Solid State Build

[BobZilla]

I’m back with an update.

I have made a significant change to the circuit. I have removed the Cap-Diode from the trigger leg. Let me explain why I did this…

These test runs take so long to complete so I built a small single transistor circuit to try changing around while the main test is underway. I wanted to try larger caps, smaller caps and higher draw vs lower etc. I figured with a small setup I could get a feel quickly for how changes would affect the larger machine.

Here is a shot of the small circuit. The coil is glued on under the circuit board, it is 44 feet of 22AWG on trigger and power.



So this post is not about this little circuit but I have to say this thing works very well at charging up to 5AH batteries. I have been using it to charge my meter batteries and some AA’s for my telescope. Works very well!

I was changing out caps on this thing and I thought to take a voltage reading across the cap. I did this by hooking the negative to the Primary negative and the pos on the pot leg with the diode leading towards the base. I noticed that at low draw the voltage was high like around 18-22V and you could see the cap fire and recharge, but as I increase the draw the voltage would drop. After a certain threshold you no longer see the cap charging and discharging but rather just steady at a relatively low voltage. I do not know if it actually is still cycling and the meter just cannot sample fast enough or if the resistance just stays fixed at that point and no longer flip flops between the pot and cap path.

Whatever was happening I decided that it cannot be good for charging. I much prefer to see that base getting hit with higher voltage. The cap-diode works great at low draws and as a matter of fact that is how Patrick was using it when he showed it in his video. I had strayed from his method of using it when I decided to crank up the juice. I would still recommend it on smaller systems or if you use low draw. It does a wonderful job of pulse triggering as long as you don’t turn up the juice to much.

Here are some pictures on the big machine where I set the draw low and then high. I went and confirmed that the big machine was behaving the same way.

With a low draw, voltage is high on the cap trigger.

*EDIT*
I attached the negative of the meter on the negative charge side in these shots because it was easier, the readings come out reversed polarity but it really doesn't matter for demonstration.




With a high draw voltage goes low on the cap trigger.





I have removed the cap and diode from the big system and have a test run underway. I will post back about those results but I wanted to share what I discovered about how the circuit was running with the cap in place. -----Bob

[BobZilla]

Ok I’m back with some results from this run. As a reminder about what is different this time I have removed the cap-diode from the trigger. I also changed the intermediate resistor to a 100 Ohm 2 watt, I think a ½ watt would have been fine here but this is what I had on hand. Nothing is getting warmer than a few degrees over ambient on the circuit.

Resting voltages prior to start of were Pri=12.8, C1=12.16, C2=12.19



Here it is flipped on, amp draw was set to 1.52A; I pretty much kept it there for the whole run.




Now these charts are pretty darn cool. I was able to get both charge legs simultaneously and it does not look anything like I though it would. Notice C1 seems to be the leader and has high spikes of voltage. C2 charged right along with C1 but at a very steady climb, see the difference?


This is the chart for C1



And for C2



Note that that big dip in both charts (around 100k) was where I turned the draw back to .6A to see how it would run. It was still charging but at a much slower rate so I put it back up to finish the next morning.

Also those with a keen eye may notice that the time scale on C2 is about 1k shorter than C1,, my meter battery ran flat on me and I didn’t notice so the chart stopped plotting until I put a fresh one in. You can see that mishap at around 45k when the chart jumps up abruptly. C1 meter never lost power and that time scale is the true time of this run. It was approxcimatly 185000 seconds or 51 hours but about 30k of those were at .6a

Before I had two meters I thought what we would see would be one climb and one drop in a flip flop fashion but that is not what happened. I’m not sure if this is how it would normally go or if my goof up on the board has caused this, remember I posted that I had to jumper some stuff in from my damaging some traces. We will find out on the next run.

Now I have another idea and it is going to change the whole idea behind this machine but I really want to make a run to find something out. I have suspicions that by making to two pairs of transistors I may be hurting the charge power.

Mr. Bedini first showed branch circuits with independent legs and for good reason. The transistors even though these are matched are probably not firing at exactly the same time. I may be zeroing out part of the potential by mixing them like that. I am going to set this board up as a two transistor unit where the power legs are truly independent, no mixing. This will also let me eliminate my goof up on the board; I just won’t use that branch. I should be able to use less draw this way too because I will only need to trigger two instead of four.

So what does that leave me with,, two extra wires right.

After a run like that and see results I can join one end to another with the extra wires and create two 400ft strands on the power legs. That will be yet another test configuration. I also have a totally different coil we can try out if the tests warrant it.

Mike should like that ;-) ,since he has asked about various affects from length. We can see what happens first from 200ft and then 400.

Here is a shot of the final resting voltage.

Pri=11.3, C1=12.71, C2=12.73



Over all I am pretty darn happy with how this thing is working but I want to improve it if possible.. We still did not push them all the way to 15v but they got a pretty good charge.

It is obvious but let’s not forget this is a two battery charger, think if we were only trying for one how good it would be. I think I may have been able to get the full charge on them but at the cost of running the Primary too low and possibly causing damage.---Bob

[John_Koorn]

Hey Bob, great report once again.

I would parallel two windings together instead of putting them in series. The transistors will be able to handle it.

John K.

[BobZilla]

Thanks John. You know I didn't think about a parallel setup for those wires. I really like that idea. That is almost what I intended originally by harnessing four wires but eliminates the transistor timing variable.

The next run will be just two of the original wires so that I can get a clean baseline of how that works and then I can try parallel wires for comparison. I think I may also change out the pot for a 5 or 10k. I am only working in the upper range of the pot with these tests and that should give me a finer dial. As the machine was setup on this run there was a jump through area on the pot where it would go from around.6A right up to 1.6 or so and could not be held in between. I dialed in to the second sweet spot I found in that high range. ---Bob

[BobZilla]

I’m back with the results from the last run.

This time we are looking at the circuit running from only two transistors and two power wires. The other wires are not connected and I removed the transistors, diodes etc for the other two branches.



These results have really got me thinking if I took a wrong turn when I took off the cap-diode mod. It appears to me that my results are basically the same on this run as the last and that indicates to me that I must be doing the majority of this charging with current. I ran about the same current through this time as last but it was on only two transistors. It kind of tells me that the current did it and not spikes. What I mean is here we have half of the wires and no real change so I must have really just been driving these two transistors twice as hard but they produced about the same as four did before.

Here are some thoughts on it. It is all just speculation but I am trying to share my experience with this machine so that others may benefit.

With the cap-diode as I have mentioned already, you get very nice pulsing from the cap as long as you keep the draw low enough that you don’t just brute force on through the cap. I think it may be key to use the cap-diode and not crank it to that point. As long as you are operating it properly you do trigger nice spikes from the sharp Pop, Pop of the cap but if you drive it to hard the cap becomes in effective at producing the pulse.

I have seen it work really well on smaller setups but it takes a long time. I may have been too impatient and trying to force this thing faster than it wants to go and take advantage of the radiant spike. I suspect that if I were to use the cap diode and keep it at a low pop, pop setting I would see really good efficiency but the problem for me is it takes far too long. With smaller batteries I didn’t mind so much but with these 75AH it could take like a week to complete. It may turn out that that is just how it goes but I just do’t know at this point. I am agoing to keep trying for a reasonable charge time but I want to be mindful of using the spikes and not just passing current around.

So I guess I am back to the drawing board with this. I think I may need to play around more with the cap-diode and try more combinations between capacitance and resistance on the pot. I want to find the fastest pop with the largest voltage without over shooting it to where the cap is not really pulsing.

Again I think as a charger this thing is doing decent, I mean I am getting 12.7 resting on two batteries from one but I know this can be done better and that is what I am shooting for. I have done similar tests with smaller setups and batteries that actually push the charge batteries all the way up from one battery of the same size so I know it can be done, my issue is scaling the whole thing up for these large batteries.

So here are the final charts and numbers from this run.

Start resting Pri=12.6, C1=12.29, C2=12.19



And started, draw is 1.46A



Final resting Pri=11.3, C1 12.70, C2=12.72

C1 Chart


C2 Chart


It may be a little while before I come back on this with another update. I want to play around some more with the cap–diode and see if I can’t get this sorted out. I do think results are valuable to look at even if they are not perfect but I don’t want to just keep posting the same thing over and over.----Bob

[BobZilla]

Back again ;-)

Ok guys I have come up with something that I think is pretty special. I don’t know if this has been done in the past but I have not heard of it being done. People have been tinkering with these circuits for a long time so who knows.

First let me explain what I wanted to accomplish with this, what prompted me to come up with it.

I really liked how Patrick’s CPD runs as I have explained before, but I like to run my machine at a higher draw. When using a CPD you reach a point pretty quickly where you zoom right past a pulse and start into a fast frequency.

I wanted to decouple the pulse rate from the resistance setting on the trigger circuit. This did not seem possible with the CPD because it is in series with the main trigger circuit. Less resistance=faster frequency and they are tied together that way.

What I wanted was a way to set the resistance on the trigger to anywhere I want it and then be able to control the pulse independently. Here is my solution.



You may be looking at that and thinking, He forgot the base resistors, the intermediate resistor! Nope I did not. I am not using them in this setup. The pot and photo resistors provide all the resistance I need. I like how that works out too because it is one less thing to have to try and match.

Here is a shot of my board with the resistors jumped-through.



I constructed an Optical Pulse Generator to control the current across the trigger. It pulses light into photo resistors to make the trigger rise and fall. Obviously you can change out the capacitor values to get different pulses but I put what I have found to work well in the schematic.

Here is a shot of the photo resistors and the LED's. I found that RED works best. They have a lower voltage than blue or white and play nice with the cap that is driving them. The color spectrum may have something to do with it too for the resistors to pick up the light. I did try Blue, White, and RED. The LED's are not the t1 style, they are larger and are capable of really bright output.

Also notice that there are TWO of the resistors in parallel. This was crucial to getting this to work properly. With only one resistor the light would not drop the resistance low enough to get a good current through the base. When you parallel two of them the same rules apply as with normal resistors, you increase their dissipation and lower their resistance. I was mostly concerned with the resistance but having two of them share the load also helps not to over drive a single one so it works out nice that way. I did not find that I needed more but you could parallel as many as you need to get the response that you want.



As is usually the case with these devices it takes a lot of care to get it tuned well. It is important to understand that the “dark” period is what we are really trying to manipulate. When the photo resistors go dark they introduce massive resistance and do not allow current to flow. The quick rise of light excites the junction and at the peak of brightness it brings the resistance down to only 100Ohm or so. Of course it all depends on your particular LED’s and photo resistors.

If you pulse too quickly you will over saturate the P junction and not get a complete switching off. The device will run this way but we want a sharp OFF (Dark) to occur AND enough time for the photo resistor to fall back They do not react at the speed of light ;-) , otherwise there is no need for a pulse at all. You could just drive the circuit normally and put in a normal resistor to accomplish the same thing.

From my experimenting I have found that if you adjust the pot of the pulse circuit just to where it pulses and then back it off a very little you will get a nice ramp up and then a dump which causes the cap to go weak and the cycle starts over. To put it another way you will see dim light for a short period as the cap charges, then it dumps and you get a very bright flash, then a dark period while it charges back up.

This is what you want because even the short dim period activates the photo resistor so you will see a draw at about 1/3 then a quick ramp up to full, then off. This works out so that we are not using a half on half off cycle but more like 1/3 dim, 1/3 bright, 1/3 off.

The off period is where we get the spike from the coil collapse. It is the equivalent of a magnet pass on a mechanical system.

Here is a short video showing what I am rambling about. I have the meter on resistance across the photo resistor. The machine is not powered on (not charging anything), we are just looking at the action of the trigger pulse. Notice we go from several million Ohm down to near zero and repeat. The meter has a hard time actually showing it properly because of the sample rate but you get the idea.

Trigger Ohm


We want the resistance to go high on the pulse, this is the equivalent of a magnet leaving the coil. If we do not go high we are not going to get nice spikes. Low resistance means the trigger is on, high means it is switched off.

Here are the LED’s without the cover to show the ramping up of the LED. Remember we don’t want a really fast pulse because it doesn’t allow the Ohms to rise high enough. If the photo resistors could switch at the speed of light it would be perfect but they cannot. They have a rise/fall time that we need to work with.

Trigger Ramp


Here is a video of it while the machine is on (charging batteries). Watch the amp draw. When it is low that is the pulse circuit charging the cap, the main SS is not drawing anything at that point. Then it climbs while the dim period begins,, then the flash where we pull +1 amp. Then the dark period and it begins again.


Trigger Running


So there you have it. I am still experimenting with small changes but so far this method seems to be really good. The huge bonus of this is that you can set your main trigger pot at whatever you like for your particular load and then control the trigger pulse independently of that. I was triggering pretty high draws (1 – 2 Amps) but what I am saying is you could turn your main pot down so that it can only draw say ½ an amp if that is what you prefer and it has nothing to do with the pulse.

I will do some runs and post back later. I just wanted to share this with the group.

By the way I did try this with photo diodes and photo transistors but they didn’t work as nicely as this. I cannot say with full confidence that this is the way to go but this is a group for experimenter’s right, so that is what I am doing and I am happy to share ideas.
---Bob

[min2oly]

Back again ;-)

Ok guys I have come up with something that I think is pretty special. I don’t know if this has been done in the past but I have not heard of it being done. People have been tinkering with these circuits for a long time so who knows.

First let me explain what I wanted to accomplish with this, what prompted me to come up with it.

I really liked how Patrick’s CPD runs as I have explained before, but I like to run my machine at a higher draw. When using a CPD you reach a point pretty quickly where you zoom right past a pulse and start into a fast frequency.

I wanted to decouple the pulse rate from the resistance setting on the trigger circuit. This did not seem possible with the CPD because it is in series with the main trigger circuit. Less resistance=faster frequency and they are tied together that way.

What I wanted was a way to set the resistance on the trigger to anywhere I want it and then be able to control the pulse independently. Here is my solution.

[ATTACH=CONFIG]1917[/ATTACH]

You may be looking at that and thinking, He forgot the base resistors, the intermediate resistor! Nope I did not. I am not using them in this setup. The pot and photo resistors provide all the resistance I need. I like how that works out too because it is one less thing to have to try and match.

Here is a shot of my board with the resistors jumped-through.

[ATTACH=CONFIG]1919[/ATTACH]

I constructed an Optical Pulse Generator to control the current across the trigger. It pulses light into photo resistors to make the trigger rise and fall. Obviously you can change out the capacitor values to get different pulses but I put what I have found to work well in the schematic.

Here is a shot of the photo resistors and the LED's. I found that RED works best. They have a lower voltage than blue or white and play nice with the cap that is driving them. The color spectrum may have something to do with it too for the resistors to pick up the light. I did try Blue, White, and RED. The LED's are not the t1 style, they are larger and are capable of really bright output.

Also notice that there are TWO of the resistors in parallel. This was crucial to getting this to work properly. With only one resistor the light would not drop the resistance low enough to get a good current through the base. When you parallel two of them the same rules apply as with normal resistors, you increase their dissipation and lower their resistance. I was mostly concerned with the resistance but having two of them share the load also helps not to over drive a single one so it works out nice that way. I did not find that I needed more but you could parallel as many as you need to get the response that you want.

[ATTACH=CONFIG]1918[/ATTACH]

As is usually the case with these devices it takes a lot of care to get it tuned well. It is important to understand that the “dark” period is what we are really trying to manipulate. When the photo resistors go dark they introduce massive resistance and do not allow current to flow. The quick rise of light excites the junction and at the peak of brightness it brings the resistance down to only 100Ohm or so. Of course it all depends on your particular LED’s and photo resistors.

If you pulse too quickly you will over saturate the P junction and not get a complete switching off. The device will run this way but we want a sharp OFF (Dark) to occur AND enough time for the photo resistor to fall back They do not react at the speed of light ;-) , otherwise there is no need for a pulse at all. You could just drive the circuit normally and put in a normal resistor to accomplish the same thing.

From my experimenting I have found that if you adjust the pot of the pulse circuit just to where it pulses and then back it off a very little you will get a nice ramp up and then a dump which causes the cap to go weak and the cycle starts over. To put it another way you will see dim light for a short period as the cap charges, then it dumps and you get a very bright flash, then a dark period while it charges back up.

This is what you want because even the short dim period activates the photo resistor so you will see a draw at about 1/3 then a quick ramp up to full, then off. This works out so that we are not using a half on half off cycle but more like 1/3 dim, 1/3 bright, 1/3 off.

The off period is where we get the spike from the coil collapse. It is the equivalent of a magnet pass on a mechanical system.

Here is a short video showing what I am rambling about. I have the meter on resistance across the photo resistor. The machine is not powered on (not charging anything), we are just looking at the action of the trigger pulse. Notice we go from several million Ohm down to near zero and repeat. The meter has a hard time actually showing it properly because of the sample rate but you get the idea.

Trigger Ohm


We want the resistance to go high on the pulse, this is the equivalent of a magnet leaving the coil. If we do not go high we are not going to get nice spikes. Low resistance means the trigger is on, high means it is switched off.

Here are the LED’s without the cover to show the ramping up of the LED. Remember we don’t want a really fast pulse because it doesn’t allow the Ohms to rise high enough. If the photo resistors could switch at the speed of light it would be perfect but they cannot. They have a rise/fall time that we need to work with.

Trigger Ramp


Here is a video of it while the machine is on (charging batteries). Watch the amp draw. When it is low that is the pulse circuit charging the cap, the main SS is not drawing anything at that point. Then it climbs while the dim period begins,, then the flash where we pull +1 amp. Then the dark period and it begins again.


Trigger Running


So there you have it. I am still experimenting with small changes but so far this method seems to be really good. The huge bonus of this is that you can set your main trigger pot at whatever you like for your particular load and then control the trigger pulse independently of that. I was triggering pretty high draws (1 – 2 Amps) but what I am saying is you could turn your main pot down so that it can only draw say ½ an amp if that is what you prefer and it has nothing to do with the pulse.

I will do some runs and post back later. I just wanted to share this with the group.

By the way I did try this with photo diodes and photo transistors but they didn’t work as nicely as this. I cannot say with full confidence that this is the way to go but this is a group for experimenter’s right, so that is what I am doing and I am happy to share ideas.
---Bob

Hi Bob,
I like your use of the negistor here!
Have you tried using a cap dump as John K suggested to monitor the charge rate? That move w/ the photoresistors is innovative.
Nice work,
Patrick A.

[BobZilla]

Hi Patrick,
I'm not sure what you mean, couldn't find what John said. Was it in another thread by chance?

I am thinking that I may be able to use this same method as a cap dumper instead of a trigger pulse. I have not tried it yet but I don't see why we couldn't place the photo resistors between a cap and the load and have a cap dumper. It would have to be gauged correctly so that we don't build up to much juice before dumping but it's worth giving a try. I have some more photo resistors on the way and I got a variety of sizes so I will tinker with it and let you know how that goes. ---Bob

[min2oly]

Hi Patrick,
I'm not sure what you mean, couldn't find what John said. Was it in another thread by chance?

I am thinking that I may be able to use this same method as a cap dumper instead of a trigger pulse. I have not tried it yet but I don't see why we couldn't place the photo resistors between a cap and the load and have a cap dumper. It would have to be gauged correctly so that we don't build up to much juice before dumping but it's worth giving a try. I have some more photo resistors on the way and I got a variety of sizes so I will tinker with it and let you know how that goes. ---Bob

Thread: Solid state SSG

excerpt:

"My idea is to basically hook up a cap pulser to the output and measure the time it takes to charge the cap against the draw current. For example, if it takes 1 second to charge a cap at a draw of 1A that would calculate out to an "X" of 1.0" - John Koorn (post #34)

Wouldn't it be nice if there was a chip that's sole purpose was to "cap-dump", and we could set the top and bottom w/ a pot or dip switch. Seems simple enough.

I hope the weekend is finding you all well and in good health.
Patrick A.

[Tom C]

Thread: Solid state SSG

excerpt:

"My idea is to basically hook up a cap pulser to the output and measure the time it takes to charge the cap against the draw current. For example, if it takes 1 second to charge a cap at a draw of 1A that would calculate out to an "X" of 1.0" - John Koorn (post #34)

Wouldn't it be nice if there was a chip that's sole purpose was to "cap-dump", and we could set the top and bottom w/ a pot or dip switch. Seems simple enough.

I hope the weekend is finding you all well and in good health.
Patrick A.

yea like the comparator does

Tom C

[BobZilla]

The experiment continues….

So I had done a side experiment using the photo resistors in a cap dumping circuit to see how well that may work. The results were pretty good so I decided to put what I learned to use on this system. There is another thread about the cap dump experiment if anyone is interested in that.

The cap dump I built for the experiment was designed a bit differently but the principal idea of using the photo resistors is the same. That circuit was meant to be a clip on unit that could be added to any existing radiant charger but for this addition to my SS I built in in permanently to the circuit. I am driving these dumpers with the same circuit I used on the trigger pulse, built with the 2n2222. There is a little energy cost on the front side to run these but the system is designed for larger batteries so anything I am using as a primary really shouldn’t get dragged down too much by a few LED’s, it’s just the cost of doing business with this design. I estimate about 20ma but they are flashing all the time so it’s difficult to get a good reading.

Ok so I will not go over the whole idea again but the specifics of the added circuitry are as follows. There are two cap-dumps each is using four photo resistors in parallel to bring the full on resistance down low enough and also to distribute the current between them to avoid over heating. I have two of the same LED’s as before to activate them.

Here is a shot of the chamber before I installed it. I used a plastic sauce cup and lined it with foil tape.



Each leg of the machine is charging up 3000uf(2x 1500) 200V caps. I left leads on the top for clipping in meters and I could also expand the capacity by clipping in external parallel banks. For now I am just running it with what I put onboard. There is a 1n007 diode coming off the positive towards the battery and the photo resistors are in-line on the negative side. This allows the voltage to build up in the caps and then release as the LED’s flash.

I have independent pots for each cap-dump so that I can adjust the frequency of the LED’s . Looking at the system as a whole I have a lot of adjustability in this thing now.


There are four pots in total which control the following.

1st the main pot on the motherboard. This adjusts the current amplitude.

2nd the trigger frequency pot. This adjusts the flash rate for the trigger pulse.

3rd is cap 1 frequency. This controls the flash rate of cap dump1. It allows for adjusting the dump rate so that voltage can be fine tuned on the dump.

4th is cap dump 2 frequency. Same as cap dump1

With so many adjustable points this system can be run in many ways and I am just now starting to get a feel for what it likes. All of the flashers can go from around 1 pulse per second all the way up to very fast like a strobe light. I can over drive the cap dumps to where the LED’s stay on brightly if I want but that was more of an accident, still it does run that way if I want it too.

On the machine itself I did parallel the power coils 2 per transistor so we have 4 power coils (200ft each 18AWG) total but distributed in pairs to two transistors. The output is still in branch mode to the segregated cap dumps.


I have been testing with small batteries to get a feel for how to adjust this thing and it charges little 17AH, 5AH really well. Next I am going to make some runs on the big batteries and I will post that. I will run a two for one just as the previous runs in this thread and we can compare.


Here is a shot of the board with all of the modifications.




Thanks-----Bob

[Nofear]

What's wrong with this picture

C-E 5800ohm.pdfC-E 100ohm.pdfB-E 100ohm.pdfB-E 5800ohm.pdf

Hi,
I am building SSG.
~ 130ft (8x23awg + 1x26awg)
- matched MJL21194, 100ohm base resistance, 1N4007's
-5K pot +100ohm resistance // 10uf,250V....my 5k pot is more like 5.7 k

Now comes the tuning: I am shooting for that elusive square wave. I am following Patrick's steps but I don't even see the transistor ON time, OFF time.
Attached are my osci screen shots.
I am measuring (C to E) and (B to E) when:
1) the pot is fully open i.e total resistance=5.8Kohm
2) when the pot is closed i.e total resistance= 100 ohm

Please look at my screen shots and tell me what I am doing wrong

Thank you

NoFear.

[Nofear]

Oops the rest of the pics

[why_me]

Battery swapper

Here is a circuit you could use for your design. It should work for any Bedini circuit using batteries that use radiant energy charging.

[Nofear]

Hey Patrick,
I was hoping your answer my post. So that's great.

So yes it is air core and I have a diodes emitter to base of course. And one on the trigger circuit following the pot/cap combo, leading up to all the trans bases. Make sense?? my 100ohm is 1/2W I beleive. They are not hot to the touch nore are my transistor, never above ambient temperature.
One more thing I didn't mention that could be relevant. I drive my circuit off a power supply...some time in parallel with a battery.


So you believe that my trigger current is low hence adding the 26awg wire. What would increasing the base resistance from 100 to 200 or 300 do, reduce the base current? voltage? I'll try everything you suggested and I'll try to rationalize it later
Where would you put the Osci pickups to measure the transistor on time/off time????
Thanks for your continuous support.
NoFear