Use for the Tesla Switch

[John_K]

3PDT Relay Tesla Switch

Looks like you are gaining on it. Great job!

Bit's

Thanks Bit's,

The last 12 hours didn't do much. All four batteries have been at the same voltage, give or take 0.01V.

I've increased the frequency from 2Hz to 20Hz to see what happens. The load is the same.

Still trying...

John K.

[StevanC]

494 sub 1Hz hacked! ("Yes my children it can be done&quot

Scalar charger like This one only 494 brains and few minor diffs.

I had 3 attempts, two "leave it rest" days and finally got it running
I had so far:
330 uF pair of caps (200V rated)
10000 uF pair (50V)
75000 uF pair (30V)

I finally used no load ("dead short")

On open circuit I had about 5.4V ((battery rests about 5V it's a 6 cell "12V" one)
My finds so far:

1. Dead time is mandatory
2. 494 can be hacked to go well below 6Hz (10k =6Hz 230k still works DMM can't read out, looks like ~0.3Hz)
3. The really interesting stuff happens when I put a really bad battery and two 330uF caps on the ultra low duty cycle:

A. The cap gets to fully charge/discharge on each cycle (10000 uF ones won't )
B. The dead time pause is obvious on the battery voltage.
C. The battery has non equal resting voltages (!) depending if it's been pulled or pushed with the pulser - peukert? - Anyhow it's thought provoking to me
D. the gain seems to be gotten in the dead time - the more dead time the more voltage gain - speed does not increase charging?
E. I left the bad battery (flooded scooter cranking 7Ah few yr. old crap ) ) exercise...

will get back with the 494 "hack" schematic

[John_K]

3PDT Relay Tesla Switch - Faster seems better

Thanks Bit's,

The last 12 hours didn't do much. All four batteries have been at the same voltage, give or take 0.01V.

I've increased the frequency from 2Hz to 20Hz to see what happens. The load is the same.

Still trying...

John K.

Running at 20Hz has made a difference:

Start
B1 12.22
B2 12.15
B3 10.81
B4 10.81

+9 hours
B1 12.20
B2 12.16
B3 10.89
B4 10.89

Slow I know, but I'll let it keep going until nothing happens.

John K.

[blackchisel97]

SC

Hi guys,

I have left my SC running overnight and after 12hrs battery remains the same. DMM is trying to climb another 0.01V once in a while which would be 0.01 over starting voltage. It seems that dead time is needed, I agree with Stevan. I tried slow 2Hz and 20Hz switching and didn't notice any difference.I can't get SG to switch slower than 1.5Hz. When they started to make those chips, they were done for military applications (just like any semi) and would operate well below subzero and over 100 deg. After a while, those made for non military purpose would work only between 0 - 70 deg. Quite a difference and probably not the only one.
Just replaced TIP41 with 2N2222, which have ability to go negative, like MJL's. At least battery isn't loosing anymore despite powering 2 bright LED's for the past 18hrs. My cap banks are 480uF photo. I'll keep tweaking with it until battery will start to climb or I drop.

Vtech

[Bit's-n-Bytes]

Update on the D-TS

Ok team, I am able to charge one way, i.e Batts1 and 3 charging Batts2 and 4. I have now changed SCR 10 (formerly D10) to a N-FET triggered by Q1 opto. I have removed SCR 9 totaly. You guy's were right. There is a direct short on this path and even when I put a N-FET in and trigger it by Q2 there is enough delay in it shutting off that it heats up to much. The SCR's as they always will, remain triggered once there is current flow. From this point, we are 2/3rds the way to victory. When we solve for the issue with D9, this thing will charge like crazy. I have gone from 11.5 to 11.9 in 20min. All componets are running cool. The battery charge light and load light are 6volt 250ma bulbs padded ea by a 100 Ohm resistor. Here is the pulse and duty cycle rates;

high 6
pause 300
low 6
pause 400
High 4
pause 300
low 4

Thoughts on what we can use for at SCR 9 (aka D9)? Solid state that is.

Thanks

Bit's

[Plazma]

Base Capacitance?

Hi Bits,

Sounds like you are REALLY close

I have maybe a wierd question - could it be possible that the combined capacitance of the base of SCR10 plus that from Q1 is causing some first order lag sufficient to hurt the sharp switching one wants? I know the SCR is prone to not shutting off when current flows. If things are bouncing around in both battery sets when Q1/SCR10 try to switch, maybe they are fighting one another. Maybe the SCR wants its own opto? This is really lame - but you did ask for a solid state possibility

Happy Hunting,

Plazma

Ok team, I am able to charge one way, i.e Batts1 and 3 charging Batts2 and 4. I have now changed SCR 10 (formerly D10) to a N-FET triggered by Q1 opto. I have removed SCR 9 totaly. You guy's were right. There is a direct short on this path and even when I put a N-FET in and trigger it by Q2 there is enough delay in it shutting off that it heats up to much. The SCR's as they always will, remain triggered once there is current flow. From this point, we are 2/3rds the way to victory. When we solve for the issue with D9, this thing will charge like crazy. I have gone from 11.5 to 11.9 in 20min. All componets are running cool. The battery charge light and load light are 6volt 250ma bulbs padded ea by a 100 Ohm resistor. Here is the pulse and duty cycle rates;

high 6
pause 300
low 6
pause 400
High 4
pause 300
low 4

Thoughts on what we can use for at SCR 9 (aka D9)? Solid state that is.

Thanks

Bit's

[Plazma]

Time Delay

@All,

As a follow on to my previous post, I wonder what,
if any, time delay might be necessary in the switching
sequence between say Q1 and SCR10. For example:

1) Q1 switches establishing a pathway for
24 volts to develop;

2) a short time delay allows that ~24volts to
fully develop and stabiize;

3) then SCR10 is switched allowing the full
potential across the batteries to be charged.

I hope this is NOT what is required as it could
require 4 IO pins instead of 2. Also, JB has not,
so far as I know, talked about such a complication.

What this approach does is eliminate interactions
between Q1 and SCR10 as to switching sequence,
if any. Q1 and scr10 would have separate optos.
Same for the other side.

Bits, et al - don't get me wrong - I sure don't want
any more complications - this is just a stab at breaking
a problem apart piece by piece to get at some dynamics
that might be hidden. Please take it with more than
a few grains of salt . . .

Best,

Plazma

[citfta]

SG3524n solution (I think)

Hi Gang,


Thanks for all the suggestions about solving the problem I was having with the SG3524n. I have tried most of them without much success. I think it was Vtech who filled in the last piece of the puzzle. He said he had problems if he tried to use 2 470nf caps together for the timing capacitor. He said it worked ok with one. I had already determined that the problem was the internal flip-flop not toggling. I looked again at the diagram of the internal circuit of the SG3524n that John B. had posted in post 551 back on page 19 of this thread. There is a small cap that feeds a pulse from the oscillator circuit to the flip-flop circuit. It appears that if the capacitor is too large the oscillator circuit cannot discharge the cap fast enough to send the proper pulse to the flip-flop circuit. The problem is made worse if the cap is an electrolytic which is why several of you have had better luck with other types of caps. The problem is also made worse by lower temperatures. My room where I do my experimenting is not heated except when I am in here. I have even gotten the circuit to work temporarily by just putting my finger on the chip and letting it warm up a little. My solution was to put a monolithic cap of 1 mf in as the timing cap and raise the resistance on pin seven by using a 500k pot instead of the 100k John recommended. This has given me a good range of adjustment and the circuit has been running all day in a very cool room. It hasn't warmed up here today.


@Bits, How about just using a couple of more MJL21194 transistors to replace the original D9 and D10? I realize you will have to add a couple of more optos but I think it will be the most reliable why to get things to work.


best regards, Carroll

[Bit's-n-Bytes]

Hi Gang,


Thanks for all the suggestions about solving the problem I was having with the SG3524n. I have tried most of them without much success. I think it was Vtech who filled in the last piece of the puzzle. He said he had problems if he tried to use 2 470nf caps together for the timing capacitor. He said it worked ok with one. I had already determined that the problem was the internal flip-flop not toggling. I looked again at the diagram of the internal circuit of the SG3524n that John B. had posted in post 551 back on page 19 of this thread. There is a small cap that feeds a pulse from the oscillator circuit to the flip-flop circuit. It appears that if the capacitor is too large the oscillator circuit cannot discharge the cap fast enough to send the proper pulse to the flip-flop circuit. The problem is made worse if the cap is an electrolytic which is why several of you have had better luck with other types of caps. The problem is also made worse by lower temperatures. My room where I do my experimenting is not heated except when I am in here. I have even gotten the circuit to work temporarily by just putting my finger on the chip and letting it warm up a little. My solution was to put a monolithic cap of 1 mf in as the timing cap and raise the resistance on pin seven by using a 500k pot instead of the 100k John recommended. This has given me a good range of adjustment and the circuit has been running all day in a very cool room. It hasn't warmed up here today.


@Bits, How about just using a couple of more MJL21194 transistors to replace the original D9 and D10? I realize you will have to add a couple of more optos but I think it will be the most reliable why to get things to work.


best regards, Carroll

Thought about that and I am sure that would work, but just trying to save a buck you know. We are so close to victory. This is a great place.

Thanks

Bit's

[Bit's-n-Bytes]

Thought about that and I am sure that would work, but just trying to save a buck you know. We are so close to victory. This is a great place.

Thanks

Bit's

Ok gang, just reveresed bias the N-FET if you will for SCR9 (aka D9) and shown a positive rate of climb (sorry, I am a pilot). All chips are cool. Batt1 and 3 are showing positive gain now. This is cool, but time will tell.

Bit's

[blackchisel97]

Hi Gang,


Thanks for all the suggestions about solving the problem I was having with the SG3524n. I have tried most of them without much success. I think it was Vtech who filled in the last piece of the puzzle. He said he had problems if he tried to use 2 470nf caps together for the timing capacitor. He said it worked ok with one. I had already determined that the problem was the internal flip-flop not toggling. I looked again at the diagram of the internal circuit of the SG3524n that John B. had posted in post 551 back on page 19 of this thread. There is a small cap that feeds a pulse from the oscillator circuit to the flip-flop circuit. It appears that if the capacitor is too large the oscillator circuit cannot discharge the cap fast enough to send the proper pulse to the flip-flop circuit. The problem is made worse if the cap is an electrolytic which is why several of you have had better luck with other types of caps. The problem is also made worse by lower temperatures. My room where I do my experimenting is not heated except when I am in here. I have even gotten the circuit to work temporarily by just putting my finger on the chip and letting it warm up a little. My solution was to put a monolithic cap of 1 mf in as the timing cap and raise the resistance on pin seven by using a 500k pot instead of the 100k John recommended. This has given me a good range of adjustment and the circuit has been running all day in a very cool room. It hasn't warmed up here today.


@Bits, How about just using a couple of more MJL21194 transistors to replace the original D9 and D10? I realize you will have to add a couple of more optos but I think it will be the most reliable why to get things to work.


best regards, Carroll

Hi Carroll, I think you have a very good idea about SG chip. I looked at the guts and it makes sense. I'm digging in my parts for bigger pot to try it. My circuit runs in warm (72F) temperature.
I had two tranny's in place of D9 and D10 when I first built TS. I didn't tuned and jumped forward too fast with serial SCR's (triac's with serial diode in my case) but circuit was working.
My SC is running steady @ 2Hz since yesterday (18hrs). 5Ah battery remains steady, no gain, no loss while powering 2 LED's. Gotta try lower freq.

@Bits

Vtech

[Matthew Jones]

StevanC posted. 1. Dead time is mandatory...

I have had a few good runs on the SCR. What I am looking at is the turn around time from the battery to the cap and back again.
I am trying to get an old laptop going so I can watch it real time but.
I have been charting it some off my meters memory.

The caps take 3 ms to charge fully from the battery. They take 8 to discharge fully back into the battery. Thats one set anyway.
Anyway. I have been running it like this then going to a pause of about 250 ms, Then starting over.

Your right a dead time is needed. If you try to pull a load from the battery before the replaced charge absorbs (or something) the energy is gone. Every time.
The caps though don't seem to care. It has it advantages from what I can tell to turn the energy around, once in the caps... immediately.

All the caps I have tried The lower the UF the better result.

I am trying to chart my turnaround time like I said. I'll post back when I have some media to show.

Cheers
Matt

[nvisser]

Smoke!!

I build the latest bedini circuit that was posted .
I blew 30 diodes and 2 scr`s!! Luckily I use cheap 1n4007`s.
The first time when some of the battery leads shorted while conecting the batteries
The second time when I conected one of the bateries the wrong way round.
And the last time when I switched the flipflop on and all short out as predicted
I made sure the duty cycle was on minimum beforehand at about 1 Hz. Only a spike but it is obvious enough to short the batteries.
I will try again tommorow with transistors as the diagram stated instead of scr`s.
If it happens again, I will have to put the 4 parralel transistors into the circuit.

[Plazma]

I surrender . . .

Hi nvisser -

Bummer - 30 diodes? Ugh!

Well, I haven't fried that many but my stack is growing.

I had to put in safety switches to isolate and help
manage battery leads - that helped. I'm using a
beefier diode than the 1n4007 (more expensive, too) but
I have the short across the diodes D9,D10 as mentioned.

You've convinced me - back to 4 trannies - I surrender!

Thanks for sharing, better days are coming -

Plazma

I build the latest bedini circuit that was posted .
I blew 30 diodes and 2 scr`s!! Luckily I use cheap 1n4007`s.
The first time when some of the battery leads shorted while conecting the batteries
The second time when I conected one of the bateries the wrong way round.
And the last time when I switched the flipflop on and all short out as predicted
I made sure the duty cycle was on minimum beforehand at about 1 Hz. Only a spike but it is obvious enough to short the batteries.
I will try again tommorow with transistors as the diagram stated instead of scr`s.
If it happens again, I will have to put the 4 parralel transistors into the circuit.

[nvisser]

Diode Killer

This is what it looks like. Not very neat as I just soldered the new diodes in everytime with long legs