Computer Control of a PWM design. work in progress.

[z.monkey]

Howdy Feynman,

So I guess that you are a printed circuit board designer also?  I have seen you mention the LTC6904 a couple of times, once in a TPU thread.  Does it have a shutdown pin?  I don't think that we need a continuous frequency, but then again were still experimenting.  If that part had a shutdown pin then we could give the electrolyzer bursts of whatever frequency.

A lot of these HHO setups just hookup the electrolyzer to DC power, that's it.  But they have problems.  The water gets ionized and starts to conduct current, then they get a current runaway situation and start blowing fuses.

Pulsed DC, like the Dave Lawton circuit, is basically switching DC on and off.  That circuit is more efficient, you don't get the current runaway, but still uses a lot of current.

Then there is the high voltage circuit, like Stanley Meyers circuit, where the pulsed DC is hooked up to an ignition coil and the high voltage pulses go to the electrolyzer, which uses less current.

Then there is the curious modified alternator setup where they take the rectifier diodes out of an alternator and feed the alternating current to the electrolyzer.

It seems we have some good electronic experience in this thread.  Of all the examples above pick out the qualities that are best and then design a circuit which has those qualities.  For starters we would want something that is programmable and adjustable for experimenting.  Then also we may want to have some amplifiers that can really kick the voltage up.  Having only 12 volts in the vehicle we would also need to have a way of generating high voltage like an inverter or a DC to DC converter.  If we need really high voltage we could always use a flyback coil or an ignition coil.

Out of all the HHO driver setups the one that makes the most sense to me is the modified alternator method.  This way you are alternating the voltages on the plates of the electrolyzer.  I am an electrical engineer, and I have experience designing audio frequency power amplifiers.  Maybe we don't need a lot of power so much as the right frequency and wave shape.  Did anyone ever see the movie Chain Reaction with Keanu Reeves?  I have a design for frequency synthesizer and biamped power amplifier that is almost perfect to do this.  If I kicked up the power supply to its limit (+/-18V) we may have something here.

OK, Mo Later...

[Feynman]

Nope, the LTC6904 has no 'shutdown' pin unfortunately. 

But To control the ON/OFF (PWM), I suspect you can just mix the output signal with a fast AND gate.   I also agree one needs pulsed width 'bunches'. (ala Meyers).

AND gate



I will reply more in a few hours

[Feynman]

Okay, I will try to elaborate as well as answer your questions z.monkey. Now, let's pretend for a minute that we are only have square waves of 50% duty cycle.  . . that's what the LTC6904 can do.  So we only have these two square wave oscillators, and we can assign them each a programmable frequency, anywhere from 1khz to 67Mhz. 


As I understand from Meyer, what he is doing is making "gated pulse trains".  I thought about how to make these.   We can accomplish this roughly with two square wave oscillators. Let's examine the image from Wikipedia entry, which comes from Stanley Meyer's water fuel cell patent.



Look closely at the part where you see the Gated Pulse Train.   Basically what we have is an underlying square wave which has parts cut out of it. So that means we can describe this with two waves.

a carrier wave:  This is the higher frequency of the two waves.  It appears to be 50% duty cycle.

a modulation wave:  This is the lower frequency of the two waves. From Meyer's patent diagram, we can see the modulation wave appears to be 2/5 the frequency of the carrier wave.  Additionally, the modulation wave has a duty cycle of 60%.


But let's put the pulse width details aside for a moment, and get back to pure square waves.  We can make these gated pulse trains synthetically, using digital logic.  It can be done simply with two programmable oscillators coupled with a high-speed AND gate.


(this is just an example, it's not same frequencies as Meyer)

The carrier wave is the higher frequency source wave. The PWM wave is the modulation wave.  They are being combined via a simple AND gate.   So basically, with two frequencies going into an AND gate and replicate the 'pulse train' bunches.
Now, we can get very close with square waves, but ideally we want that 60% duty cycle for the modulation wave, if that's what Meyer was using. 

Meyer electrolysis:

carrier wave: 42khz
modulation wave: 16.8khz     (2/5 of carrier).

These frequencies are easily achievable by the LTC6904 programmable oscillator.   

[Lanmasterd]

WOW nice responces were deffinatly hitting the rite notes here lads..

Ok firstly the arduino nice tool it is.. i actually dont have one yet but my usb to uart board design with Atmege88 28pin and 5v deditated supply will do me its basiclly the same just more raw..ive learned from the past that i test and do better with prototypes .ie build a bit perfect it and move on.. so im working on this like from scratch so a board with pc controler and regulater power .. Assuming we only have a 12v 7Ah battery to use,,
yep you heard me..

12v 7Ah  you now must be thinking im nuts.. but wait.. there reason behind my nutty {LOW} battery power ..

Firstly the idea will be like as in a slightly older deisel car u turn key it heats up then starts ... the same thing ill want on my Test MotorBike.. i get on .. flick switch that start up my {reactor} and once a head of pressure is detected by the pcd control system it will allow me start the engine..

who needs to be rushing to get started like a petrol.. im looking at about 30sec to a min prestart time.. and like wise when i power down i want a safty feature so ill use a pressure bleeder into the exhuast system,, nice and simple..

my cell design will be simple yet effeective high reaction area .. still looking for plate design to do this.. effectivly.. the cell is importent dont get me wrong but control is vital..

Nice work so far on the java setup  id be interested in knowing more.. jave might be handy i was thinking my self that the atmel atmega88 and 168 both have good lcd ability which i was gonna include for monitoring..

the unit im building for all the control and monitor is so far 6.5 inch by 5 inch and thickness well not sure yet depends on some componts

the resonant charging choke.. i dont think we will need .. im leaning more towards Very High Voltage + very tight plate Packing + PURE WATER [OR RAIN WATER] IF U ARE STUCK TO RUN THE CELL..


the high voltage i got an idea last nite when i was brushing my dog lol yeah odd i know she is a german shepard and i use a Ion brush to keep her coat good.. and me being me i put the brush in my back pocket and forgot it was there.. anyway i jumped into the armchair to hear a crack and smash lol.. low and behold i saw a nice inside view of my dog brush .. and wat did i find.. a battery 2 aa's and a switch and a simple Hv circuit.. steping it up to 600v low amps by means of a sinply invertor and cap diode lader like below


{------}
{------} transformer / invertor not sure not labled..
{------}
       |
--[ ]--- cap B 471K 3KV
|--->|-- DIODE R3000F
--[ ]---| cap B 471K 3KV
|--|<--- DIODE R3000F
--[ ]---| cap B 471K 3KV
|-->|--- DIODE R3000F
--[ ]---| cap B 471K 3KV
|--|<-- DIODE R3000F
--[ ]--| cap B 471K 3KV
|-->|------ DIODE R3000F
--MVM--|  RESISTOR 2R3
|
|
|
HIGH VOLTAGE OUT

Anyway im looking into the parts in this design and seeing wat kind changes i need to make to have it support my 12v supply and little higher current..
-------------------------------------------------------
i also have one other thing guys but im keeping it under wraps till i have my proven test done be for i go ahead with it..  not so much a secert but if it didnt work no point mentionsing it..
---------------------------------------------------------

Anyway thats all for now.. feel free to PM me if u need any questions answered or wanna share ideas..

Lanmasterd

[Feynman]

Thanks lanmasterd.

Some of you may point out that finding an AND gate of sufficient switching speed may be difficult.  There is no way you will find a suitable component at Radio Shack. . . they are just too slow.  Well , problem solved!   Introducing the 74G08 high-speed gigahertz AND gate, from Potato Semiconductor.  These guys make the fastest logic chips available.  It's just a simple AND gate that has rise  time under 1ns and can switch up to 1.1Ghz.






You can get these chips on ebay for about $1 apiece.
http://cgi.ebay.com/7408-G-Series-GHz-TTL-CMOS-logic-IC-14pin-SOIC-QTY-10_W0QQitemZ330223129020QQcmdZViewItem?IMSfp=TL0803270830a24281