This is it !

[gmeast]

grrrrrrr

I want my parts! GRRRRRRRRRR!

The only modification left to make to my FET board will be to split the 'inhibit bus' into 3 separate inhibit inputs enabling me to (individually) inhibit the gate driver outputs to the FETs/coils separately. Presently, all 3 inhibit inputs are bussed together and serviced by one signal.

Question: Why would I want to do this?
Answer: I haven't the slightest idea why!

However ... there may still be some sort of modification to the driver IC-parts of the circuit to get rid of that small amount of HF 'ringing' I showed in some recent scope captures. As I said, that ringing is only there when the PPC signal is connected to a FET that's doing some work.

Later

[gmeast]

up and running

Hi all,

Ok ... I'm back up and running. My board mods were successful. I'm glad my UCC27322's are in IC sockets now & my FETs are easier to replace if I have to.

Scope, DVM, and any other instruments have been removed. Operational frequency is around 28kHz, and phase angle is around 120deg ... tuned for 'quietest' phases (no jumping around). PW is 1usec.

I'll let it run all night. It's midnight now. Battery voltage is 12.66VDC. I started testing about 5 hours ago ... it was 12.66VDC then also. I had hit some 'bad' resonances that yanked it down to 12.59/12.60VDC and over time was pulled back up to 12.66.

The FETs are 'stone cold'. Last measured, the PPC was drawing about 75mA, so it's keeping up with itself. To tell you the truth, I have no idea of where to go from here since this LE stuff is new to me and I haven't laid eyes on any working theories yet. Maybe Bob Boyce will chime in soon.

I'll post a video ASAP.

Later

[Farmhand]

Hi Greg, Good to see you have your setup going. Can't wait for the video.

I have to admit that I was still a bit confused about your setup and how to build it. So I went ahead and designed and built my own. I can adjust the Phases the PW the frequency and I can have a extra space after each cycle, which is what I want. I admit it is probably a lot less advanced than yours, but I designed and built it in about 24 hours. It's a work in progress.

Minimum PW I hve used so far is about 877 ns. Maximum frequency I have used is about 80 Khz. I am only using the pulses directly from the CD4047's to drive the Mosfets with the diode and PNP turn off sharpener.

Here are some picture's.

http://koynpq.bay.livefilestore.com/...006.JPG?psid=1

http://koynpq.bay.livefilestore.com/...003.JPG?psid=1

Drive pulse
http://koynpq.bay.livefilestore.com/...002.JPG?psid=1

120 phase
http://koynpq.bay.livefilestore.com/...002.JPG?psid=1

120 phase
http://koynpq.bay.livefilestore.com/...003.JPG?psid=1

Grouped
http://koynpq.bay.livefilestore.com/...007.JPG?psid=1

Grouped
http://koynpq.bay.livefilestore.com/...006.JPG?psid=1

Clock pulse.
http://koynpq.bay.livefilestore.com/...003.JPG?psid=1

I have done the one battery thing for a while and did see some volage rising but very little. Once I get the design finalised and boarded up I'm sure it will work much better.

I have already designed my Mosfet and driver board. And etched the PCB.


Double sided version.
http://wv2uga.bay.livefilestore.com/...hot.jpg?psid=1

Single sided version.
http://wv2uga.bay.livefilestore.com/...hot.jpg?psid=1

I have some drawings for the Boards, but I don't have a drawing for the controller circuit. I'll make one soon though.

Cheers

[gmeast]

update

Hi all,

Well I guess Farmhand was first to post a poly phase setup. I guess he couldn't wait. He says it only took him 24 hrs to do it. Good for him.

I'll still post mine anyway.

The operational frequency is 28kHz and the PW is 1usec. The green is one of the 3 PPC signals to the driver chips (used to 'synch' on the pulse train) and the yellow is the output measured between the diode and (recovery) choke.

3-ph1a shows poly-phase operation at approximately 120deg phase angle separation.
3-ph1b shows all phases inhibited and the mean battery voltage.
3-ph1c shows a closeup of the 3rd phase in the pulse train.
3-ph1d shows 3-ph1c but with phases inhibited.
3-ph1e shows the 3rd phase in the pulse train being similar to 3-ph1c
3-ph1f is the same as 3-ph1e but showing the actual power pulse taken on the Mosfet's Drain ahead of the (recovery) diode.


http://02d1852.netsolhost.com/radian...ing/3-ph1a.png
http://02d1852.netsolhost.com/radian...ing/3-ph1b.png
http://02d1852.netsolhost.com/radian...ing/3-ph1c.png
http://02d1852.netsolhost.com/radian...ing/3-ph1d.png
http://02d1852.netsolhost.com/radian...ing/3-ph1e.png
http://02d1852.netsolhost.com/radian...ing/3-ph1f.png

For all of my hard work, it would have been nice NOT to have been shown up. I guess I'm done here.

Later

[Farmhand]

You Were First

No Greg you were first. Anyway i'm not trying to compete, I am trying to participate. I thought we were a team.

Congrats to both of us, I got my idea to cascade from you're setup, but I am still a stranger to the comparator. After building this though now I think I might understand it better.

I still want to build you're setup. When you make a drawing, that is. I understand about the drawing because I did the same thing. I don't have one either.

So, are you seeing some charging ? I think we have some note's to compare if you would like to Email me.

I was wrong about the second battery for some reason with this setup if I want to charge a second battery, the charge battery negative is better going to the source negative or ground rather than the source positive like I was saying. Strange, it still works but to ground looks better.

I'm using ferrite chokes, and they are working too with one battery, sweet.

Good job.

Cheers

[gmeast]

update

It's very difficult to determine (definitively) if charging is occurring, I might be able to 'imply' that it is charging only because we know that there is ohmic/impedance losses in the transformer, power consumption by the controller, probe-load when they're connected, etc, so since the battery is still at 12.66VDC after running for 8-1/2 hours (unhooked from everything) it's safe to say it is at least holding its own. It also means that a 'tuning' strategy is called for ... if I only 'really' knew what to look for??

I know this ... if you crunch them up too much, it consumes power and the same goes for too-high a frequency.

I haven't detected a difference between the 'swirl' directions yet. Now that I have some time on the 'CCW swirl' I'll change that ... it's the easiest thing to do.

I detected NO real affect there. HOWEVER, when I went from the CCW (recovery) choke to the CW choke ... it did NOT like it at the higher frequencies ... 42kHz for example ... the battery drained pretty fast. So I have started to see that the choke affects the direction of LE flow for a given frequency and phase angle(s).

I'll just keep trying various combinations of things for now until I figure out more about 'cause and effect' and other stuff.

Later

[Farmhand]

Yes I know what you mean. Mine looks like it really wants to go up, and it did a couple 0.02v but then it stopped, that was on a rested battery too. But my circuit has very small overhead, so I am not seeing the same amount of "holding it's own" you are.

I still don't have proper mosfet drivers or diode's yet either, I have the fast recovery diode's but the driver's haven't arrived yet. I'm sure we will see better function when we have the circuits boarded up.

Ok i'll check my choke direction, that's what else I need is the right choke's.

Maybe i'll make a copper egg and use a good wide pulse while adjusting the frequency up, see if it spins. Could be fun.

Cheers

[gmeast]

29kHz

For some reason, my stuff likes 29kHz, because that's where all of the phase FET drains are uniform and of lowest magnitude. I have been told that it's NOT the BEMF spikes or any of that other stuff that causes battery charging, so I'm tuning for 'minimum', spikes and phase uniformity.

Since you can't see Longitudinal waves with any known instruments, it's like 'flying blind'.

I have entertained the the thought of adjusting the phases to produce a sort of 'peristaltic' action to provide an easier conduit for the LE. That would entail either successively Increasing spacing pulse-to-pulse in the pulse-train or successively Decreasing the spacing pulse-to-pulse in the pulse-train.

I will try that.

Later

[Farmhand]

Ok it's almost morning again, so i'll stay up to feed the hogs and chooks before I have a nap. Come to think of it I think I lost a day or two maybe it took me longer than I thought to build my circuit. Bit hazy.

Yeah that might be worth a try, anything I do with these choke's won't be much help anyway. They are pretty loud. Do your chok'e ring ? It's a bit annoying. Thing is they are loud when the waveform is neatest and lined up really good. Oh well. I can't think of anything other than the order of the recovery diodes, mine the recovery bus go's from left to right, right being the output, but the first coil is on the left, I think I should change it so the first coil to fire is on the output end of the recovery, for when I'm using a space mainly, that would ensure the second or third pulse didn't interfere with the first maybe. It looks strange on the scope, if I go from diode to diode the interferance looks different.

Oh yeah at the moment mine is lining up at about 21 Khz the base oscillator/clock is 7 080 Hz.

[gmeast]

my setup

I forgot to show my setup.

The first picture shows the overall setup. You can see the battery at right on the table, toroid in front of the battery, DVM also on the table. In the background you see the PPC prototype & switching power supply (front edge, right-side of my desk) and my Net Book (behind the PPC prototype) running the PC Oscilloscope (far right on the desk behind the soldering iron).

The second picture shows the toroid and its wires (radially spread) and the FET board in the middle of the toroid's hole. Beneath the FET board is the B+ choke with one lead exiting under the table to the 3 toroid 'commons' and that choke's other lead flopping out radially & going to the B+ common. Also going to the B+ common is the (recovery) choke that exits up and out the corner of the FET board (white wire).

http://02d1852.netsolhost.com/radian...ng/overall.JPG
http://02d1852.netsolhost.com/radian...ing/toroid.JPG

Later

[gmeast]

timing

Farmhand,

How are you 'locking' the phases in place?

[Farmhand]

I just sneak up on it and it kinda jumps in and sticks there a bit sometime's, it's wierd because sometime's it doesn't jump and stick, it just runs right up real smooth and like you said starts to use more power. Maybe it had to do with the different way I was using it. After it jumps ahead I keep going till I see it start scrunch them up then I back it off to where I reckon halfway is. And the sound flutters too when just out of phase, maybe thats good. Don't know. But it's fun.

Oh or did you mean to get the gaps inbetween the pulses ?

I can explain how I got the phases if that's what you mean.

[gmeast]

3-phase

I just sneak up on it and it kinda jumps in and sticks there a bit sometime's, it's wierd because sometime's it doesn't jump and stick, it just runs right up real smooth and like you said starts to use more power. Maybe it had to do with the different way I was using it. After it jumps ahead I keep going till I see it start scrunch them up then I back it off to where I reckon halfway is. And the sound flutters too when just out of phase, maybe thats good. Don't know. But it's fun.

Oh or did you mean to get the gaps inbetween the pulses ?

I can explain how I got the phases if that's what you mean.

Looking at the board, it looks almost like 3 separate circuits that are independent.

Otherwise: "I can explain how I got the phases if that's what you mean." ok ... that.

In the PPC you get a pulse at a precise time always.

Later

[Farmhand]

Yes thats the advantage/disavantage of both arrangements when you think about it.

Ok here go's, I have the basic 4011 oscillator circuit and three CD4047's in monostable mode, each 4047 has three output's I use two of the outputs one much smaller than the other, so the clock oscillator trigger's the first 4047, it's two pulse's fire simultaniously, a very short one go's to the first fet and the second one when it finishes, trigger's the next 4047 with the negative going edge "edge triggered" , the next one does the same.

The pulse width and the space between the pulses is determined by the timing caps and resistors, I used the 555 PWM trick to determin the PW for the Fets, it is two resistors in series with a diode in parallel with one resistor so that the timing cap must be charged through both resisters but is discharged through only one resistor. I can adjust the space between the phases by one resistor and the PW with the other and the cap which also changes the frequency, the 4047 chips are oscillating one shot per trigger so the faster the oscillating frequency "relative" the shorter the PW and the space.



All this happens well within one cycle of the clock at low frequency which leave's the remaing time as a gap after the last pulse but before the first, so I start it up with three even pulses then a big gap then adjust the frequency of the clock to close the gap to where I want it adjusting the PW of the clock is like a very fine adjustment for frequency and can limit the scrunching room availiable too I think. All the drive pulses are 14 volts or so.

[Farmhand]

I think when using timing componants resonance of the actual controller and circuit is likely, where with a digital preciscion controller the resonance may be limited to other parts of the circuit and difficult to find. The timing resistors and caps can all fall into harmonious syncronisity which promotes phase locking I think.

Sounds good anyway

Ok found it, here is a shot of the three pulses before I connected the fets, I bussed the three outputs together through some 1N4148 diode's and some 330 ohm resistors with a 2 k resistor to ground probed between the resistors.

http://public.bay.livefilestore.com/...004.JPG?psid=1

That was a small gap and wide pulses. Before i worked out how to trim the differences. The 4047 chips and 4011's and stuff need to have at least 1 K resistance for the timing section. So The PW is restricted by that and the timing cap, 1 k and 0.001 uf give a PW of about 600ns PW, 2k and 0.001 uf gives about 877ns PW and 2k and 2 x 0.001 uf give's about 1.6 us PW and so forth, it works out pretty neat, a 50 K reistor for the frequency gives a space (forgot the time) which works out to a phase locked frequency of about 21 Khz with any of those PW's.