Fixing the alien translators mistake

Design for highest temp operation possible and sell to Baker Hughes, Halliburton, Schlumberger, etc..

On a sunny day (Sun, 16 Oct 2011 13:05:33 -0800) it happened Robert Baer wrote in :

Thanks, good idea :-)

On a sunny day (Sun, 16 Oct 2011 11:06:42 -0700) it happened John Larkin wrote in :

I build it, and ran quite a few measurements on it. It works very well. At 3.3V supply I get about 2 Vpp at the emitter, and more then 6.6 Vpp at the collector. the base seems to stabilise at about -300 mV. The electrolytic capacitor is negative at the top. I used 1 turn feedback, 5 turns primary, and a 2k7 base resistor, and 1 uF base decoupling. Frequency is close to 100 kHz (the maximum recommended for these E cores). In some circumstances, like low load or low bias, it can go into blocking mode it seems. Compared to my original circuit that has the emitter grounded, the voltage was 250 V out on stage one for 3.3 V DC power supply, while yours gives 300 V out for the same 3.3 V DC. I was a bit afraid you circuit would do bad things to Vbe, and I am amazed it stabilises Vbe at -300 mV (for a Si transistor). The Vbe breakdown is never reached, measure about -1.5 V max at 3.3 V supply, clipping at about +.75 V where the Vbe diode conducts. Cannot find anything wrong with this circuit, may use it as it provides more output for the same copper (turns). May as well call this HV generator a JJLarkin design :-)

My little oscillator is happier with a small amount of feedback in the emitter winding, just a few tenths of a volt p-p. That shouldn't block. Of course, you can't go below 1 turn on the feedback winding.

The one turn in the emitter is effectively removing one turn from the primary! In other words, you are getting more than (2 * V+) peak-to-peak at the collector, because the emitter is swinging so far negative.

This AC bootstrap effect approaches infinite as the emitter feedback ratio approaches 1, at least until something else breaks. Interesting.

You need two time delays around the loop to block. The LC resonance is one, and the base cap is another. A bigger base cap should prevent blocking, as would the higher feedback ratio.

Parts are fun.

John

Larkin

cores).

supply,

Decent little circuit noted, John L. I will remember.

?-)

On a sunny day (Mon, 17 Oct 2011 07:19:35 -0700) it happened John Larkin wrote in :

Never too old to learn a new trick :-)

On a sunny day (Sat, 15 Oct 2011 13:36:44 -0700 (PDT)) it happened George Herold wrote in :

Yes, I wound a new transformer with fewer turns, and am now at > 100kHz.

Parts arrived [on Saturday!] tried to build a small proto PCB, to see how it would work out with a single layer PCB and wiring, arc distances etc. This is the size I could make one in (or an 8 anode PMT, add some for each extra dynode):

There are 2 reasons the Wima capacitors cannot be closer together,

1 arc space between tracks, and 2 the pins are not exactly centre in the red encapsulation, so sometimes they do not fit on 1/10 inch raster close against each other. (Hint to Wima). The empty space bottom left will get the stabiliser, feedback via a 1G resistor. I have a lot of 120 MOhm SMDs to make a large 1.2 GOhm resistor, need more than one to stay below arc over... More soldering on the Sunday... Works incredibly well so far :-) I used 400 V capacitors, so this one *could* get up to much more than 2 kV, and that makes it rather big. For some PMTs 250 V capacitors would do. Maybe it could or should be potted too.

This thing will be in a pipe, with the PMT, scintillation crystal, and a nice cable with connector to the main unit with LCD display and battery.

-------------------------- | ----------- 0 0 | | | | | | 0 0 | | |||..|||. | 0 0 | | ----------- 0 0 |

--------------------------------------- | 0 0 =-------------cable -----------| HV generator - preamp

- PMT- [crystal] | 0 0 | connector

--------------------------------------- ------------------------