Gary's "Complete Advanced" SSG Build

[Gary Hammond]

Hi Rodolphe,

Just a theoretical (since I’m not going to use it now) question on mode 3 (C.G. mode + cap pulser):
For the Cap pulser that we both build*, can that circuit handle mode 3? In other words; would no components be overloaded when the coil discharges right at the moment when the cap discharges as well? I assume it can handle mode 3 since it can also handle mode 2, but just wanted to double check.

It handles mode 3 just fine and charges faster than mode 2.

Would it be more correct to say: The height of the spike is determined by the speed the Transistor can shut off (and the speed with which the magnetic field can collapse based on the inductance of the medium, iron in our case), the speed with which the diode can switch on determines how much of that spike is being caught.

As I understand it, the faster the switching speed (both on and off) of the transistor and/or the capture diode the higher the potential of the spike can go. The spike will go as high as it can depending on the load impedance of the battery and associated wiring. The load becomes the limiting factor of the spike voltage.

If I read on the top of page 30 of the advanced manual, Peter calculates pulses of 72.28 amps to the output battery (based on his scope screenshot from page 29). That is a high number (however average amperage on the output 0.7amp.)

The 72.28 amps is the cap dump current to the charge battery. The high voltage, low current reference was only to the capture diodes as regards their forward voltage drop being slightly higher for the faster switching diodes. The faster switching speed will more than make up for the slightly greater (.7) forward voltage drop when capturing high voltage, low current spikes.

Do you have an explanation why this is so; that the amp draw went down (and the charge time increased, so same COP? It’s like these UF4007 have the same effect as if I’d increased the base resistance… If the just the amp draw would have gone down, without the increase in charge time, the COP would have improved, which was what I hoped for.

I'm just guessing here that the clamping diodes faster switching and higher forward drop causes the transistor to switch off faster and limit the total current draw to a lower value. Slightly reducing the resistance of the base trigger pot should bring the current draw back up. ..............Interesting that this did not increase the COP.

So with FETs (faster switching) and the hall sensor (adjustable timing on trigger circuit), you say you got better result? More than COP 1.25 which you got on the SG with MJLs?
This is also what Paul said in his video, I’ll post the links we’ve been talking about here too so others have access to it:

I did this on both my attraction (switched reluctance) motor and the wheel driving my two stage mechanical oscillator. These are different type machines with a different type load so I can't really compare them. I also modified my first SSG (which never did over a .8 COP) with adjustable hall switching, but still using the MJL21194 transistors. I don't think that increased the COP much on that machine.

p.s. For the CBA software you use, do you have the "extended sofware licence"m with the extra features by any chance?

No. I do not have the extended license. I have an older CBA3 unit with the standard license that came with it.

Gary Hammond,

[RS_]

Hi guys, Sorry I have not been on in a while. I have been extremely busy with work, and my 300A solar charge controller replacement for JB's 300A model that blew up, and he did not get it replaced before he passed..... The Software to control it is killing me. I'm a hardware man, not a software man. I've been learning to code Ardiunos for a while now, it's still not easy to make them do everything I want, and I want a LOT..... LOL Been sick with various things in between, slowing things down too..... I'm Back to it.... Y'all carry on....

[Gary Hammond]

Hi RS,

Glad to see you back on here, and hope you're back to good health.

I'll be off the board for a few days myself. Found out we had some storm damage to our mobile in Florida from hurricane Ian. Going down to do some repairs.

Gary Hammond,

[pearldragon]

Hi RS,
Glad to hear your feeling better and back!
Quick update about the major things that Gary and I’ve been discussing and what I’ve done in the meanwhile:

• Still trying to figure out were the huge performance difference between Gary (Cop >1) and my COP +/-0.7) machine comes from.
• a couple of months ago I received my long awaited order from TeslaGenX with a new PCB kit (including their matched transistors, diodes, resistors, etc) and a new coil. -->> Swapping all these components did not make any difference.
• Gary and I went then into detail about the cables; I changed cables to pure copper ones, (was a mix with alu), put an automotive on/off switch in the negative cable from/to the run battery (in like Gary has), cleaned up the wires/cables underneath the PCB. For pictures see #162. -->> Again, all to no avail.
• Then based on what Nick told me and what I saw in the Paul Babcock 2022 ESTC presentation, I thought that the 1N4007s might be the cause (although Gary used them in some of his built as well), so I swapped them for UF4007 -->> To no avail; COP still the same (#179).

Those are the major things that happened in your absence. Now in the next post I’ll respond to post #176 of Gary.

Regards,
Rodolphe

[pearldragon]

Hi Gary,

The 72.28 amps is the cap dump current to the charge battery.

You’re right, my bad; that 72.28 amps relate to a cap dump, not the situation we’re discussing.

I'm just guessing here that the clamping diodes faster switching and higher forward drop causes the transistor to switch off faster and limit the total current draw to a lower value.” I was wondering about that too: that these diodes cause the transistor to shut off a bit sooner, resulting in a lower amp draw.
Interesting that this did not increase the COP.

Yes, interesting for sure… Could mean that the transistors are so much slower than the diodes, that changing the diodes has very little effect at this point…

I did this on both my attraction (switched reluctance) motor and the wheel driving my two stage mechanical oscillator. These are different type machines with a different type load so I can't really compare them. I also modified my first SSG (which never did over a .8 COP) with adjustable hall switching, but still using the MJL21194 transistors. I don't think that increased the COP much on that machine.

Hmm… I was considering exactly that: moving away from the ‘trigger coil’ triggering that is now wound together with the main coil, to see if with more freedom to change the timing of this trigger signal I could get the COP up… But I also saw a potential down side in that:
Since the magnets don’t have all ‘exactly’ the same shape, and not spaced ‘super super’ precise (I did it as best as I could), and then assuming that I’d make the (hall)trigger, trigger on the magnet preceding the magnet that is approaching the coil, the timing might be always a bit off for some of the magnets. (the ones that vary a bit on size or spacing). This problem is not there with a trigger winding around the maincoil, but with the tradeoff of not having adjustability of the timing.
And now reading that you already tried this with no effect on the COP (with the MLJs), is another reason to hold off on that for now…

I also modified my first SSG (which never did over a .8 COP)

So this is a very interesting thing you mention here! Were you never able to pinpoint this cause? Have you not swapped out parts (PCB/Rotor/coil/batteries/etc) between the machines that got a COP of >1??? I mean, if you still have that COP 0.8 machine, if you’d start swapping out parts, at some point you will swap out a part/combination of parts that will make that machine have a COP of >1 as well, and might solve my mystery at the same time.

I asked about whether you have an extended license software for your CBA for the following reason: The extended license has the option to measure internal resistance of the battery. I’m in the process of obtaining such a license and are curious to see what the internal resistances of my batteries are. Would have been interested to compare them to yours, and probably would give the answers whether your batteries are partly responsible for the COP difference. Although, we all agreed already that that probably is not the case, at least not to a big extend, would still be interesting to compare.

Best regards,
Rodolphe

[Gary Hammond]

Hi Rodolphe,

So this is a very interesting thing you mention here! Were you never able to pinpoint this cause? Have you not swapped out parts (PCB/Rotor/coil/batteries/etc) between the machines that got a COP of >1??? I mean, if you still have that COP 0.8 machine, if you’d start swapping out parts, at some point you will swap out a part/combination of parts that will make that machine have a COP of >1 as well, and might solve my mystery at the same time.

I still have that machine, although it has gone through several changes since I first built it. Here's a little history of it ................It started out with 5 AGW18 power windings each 150' long twisted together with a AGW20 trigger strand also 150' long. The circuit was open wired in space (like John Bedini's in EFTV videos #24 & #25) using 5 MJL21194 transistors that I tried to match myself. (No circuit board used) It originally ran in repulsion mode using 18 magnets on a 26" aluminum bicycle wheel with its original wheel bearings. I think the rundown time was about 10 minutes The first change to it was a rewiring to run in attraction mode. As best I can recall it gave around a .8 COP.

The next change was to convert it over to hall switching. The halls are powered by the run battery and the trigger winding is disconnected. It uses two movable hall switches working in series located on the back side of whichever magnet is passing over the coil core. This makes both the timing and pulse width adjustable and is triggered by the same magnet that is being attracted in. With this set up it can only run in one direction. I also rigged up an auxillary generator coil running some LEDs which didn't put out much and was later removed.

Then I set it up with a 4 battery, split the positive battery switching arrangement using only 4 transistors and windings to drive it and taking the high voltage pulses from the 5th winding which was now isolated from the run circuit. I didn't get the results I wanted so even tried running it on 3 windings and harvesting the other 2 in isolation. This didn't make much difference, so I changed it back to 4 powered and 1 isolated winding. I added a full wave bridge rectifier to the isolated winding for the charge battery and use only one run battery. This is how it is currently set up and I don't want to change it because it is somewhat novel.

I wanted to do the 4 battery swapping with my newer machine that gets the 1.25COP, but I didn't want to isolate a coil like I did for the split the positive method. So I totally rewired my mechanical switches to run in a split the negative method which let me run my good machine as it was. This set up will run for long periods of time bringing the batteries back to full charge each cycle all while lighting 84 LEDs from the 2 auxillary generator coils. This is with 105AH batteries and manual switching. I can also run this machine with smaller batteries and the automatic swapper using the auxillary generator coils to power the automatic swapper.for continuous running.

The two machines are configured differently, as I described, and I really don't want to change either one by swapping out parts. They both run and demonstrate their own characteristics.

Gary Hammond,

[pearldragon]

Hi Gary,

How did the repairs go? All fixed again?

Thanks for your explanation about both machines. I understand that you cannot swap out parts that easily with what you told.

Zero point energy gathering/conference in the Netherlands
In the end I did decide to go to the conference and did a presentation about the Bedini SG and although I was a bit reluctant to do I when Nick told me he would not join, it was a really nice day! My presentation was received well and was even approached by a guy afterwards who built an 8 coiler himself 10 years ago. He never managed to get to over unity and abandoned the project. But had a good talk with him, exchanged contact details, and who knows he might pick it up again. Besides him I made contact with some a couple of more interesting people.

More thoughts on my missing COP
I’ve been thinking a lot this week what else could be a culprit for my missing COP, and I was thinking about the trigger circuit. Could the fixed/variable resistances (so not the 100ohm / 470ohm ones) be the cause for the following reason:
The resistances I got to start off with (with my last order of TGX) are 3x 12ohm 10W resistances (Yageo) wire wound resistances… which I put in series (36ohm) but then had to add 3 more wire wound resistances (of different values) in series as well to get the amp draw in the area where I wanted it: max 2amps. Total resistance being around 58ohm.
Now here is my thought: If one of the key important things is a sharp shut off of the Transistor… 6 WIREWOUND resistors do not seem like the most wise choice: they have inductance and hence slow down the falling away of the current through the base of the transistor, resulting in a slow shutting down of the transistor…
The only thing that I’m not sure about is if this ‘resistor’ inductance has any weight in comparison to the trigger coil inductance itself…
But when I look back into earlier results/measurements, when I installed the components of my last TGX package, and started with the 36ohm… although the amp draw was way too high, the COP dropped down with every resister I added… Before I swapped out the components from this last TGX package, I had a 2W potmeter installed. I looked up the spec of that one: also wire wound.
On the one hand this gives me some hope, on the other hand: in the handbook they show also these 3 wirewound resistances that I have from TGX in package (see attachment)… and I think you used them too at some point? What is the technology of your pot meter?
However, worth a shot I think.

Regarding reduced amp draw, post #181

I'm just guessing here that the clamping diodes faster switching and higher forward drop causes the transistor to switch off faster and limit the total current draw to a lower value. Slightly reducing the resistance of the base trigger pot should bring the current draw back up. ..............Interesting that this did not increase the COP.

I’ve been thinking about this some more. I don’t think the clamping diodes should have anything to do with it. They just prevent the transistors from possible damage when the magnet swings past the coil, and the current in the trigger coil circuit reverses.
However, the ‘spike’ diodes, they do shut off earlier with their increased forward voltage, which would result in less energy from the coil being discharged into the output battery, see attachment. That would be the cause of the lower amp draw if you ask me. Is this correct like I state it here?
(If the trigger coil resistances are the ‘slowest’ factor/bottle neck, it is not so surprising that the COP didn’t change by using faster diodes).

Radiant spike & coil discharge
It is this last part of the previous topic that I still find a bit hard to grasp, I mean the following: I’d expect (in my own logic ), that if the transistor shuts down, ALL the energy in the coil would be discharged in one big hit, and not a ‘spike part’ and a ‘coil discharge’ part, as can be seen in the attachment/graph… So do I then have to look at this as two different processes taking place in parallel, both also not having exactly/completely the same cause:
1 The collapse of the magnetic field, resulting in the Spike (radiant part)
2 The falling away of the current flow through the coil, resulting in the stored energy there being discharged (in the output battery), (normal part)

Best regards,
Rodolphe

187 - 2022-10-16 - Attachment.pdf

[Gary Hammond]

Hi Rodolphe,

How did the repairs go? All fixed again?

No. We just made a quick trip to access damages, do some temporary repairs, and clean up the stuff blown into our unit. Repair parts are not readily available yet because there is so much devastation 60 miles farther south. Our actual repairs can wait til we go down for an extended stay.

The resistances I got to start off with (with my last order of TGX) are 3x 12ohm 10W resistances (Yageo) wire wound resistances… which I put in series (36ohm) but then had to add 3 more wire wound resistances (of different values) in series as well to get the amp draw in the area where I wanted it: max 2amps. Total resistance being around 58ohm.

58 ohms sounds like too much resistance for the common trigger resistor. Are you using the 470 ohm/one watt resistors on the trigger branch circuits that came with the kit? Or are you still using the 100 ohm ones you used originally?

I’ve been thinking about this some more. I don’t think the clamping diodes should have anything to do with it. They just prevent the transistors from possible damage when the magnet swings past the coil, and the current in the trigger coil circuit reverses.

That's not entirely correct. When the clamping diodes forward conduct the current from the reverse voltage swing of the trigger coil, they are biasing off the transistor by the value of the forward voltage drop of the diodes. So the higher forward voltage drop and faster switching speed of the UF4007s will cause the transistors to turn off faster and more completely. RS frequently uses two 1N4007 diodes in series for this very purpose. That is called a "Fogal" modification and results is faster and more complete shutoff of the transistor.

However, the ‘spike’ diodes, they do shut off earlier with their increased forward voltage, which would result in less energy from the coil being discharged into the output battery, see attachment. That would be the cause of the lower amp draw if you ask me. Is this correct like I state it here?

The spike collection diodes shouldn't affect the "on current" draw as they only conduct when the transistor turns off. The extra forward voltage of these diodes is immaterial because the total voltage of the spike is so high, but the faster switching speed is helpful to collect more of it in the short time available.

Radiant spike & coil discharge
It is this last part of the previous topic that I still find a bit hard to grasp, I mean the following: I’d expect (in my own logic ), that if the transistor shuts down, ALL the energy in the coil would be discharged in one big hit, and not a ‘spike part’ and a ‘coil discharge’ part, as can be seen in the attachment/graph… So do I then have to look at this as two different processes taking place in parallel, both also not having exactly/completely the same cause:
1 The collapse of the magnetic field, resulting in the Spike (radiant part)
2 The falling away of the current flow through the coil, resulting in the stored energy there being discharged (in the output battery), (normal part)

I think the way John Bedini explained this is that yes there are two separate but related events taking place here...........

#1 The falling away of the current flow through the coil results in a back EMF being applied backwards across the transistor and run battery. The value of this is only about 2/3 the value of the applied voltage and can therefore not recharge anything. BMEF is best understood by studying conventional electric motors. Peter Lindemann explains this well in this video. http://www.electricmotorsecrets.com

#2 When the current flow through the transistor turns off abruptly, the surrounding magnetic field collapses and tries to keep the current flowing in the same direction through an inductive process similar to inertia. Now the coil changes from being an energy sink to being an energy source and pulling in some free energy with it. This results in the voltage across the coil reversing polarity and rising as high as it must to find a conductive path to dissipate the energy. If there is no conductive path available this voltage can reach thousands of volts until something breaks! This is the energy we harvest with fast diodes and a battery.

[pearldragon]

Hi Gary,

58 ohms sounds like too much resistance for the common trigger resistor. Are you using the 470 ohm/one watt resistors on the trigger branch circuits that came with the kit? Or are you still using the 100 ohm ones you used originally?

I was thinking about this too last week. I’m using the 470 ohms/one watt that came with the last shipment from TGX (so when I swapped my PCB and all the PCB components). What I figured was that although both our transistors are matched per set (by TGX), your set still differs from mine. But I’m not sure whether the difference in our sets should be responsible for this big of a difference (also see attachment).
I’m running a bit smaller gap (6mm) at the moment than you, but ran with 8mm before as well. Before changing the 1N4007 diodes, with the fixed resistance at 58 Ohm (57.9), I had an amp draw @ 2min into a run of about 1.9 - 2 amps… so even the 58 ohms were actually a bit low. Now with the UF4007 it is around 1.7 amp @ 2 min into the run, which is more in range with what you have I think.

I’ll come back to some other parts of your post #188 at a later point; I’m not understanding some of it at the moment, but I’ll try to analyze some signals in the circuit a bit more with my scope and then have a call with a friend of mine who is an electrical engineer, hoping that that will clear things up a bit, saving a couple of posts between you and me .

Best regards,
Rodolphe
189 - 2022-10-20 - Attachment.pdf

[Jeffrey Joseph]

Aloha from Hawaii. Jeffrey Joseph Here. can any one help me with John's transistor matching circuit? The wife rearranged my things when I was at the 2022 conference and I cant find my circuit or a diagram of it. Also I got confused and burned up all but one of my wheels. I ordered more 3055's and need to match them. Will someone kindly supply the schematic? Thank you in advance for your help. Happiness and Joy Jeff.

[Jeffrey Joseph]

I have had success with increasing the speed ; I use a decade box on the trigger and tune to one ohm resistor least amount of current. Increase the resistance slowly and watch the speed increase. Also if you Has if you get every thing matched you can double the run voltage with a power supply ; limit the current so you don't burn up the transistors. Have fun. I am studying the mosfets and my brain hurts now. Happiness and Joy Jeff

[pearldragon]

Aloha Jeffrey,

Assuming you have the 3 handbooks (http://www.bedinisg.com/), in the Intermediate Handbook, page 11 the matching circuit is explained in detail,.

But maybe the questions regarding this circuit are a bit off topic in this specific thread, would maybe fit better in a thread regarding the advanced handbook, in some of these threads:
https://www.energyscienceforum.com/f...2-intermediate

Best regards,
Rodolphe

[pearldragon]

Hi Gary,

That's not entirely correct. When the clamping diodes forward conduct the current from the reverse voltage swing of the trigger coil, they are biasing off the transistor by the value of the forward voltage drop of the diodes. So the higher forward voltage drop and faster switching speed of the UF4007s will cause the transistors to turn off faster and more completely. RS frequently uses two 1N4007 diodes in series for this very purpose. That is called a "Fogal" modification and results is faster and more complete shutoff of the transistor.

Ok, this part I completely understand now. What made me misunderstand this until now were the pictures in the handbooks (e.g. page 19/22 / 45 (top image) / 73 (top image) beginners handbook): The reverse current also goes through the Transistor from the Emitter to the Base, not just through the diode…

A small test I want to do in the near future is to see how much ‘reverse’ voltage the transistor in ‘reverse’ will allow (before it starts to act as some sort of Zener) -> It might show that 3 diodes in parallel are even possible there… However for me at this moment also my issue is not in those diodes I conclude on my previous test runs, but it is a potential worthwhile optimalisation step in general.

Wirewound to ceramic
I replaced my fixed resistors. They were wire wound (same as in handbook), and now they are ceramic ones. Since the input voltage dropped a bit when I swapped the 1N4007 diodes for the UF4007 versions, I decreased the fixed resistance value a bit to 49.7 ohm (was 58 Ohm), which brought me back to my earlier 1.9-2 input amp draw, measured 2min into a charge cycle with the SG. I did a couple of runs, see Attachment 1.
What I saw in the results is something I noticed actually for a while already; that quite a couple of values go down in consecutive cycles. The COP goes down from Cycle IDs C221001, C221002, C221003… Then in Cycles C221004 C22105 I changed the following parameter: instead of charging the input batteries for 1h with the TGX charger before the a new cycle, I charged them for 1.5h. Although you see the COP goes up initially when I make this change (C221004), it goes down again in the following cycle: C221005 again… And even though it jumped up a bit in cycle C221004 it is still below the first and the second cycle (C221001 & C221002).
And so more values go down:
-Output voltage @ rest, before testing
-RPM
It gives me the impression that (all) the batteries that I use, are not liking to be charged in CG mode consecutively… Have you done 3, 4 or 5 cycles in a row (every day 1 cycle) with steady parameter (especially COP) results? Or just one offs?

If my suspicion is correct, then many of the tests I did before trying to determine the influence of certain parameters are actually invalid/not trustworthy. What I’m thinking now is hooking up the Cap pulser (comparator) in between the SG output and the output batteries, but still running in CG mode. So basically how it is described in the advanced handbook. If my output gets steady, I need to redo a couple of tests…

Resistance/Rotor/Rotor gap/RPM
As mentioned I might not have drawn the right conclusions yet about parameters like optimum rotor and gap, still my current rotor (rotor 3) is fairly similar to yours, see Attachment 2. However, your (fixed) resistance is significantly lower as you pointed out already, and your RPMs are way higher (based on your post #29), while I have a slightly smaller rotor gap at the moment (6mm). (smaller gap = more RPM).
Even with my low COP I find these differences peculiar…
The faster the RPM, the faster the transistor is shut down, the sharper the spike. So also my low RPM does not work in my advantage…
While writing this: I think you have dedicated bearings right? I use the bearings that are included in the bike wheel, but a bike would not get to such a high rpm normally I guess… Although my free spinning time is very good following your advice about removing the grease, maybe those bearing run into higher resistance going into higher rpm, just a thought… But on the other hand; it would not explain the COP going down over several tests...

Internal resistance batteries
With the extended license for my CBA I also did some quick measurement on the internal resistance of my batteries, see attachment 3.

Paul Babcock (PB)
In the presentation of PB, he indicates that battery swapping is essential. Although I’m not concerned with battery swapping yet, there is a question that I have about that:
Have you any idea in which mode his machine(s) run? I’m guessing it is Radiant mode or otherwise Radiant + Cap dump, but I guess it is the first… If that is true, how does this hold op to what JB said about not being able to swap the batteries in radiant mode?
I do not know if PB swaps out just 2 batteries, or uses a setup with 3 or 4 batteries, in different positions (e.g. page 90 advanced handbook).

Best regards,
Rodolphe
193 - 2022-10-29 - Attachment 1.pdf
193 - 2022-10-29 - Attachment 2.pdf
193 - 2022-10-29 - Attachment 3.pdf

[Gary Hammond]

Hi Rodolphe,

Ok, this part I completely understand now. What made me misunderstand this until now were the pictures in the handbooks (e.g. page 19/22 / 45 (top image) / 73 (top image) beginners handbook): The reverse current also goes through the Transistor from the Emitter to the Base, not just through the diode…

The reverse current only goes through the diode causing a .4 to .6 volt drop. This is the reverse bias across the transistor base/emitter junction that turns it off. There is no appreciable current flow in the reverse direction through the transistor except for a few micro-amps forcing the electrons and holes away from the PN junction. The higher this reverse bias shutoff voltage, the faster and more completely the collector current is shut off. The emitter/base break down voltage is 5 volts. If this is exceeded damage will occur. This device is not designed to operate as a zener!

Have you done 3, 4 or 5 cycles in a row (every day 1 cycle) with steady parameter (especially COP) results? Or just one offs?

It has been long enough ago that I don't recall how many days in a row, one day apart, I ran these tests. But I think the results were fairly consistent, as best I can remember.

I did however do several continuous CG mode back to back runs, by manually switching the batteries. This is how I determined that the COP was actually less than 1 because each run would take a little longer than the previous one to fully recharge the receiving battery. I've done this with several of my machines with somewhat the same results.

But I have managed to get two of my machines to run continuously with the four battery split the negative swapping arrangement. This appears to be actually running at unity as the batteries always recharge to the same voltage in about the same time. This method uses the run current plus an inductive spike to charge the one battery that's in the charge position. Then the next cycle it is at rest. So each cycle one battery is resting, two batteries are supplying the run current, and one battery is being charged. Every cycle they each advance to the next position.

While writing this: I think you have dedicated bearings right?

I'm using precision steel ball bearings with the grease removed and lubricated with Tri-Flow in the machine you are comparing to.

Paul Babcock (PB)
In the presentation of PB, he indicates that battery swapping is essential. Although I’m not concerned with battery swapping yet, there is a question that I have about that:
Have you any idea in which mode his machine(s) run? I’m guessing it is Radiant mode or otherwise Radiant + Cap dump, but I guess it is the first… If that is true, how does this hold op to what JB said about not being able to swap the batteries in radiant mode?
I do not know if PB swaps out just 2 batteries, or uses a setup with 3 or 4 batteries, in different positions (e.g. page 90 advanced handbook).

I think he was running in straight radiant mode. No cap-dump. This is a very different machine than what we are using and he mechanically swapped only the two batteries every 3 minutes. They didn't have time for the run battery voltage to drop between cycles. This was all done at higher voltage and higher speeds with multiple air wound coils and faster, lower impedance switching. Totally different animal!

Gary Hammond,

[pearldragon]

Hi Gary,

Just a very brief post, will respond more elaborately at a later point:
If I just consider the first cycle of a series... then I now and then are already at COP > 1... Just started a new series today with on of my AGM batteries in the output, based on the Internal Resistance tests I choose AGM-A -> COP 1.08...
However, I'm sure that tomorrow in the second cycle it will be below 1... I'll finish this series (probably 3 or 4 tests) and then change to putting the Comparator (cap dump) in, and see if the COPs stay stable, after that I might dive into to battery swapper...

Best regards Rodolphe