just tested 7 transistors in the reverse voltage test: 2X E13007-2, 3X E13007-1 and 2X D13009
They all behave the same - the bulb lights up at about half brightness (in my previous post I wrongly stated that the bulb lit up at almost full brightness - it was at about half brightness).
I very much doubt that all of the transistors are defective, what I think is that they work in the negative resistance region as explained in this video: https://www.youtube.com/watch?v=rpGOKGrcpAk
Thank you for all the links and suggestions, I think I'll order two or four IGBTs and see how that goes.
So that leaves me with nothing to experiment with this circuit for this weekend, well, I'll have fun with some other things to test then.
Others are still welcome to post any thoughts and Ideas.
THX!
Dann
-
You must know that circuit in simulation is not the same as circuit on workbench. I learned that hard way when I was unable to switch off my circuit without receiving a bolt of electrostatic discharge.
Circuits are not isolated in space and sometimes you find the strange factors which enable external energy filling space around ;-) You posted no picture of your circuit , that's fine. Now your turn to find the anomaly
-
Hi Dann,
Sorry for my rusty brain, it did not 'click' for the negative resistance behaviour possibility with the inverse supply polarity.
So I agree, that with this test the 7 transistors cannot be all bad, nevertheless you may wish to test the transistors
(unless you already did so) with a DMM as suggested here
https://vetco.net/blog/test-a-transistor-with-a-multimeter/2017-05-04-12-25-37-07 (https://vetco.net/blog/test-a-transistor-with-a-multimeter/2017-05-04-12-25-37-07)
And there remains the question what is the explanation for only the 30 mV difference
between the on and off states you found between the base-emitter of the transistor with the scope?
Now there is another question with the IGBT: what if its collector-emitter behaves similarly for the inverse polarity voltage?
Afterall, the output side of it is that of a bipolar transistor....
I would suggest a combination of two N channel MOSFETs for switching if you have two power MOSFETs with 200 -300V drain source breakdown voltage ratings for each. Called also an AC switch because AC current can also be switched with it.
The drawback is the need for using isolated gate-source drive circuit because there can be no common in and out points
directly connected. The source pins of the two MOSFETs are connected together and their gates too so you can get a switch between the
two drains and current can flow through them in both directions when the MOSFETs are on. And when they are off, then
no current can flow through any of their body diodes because they are mutually in blocking direction with respect to each other.
I attached such circuit which uses a 9V alkaline battery for feeding the control bias voltage via a reed switch to the common
gates and sources. This was used to short circuit a coil in a pulse motor.
With your opto coupler and the 9V alkaline you can build a driver circuit which is ground independent from the output drains
side where your coil will be with the 12V supply voltage. There are isolated driver ICs too.
-
Hi Forest!
Sorry for not posting the picture, I thought I described the circuit well enough... will post it tomorrow, THX!
Dann
-
Hi Gyula!
Wonderfull, this really looks like it should work. So I just have to put in the opto instead of the reed switch, the correct resistors and the 9v battery? I have some IRF740 mosfets... I'll put it together tomorrow and see what it does.
Dann
-
Yes, the opto output can directly replace the reed switch or the opto output can be connected across the gates and sources to discharge the gate - source capacitance whenever the opto is on so the MOSFETs can switch off. The series resistor R1 (1-2-4.7 kOhm or so) remains of course to limit current via the opto output from the 9V alkaline. Of course a wall plug-in DC supply can also be used with up to 15-18V DC output maximum, it is ground independent from your 12V supply for most types.
Gyula
-
Hi Gyula!
I saw jour post just now, but I already put together the circuit - it is exactly like you have drawn it. I used a 9V battery and a 4.7k resistor (with smaller resistance the opto was unable to sink the current fully). I first tried to switch a 10W 12V bulb on and off with the circuit. It worked, then I reversed the polarity on the 12v battery and it also worked. Then I connected it to the tank circuit and the voltage and current oscillations look good. Awesome! I did not test it for an extended time, I only hope it keeps working.
The next step is to make another channel with another identical LC tank with the second channel from arduino phase shifted by 180°.
I'll post again in the evening, now I have to do something else...
Thank you very much for all the help, it is much appreciated!
Dann
-
Okay Dann, that is fine. The MOSFETs will work as long as their drain voltage and current ratings are violated. In case you use a tank circuit in which the coil will have but 1 - 2 Ohm DC resistance or even less, then the drain current limits may easily be reached (depending on your supply votage amplitude of course). Drain voltage ratings are added due to the series connection of the FETs so it has higher safe margin. Unfortunately the RDS 'on' resistances are also in series and added, doubling the loss of the switch.
One notice with respect to the bipolar power transistor you used so far: if a diode is connected in series with the collector in the forward direction for your 12V supply voltage, then this diode could block any current whenever the collector - emitter get the inverse voltage from the tank circuit swinging so that would also be a solution a few days ago. Of course, the 0.7V forward voltage drop of the series diode in the on state of the transistor would add to the switching losses if that were important. But anyway the opto device could still have issues to drive the base-emitter correctly.
Gyula
-
Hi Gyula!
Before I started this thread I did play around a bit with the optos and mosfets and there was one configuration with the diode after the emitter that worked... well... for about 30 seconds ;D , after that the oscillations gradually decreased and stopped. :( The mosfet became faulty and after that I did not want to sacrifice any more components.
Anyway, now I added the second channel on the protoboard (identical to the first) and when I tested it, it would not switch - the opto was not switching, then I replaced it with the opto from the first channel and everything worked as expected.
The optos are identical and I tested that opto before, I know it works, maybe just with a different resistor. So it might be that the optos are worn out to some extent, tomorrow I'll try some other ones in the circuit just to see their performance.
Will also post some scopeshots and a picture of the "rats nest" ;D
Dann
-
OK!
Here are some scope shots and a picture of the circuit (not the best picture) : https://postimg.cc/gallery/mpgxxk12/
I rechecked the opto's datasheet and figured out that the resistor for driving the opto's LED was not right (1k ohm - allowing only 5mA of current) and the opto was barely switching on the HV side. By measuring the current through the opto's LED I selected a 220 ohm resistor (for around 20mA of current, the forward voltage drop on the LED was about 1.24V).
The pull up resistor for the mosfet gates was also allowing too small current, so I replaced it with 470 ohm (about 18mA through the opto's transistor) and now the switching is more solid.
In the picture (of my bench) there are two channels - only the second one is used right now connected to only one LC tank (There is another LC tank on the right - not used).
The double coil in the middle (connected to a small lightbulb) was just used to check the output when all the coils are lined up and close to one another (Figuera setup)
That is all for now, I'm headed in the Figuera direction with this setup, so I'm not sure about posting here or in a new thread under the solid state devices section in the future. Are there any moderators there?
THX all!
Dann
-
Okay Dann, thanks for reporting back.
Well, opto devices are different, some types are able to work fine with less than 10 mA input diode current, other types need higher currents. (Maybe this was part of the problem when driving the bipolar transistor some days ago, maybe not.)
Also, within the same types a single opto may work inferior or better to an another one of the same type.
I wish you further good luck.
I do not know about moderators.
Gyula