Use for the Tesla Switch

[nvisser]

@ vissie
If you are concerned about the triggering I would consider replacing the MPS8599s with MPSA06s then you could just connect a H11d1 between the collector and base of it with the diode and cap just like Bits is doing with his build and I also am doing with mine.
Carroll

Of coarse, Opto`s. How could I forget about them?
O yes. I saw that circuit to with the MPSA06. A npn.
Thank you for your response
I still dont see where the gain will come from in a setup like this.
I could not get the scaler charger to work .

[Bit's-n-Bytes]

Back to basics

Hi gang, well I have modified the D-TS to resemble more of the contemporay one. I changed Q9 and Q10 to NPN. I am now using this code;

high 4
pause 2000
high 5
pause 500
low 4, 5
pause 50
high 6
pause 2000
high 7
pause 500
low 6, 7
pause 50

Pin 4 = Q1
Pin 5 = Q10
Pin 6 = Q2
Pin 7 = Q9

R7 of Q10 is now connected to the collector of Q10 and R8 is now connected to the collector of Q9.

The code basically brings the batts in series longer before we fire Q9 or Q10. This creates the likeness of a surge going into the opposite side. The load I have on this test is the 1157's.

Bit's

[ldissing]

Limiting

Hi gang, well I have modified the D-TS to resemble more of the contemporay one. I changed Q9 and Q10 to NPN. I am now using this code;

high 4
pause 2000
high 5
pause 500
low 4, 5
pause 50
high 6
pause 2000
high 7
pause 500
low 6, 7
pause 50

Pin 4 = Q1
Pin 5 = Q10
Pin 6 = Q2
Pin 7 = Q9

R7 of Q10 is now connected to the collector of Q10 and R8 is now connected to the collector of Q9.

The code basically brings the batts in series longer before we fire Q9 or Q10. This creates the likeness of a surge going into the opposite side. The load I have on this test is the 1157's.

Bit's

The purpose of the resistors on Q10 and Q9 is to limit the current through them, no? That will, of course, not stop the 24V upper batteries from discharging through the load to the lower battery on the 24V side, hence, both loads appear to be on...and indeed they are. I still believe the term "relief valve" is an incorrect term, but here you use the term "surge", so you may have decided that your PNPs were really off all that time? The latest code will indeed create a surge, and I agree with you wholeheartedly on that term. If you "really" wanted to create a surge, you'd put back in the lower transistors but you'd lose both lights being lit all the time (except when everything is "off", because nothing is lit during that 50ms in the code above).

My batteries are still losing potential, but I have a new switching scenario in mind that seems promising in my head...working on it still.

Leroy

P.S. JB has not told the whole story on the switching technique he is using...IMHO. He is an audio engineer after all...out of my league, as are most of you on this list!

[Bit's-n-Bytes]

My batteries are still losing potential, but I have a new switching scenario in mind that seems promising in my head...working on it still.

Would love to see your new switching scenario.

Bit's

[blackchisel97]

It was nice talking to you Bit's
Still looking for best charging spot. Indeed, it is much simpler to change a code than play with 4 pots
For now I'm stuck with 7.2Ah SLA batteries as I don't have a four LAB to try it.
Watching DTS working tells me that my analog version did work and all was needed is the load to be balanced with proper timing.
I think I got some understanding (with Bit's help ) what we need to look for but despite being able to change timing "on the fly" it isn't simple.
My bulbs are automotive 12V 5W. I'm just trying to get them glow. If they're getting too bright it means we're allowing current to flow. What I understand is that we have to abruptly dump over 12V bank before current starts to build up.
So, turning Q1 on, connecting batt.1 & 3 in series, hold them just long enough to build 18 -19V and pulse Q9 (PNP version) before current starts to flow. Turn Q1 off and Q9 back on at the same time, than pause before repeating the same pattern on Q2 and Q10.
I didn't determine perfect "dead time" length (pause between batt. bank 1 & 3 and 2 & 4). I know it's important but I want to determine two variables; Q1 and Q2 ON times and "cut off" time first. DMM reading can be confusing at times. Just because it seems like a battery is dropping doesn't have to be a bad sign. After a while it may start to rise. Sometimes it looks like it just sits and does nothing. Every time I'm confused I can see John's reply to my enthusiasm over TS in post 454 - " The Tesla Switch is very difficult to build..." No kiddin'


Vtech

[JohnStone]

Hi folks,
Reading this thread since many month.... Great!
Until now I had not the funds for experiments – bat this state will end soon :-)

Preparing my experiments I try to understand the components and circuits posted here.
Some questions and hints:

@Bits
Congratulations to your success!
1.
Your schematic contains a not understandable area (for me) – driving Q9 an Q10. You corrected this part (emitter of the transistor need to be connected in the same direction (arrows) but it is still not understandable for me. I may be wrong – but maybe you can enlighten my brain!

The opto here should be a low side switch and should enable the current flow from high side emitter along the transistor base through the opto, resistor, to low side collector voltage level. So far my comprehension. This seems not to be realized in the schematic ?:-(
Nevertheless a transistor can work if collector and emitter exchanged. The voltage CE is less than in normal mode and the current amplification is extremely lousy. This operation mode was sometimes used before the world was blessed with low impedance FETs.
2.
Where is the high side when switching on?

@ ALL
Congratulations to you all with your efforts and success!
Studying the thread I’d like to direct your focus to a quote of JB reposted recently:

“If you look at the Tesla Switch you will find that some of the devices are inverted in parts of the switch, you will get a loss if not switched correct. Some Tricks, to get the power through the device you must use a base discharge circuit. The audio transformers did this as it was connected between the base and emitter. “

It is all about switching and switching speed. That rings in my ears! But this is not expressed directly here. Here I give some of my thoughts to think on. I do not know if this is relevant in the TS design - but please check and make the best out of it.

Commonly known is that a transistor is a current amplifier and it does this by a base current. (we will come back later to this …)

Let’s have a look first to some not commonly known but essential properties of a transistor:
Commonly unknown is that the Base contains a build in capacitance (BE and BC as well and CE). In lousy normal conditions this is of no importance. But high skilled schematics like TS need to perform fast – sorry very fast – switching and therefore these capacitances is of great importance.
(Think of Nicola Tesla: He experienced a great advance in his science when he added strong magnet fields to his commutator or spark gaps in order to gain much faster switch transitions.)

Let’s take as example a NPN transistor (PNP behaves similar but special to PNP functionality).

The CB capacitance performs a direct connection to the collector (for short time). In switch off transitions this can make the base switch on again for a short pulse. I red in a post something about double pulses and ringing – singing transitors… !). Apart from increased heat dissipation this can make a setup non functional.
In switch on conditions this is a small coulomb thief (out of base current) and reduces the speed of switching transition (see below).

The BE capacitance reduces the switch on and off speed – especially when using power transistors witch can have a high capacitance. This capacitance needs to be loaded first before the base current starts to flow to the emitter in order to open the CE current flow. Of course – the voltage is 0,7 V only but this is essential when fast transitions are intended.
To minimize this effect the bias resistor can be divided in two elements. A well tuned capacitance in parallel to one of the two resistors will make things better. But please check if the opto can drive this inrush current. All needs to be tuned well.

In case of switch off condition this well loaded BE capacitance will feed the base current for a short while. Fast transitions can obtained only if this capacitance is being discharged as fast as possible. Refer to JB –“discharge circuit” above.

In the current TS the opto charges the BE capacitance but there is provided no means to speed it up or discharge it. For faster off transitions this is a must. Apparently this is the reason someone mentioned bipolar totem pole switches.
(By the way: FETs suffer much more from this effect but this effect is well known and considered there - fast and powerful high and low side drivers.)

Now listen to JB: “The audio transformers did this as it was connected between the base and emitter” Did you get it?

The transformers performed a fast and high rush in current to the base and a fast rush out as well. You need for a short time POWER and voltage for charging / discharging the base capcitance in order to obtain fast transitions. Amazing that in the cigar box TS there was no protection for reverse CB voltage in order to protect the BE junction of transistors. Some diodes in series will perfom well (1N4148). (No zehners please because they contain a high capacitance). Usually 7V reverse is the maximum allowed for transistors while initial higher voltages to switch on will be helpful.
Please check and test if a 2nd opto for discharging the base capacitance will help for faster transitions. I never red of some measurements regarding transition time!

An other concern is the base current applied. The transistor is a current amplifier. When using NPN transistors this current is a waist current while using PNPs the current is part of the intended load current. But there are other trade offs in the last case. When switching relatively small currents the transistor performs a very good and effective amplification. Increased collector currents diminish the amplification factor tremendously sometimes. See data sheets. This tells us that the base current needs to be well tuned to the current to be switched. Please take in account the maximum peak collector currents. So please tune transistors.

There is an other trade off regarding the base current:
High base currents (necessary when collectro currents need to be high) makes the switching of the transistor slow. In this case the base capacitance is to be discharged – it is a MUST.
This tells us that a TS needs to be tuned to the intended current flow. Do not inject a too low or o too high base current!

This is what I know of transistors. As far I understand JB this comprehension is not diminished if we need to think “off road” in order to grasp some experience in scalar energy. Please correct me if you have better knowledge. This will the way for all of us to learn.

I do not know if this helps but please check your setups, measure and decide yourself.

I hope that my thoughts are expressed clearly enough because English is not my native language. Please ask for clarification if necessary.

JohnStone

[ldissing]

Would love to see your new switching scenario.

Bit's

We are dealing with many variations of the "switch". It is the switching, er...rather the devices...that matter according to JB. I believe that he said..."the timing is not the key, but the devices are the key".

So, my switching scenario won't mean a hell o' beans if that statement is true...so, are we to believe the above. I actually believe it is the switching of the device, but it is not "just" the cap and diode in parallel that causes the "magic" to happen. There is something more, and if, when, JB decides to "let" us in on the "secret" then we will have something. The switch by and of itself it too simple for it not to "work". We are not all idiots (myself included in the idiot category), but until we figure that out....the scenario.

My scenario is based on your work and my work with NPNs. I can light both lights with NPNs just as you do with PNPs. The flow of "charge" is not what we initially expect. When a single "serial" transistor is thrown and no others...we get both loads to light up. One brighter than the other. Then, we can switch that one off and switch the other serial one on, and the other load is brighter than the other. SO, the potential is flowing in two directions, one from the two parallel batteries to the lower battery on the 24V side...easy enough. The other flow is the interesting one. It flows from the 12V side of the 24V side through the diode connected to the "12V" side and then on to the bottom battery on the 24V side...less current than the "other" load, as those currents are additive, hence more light on one side than the other.

The above is true of the "no transistor on top", "just diode, two serial transistor" version too, although a little more comes into play with that version. There is an additional current coming from the 24V top battery in that version, in my mind I can see it flowing, but have not "tested" for it.

So, we want all these currents flowing, the concept of open paths and all. So, we only need to turn on and off the serial transistors and get rid of the transistors on top all together. We can pulse the serial transistors to keep current flowing I think.

It would go something like this (not in any particular language, I made it up):

I'm not sure about length of time yet!

-> start of program

-> both side start off
left side off
right side off

; start currents flowing both directions, left first
left side on
right side on

; this is the location to jump to to re-start the cycle
:jump1

; now only the right side, so turn left side off
left side off

; start currents on left side again
left side on

; now only the left side, so turn right side off
right side off

; start currents on right side again
right side on

goto jump1

My theory is, that starting with both serial transistors (Q1,Q2) on, that when the one serial transistor is turned off (assume Q1 in your schematic) with the other one (Q2) on, it will cause that transistor which has current flowing from the positive of the 12V bottom battery (bat 3) through the transistor out the diode to the other side (and to the other battery above it (bat 1)) to freak out just a little bit. At a minimum, current that was flowing basically from the negative of bat 1 to the other side is abruptly shut off, so we will get surge of current to the positive of bat 1 and to bat 3 AND if the transistor were to "freak" out and go negative, then current is flowing in the opposite direction through the tranny which is an abrupt current flow reversal. I don't really know that it will work, but it is a switching variation on the diode, two serial tranny TS.

It is just a theory right now, as it is colder than a witch's boob in a brass brazier in the garage. Plan on testing it when my batteries are fully charged and in excellent condition and it is slightly warmer in there.

Leroy

P.S. Probably just another rabbit hole, but I've tried about everything else. I just think that the switching is not right, and "just" having it controlled digitally is "not enough", it is in the connections and I do not know what the connections are. Of course, I'd have never thought to put a diode and cap in parallel there either, so I'm closer than I've ever been.

[Bit's-n-Bytes]

!

Your schematic contains a not understandable area (for me) – driving Q9 an Q10. You corrected this part (emitter of the transistor need to be connected in the same direction (arrows) but it is still not understandable for me. I may be wrong – but maybe you can enlighten my brain!


JohnStone

Thanks JohnStone, attached is the updated schematic to the way I presently have this circuit. R7 and R8 are to control base current and limit the opening of Q9 and Q10. I am still experminting with this configuration and may yet change it again. I am preparing to experment further with putting the caps in as I earlier posted and expect a huge surge from them if it works out right. Caps will be attached to the collectors of Q9, Q10.

Hope this helps.

Bit's

[Bit's-n-Bytes]

We are dealing with many variations of the "switch". It is the switching, er...rather the devices...that matter according to JB. I believe that he said..."the timing is not the key, but the devices are the key".

So, my switching scenario won't mean a hell o' beans if that statement is true...so, are we to believe the above. I actually believe it is the switching of the device, but it is not "just" the cap and diode in parallel that causes the "magic" to happen. There is something more, and if, when, JB decides to "let" us in on the "secret" then we will have something. The switch by and of itself it too simple for it not to "work". We are not all idiots (myself included in the idiot category), but until we figure that out....the scenario.

My scenario is based on your work and my work with NPNs. I can light both lights with NPNs just as you do with PNPs. The flow of "charge" is not what we initially expect. When a single "serial" transistor is thrown and no others...we get both loads to light up. One brighter than the other. Then, we can switch that one off and switch the other serial one on, and the other load is brighter than the other. SO, the potential is flowing in two directions, one from the two parallel batteries to the lower battery on the 24V side...easy enough. The other flow is the interesting one. It flows from the 12V side of the 24V side through the diode connected to the "12V" side and then on to the bottom battery on the 24V side...less current than the "other" load, as those currents are additive, hence more light on one side than the other.

The above is true of the "no transistor on top", "just diode, two serial transistor" version too, although a little more comes into play with that version. There is an additional current coming from the 24V top battery in that version, in my mind I can see it flowing, but have not "tested" for it.

So, we want all these currents flowing, the concept of open paths and all. So, we only need to turn on and off the serial transistors and get rid of the transistors on top all together. We can pulse the serial transistors to keep current flowing I think.

It would go something like this (not in any particular language, I made it up):

I'm not sure about length of time yet!

-> start of program

-> both side start off
left side off
right side off

; start currents flowing both directions, left first
left side on
right side on

; this is the location to jump to to re-start the cycle
:jump1

; now only the right side, so turn left side off
left side off

; start currents on left side again
left side on

; now only the left side, so turn right side off
right side off

; start currents on right side again
right side on

goto jump1

My theory is, that starting with both serial transistors (Q1,Q2) on, that when the one serial transistor is turned off (assume Q1 in your schematic) with the other one (Q2) on, it will cause that transistor which has current flowing from the positive of the 12V bottom battery (bat 3) through the transistor out the diode to the other side (and to the other battery above it (bat 1)) to freak out just a little bit. At a minimum, current that was flowing basically from the negative of bat 1 to the other side is abruptly shut off, so we will get surge of current to the positive of bat 1 and to bat 3 AND if the transistor were to "freak" out and go negative, then current is flowing in the opposite direction through the tranny which is an abrupt current flow reversal. I don't really know that it will work, but it is a switching variation on the diode, two serial tranny TS.

It is just a theory right now, as it is colder than a witch's boob in a brass brazier in the garage. Plan on testing it when my batteries are fully charged and in excellent condition and it is slightly warmer in there.

Leroy

P.S. Probably just another rabbit hole, but I've tried about everything else. I just think that the switching is not right, and "just" having it controlled digitally is "not enough", it is in the connections and I do not know what the connections are. Of course, I'd have never thought to put a diode and cap in parallel there either, so I'm closer than I've ever been.

Interesting, basically creating a "wave" then cutting just the top portion of it off before it changes direction.

[ldissing]

Interesting, basically creating a "wave" then cutting just the top portion of it off before it changes direction.

Well, I keep thinking about the diode only version (two serial transistors). How can that possibly work....right? The flows will be continuous as long as the serial side is on, and we have yet one more current flow, from the 24V side (bat 1 and 3) to bat 4 positive to think of here. So, we NEED to keep it from frying our components, I mean, I can blow a 30 amp fuse with these batteries in a heart beat. So, can I also fry 35 amp diodes....sure...and 16 amp MJLs....you bet ya. Need to keep the "charge" from getting up to full speed...switching them constantly. Or, letting them rest for a while and then jumping back on them with some more "charge" before they are fully at rest. It is an interesting phenomenon. This charge bouncing back and forth. JB once told you (Bits), to "pulse one side three times", to get the battery in recharge mode, I assume...then switch it while it is in the recharge mode..."stealing" some of the ion energy....we could do this in this pulsing scenario too, but only long enough that the "re-charge" is continuing. JB is the master at this stuff.

I have some ideas on the "analog" monitoring, but Bits, please think about moving your "computer" to the lower battery...not the upper one.


Leroy

[Inquorate]

Humble request

The 'magic' happens in the batteries, right? Do you have a scope, bits? I'd love to see the battery waveform alongside the switch waveform.

[ldissing]

The 'magic' happens in the batteries, right? Do you have a scope, bits? I'd love to see the battery waveform alongside the switch waveform.

The magic "does" happen in the battery, but the magic also comes from the "devices" according to JB.

Jack Handy....laugh at me all you want...I'm going to get this if it is the last thing I do and it may be You know....Stubblefield quit working on his farm and died alone in a house over this..."free energy"....I'm not far behind him.

[blackchisel97]

I don't have DMM with a bar but found analog V-meter across banks quite helpful. I can see if I have enough potential across the serial bank and the "snap". A bit slow switching to get a good reading on my scope. I reduced ON time from 500 to 300ms and still have enough to dump across 12V bank. I tried changing load and small inductors as a load as well as jumper in place of loads. I found that I need to make ON time slightly longer with jumper than with a bulb. It seems that my 5W bulbs are good enough but more tweaking is needed. I also found Q9 colder than the rest. I mean, they're all cold but Q9 is just as cold as a metal in room temperature (19deg C). I'm tempted to try Bit's setup with caps but not until I'll get a good "feeling" of this "vehicle".


Vtech

[Inquorate]

My idea of using a relay didn't work. Tried a self oscillating relay to drive a coil in my 3 battery setup, that didn't stay on long enough to build a magnetic field.

I realised I have a solid state relay though, and found another 555 timer hiding somewhere it shouldn't have been. I reckon the solid state relay has an optoisolator.

So, when I get some more time, will play.

[blackchisel97]

My idea of using a relay didn't work. Tried a self oscillating relay to drive a coil in my 3 battery setup, that didn't stay on long enough to build a magnetic field.

I realised I have a solid state relay though, and found another 555 timer hiding somewhere it shouldn't have been. I reckon the solid state relay has an optoisolator.

So, when I get some more time, will play.

You'll get it, one way or the other I think you're on the right track.


Vtech