Donald Smith Devices too good to be true

[dragon]

You could use an SCR in the same configuration at almost any amperage needed... also, the SSR's are quite inexpensive with large amp capacities

[Wistiti]

Well, I have to find another flyback. Mine seem to not output enough. If someone have a suggestion for another easily available hv source im all ears open. I have another Tv to dismantle...

Bruce I see in your last schematic you use lithium battery instead of cap... witch kind are them?

Thank you!

[Wistiti]

No cap on secondary at the moment.

You made the same mistake i did in the beginning.. You mounted Hv-wire in wood.

Mwtj, Tswift, What have you experienced with wood as the support of the project? Does it conduct the hv?

What should I use instead?

thank you!

[tswift]

You could use an SCR in the same configuration at almost any amperage needed... also, the SSR's are quite inexpensive with large amp capacities

I think maybe you're thinking of a crowbar circuit, which looks very similar. I haven't actually rigged up either circuit to test it, but I think with the SCR it will just gate on into full conduction and not turn off or proportion at all. Using a bipolar transistor it will act like a bigger, more powerful zener diode. If the voltage rises enough so current starts flowing into the transistor base, it will turn on just enough to reduce the voltage back down to right around the zener threshold. It will also dynamically adjust with the external load applied to the circuit. For instance, if the secondary is producing 30A of current but only 2A are needed to drive the HVM for looping, the other 28A net gain will charge the battery or cap bank. Once the bank voltage reaches the 12V zener threshold plus the 0.7 diode drop of the transistor base p-n junction, gate current will begin to flow and the transistor will conduct. Since the current gain of a big, slow power transistor like a 2N3055 is very low, the base current might be as much as 1/10 of the emitter current. In other words, something like nearly 3A will pass through the zener and the other nearly 30A will go through the transistor. The zener will have to be appropriately rated and more likely using a darlington pair for the pass transistor might be a good idea, perhaps a 2N3904 or some cheap readily available PNP transistor could be used. Hmm, actually those aren't good for 3A of current either, perhaps just a 2955 would be the simplest solution. I like to design with the simplest, cheapest, most commonly available parts.

N.B. I know this is in-depth and there are a lot of experienced experimenters here, I don't want to sound pedantic. I want to include enough technical details for those who are learning and might benefit.
Anyway, if an external load of 10A is applied, the zener/transistor combination will almost instantly go from passing 28A of current to passing just 18A of current, and go back again if the load is removed. The 12.7V voltage will change just a teeny tiny amount because of the zener characteristic curve and the transistor current gain. It's just a shunt regulator with a dump load instead of the more common series regulator. It's also a really easy way to measure the excess power being produced, because all you have to do is put an ammeter on the shunt regulator to measure the exact amount of excess power gain above what it is required to run the HVM. Measuring the power at DC is much easier to do accurately than to try and make accurate power measurements at HF/HV. Perhaps it's easier to show than tell, I'll try to get an initial version wired up soon. Also, whether the transistor alone is the dissipation element or additional resistors are used, a lot of heat is going to be generated. The maximum ratings of a 2N3055 are 15A collector current and 115W. Hmm, ok so you'd essentially have to make individual units with one transistor for 10A at 12V, or 15A if you use some resistors to help it. Dissipating nearly 120W continuously is going to require a big heatsink and possibly active cooling with a fan to keep the heatsink requirements more moderate.

Sorry if all this is off-topic, I don't want to derail the discussion. I had done some research and design into the shunt regulator because I didn't want to blow my $50 cap bank and have to replace it because of excess power in the event some shows up! The little balancing circuit on the back of the PC board connecting the 6 caps has some little surface mount resistors, but no way are those going to be able to dissipate more than a couple of watts of power. If more current than this is applied to an already full cap bank, those are going to smoke and the caps are going to be destroyed.

If I didn't also want a way to be able to measure the excess power accurately it might make more sense to use a normal series regulator of some kind. The secondary would drive a diode bridge directly, and that would charge a sizeable smoothing cap rated for the full no-load voltage that could be expected from the secondary. Then something like a 723 regulator IC to drive a pass transistor to reduce it to 12V or whatever is desired. Easy but inefficient if the no-load voltage is significantly higher than 12V.

[radioionics]

Well, I have to find another flyback. Mine seem to not output enough. If someone have a suggestion for another easily available hv source im all ears open. I have another Tv to dismantle...

Bruce I see in your last schematic you use lithium battery instead of cap... witch kind are them?

Thank you!

LiPO lithium ion | eBay

[soundiceuk]

Don't use speaker cable for secondary.

Stick to HV wire!

It won't get hot!

[Mwtj]

It's really simple, I pretty much described it. Even the base resistor is essentially optional. External pass resistors would go between the transistor collector and +V, and should be suitable resistance and power dissipation for the application. In this case, at 12V the desired dissipation might be 30A, so Ohm's law means we need 0.4 Ohms resistance and 360W power handling. Of course all the cables and connectors need to handle 30A as well, and the traces on the PCB need to be very beefy and probably solder reinforced.



Full discussion:
design - High power equvalient for a zener - Electrical Engineering Stack Exchange

For even more power dissipation these could be made in single 30A modules, like a brick. Just parallel multiple units to reach the desired dissipation. Designing a PCB for it would be great!

Again load the file. Takes a while to charge the Caps. Adjust the speed of the simulator.

Circuit Simulator Applet

[soundiceuk]

Basic circuit corrected.

[tswift]

Mwtj, Tswift, What have you experienced with wood as the support of the project? Does it conduct the hv?
What should I use instead?
thank you!

Sadly, yes wood conducts much better than you would think. At low voltages 12V up to house current 120V/240V it's not enough to matter but as you reach tens of kilovolts it definitely does matter. You need some kind of plastic or ceramic material. Sadly pretty much all the substitutes cost more and/or are harder to work with than wood. You can use a wood baseplate and build upward from it with good enough insulators like Mwtj is doing. You might try using a non-conductive baseplate like a big HDPE (polyethylene) kitchen cutting board. You can get these up to maybe 18x24 inches in size and at least 1/2 inch thick. Then you can drill holes in it just like wood to mount your components, that should work OK for minimal cost. I like acrylic because it also looks good and has desirable properties at HV/HF (high dielectric strength and low dissipation factor), but (A) it's expensive and not available at your local store like a cutting board and (B) it takes a lot of time and care when constructing to make sure it looks pretty when you're done. So probably I would suggest polyethylene cutting boards, you could use one big one, or if all you can find are smaller ones, use a bigger wood backplate and mount the smaller HDPE cutting boards on it, even better would be to use standoffs so there is an air gap of 1 inch or so between the wood and the plastic boards. Divide your circuit into components (the HVM driver, multiplier, PPV's, and LV section) like I have done and mount each part on its own cutting board. As long as there is 1/2 inch of polyethylene between the wood and anything with HV it will probably work OK at voltages up to maybe 100KV. Adding a 1 inch air gap and ensuring at least 2 inches between any HV surface and anything metal or wood should let you get closer to 200KV.

[Wistiti]

Sadly, yes wood conducts much better than you would think. At low voltages 12V up to house current 120V/240V it's not enough to matter but as you reach tens of kilovolts it definitely does matter. You need some kind of plastic or ceramic material. Sadly pretty much all the substitutes cost more and/or are harder to work with than wood. You can use a wood baseplate and build upward from it with good enough insulators like Mwtj is doing. You might try using a non-conductive baseplate like a big HDPE (polyethylene) kitchen cutting board. You can get these up to maybe 18x24 inches in size and at least 1/2 inch thick. Then you can drill holes in it just like wood to mount your components, that should work OK for minimal cost. I like acrylic because it also looks good and has desirable properties at HV/HF (high dielectric strength and low dissipation factor), but (A) it's expensive and not available at your local store like a cutting board and (B) it takes a lot of time and care when constructing to make sure it looks pretty when you're done. So probably I would suggest polyethylene cutting boards, you could use one big one, or if all you can find are smaller ones, use a bigger wood backplate and mount the smaller HDPE cutting boards on it, even better would be to use standoffs so there is an air gap of 1 inch or so between the wood and the plastic boards. Divide your circuit into components (the HVM driver, multiplier, PPV's, and LV section) like I have done and mount each part on its own cutting board. As long as there is 1/2 inch of polyethylene between the wood and anything with HV it will probably work OK at voltages up to maybe 100KV. Adding a 1 inch air gap and ensuring at least 2 inches between any HV surface and anything metal or wood should let you get closer to 200KV.

thank you for the hint Twsift! Maybe it is the reason why my side seems to not produced so much hv...

I'll have to go shopping for some cutting board!

[tswift]

Again load the file. Takes a while to charge the Caps. Adjust the speed of the simulator.

Circuit Simulator Applet

Man, that circuit simulator is dynamite! I can't believe I didn't know about that before. Anyway, you have the idea basically correct but I corrected a couple of things. I adjusted the connections to the transistor emitters, they go to ground. The whole shunt regulator literally just sits across the cap bank and shorts out excess current. I also changed the voltage source to 20V and the series resistor to 0.25 ohms. Yes the cap bank takes a large amount of time to charge when you start the simulation, this is not unlike real life when you're charging them from a small bench supply. The RC time constant is (83F * 0.25 ohms) so about 20 seconds. Go get a cup of coffee while the simulation runs. Sure enough, after about 20 seconds of simulated time, the voltage on the cap bank gets close to 12.8V and the transistors begin to gate on. If you let it continue to run it reaches a final steady-state where all the current is going through the two transistors, 13.71A each. The final voltage across the cap bank is 13.08V.

I wouldn't consider this a final circuit just yet. The zener current is not realistic, I know from experience that big power transistors like the 3055 have a very low current gain (more like 10-20) than the 100 shown in the simulation. This will burn up the zener diode. I'm sure there are other better transistor choices but it's easier to just add another NPN transistor to each one as a darlington pair. That will add another diode drop (~0.7V) to the voltage but that shouldn't really matter, it will greatly reduce the necessary base current that the zener has to supply.

N.B. to every experimenter: if you don't already have this book on your bookshelf, you are making a BIG mistake. No other source I have ever found does as good a job of explaining basic circuits all the way up through how transistors really work, into digital logic design. It's well worth your time and money.

https://www.amazon.com/The-Art-Elect...=maggicom0e-20

[tswift]

Hopefully getting close to a workable version of the shunt regulator circuit. I changed the transistors to darlington pairs. Making the assumption that all of them are 3055's, the two being driven by the zener directly are running at about 1A collector current and the other two are running at about 11A current and about 85W. The 0.5 ohm resistors dissipate 77W. I have some 100W 1 ohm resistors on order, I can parallel these for 0.5 ohms and suitable power handling. Now the zener current is down to just over 100 mA and a 2W zener should do ok, I have some 12V 2W zeners on hand. So I think this configuration will work, I will make a breadboard version and test it.

[tswift]

Don't use speaker cable for secondary.
Stick to HV wire!
It won't get hot!

OK, not planning on changing anything. I remember what happened last time I tried to substitute wire....

[dragon]

TSwift said; "I think maybe you're thinking of a crowbar circuit"

Your right, my bad... split focus... that's what I get for mixing my projects with others - I'm dealing with high amperage output conversions using 350 amp 1000 volt IGBT's...

I tend to pop in and out of the forum when I get stuck on a problem to divert my attention away from the problem. Sometimes things can be seen more clearly by not looking directly at it.... anyway, I should probably refrain from posting without a clear view of the overall conversation....

Edit; On another note, I've used 8ga 700 count speaker wire for 100kv + projects with great success, I wouldn't be afraid to experiment with it.... expecially for high amperage outputs....

https://www.youtube.com/watch?v=V4l1IJJgjNc
https://www.youtube.com/watch?v=3t4pZpKSpP4
https://www.youtube.com/watch?v=WBpUSDZ3CmQ

[tswift]

OK, latest build progress. How to mount the meters: go vertical! This took me more hours than I intended to fabricate, but it works. Still have to mount the meters on the HVM board. Also, having the metal rod this close to the HV section is not good. I need to get some nylon fasteners and threaded rod instead, it exists but not at the local hardware store....