Use for the Tesla Switch

[ldissing]

Bit's-n-Bytes, Leroy,
I think what we must do is replace the SG1524 with a Pic and just have the pic do some analog functions with a pot at the ADCIN pin by turning the pot we should be able to adjust the switching speed. I think we can go from 1Hz to 5Khz and get the correct time. I know that most will not be able to program the chip so we must make it very simple to do. I also notice these chips have a de-bug pin where we can watch the program run. I'm not sure it can be done that simple but I'm willing to give it a try with you guys if you want to do that. It also looks like we could build a circuit to charge a capacitor to run the chip at the same time. If I'm correct nothing over 5 volts on any of the pins. Let me know if you guys have the spare time to do this, as I know we all must work.
John B

Reply to me in private your requirements, and I'll write a program to do what you need. I'll research the chip in the morning if want to use this particular chip and be ready for whatever you want. Just let me know so I know what chip to research. This will be a fun project for me. I am not working, so I have plenty of time to devote.

Leroy

[nvisser]

Bit's-n-Bytes, Leroy,
It also looks like we could build a circuit to charge a capacitor to run the chip at the same time. John B

Is that the reason for that large 27000uf cap on your TS photo?

[nvisser]

From what I have seen on a Etmega pic used by Bob Boyce in his PWM you have to switch off and connect to the pc before you can adjust frequency or duty cycle
It will be good if the pic used can use a pot to adjust parameters as John said.
They do draw much less current than the SG chip and run from 5V
I am currently busy laying out a pc board for the scaler charger posted by John. Unfortunately I only got Pcb express. But te latest version is printable. I have to go and widen all the pads with a thicker track otherwise it prints to narrow
I will see today if my TL494 circuit can do the lower frequencies properly and than decide what to use for the flipflop as it sounds like the newer Sg chips does not work properly in the lower region?
I noticed on the original scaler wave controler the supply comes through a diode to a large capacitor from the main battery.

[nvisser]

Interesting results

I ran John`s 3 battery test circuit yesterday for 5 Hours. I swopped the batteries 3 times. I used 12v 7Ah sealed lead acid batteries, made in Taiwan in a hurry I think
The resting battery voltages before the test were :
Batt 1: 13.10v
Batt 2: 13.02v
Batt 3: 12.16v Sum of 38.28v
I then disconnected everything and this morning checked the resting voltages again
Batt 1: 13.12v
Batt 2: 13.16v
Batt 3: 12.30v Sum of 38.58v

Now who would have thought!!

[morpher44]

I can recommend the PIC32

Bit's-n-Bytes,
I have a PIC 18F2423 which I'm going to try and use. I do have the programmer for this type do you think this one will work the same.
I'm not the worlds best programmer but I'm going to give it a try. This PIC has more functions I think, may need some help on this one. Forgive me I'm just an analog guy.
John B

I can highly recommend the PIC32. Very cheap. 32-bit CPU.
Just hook it up to your laptop's USB ... no programmer necessary.
You can write code in assembly or C -- easy as pie.
It has I/O up the wazzu.
I would drive the optos with the PIC to isolate them from
any thing back EMFffy ...
To avoid any thing happening to your laptop, you can basically
write the flash on the PIC32 ... unplug from your laptop ...
plug into your circuit under test -- with expansion I/O board --
and you are completely embedded at that point with no
connection back to your precious laptop.

To do anything intelligent ... other than just a timing trick, you
might need to utilize sensors that can sense voltage, current
and or power. They sell these now too.
But certainly you can just mess around with pulsing the optos
at various duty cycles under programmatic control.
I'm very pleased with the PIC32.

It would be perfect for the Tesla Switch project ... for those of you
considering a microcontroller.

Meanwhile, there seemed to be some confusion about the
switch ... and the timing of it.
If I'm not mistaken, you want to switch rapidly back and forth
to create -- essentially AC (or pulsed DC) through your load.
Think Tesla here.
You do NOT want to switch it one way and run the load
on a set of batteries until they drain down for several hours ... and then
switch it the other way ... etc.
The magic is in the pulsing to fool ALL batteries to keep charging ..
even though a load is present. ... but also to have essentially AC
to your load (like inverter technology).
As for frequency to switch ... you have to find that "sweet spot"
which will vary with load and batteries .. and circuit losses, etc.
Do I have that correct John?

re: Pot
Think digital. Pots are old school.
They have devices that look like pots called "digital pots".
They give the CPU a "he/she is turning it left now" and
"he/she is turning it right now" input. I forget what the
are called. OR ... you can take input from a serial
port to command the PIC to do whatever you want.
So your pot is essentially a command you type on
a dumb terminal connected via serial port.
OR ... there are ways to do that with push buttons ...
one to go "left" say ... and another to go "right" ...
or "up" or "down". Etc.

URL to some sensors for PIC:
Microchip - Analog & Interface Products

[John_Bedini]

Tesla Switch

Leroy,
I think you can go to a low frequency on that chip, I have one also. I can use some other crystal controlled osc and then use other chips that divide the frequencies up and then use a jk flip flop but this is really the hard way. I could not answer because I'm getting my snow plow ready as we are going to get hammered tonight. But the PIC I have to work with is the 18F2423 I can get this chip to run real low frequency in a basic program like Bit's did. And yes the 27000Uf cap was an attempt to have it self charge but I goofed up and did not ground it right so I disabled it and have run the oscillator separate for now. I use the program Aries as I can print my circuit boards from my printer and then etch them.
I think you can get Aries light for free it looks through the board.
I'm going to think about this problem tonight before I waste everybody's time.
John B

Is that the reason for that large 27000uf cap on your TS photo?

[nvisser]

I can highly recommend the PIC32. Very cheap. 32-bit CPU.
Just hook it up to your laptop's USB ... no programmer necessary.

You surely need some sort of pc board interface with a ic socket that plugs into your usb?

[morpher44]

microcontrollers/Tesla Switch continued...

You surely need some sort of pc board interface with a ic socket that plugs into your usb?

The PIC32 "starter kit" board is the board ... and there is just a
usb cable. I think the sucker cost around $50 USD or so.
Cheap!
ww1.microchip.com/downloads/en/DeviceDoc/PIC32_UsersGuide.pdf
Product Search - microchipDIRECT

There are 3 user LEDs that could be wired directly over to opto.
There is more I/O for even more optos.

You can then also buy "optionally" two other useful boards.
A Base board giving you more slots ... and an I/O proto board giving
you a prototype area. You can buy lots of these proto boards
and mess around with all sorts of circuits to plug into your base board.
Its a slick packaging ... and the cost is low -- which I like ...
and its C code.

Not to get too vendor specific here, though, there are some
other amazing microcontroller packages out there. The prices
have really gone down & functionality is amazing.
Many people like the STAMP. Especially Robotics crowd.
If you like Verlog/VHDL ... and hard core digital electronics ..
go FPGA board from Xilinx or Altera.
FPGA offers high-speed and a platform for "ultimate" hardware
prototype ... affording a path to ASIC development for mass production.
I've played with some Xilinx tools and parts.
As a programmer though, I prefer these easy to use microcontrollers.
Must plug them in via USB ... and your in business.
I wish I had something like that when I was in college.

Read about the "Energy Measurement" technology on
Microchip's website. Also under "Power Management",
they have Voltage detectors, battery management, PWM ..
and so other cool stuff. This is right up the Tesla Switch alley.
Ok ... enough on that. Sorry to interrupt.
PS> I am a hobbiest ... not in sales ... forgive me.

[rave154]

MORPHER, is this the product to which youre referring ( from a european source ...

Microchip | Semiconductors | Evaluation Kits And Development Boards | Microcontrollers & Peripherals | PIC32 Microcontrollers |DM320003

if so, for only £33 i might get one to have a play with. ( though i HATE and DETEST C, toooo many curly brackets !!!....give me Delphi anyday :-)


Thanks,

David. D

[lamare]

Hi John
Anyway, this circuit idea is fascinating ... I was able to run
motors, incandescent, and various low-resistive loads --
after using full-wave bridge rectifier & cap to convert back to DC.
But I found that when tuning is wrong, ALL your batteries go dead
pretty rapidly. So "tuning" is crucial ... and really what is needed
is something whereby the "tuning" could adapt and be smarter.
So I thinking a microcontroller might be the best way to
fire your "optos". That way a cute little software algorithm
could "tune" and "adapt" ... Some sensors would be needed
to measure currents and voltages.

Given that you talk about low-resistive loads and what John stated earlier on, it appears the most important criterium for being able to charge a battery (or: keep the batteries charged) is the voltage that is left over for charging.

If that is the case, then "measuring" the voltage over the cap at the rectifier should be the most important. If you would use two sets of zener diodes in series with an opto coupler, you could easily get a "high" and "low" indicator.
When the cap is charged -- "high" -- switch all transistors of.
When the cap is dischaged -- "low" -- switch one set of transistors on.

That way, the circuit could autimagically adapt to the load and keep your batteries charged, because no power is drawn from them when the cap has sufficient charge to power the load.

[nilrehob]

John B,

I'm really trying to understand the circuit, and I hope that You don't mind my questions.

I cant' figure out why not do the circuit as proposed by me, redeagle and others with six switches and no diodes:
http://www.energeticforum.com/renewa...html#post65249

So i started writing out the potentials on the two different circuits (assuming 1V across collector emitter drops and diode junctions for simplicity).

What i found out is that in Your circuit (to the left) the batteries in parallel have the same potentials, both have negative on 8V,
while on the no-diodes-circuit (to the right) the batteries in parallel have different potentials, one neg on 10V and one neg on 11V.

Is this the reason for not going for the no-diodes-circuit?

/Hob

[Matthew Jones]

@Hob

Without the diodes the charge does not spread evenly, especially with 4 batteries.
redeagle's circuit will load one battery per side alot higher the its partner battery. I am not sure of the outcome but most likely it will lower the over all performance in some manner.

Matt

[nvisser]

I made a pcb layout of the Scaler Charger on Pcb Express using a TL494 as flipflop. I havent build it yet. Anybody that wants a copy can mail me. You will have to download pcb express to use it.
I cannot upload that type of file on this forum.

[Bit's-n-Bytes]

Here is the first of many building blocks

Team, here is a first attempt to illustrate the code for the oscillator using a PICAXE 18X chip. I have also included a pinout drawing. Not sure if this is going to do what you need it to John, but it is a start none the less. Code and drawing are only theoretical at this stage. Optical couplers and hall effects are not shown.

Thanks

Bit's
Start:
SYMBOL TrimPot = W3 '########## Reserves buffer for User adjustment
SYMBOL Amp = W1 '########## Reserved for Hall Effect
SYMBOL Volt1 = W0 '########## Reserves buffer for D9, D10
SYMBOL Volt2 = W2 '########## Reserves buffer for D1, D5

Main:
readadc10 6, Volt1 '##### Reads voltage at pin 15 an assign value to Volt1
readadc10 0, Volt2 '#####Reads voltage at pin 17 an assign value to Volt2
readadc10 7, TrimPot '#### Reads voltage at pin 16 an assign value to TrimPot
'readadc 2, Amp
'if Amp < 126 then '########## Hall effect used can read + or -
'let Amp = 127 '########## any (-) value make zero
'endif

if Volt1 > Volt2 then SwitchGroup1
if Volt2 > Volt1 then SwitchGroup2
goto main

' ########## Oscillates S1, S3, & S6

SwitchGroup1:
pulsout 0,10000 ' send pulses out of pin 6
pause 20 ' pause 20 ms
goto Main

' ########## Oscillates S2, S4, & S5*

SwitchGroup2:
pulsout 7,10000 ' send pulses out of pin 13
pause 20 ' pause 20 ms
goto Main

'################# Need Help here to handle TrimPot setting and it's
'################# affect on the Oscillation
'DO
'{code}
'LOOP UNTIL/WHILE VAR ?? COND

[nvisser]

On R Brandt circuit what is meant by:
"D9 D10 each one must be equal to D1 D5 together"
I see on the pic diagram that you got sensing inputs for this diodes.
Can somebody please explain to me what this is for