Gary's "Complete Advanced" SSG Build

[pearldragon]

Hi Gary,

The batteries I show on my home page are NOT maintenance free. They had a pair of 3 individual caps connected together at the top and did occasionally require having some distilled water added. Here's a link to them. Look at the picture at the top of the linked page.

Yeah, I saw it… that’s why I was in doubt whether I should go for the Varta’s, since they looked exactly the same as yours, construction wise… but I found them too bloody expensive… I’ll test with the Yuasa’s first, see what they yield, maybe I’m lucky! If not, I’ll look for a cheaper alternative, but look construction wise like yours with the caps on top.

I think it is also best to let the batteries rest overnight between tests. They will recover some on their own, especially when using the Teslagenx charger. Start off by completely charging both batteries with the TGX charger. Then after an hour's rest, discharge the secondary battery (the one to be recharged with the SSG) using the CBA to remove 1 AH at a 1 amp rate.

Next let both batteries rest for another 10 to 12 hours, and then recharge the receiving battery with the SSG while monitoring the charge with the CBA.

If the TGX charger or SG charging has an effect that works even after the charger/SG is removed, I’d reckon that I’d wait also a while after I charged with it, rather than letting the output battery only rest for an hour before discharging. Otherwise the ‘after-effect’ of the charger/SG tries to charge while the CBA is discharging the battery. But that is just my theoretical reasoning. I mean if you say that out of your experience the net. result of discharging the output battery an hour after is has been charged, still results in the fact that the ‘after effect” of the SG/charger extends even beyond this time frame and continues charging after the CBA discharge is done, than that is probably how it is (regardless of my reasoning).
But if this is true, than I would expect that after the CBA discharge is done… and after a short rest, recharging again with the SG (or charger) would boost the ‘after effect’ of the previous charge cycle… Anyhow, I’ll try again doing one charge a day (or at least let 10 to 12 hours in between).

Regarding the discharge rate; although did discharge at my 24Ah FLAs @ 1A rate most of the times (0.6A for the 12Ah AGMs), I'd reckon that C20 should be what I can go for no? So that is 1.5A for my new 30Ah batteries.

I’ll post here once I have new results, hopefully somewhere next week.

If you try to hurry the process, you'll not be able to take advantage of the natural recovery process of the batteries. This recovery process is enhanced by using the TGX charger.

How does this than work in that battery swapper? There one battery is left out for a rest, but that rest is not 10 to 12h if I remember correctly, was more in the ballpark of an hour I think…

Best regards,
Rodolphe

[pearldragon]

Hi Gary,
While starting my initial cycles with my new Yuasa batteries, I’m reading through your posts here paying extra attention to the resistance in the trigger coil circuit (TCC) and I notice a huge difference between your resistance values and mine.

Gary - 100*** base resistors, 7 power windings
#7/#10 / 12 Ohms / 0.375” (9.5mm) gap / FLA batteries (with fans)
#29 / 26 Ohms / 0.310” (7.9mm) gap/ AGM batteries / 1.6-1.7A draw / 276rpm@end of run

Rodolphe - 100*** base resistors, 8 power windings (This is with the new PCB, matched transistors from TGX and also resistors from TGX)
C210713 / 36 Ohms / +/-8mm gap / FLA@output, FLA@input / 2.55-2.19amp draw / 255rpm@end of run / COP 63%
Cycle C210722* / 57.9 Ohms** / +/-8mm gap / FLA@output, FLA@input / 1.8-1.42amp draw / 264rpm@end of run / COP 43%
Cycle C210731* / 57.9 Ohms** / +/-8mm gap / AGM@output, FLA@input / 1.64-1.48amp draw / 260rpm@end of run / COP 85%
Cycle C210802* / 57.9 Ohms** / +/-8mm gap / AGM@output, AGM@input / 1.79-1.58amp draw / 253rpm@end of run / COP 80%

*See attachment from my post #119
**3x12 Ohm + 3x8 Ohm, so in theory 60 Ohm
***UPDATED 2021-08-16 -> 100 is incorrect, this should be 470, see posts #123/#124

What I find peculiar:
-If I compare C210713 with your #7/#10, where we both use FLAs @ input & output, although I have a way higher base resistance, my amp draw still is way higher too…
-If I compare C210802 with your #29, where we both use AGMs @ input & output, I have a way higher base resistance (almost 5x!!), to get my amp draw somehow in your ballpark….
-Furthermore, comparing my C210713 to C210722, you see that my COP goes up when I use FLAs, with a lower base resistance (but a way higher amp draw)…

First cycles with the new FLAs(Yuasa) -> / 57.9 Ohms / 1.8-1.6 amp draw

What does it say, that your TCC resistance is so much lower than mine? That when I test with batteries of 3 different brands, they all need such high resistance in my TCC to get the amp draw somehow in the ballpark of 1.6amp draw? Is this just a major coincidence, that I just happened to buy batteries with an internal resistance/structure that results in this high TCC? Or is did I still overlook something in my machine that causes this difference between your machine and mine??

Best regards,
Rodolphe

[Gary Hammond]

Hi Rodolphe,

Hi Gary,
While starting my initial cycles with my new Yuasa batteries, I’m reading through your posts here paying extra attention to the resistance in the trigger coil circuit (TCC) and I notice a huge difference between your resistance values and mine.

Gary - 100 base resistors, 7 power windings
#7/#10 / 12 Ohms / 0.375” (9.5mm) gap / FLA batteries (with fans)
#29 / 26 Ohms / 0.310” (7.9mm) gap/ AGM batteries / 1.6-1.7A draw / 276rpm@end of run

My base branch resistors are seven matched ones of 470 ohms, not 100 ohms each.

If you have a new complete kit from Teslagenx it should also have 470 ohm resistors. http://teslagenx.com/kits/tx-sg8.html?category=kits

Rodolphe - 100 base resistors, 8 power windings (This is with the new PCB, matched transistors from TGX and also resistors from TGX)
C210713 / 36 Ohms / +/-8mm gap / FLA@output, FLA@input / 2.55-2.19amp draw / 255rpm@end of run / COP 63%
Cycle C210722* / 57.9 Ohms** / +/-8mm gap / FLA@output, FLA@input / 1.8-1.42amp draw / 264rpm@end of run / COP 43%
Cycle C210731* / 57.9 Ohms** / +/-8mm gap / AGM@output, FLA@input / 1.64-1.48amp draw / 260rpm@end of run / COP 85%
Cycle C210802* / 57.9 Ohms** / +/-8mm gap / AGM@output, AGM@input / 1.79-1.58amp draw / 253rpm@end of run / COP 80%

Did you replace the 470 ohm branch resistors in the kit with the 100 ohm resistors you used in previous tests? We had a lot of discussion about branch resistor values back in posts 55, 56, 63, 64, 78, 82, 87, 90, 91. 92, and 93. Attachment 3 in your post #114 appears to show 470 ohm resistors installed on the new board as best I can tell.

What does it say, that your TCC resistance is so much lower than mine? That when I test with batteries of 3 different brands, they all need such high resistance in my TCC to get the amp draw somehow in the ballpark of 1.6amp draw? Is this just a major coincidence, that I just happened to buy batteries with an internal resistance/structure that results in this high TCC? Or is did I still overlook something in my machine that causes this difference between your machine and mine??

Assuming that you accidentally misspoke saying 100 instead of 470 ohms, the biggest other difference I see that would result in your having more base resistance than me, would be that you are using 8 power windings on your coil, while my coil only has 7 power windings. The extra branch circuit in parallel with the other 7 branch circuits will require extra common base resistance to throttle down the total current draw from the primary battery.

One additional potential difference could be the DC gain of the matched transistors between my board and your board. Even though all the transistors on each board are all matched, the gain values of one matched set are probably different than the gain values of a different matched set.

Another difference could be the accuracy, type (analog or digital), and method of attaching our ammeters to measure current flow from the primary battery. DC pulsed current with a back spike and induced wave form is difficult to measure as an "average" value. These values are approximations at best and may well be 30% or more in error. It also makes a difference whether you measure current draw on the positive lead or the negative lead. These values will differ because of the inductive kick back (i.e. -- radiant energy).

Gary Hammond,

[pearldragon]

Hi Gary, Thanks for your input!

Assuming that you accidentally misspoke saying 100 instead of 470 ohms, the biggest other difference I see that would result in your having more base resistance than me, would be that you are using 8 power windings on your coil, while my coil only has 7 power windings. The extra branch circuit in parallel with the other 7 branch circuits will require extra common base resistance to throttle down the total current draw from the primary battery.

You’re compete right; this new PCB has 470 Ohm base resistors not 100 Ohm; I misspoke/miswrote. (My old PCB had the 100 Ohms. I switched them around in my head, sorry for causing confusion.)

Regarding your 7 base resistors in parallel vs my 8 resistors, see attachment:
At the moment my total resistance is 116.7 Ohm. If I use that as the total resistance that I would have with a 7 power coils, and then calculate back how much my variable/fixed resistor value would be, I come to 49.5 Ohm, that is still almost double of what you have in post #29 and 4 times more than in #7 and #10…. Seems a lot to me…

One additional potential difference could be the DC gain of the matched transistors between my board and your board. Even though all the transistors on each board are all matched, the gain values of one matched set are probably different than the gain values of a different matched set.

This is very true, but also there a factor 2 seems a lot to me: I just quickly looked at my box of transistors that I used for the matching that I did myself for my original PCB: lowest value =100 highest value is 120, so that is a 20%. But that is 20% in the extreme values, meaning that most values are in between, so less one might expect in general that the spread between matched sets is smaller than 20%.

Another difference could be the accuracy, type (analog or digital), and method of attaching our ammeters to measure current flow from the primary battery. DC pulsed current with a back spike and induced wave form is difficult to measure as an "average" value. These values are approximations at best and may well be 30% or more in error. It also makes a difference whether you measure current draw on the positive lead or the negative lead. These values will differ because of the inductive kick back (i.e. -- radiant energy).

Also a good point. I have a analog meter, 2.5% accuracy type (applies of course for normal current), connected to in the positive(+) wire.
But if indeed the accuracy for measuring current at the input can be 30% off (or more….) than only COP of 1.3 (or more) would start becoming of interest… But to be really sure, you’d need a battery swapper/carousel so the machine keeps on going (with a load)…

Regarding that battery swapper/carousel, I still have a question about that from my post 121, at the bottom. The question relates to the process that you described by leaving 10h to 12h between cycles:

How does this than work in that battery swapper? There one battery is left out for a rest, but that rest is not 10 to 12h if I remember correctly, was more in the ballpark of an hour I think…

Lastly, what do you make/say of this observation:

Furthermore, comparing my C210713 to C210722, you see that my COP goes up when I use FLAs, with a lower base resistance (but a way higher amp draw)…

Did you experience the same?

Best regards,
Rodolphe
124 - 2021-08-12 - Attachment 1.pdf

[Gary Hammond]

Hi Rodolphe,

Regarding your 7 base resistors in parallel vs my 8 resistors, see attachment:
At the moment my total resistance is 116.7 Ohm. If I use that as the total resistance that I would have with a 7 power coils, and then calculate back how much my variable/fixed resistor value would be, I come to 49.5 Ohm, that is still almost double of what you have in post #29 and 4 times more than in #7 and #10…. Seems a lot to me…

Your observation here is basically correct. The only things I can theorize that would make sense to me would be that your machine has a higher trigger coil voltage and/or higher transistor gain.than mine?

I do know, however, that the base resistances on my machine are the nearly same as those shown by both John Bedini and Peter Lindemann in their presentations.


Looking back at your post 119 attachment 1 the test labeled C210725 is very interesting and close to the results I was getting at .82 COP. The 24AH FLA battery on the primary had been charged on the TGX charger and started the cycle at 12.95 volts. This is similar to my original test where my primary was two 13AH FLA batteries in parallel starting out at 13.08 volts. (These batteries were well conditioned from dozens of repeated chargings with the TGX charger.) Your primaries started out under load at 12.4 volts and dropped to 12.29 volts under load at the end. Mine started out under load at 12.54 volts and only dropped to 12.52 volts at the end. Your RPM started at 240 and ended at 260. (From memory, not recorded) my RPM started at 215 and finished at 265 pulling two fans as added load. Your average current draw was 1.55 amps measured by analog meter on the positive battery lead,.and mine was 1.6 amps measured by analog multimeter on the negative battery lead. You have an isolation diode between the negative battery leads and I had mine removed for faster charging.

My suggestion would be to get your 24 AH FLA's better conditioned, use two of them in parallel for the primary, temporarily remove the isolation diode, and then charge back a single 12 AH AGM to see if your results improve using the same 57.9 ohm primary resistance.

Regarding that battery swapper/carousel, I still have a question about that from my post 121, at the bottom. The question relates to the process that you described by leaving 10h to 12h between cycles:
How does this than work in that battery swapper? There one battery is left out for a rest, but that rest is not 10 to 12h if I remember correctly, was more in the ballpark of an hour I think…

My 10 to 12 hour rest was only meant to show how I did my initial test and the possible effect that had. This doesn't fully apply to the battery swapper as the cycle times are only from a few minutes to maybe an hour depending on setup and batteries used. A rest period is important but is somewhat compromised in the swapper.

Lastly, what do you make/say of this observation:
Furthermore, comparing my C210713 to C210722, you see that my COP goes up when I use FLAs, with a lower base resistance (but a way higher amp draw)…

Did you experience the same?

Your SLA's are more conditioned than your FLA's because they have been charged on the SSG for more cycles than your FLA,s. They also sit at a higher voltage which means that they will recharge faster and not pull the voltage of the primary battery as low. And yes, using more current will charge faster, but current charging produces heat and wastes energy. The CG mode is a combination of current and radiant which is a little tricky to get balanced for the best results.

The FLA,s have a lower internal resistance so will work better on the input side than the SLA,s even though they are not as well conditioned yet. Low input resistance in the primary battery and primary circuit are very important. These are things I've learned from my machine and batteries. That's why I suggested what I did above.

Gary Hammond,

[pearldragon]

Hi Gary,

My suggestion would be to get your 24 AH FLA's better conditioned, use two of them in parallel for the primary, temporarily remove the isolation diode, and then charge back a single 12 AH AGM to see if your results improve using the same 57.9 ohm primary resistance.

Although in my previous post I did too referred to your post #7 & #10, for now I prefer to stay focused on your results from #29, since it is easier for me to more closely match those parameters, since I do not have fans (yet). In that post #29 you used the isolations diode(s), so prefer to keep them in for now: You explained to me earlier why they will improve the COP a bit, but before adding a “Turbo” to my engine, I first want to find out why I can’t get the max HP out of it with Natural Aspiration . The only reason that I referred to posts #7 & #10 is because I wanted to have a reference boundaries for where you were with your TCC resistances, drawing the conclusion that we’re far apart (as described in my previous post).
Matching your results of post #7 & #10, is definitely in the pipeline, also using my “spare main coil” underneath my main coil, but as mentioned, I want first max HP out of my current engine before adding more parameters to the mix. In your post #29 you got a COP of 0.99, While I’m still at 0.85 at the absolute max (C210731), so still really far off.
But part of your suggestion I already tried, see cycles C210731/C210732/C210733/C210801: there is have FLA1&2 paralleled at the input, with the highest input voltages under load and at rest as a result of this parallel setup.

Regarding the conditioning
I’m pretty sure my FLAs (LA1&2) where pretty well conditioned after cycle C210703. Before C210603 I charged (and discharged) battery LA1 a couple of times with the TGX charger, and did the same with LA2 before cycle C210701. Once one of these battery is used for the output battery, it took about 3 cycles for the COP to get relatively stable (which I conclude as conditioned). I see the same process with my new batteries at the moment; after a couple of initial TGX charges and then putting one of the batteries at the output, within 3 cycles the COP stabilizes.
Another indication (to me) that they are conditioned is that the batteries that I use for the input do not go up any further in input voltage.

My 10 to 12 hour rest was only meant to show how I did my initial test and the possible effect that had. This doesn't fully apply to the battery swapper as the cycle times are only from a few minutes to maybe an hour depending on setup and batteries used. A rest period is important but is somewhat compromised in the swapper.

Ok, all clear, thanks.

Internal resistance

The FLA,s have a lower internal resistance so will work better on the input side than the SLA,s even though they are not as well conditioned yet. Low input resistance in the primary battery and primary circuit are very important. These are things I've learned from my machine and batteries. That's why I suggested what I did above.

Since I’m under the impression that I the FLAs where conditioned already after the cycles mentioned above, and comparing cycles with FLA@output versus AGM@output (C210725 up and including C210730), I drew initially the conclusion that the AFMs have a lower internal resistance; they yield the highest COP. I then expected that paralleled AGMs@input and AGM@output would yield the highest COP, but it stuck at 0.8 (C210802). I do have to note that I did only one cyle in this setup, and when seeing the COP, didn’t saw a need to investigate/refine/verify the COP further with more cycles. However, while typing this, remembering what you wrote in #120, I probably will redo a couple of cycles with this setup. Will update once done.

Internal resistance in graphs
Another indication regarding the internal resistances might be are the graphs indicated in attachment 1. Note that in post # 119 I mentioned a COP of 0.64 (C210709) for the FLA (LA2 / VMF), this was with lower TCC resistance. So in attachment 1 took the highest measured COPs of the 3 different batteries taken from cycles with the same TCC resistance. As a comparison, I also included the graph the graph from your website. As you can see, your charge graph has the most resemblance with my AGM battery graph.
As a further comparison I also included some discharge graphs. Here I cannot make a good comparison with your discharge graph, since only my AGM has the same capacitance as the battery in your graph, but you discharged at 1A, and I discharged my AGM at 0.6A (=C20), see attachment 2.

Results with new batteries
I finalized some cycles with the new batteries, where I constantly used 2x 30AH (=60Ah) at the input, and a singe 30Ah at the output, seen attachmen3. The COP falls in between my previous FLAs and the AGMs.

More in dept battery analysis
In an attempt trying to explain where the difference in performance com originate between my old FLA and my new ones (and yours). I contacted the supplier/looked up the information:
LA1&2 24Ah/293CCA (VFM 52805 / 12N24-4) -> Lead-antimony (Sb)
S1,2,3 30Ah/330CCA (Yuasa YBX U1) -> Lead-calcium (Ca)
http://www.necon.co.za/batttech.php?w=1680&h=1050#calc

Amp Meter analyses
Considering what you wrote about the amp meter in post #123 & #125, I tested with a different amp meter (cycle C210808), see attachment 3: more or less same COP. I then read a bit about amp meters, learning that there are “moving iron” and “moving coil” types, and looked up which ones I had: both “moving coil” (more accurate, and generally only used for DC current). Which type do you have?
This morning I then started a new cycle (C210809) where I put my original amp meter back in, but this time in the negative wire (instead of in the positive). Since Sunday I did no test at all, the batteries have been resting a bit longer than normal and will do another cycle in the evening and if necessary another one in the morning as well, but it looks like that also having the amp meter in the negative or positive wire does not have a noticeable impact on the COP, at least not with the amp meters that I have…
UPDATE 22:31, just finished another run with the amp meter in the negative wire: COP 0.67, so same as in cycle C210805, so no difference whether I have the amp meter in the positive or negative wire.

Thanks once again for spending so much time analyzing all these parameters/results with me Gary!!

Best regards,
Rodolphe
126 - 2021-08-16 - Attachment 1, 2.pdf
126 - 2021-08-16 - Attachment 3.pdf

[Gary Hammond]

Hi Rodolphe,

Amp Meter analyses ....................Which type do you have?

On my initial tests I was using a FET multimeter with moving coil. (The FET multimeter amplifies the input with a FET. It has selectable, built in shunt resistors for scale selection. It's normal input impedance for voltage measurement is 10meg ohms.) I have since been using a dedicated 0 to3 amp ammeter with a moving coil. Don't remember for sure, but I think I used the dedicated one in the post 29 tests.

[pearldragon]

Hi Gary,

Another thought for the (near) future:
Could I put two input batteries in series, so they add up to 24V, and then adjust the TCC resistance so that the amperage is still in the range as in my current cycles (between 1.6A-2A)?

Best regards,
Rodolphe

[Gary Hammond]

Hi Rodolphe,

Could I put two input batteries in series, so they add up to 24V, and then adjust the TCC resistance so that the amperage is still in the range as in my current cycles (between 1.6A-2A)?

Yes. The SSG will run on 24 volts by adjusting the common base resistor to a higher value. I think you will also need to charge the output batteries in series at 24 volts as well. I've only ran mine at 24 volts a couple of times, so have no experience to share with you on this. Some people have gone up to 36 volts and the inductive spikes can get pretty high requiring extreme caution to keep from blowing the transistors. The branch resistors may also have to be a higher value as well?

I think RS has run several machines at these higher voltages, and could give you way more guidance than I can.

Gary Hammond,

[pearldragon]

Hi Gary,

24V system

I think you will also need to charge the output batteries in series at 24 volts as well.

Is this because otherwise in section C of the signal (see attachment 1) the output battery gets charged with 24V, which might be too much if this is for a prolonged time frame? (A prolonged time frame as in section C compared to the spike (X) which voltage is higher, but just for an instant). If that is correct, I see three options:
1-have 2 batteries in series as well in the output (as you suggested)
2-Use the machine in radiant mode
3-Use the machine in capacitor dump/comparator mode
(4-Use it in SG mode with capacitor dump/comparator mode connected behind it)

Some people have gone up to 36 volts and the inductive spikes can get pretty high requiring extreme caution to keep from blowing the transistors.

What precaution could I take to protect these specific transistors other than making sure that I increase the TCC resistor accordingly to maintain the same current as I have with the 12V system? If the voltage spike itself will become an issue, I could change to higher voltage transistors, but that will probably go at the sacrifice of amp rating…

The branch resistors may also have to be a higher value as well?

Whether I limit the current through (the base of the transistor with the branch resistor or the TCC base resistors, has the same net effect, or am I overlooking something here?
The few times you did run it with 24V, what where you’re findings COP wise? I’d expect it to go up, the higher the input voltage is (but as you mentioned yourself already there will be a limit to which the transistors will hold).

I’ll ask to RS to respond on this issue too.

Another future question: 2nd coil underneath the main coil
If I want to use the my spare coil under my main coil (as you did in post#7), and connect it to my comparator, can I just connect the trigger & main coil wires together on one end, do the same on the other end of the coil, put diodes in and then connect it straight to the comparator? See attachment 2. Or do I need to add more components/circuitry?

Update on testing/cycles
Taking into account what you said about the charging/discharging process in post# 120, I use a slight variation:
1-Do a charge cycle with the SG
2-wait for an hour
3-discharge 1Ah from the output battery with the CBA
4-wait for a couple of hours (in case I did a cycle in the morning) or let all rest overnight (in case I did a cycle in the evening)
5-charge the input batteries with the TGX charger for an hour
6-wait for an hour
7-Do a charge cycle with the SG (so back at the start, 1)

I’m still testing with the Yuasas, since with the above mentioned procedure and their big capacity I get fairly stable results. I’m now re-evaluating some parameters so I can determine their impact much more clearly now I know the charge/discharge procedure has impact too, which I now more accurately control.
Increasing the gap to 10mm I reached 0.70 COP (highest measured so far with the Yuasas), but increasing it to 12mm brought it down quite strong to 0.63. In the next run I want bring the gap down to 6mm. Next on the agenda is than re-evaluating the impact of my different rotors. When that is all done, I want to switch back to AGMs (at least in the output).


Best regards,
Rodolphe
130 - 2021-08-19 - Attachments.pdf

[pearldragon]

Hi Gary,

Rotors
Before my holidays, I finished up some testing with the different rotors/gaps, see attachment:
As you can see, Rotor 3 (21 magnets, 22mm wide, 25% ON-time), cycle C210814, yields more or less the same results as Rotor 1 (24 magnets, 20mm, 25.1% ON-time), cycles C210821/C210822.
Rotor 2 (20 magnets, 20mm, 21% ON-time) performed less, but the current draw was less too, so it might perform fairly equal if I’d reduce the base resistance. But I did not since I want to keep that stable for now with the fixed resistances.

Output Battery Size
Reading on your website and the different forum threads where you posted about your performance, I saw that you have a 13Ah battery.
I think this is another key parameter in the performance: the output battery should be sized to the machine. As you suggested already, I’ll will test again with one of my 12Ah AGMs at the output. With all the other parameters fairly optimized (mainly with the procedure that I use now mentioned in the previous post), I’m curious to see of I will get over the 85% that I previously achieved with it.
But as we discussed already, size is not all; also the internal structure of the battery matters: My 30Ah Yuasa outperforms my 24Ah VMF… So apart from the AGM that I’ll use in the output, I also ordered this battery:
https://www.yuasa.de/batteries/motor...n/yb14l-a.html

I actually ordered this one too to do even further comparisons
https://www.yuasa.de/batteries/motor.../12n14-3a.html
but got a message later on that they did not have it anymore, so will first just test with the above mentioned battery and the AGM.

24V & coil under main coil
As for as possible, would you still be able to give some feedback/comment on the 24V system from the previous post, but more importantly on the “2nd coil underneath the main coil” from the previous post?

Best regards,
Rodolphe
131 - 2021-09-13 - Attachments.pdf

[Gary Hammond]

Hi Rodolphe,

24V & coil under main coil
As for as possible, would you still be able to give some feedback/comment on the 24V system from the previous post, but more importantly on the “2nd coil underneath the main coil” from the previous post?

Is this because otherwise in section C of the signal (see attachment 1) the output battery gets charged with 24V, which might be too much if this is for a prolonged time frame? (A prolonged time frame as in section C compared to the spike (X) which voltage is higher, but just for an instant). If that is correct, I see three options:
1-have 2 batteries in series as well in the output (as you suggested)
2-Use the machine in radiant mode
3-Use the machine in capacitor dump/comparator mode
(4-Use it in SG mode with capacitor dump/comparator mode connected behind it)

Except for option 1, I have not tried doing any of what you suggest with 24v input. .......................When running on 24v and outputting into a 12v battery (options 2,3,&4), my intuition suggests the current flow would be so great as to be destructive to the circuit and/or batteries. If you try any of this, I would interested in seeing how it turns out.

You can directly charge a 12v battery from a 24v source using a proper resistor with no SSG required, but this would be a waste of energy as heat.

Another future question: 2nd coil underneath the main coil
If I want to use the my spare coil under my main coil (as you did in post#7), and connect it to my comparator, can I just connect the trigger & main coil wires together on one end, do the same on the other end of the coil, put diodes in and then connect it straight to the comparator? See attachment 2. Or do I need to add more components/circuitry?

I didn't use a spare SSG power coil for this at all! ................... I used a single strand generator coil with 4026 turns of #23 wire at 38.8 ohms resistance with it's core spaced approximately 3/8" below the SSG power coil. This was inductive coupling thru an air gap. It's unloaded voltage image in the o-scope looked exactly like the voltage image across the power coil. Under load it looked different because I tapped off the output with a voltage doubling diode circuit supplying current to 42 LEDs.

If you were to use a spare power coil, you would need to connect all the windings (including the trigger winding) in series in order to get a high enough voltage to be useful. I wouldn't connect it to the SSG output at all. But like you show, it could drive a separate, small comparator to charge another small battery. Or you could also use it to back pop the run battery for reduced run current draw.

The circuit you have shown is a half wave rectifier. I would either use a full wave rectifier or else a voltage doubling diode/cap arrangement like I used.

Gary Hammond,

[pearldragon]

Hi Gary,

24V system

Except for option 1, I have not tried doing any of what you suggest with 24v input. .......................When running on 24v and outputting into a 12v battery (options 2,3,&4), my intuition suggests the current flow would be so great as to be destructive to the circuit and/or batteries. If you try any of this, I would interested in seeing how it turns out.

The component most prone to damage is the transistor I guess, I would dial the amperage down the same level as with a 12V system, by adding more resistance to the trigger circuit. I need to check the specs of the transistor if it can deal with the 24V.
In the options 2, 3 & 4 I thought that the output battery would not experience the 24 volt, so no potential damage to the output battery I’d reckon?
I’d like to gather some more information on these higher volts systems, but considering testing with it in the future, will update you then about my findings of course. I hope that RS will chip in at some point.

2^nd Coil under main coil

I didn't use a spare SSG power coil for this at all! ................... I used a single strand generator coil with 4026 turns of #23 wire at 38.8 ohms resistance with its core spaced approximately 3/8" below the SSG power coil. This was inductive coupling thru an air gap. It's unloaded voltage image in the o-scope looked exactly like the voltage image across the power coil. Under load it looked different because I tapped off the output with a voltage doubling diode circuit supplying current to 42 LEDs.

Clear, thanks. Why did you use 3/8” gap and not a gap ‘as small as possible’, say 1 or 2mm?

If you were to use a spare power coil, you would need to connect all the windings (including the trigger winding) in series in order to get a high enough voltage to be useful. I wouldn't connect it to the SSG output at all. But like you show, it could drive a separate, small comparator to charge another small battery. Or you could also use it to back pop the run battery for reduced run current draw.

If the magnetization of the secondary (in my case a spare power) coil/core placed underneath the main core/coil would be of the same magnitude as this main coil, I’d assume the spike would be in the same amplitude range to no? And so I reasoned that the windings on this secondary power coil/core could stay parallel, rather than connected into series. Bur reading your answer I guess there is an error in this reasoning and I should connect the windings in series.

Input connections
In an earlier post you mentioned that the way you connected your input connections made a big difference in the performance of your SG. Could you tell me percentage wise what kind of difference this made (ballpark), and also how you connected them before and after?
As you might remember I changed some of the connections before, but without a significant change. Until recently I connected to my input and output batteries via alligator clips, and then at some point switched to banana plug connections, but without any noticeable difference,see attachment.

Update about results/performance
After I came back from holidays I tried started to test again with 2x 30Ah parallel batteries at the input and one of my AGMs 12Ah at the output: first run yielded 0.91 COP, highest COP measured so far, but when I did another run the next morning and again in the evening, COP dropped to somewhere in the 0.7x. I then stopped testing since my Radiant charger broke down so it seems, in the process with TeslaGenX of shipping me a new one. Will try again/continue once the new charger arrives. As mentioned before I also bought a new 14.7Ah FLA battery to test with in the output, but because of the breakdown of the charger, haven’t got around to test with it yet.

Li-ion
Have you done any testing with Lithium (Li-Ion) Batteries?

Light bulb test
In EFV DVD 2, @ +/- 2min30sec, JB shows a little light bulb test/experiment. I wanted to verify this, but now am waiting for the charger. I was wondering if you’ve tested this yourself/verified by any chance?

Best regards,
Rodolphe
133 - 2021-09-22 - Attachment.pdf

[Gary Hammond]

Hi Rodolphe,

Clear, thanks. Why did you use 3/8” gap and not a gap ‘as small as possible’, say 1 or 2mm?

I didn't want to couple directly and place an added load on the main coil. I was only seeing if I could capture some of the stray, lost flux that would otherwise be unrecoverable. I got the results I was looking for, so didn't experiment any with the gap. I'm only guessing at the gap. May be as small as 1/4" or somewhere between 1/4" and 1/2". The cores of the coil spools extended maybe 1/8" to 1/4" past the spools through a common hole in a 3/4" board with the spools directly across from each other. It was easy to mount the spools directly to the board, so that's what I did.

Input connections
In an earlier post you mentioned that the way you connected your input connections made a big difference in the performance of your SG. Could you tell me percentage wise what kind of difference this made (ballpark), and also how you connected them before and after?

As best I recall, I started out with 12AWG stranded wire with crimped and soldered ring terminal ends at the terminal strips of the machine. And on the battery ends I had crimped and soldered battery clips like used on battery chargers. I then changed to shorter 8AWG stranded battery leads with large ring tongue terminals soldered on each end, and bolted the battery ends directly to the battery terminal lugs with 1/4" bolts.

Don't remember the results exactly, but think I got maybe 10% faster wheel speed, 10 to 30% higher voltage spikes on the o-scope, and maybe a 10 to 20% reduction in charging time. Been a while since I did this, so am a little fuzzy about the differences I observed. I just know that it was well worth the small effort required to reduce the impedance losses.

Li-ion
Have you done any testing with Lithium (Li-Ion) Batteries?

I haven't experimented with Li-ion batteries at all. John Bedini did and reported that they don't like being charged by the SSG at all. They require a highly monitored hot charger and will not tolerate much over voltage. The temperature and upper charging voltage are critical. If exceeded, the battery may be permanently damaged or catch fire.

He did, however, experiment with LiFePo batteries and developed a special circuit to charge them with either an SSG or a plugin charger. These batteries are superior to Li-ions, but are very expensive to buy.

Light bulb test
In EFV DVD 2, @ +/- 2min30sec, JB shows a little light bulb test/experiment. I wanted to verify this, but now am waiting for the charger. I was wondering if you’ve tested this yourself/verified by any chance?

No. I don't have that DVD and have not done that experiment. Is that where the bulb is placed in the trigger lead, or where he checks the output of the machine?

Gary Hammond,

[pearldragon]

Hi Gary,

2^nd Coil under main coil

If you were to use a spare power coil, you would need to connect all the windings (including the trigger winding) in series in order to get a high enough voltage to be useful. I wouldn't connect it to the SSG output at all. But like you show, it could drive a separate, small comparator to charge another small battery. Or you could also use it to back pop the run battery for reduced run current draw.
The circuit you have shown is a half wave rectifier. I would either use a full wave rectifier or else a voltage doubling diode/cap arrangement like I used.

1-)Did I indicate the FWB and the doubling voltage arrangement (DVA) correctly in attachment 1?
2-)If correct, what advantage would the DVA have over the FWB version? Higher voltage output?
(The downside of the DVA would be that I need to find the right size for the caps for the DVA and the comparator, while with the FWB version I only need to size the cap or the comparator).
3-)Why would you not connect the output of the comparator directly to the output battery of the SG, see attachment 2? Would the two systems (CG mode & comparator) interfere with each other, sending CG output into the comparator and visa versa?

I didn't want to couple directly and place an added load on the main coil. I was only seeing if I could capture some of the stray, lost flux that would otherwise be unrecoverable. I got the results I was looking for, so didn't experiment any with the gap. I'm only guessing at the gap. May be as small as 1/4" or somewhere between 1/4" and 1/2". The cores of the coil spools extended maybe 1/8" to 1/4" past the spools through a common hole in a 3/4" board with the spools directly across from each other. It was easy to mount the spools directly to the board, so that's what I did.

What you’re saying is: if the gap is too close, you couple the coils, and the drag/load will be noticeable in the SG output performance, whilst when sufficient space is between the coil, you only capture stray flux, and the effect is not noticeable on the SG output. And because you/I only capture the stray flux, I do need to put the windings in series to get an high enough output of the 2^nd coil. Correctly summarized like that?

Connections

Don't remember the results exactly, but think I got maybe 10% faster wheel speed, 10 to 30% higher voltage spikes on the o-scope, and maybe a 10 to 20% reduction in charging time. Been a while since I did this, so am a little fuzzy about the differences I observed. I just know that it was well worth the small effort required to reduce the impedance losses.

I understand what you say. Since I didn’t notice any difference with the different connection options or wire length so far with my 10AWG wire, I’m not sure if I can get much better results with my 10 AWG wire. However, going to 8 AWG might make a difference.

Lithium / Li-ion

I haven't experimented with Li-ion batteries at all. John Bedini did and reported that they don't like being charged by the SSG at all. They require a highly monitored hot charger and will not tolerate much over voltage. The temperature and upper charging voltage are critical. If exceeded, the battery may be permanently damaged or catch fire.
He did, however, experiment with LiFePo batteries and developed a special circuit to charge them with either an SSG or a plugin charger. These batteries are superior to Li-ions, but are very expensive to buy.

I asked Erik from TGX the same question a while ago and he said that hi had very little experience with it, that the Li-Ion batteries didn’t seem to like the pulse/cap dump charging.
Via via I heard from somebody else too that the LiFePo batteries seemed to be able to handle to charge with an SG, don’t have details yet.

Light bulb test

No. I don't have that DVD and have not done that experiment. Is that where the bulb is placed in the trigger lead, or where he checks the output of the machine?

No, this is where he connects one end of the little light to a frame (not connected to the battery terminals) and the other end of the light he touches on of the plastic caps of the battery and the light goes on, see attachment 3. The batteries are charged with a solid state charger (so it seems).

Best regards,
Rodolphe
135 - 2021-09-28 - Attachments.pdf