Photo analysis
In the attached schematic, I scetched how Hack could have done that, whereby I speculate that the RF choke *could* have been that extra winding which seems to be present in the box. I only have like 3 high-res pictures, but I noted some things:
1) The 3d diode in the row attached to the collectors of the transistors appears to be isolated. There's a plastic ring and some tape or something at the other side.
2) The thick wires from the primary going to the diodes are soldered pretty badly and underneath the shrink wrap it seems there is a solid connector.
3) There seem to be older wires, covered with linen, and newer, modern wires. All the older ones have been soldered, while the newer connections have more modern non-soldered connectors.
4) From EV38.jpg, it looks like there is a 5th diode, to which the wire from the diode with the red wire to the snubbler caps leads to.
5) It seems like the other primary connector has also been shrink-wrapped and thus tampered with.
IMG2915.jpg: The transistors are MJ10021 NPN Silicon Power Darlington Transistors with Base-Emitter Speedup Diode. The MJ10020 and MJ10021 Darlington transistors are designed for high–voltage, high–speed, power switching in inductive circuits where fall time is critical.
100_1106.jpg :
3 Resistors DALE RH-50 50W 150KOhm 1% 3 Capacitors .036 10% 2000V Diodes 1N 5388 8148 -> 200V zener 36.6 mA, 900mA surge, 5W
100_1184.jpg
The bar connecting the capacitors positives appears to be iron, while the others are copper. -> tampering. This makes no sense, given the separate fuses. -> capacitors were separate in the original
Capacitors are 78,000 MFD 40 VDC
There appears to be the same resistor (DALE) in the snubber circuit.
Another (primary?) lead is visible at the bottom of the transformer at the red block
100_1193.jpg
The core reads H7c1 T.... These are made by TDK and currently available on eBay, still 10 available:
http://www.ebay.com/itm/FERRITE-E-CORES-H7C1-EE-55-55-2-NEW-2-PAIRS-4-PCS-/310352364736?rmvSB=true
Images saved as H7C1_TDK_*
100_1195.jpg The top resistors appears to read: AMF L05 .02 Ohm 3% 8117
100_1195.jpg The snubbler resistor appears to be DALE RH-50 50W 30 Ohm 5%
100_1189.jpg Snubbler capacitors read: ELPAC 5.0 +/- 10%
So, these would be 5 uF.
Datasheets: https://www.westfloridacomponents.com/mm5/graphics/ds7/Elpac-Capacitors.pdf
Similar types:
https://www.tedss.com/2020006782 saved as: ZD8684 ELPAC capacitor 0.25uF 1600V Film Axial Part Number: ZD8684 Manufacturer: ELPAC Item ID: 2020006782
capacitance: 0.25 uf voltage: 1600 v tolerance: 10% lead length: 47.75 mm (1.880 in) lead type: axial lead style: straight package: bulk item per pack: 1 Diameter: 0.79 in, Length: 1.94 in
https://www.tedss.com/2020060136 Saved as: ZD4A505K EL PAC capacitor 5uF 400V Film Axial.jpg
Part Number: ZD4A505K Manufacturer: EL PAC Item ID: 2020060136
capacitance: 5 uf voltage: 400 vdc tolerance: 10% lead type: axial package: bulk item per pack: 1 Width: 0.6 in, Height: 0.9 in, Length: 1.95 in
https://www.tedss.com/2020024003 Saved as: Z5B475J EL PAC capacitor 4.7uF 50V Film Axial.jpg
Part Number: Z5B475J Manufacturer: EL PAC Item ID: 2020024003
capacitance: 4.7 uf voltage: 50 v tolerance: 5 package: bulk item per pack: 1 diameter: 0.4000 IN, length: 1.2000 IN
https://www.tedss.com/2020046724 Saved as: Z4A154K ELPAC capacitor 0.15uF 400V Film Axial.jpg Part Number: Z4A154K Manufacturer: ELPAC Item ID: 2020046724
capacitance: 0.15 uf voltage: 400 v tolerance: 10 % lead length: 50.8 mm (2.000 in) lead type: axial package: bulk item per pack: 1 Width: 0.24 in, Height: 0.35 in, Length: 0.63 in
http://www.bmisurplus.com/products/eb1693-elpac-zd6a504k-capacitor Elpac ZD6A504K Capacitor specifications: - .50 +/- 10% - 600V Saved as: Elpac ZD6A504K Capacitor.jpg
https://www.westfloridacomponents.com/K124APM01/2.16uF+200V+Axial+Polyester+Film+Capacitor+Elpac+ZB9637.html Manufacturer: Elpac Part Number: ZB9637 Capacitance (uF): 2.16 Voltage: 200 Tolerance (%): 5 Mounting: Through Hole Lead/Terminal Type: Axial Number Leads/Terminals: 2 Material: Polyester Color: Yellow Shape: Round (Cylindrical) Termination Method: Solder
http://www.tedss.com/B1A223H saved as: B1A223H ELPAC capacitor 0.022uF 100V Film Axial.jpg
Part Number: B1A223H Manufacturer: ELPAC Item ID: 2020046718
capacitance: 0.022 uf voltage: 100 v tolerance: 2.5% lead length: 76.2 mm (3.000 in) lead type: axial package: BULK item per pack: 1 Width: 0.13 in, Height: 0.2 in, Length: 0.38 in
http://www.tedss.com/2020007625 Saved as: ZA8401-2 ELPAC capacitor 3uF 600V Film Polyester Axial.jpg Part Number: ZA8401-2 Manufacturer: ELPAC Item ID: 2020007625
capacitance: 3 uf voltage: 600 v tolerance: 20% lead length: 42.16 mm (1.660 in) lead type: axial package: bulk item per pack: 1 Width: 1 in, Height: 0.87 in, Length: 1.75 in
Some Matt Jones posts:
http://www.energeticforum.com/185885-post4055.html
http://www.energeticforum.com/186078-post4061.html
http://www.energeticforum.com/263946-post4169.html
http://www.energeticforum.com/275869-post4192.html
http://www.energeticforum.com/281604-post4206.html
I wouldn't mind having a discussion but there is not much to talk about. The article said it all. The transformer can be tuned to perfection between the load and a capacitor to maximize the conversion of back and forth DC to a better AC sine so the transformer no matter what it is at its most efficient. The batteries need to be in very good shape.
If everything is working correctly and you are not asking for huge load the system will balance out for a long period of time. I found 300 amp hour batteries can supply a 100 watt load for real long time without any consumption from the battery. The trick is keeping the battery temperature stable at or around 72 degrees. If you get big temp swings then you throw everything out. This is also true in several systems I have come up with since then.
Internal resistance in a battery is very important.
But out of it all the newest systems I have built are tuned to the response of an isolated inverter. When I say Isolated I mean the AC side and DC side are not grounded in common which is what you'll find in cheap inverters. Like wise the grounds are earth ground on the AC side, or neutral like normal grid power.
Inverters SMA's Sunny Boy Islands are so capable of tuning into a Tesla switch type system it gets ridiculous. All you need is to be able to turn the pulse into a flat DC with no more than 1% ripple at any point. Its not hard though with 3 banks. Discharging, charging and resting. Generally if you able to get this system going you can calculate loss at switching cost, but you have to have big enough banks to take the power and raise the potential difference`while providing adequate current. This current also needs to be regulated to make sure the batteries are at a peak point of about 14.5 volt, or 29 volt, ect.
To truly do some real work you are looking at an investment of around 8 - 10 dollars a watt. This makes it hard in the face of solar which is well below 1 dollar a watt for whole systems, if you can install it yourself.
http://www.energeticforum.com/281745-post4211.html
To tall ya the truth I think he had it wired wrong. The best performance I have seen out of his setup came from driving an AC EI core in singular directional pulses. Instead going back and forth on seperate windings to drive a third winding you can pulse each side in the same direction according to the geometry of the transformer. Toroids don't work. They are design to suppress transient responses. EI Cores put out huge spikes. Coils work too but not as well, you need that magnetic loop. Laminated toroid AC core usually have silicon steel and that doesn't appear to work either. At least from what I have seen. On short runs (thats the most I have been able to do) I have seen all four batteries charge. Thats the good side of it... The noise alone from the transformer is near unbearable and radio for 100 yards +- is impossible. The worst thing though is the transistors or Fets fail real quick. But it works, I am just not up to the task of building a switch thats good enough to handle it.
I did this about 4 years ago and tried real hard to make it work but with no resolve and eventually ran out of money.
Later, while looking at Babcock's patent I could kind of see a way through all this but I didn't have the resources to try. It didn't follow the simple outline of the transformer setup shown in the Brandt article but It could work. Looping the inductive response back through the system at higher frequencies to emulate a low frequency pulse, similar to what happens in a modified sine wave inverter. It was either this or find that "Special Transistor" in the Brandt drawing that could handle that inductive response with out damaging heat.
So some of the tests I have done show weird results. The largest test I used 4 1500 amp hour starter bats. I was able to push 12 volt +- at 100 amp into one side of the transformer at a pulse width of about 30 ms per side with a 10 second off time between per side pulses and the output of the transformer was 10 volt at 150 amp for 35ms. The 10 volt +- was above the charging bank. The output on the third winding was at 24 volt +- dc at 54 amps with 38 ms pulse width. I built this one with 6x 400 volt 20 amp mosfet's per switch. The end result was a foul stench and a lot of smoke. But all the batteries went up after a 1 hour rest
I have always wished I could crack that nut and this is the first time I have ever told anyone publicly the results. In my mind it could have been as simple as bad wiring on the transformer, if he had one in the car and some kind of super transistor that could handle the noise. But who really knows.
http://www.energeticforum.com/281774-post4212.html
So couple of things come to mind right off the bat. Just pointing them out not criticizing. That transformer looks a little loose and wound heavy like that may have a little to much resistance while having a lot of induction. Depending on the material in the core you might need higher speeds but the induction and resistance in the windings will inhibit this hence 40 hz, when generally I have seen speeds ranging from 60 to 400hz which seems to work best depending on the application. Thats also with an EI core. I have not seen much luck with toroids and they are pain to wind as you probably now know. LOL
Even though your load is low one thing you need to look at is how much current passes on the primary windings. Input and output. So a good transformer passes 10 amp to do 5-7 amp of work on the output winding. Its never equal but should be 50% at least.
The transistors you are using may suit the system but I would shy away from using to much resistance on the base. Let the base catch a little current. This also may cool down the transistor overall. After all they can handle 2 amp (MAX) on the base. Not sure but you may not be opening up the junction enough. To know that though you have to push it and see its performance heat wise. Also the transformer may effect the heat situation with spikes as per your induction level in the windings. Lower the amount of windings to decreases the size of the spike. Toroids don't require as much winding to couple the magnetic field. Evenly wound shorter wires perform better. If the transformer is the culprit for heat mosfets won't change this. Transients are also the reason I encourage DC output to a capacitor. The transient will leave the transformer on the output line and be negated by the cap. As far as the diodes go I like "No Recovery Time" Diodes which are generally Schottky but not always as they quikly move spikes away, No heat and in situation where they are over rated for the job you get no voltage drop. This is a typical one I use. I would run that up to 5 amp, and at 5 amp it would drop the voltage only a small portion at 1 amp you would see no voltage drop. Depending on the application and my expectation I would wind the transformer output to compensate for that drop. I always size my diodes high to avoid voltage drop, after all this is research so the budget should have to count, but I know it does.
Also what people never thin of is trying to capture the OUTPUT power off the transformer and shuttle it to the batteries to prolong runtime. In your case with the small batts that might make all the difference. Its trick though.
Keep me up to date, or post more questions and results.