Agentgates´s TPU setup with strange wavehill hump

[zerotensor]

Hmm, the positive hump wavehill is mysterious there.

I think the shape of the waveform is partially due to the load being a resistor and diode in parallel.  On the positive swings, the ohmic load is distributed across the resistor and the LED;  1/R_total = 1/R_resistor + 1/R_diode, while on negative swings, the coil only sees the resistor.  One way to mitigate this without removing the LED entirely would be to wire two LEDs in parallel, but with opposite polarities so that one or the other diode will pass current regardless of the polarity of the applied pulse.  Even then, the overall resistance will still fluctuate because of the nature of the diodes' response curves, (esp. at high frequencies).

[agentgates]

Hi Marek

Thank you for your post.

@ Tony
Is this the effect You are observing driving your TPU?

Similar, but I don't have those ugly "afterhumps". Only the big ones. You perhaps driving it in current mode. (that is not good) The pulse length should be around 500ns depending on your setup but it never changed proportionally with my setups.

I meant "not freq sensitive" that those spikes come at any frequencies at approximatelly the same length. You can make those powerful spikes even without generators. Also they will be far stronger when you'll fire the 2 other coils in sequence.

These spikes won't even change in my previous setup (3 primary with the same one turn secondary, 20-30V 100-200mA, 87Hz) when I put a 50W 240 halogene bulb on it. They just don't care about your load.

Some interesting things I noticed:
1. There are not only stronger but lot more spikes when you drive all the 3 coils.
2. By only increasing the input voltage they are coming closer and closer a bit, some kind of spin-up effect.
3. I also observed a "jump-together" effect. When I increased the input voltage I zoomed the scope to observe closer the growing of a spike. A single spike grows out gently from the ground line. By just keep on increasing the input voltage, after reaching a certain level other spikes RAPIDLY jump out. So there is some kind of critical level that when you reach with the input voltage the number of these pretty curvy spikes rapidly multiplies. (it happened somewhere between 15V-20V)
4. I kept on increasing the input voltage and found another thing that might be interesting. There is a next voltage limit that when you reach they stop growing immediately. (IMHO the secondary gets saturated magnetically)

experiment description: Tpu with one driving coil, ONE TURN of collector loaded with 2.2ohm/3W resistor and white LED parallel. Diode light is very bright, voltage measured on the 2.2ohm load You can see on the scope. Input power about 12W, output about 1W. Still far from from OU, but for me there is something goin on in this device (Im an electronic engineer). Btw, this setup IS frequency dependent. The highest power occurs when freq=33kHz

Yes you can not make OU with one primary without other parts like a good LC pair.

12W is HUGE in consumption Marek. You need a current limiter resistor in series with your primary as this is not a current device. It is strongly a voltage device - as I found in the meantime. I think this is one of the many of those things were the other lads failed over the years, they just instinctly tried to handle it as a transformer where you have to punch the current hard through the wires if you want result. (I also did it ) Set your primary resistance somewhere between 15-22ohms to find the optimum of your coil. My present setup likes 15 and 18ohms 3-5W resistor in series on the common ground.

I think the latter will eliminate those ugly "afterhumps".

I hope I could help. Good luck.

Kind Regards
Tony

PS: I am currently writing the code for the PG to tickle this thing from as many angles I can and opefully I can tell about the results tomorrow.

UPDATE

Set your primary resistance to an optimum level when you can drive it between 100-200mA up to ~25V. (there is no point to go above as I noticed)

Another thing I have forgotten, when you get that "jump-together" thingy your power consumption will fall back rapidly in the same time. About to the half.

[agentgates]

@zerotensor

Yes you're right but he also drives it in current mode and that thing is tring to be a back EMF pulse due to the high current he puts in and perhaps the res+diode pair rectifies it.

[broli]

Ok the revised page is up, I hope all the parameters are correct now. The link is at:

http://ziosproject.com/NJ/AAcoilPres/index.htm

Perhaps this can be put in the first post so new comers can directly see the latest design?

[agentgates]

@broli

Thank you for your ambitious work again. However I can not see change on the page.

I can not modify my first post. There are many things I would like to change there but somehow the "Modify" button disappeared.

UPDATE

Code is ready but I feel dizzy so I have a rest. Think this glue is a poison to smell.