DC Conversion

Home Up


I'm working on a subsystem which will allow my 6.5" coil to run reliably in a DC mode of operation ( i.e. primary capacitor is charged using high voltage DC instead of AC). I'd eventually like to incorporate these findings into the design of a large coil. Following is a log of my progress.


The first steps were taken in the Spring of 2000 when I had a number of discussions with Kevin Ottalini (one of the Tesla Coil DC guru's).  I also received input directly and indirectly from Ross Overstreet, Jim Lux, and Jeff Parisse.


First thing I needed was to build a DC Bridge Rectifier capable of handling a minimum of 500ma at 32kv (15kv*1.414*1.5). This resulted in the construction of two bridges.


Kevin provided me with eight H2612-22's which I  configured as a bridge with two series diodes per leg. This configuration yields 330ma @ 24kv.

brrect1.jpg (17519 bytes)


I also purchased a bag of one hundred 1N5408's (3a @ 1000v).  I used these to build four strings of twenty diodes. Each string fits into a PVC pipe for insulation and easy replacement. The assembled unit is capable of 3a @ 40kv.

brrect4.jpg (49267 bytes)    brrect2.jpg (49720 bytes)


The DC discussion evolved into a small get together at my house in May of 2000. We looked at Kevin's redesigned DC Coil and talked about various aspects of DC operation. I also purchased a bunch of high power resistors from Kevin.


To move further I needed to go through all my test equipment and compete the many partially finished projects.  This motivated me to clean up my Tektronix 466 Analog Storage Scope. I also finished the Fiber Probe kit I received from Terry Fritz, and threw together a 15kv DC meter.

Fiber_Probe_a_small.jpg (1479 bytes)


Kevin suggested using a limiting resistance of 10K to 20k-ohms and slowly reducing it in steps down to 5k-ohms. Many e-mails later I decided to build a limiting resistor of 2.5K-ohms @ 1,300w (The "More Power is Better" bug took over).

Resistors1.jpg (20759 bytes)


I put the 3a bridge and 2.5k-ohm resistance in my charging circuit (between the PT and the RSG). Low power tests at 60% were a great success. Everything ran very smoothly (without the need for any ballasting) and measurements on the running coil were comprehendible. Break rates from 240pps down to 1pps were obtained by letting the RSG spin down while the coil ran.

DC Conv Test 2.jpg (37305 bytes)   DC Conv Test 1.jpg (27994 bytes)   DC Conv Test 3.jpg (44731 bytes)


Ross came over on the 5th of August, 2000 and helped me document, measure, and then test the current set-up at full power.

Measurements were made using the Fiber Probe, appropriate transducers and my Tektronix 466 analog storage scope. Ross captured the images using his new Olympus C3000z digital camera.

P1010003-web.jpg (29158 bytes) Primary Current - 100A/div, 20uS/div, input voltage at 67% (100vac)
P1010004-web.jpg (34128 bytes) Secondary Voltage - 5V/div, 20uS/div, 10x probe,   1-foot wire antenna, input voltage at 47% (70vac).
P1010006-retouched-web.jpg (10817 bytes) Charging Current - 100mA/div, 1ms/div. The two peaks are 8.3mS apart which equates to 120pps. Strange since we were running at 240pps on the RSG.
P1010006-web.jpg (24611 bytes) Charging Current - With grid lines visible (I've now fixed the storage scope and eliminated the background scatter)
P1010009-web.jpg (41916 bytes) DC Set-up - The PT was simply moved away from the base and the bridge and resistor bank inserted. Ready for High Power testing.
P1010016-web.jpg (28549 bytes) DC vs AC  - Both configurations performed equally well. DC reliability is an issue that will need to be worked out.
P1010018-web.jpg (22936 bytes) Small Variac - Running on a 20a variac with no current limiter. The elimination of the current limiter is a Big plus for consideration when building a DC coil.


bulletFull Power Runs, Results, and Consequences - 
bulletWe were able to see the energy transfer and third notch quenching while looking at the primary current.
bulletWe saw the ring-up while viewing the secondary voltage.
bulletMonitoring charge current we were able to see that the caps were reaching full charge (at the reduced voltage we were using) before the next 240bps bang.
bulletThe first high power DC runs were with the H2612-22 diode bridge. Performance was about the same as using AC.
bulletThe primary cable running to the tap, which I had temporally re-routed, took a direct strike blowing the bridge. Arcs stopped and the input current went sky high.
bulletReplaced the bridge with the one made of 80 1N5408's and performance was back to where it should be.
bulletWe plugged the PT into a 20-amp variac and ran the coil again. I was able to get the same performance from the 20lb variac as with my 400lb controller.
bulletWe got out my 2100w generator and plugged the small variac into it. The coil ran but the generator was struggling with the rapidly varying load and the current requirement was at the generators maximum.
bulletAfter a bit of runtime on the generator the second bridge blew without warning.
bulletThere was never any indication of heating in either bridge.
bulletThe resistors barely got warm enough to sense (with a calibrated finger).
bulletI probably went way too low on the resistance 2.5k-ohms and should I bump it up to 5K or more next time.
bulletSince we were out of DC parts we switched back to A/C and took a bunch (75) pictures with Ross' new camera.
bulletBenefits of DC (as I currently see them) -
bulletDC eliminates the need for current limiting/ballasting on the low voltage side of the PT/Pig.
bulletBreak rates can be varied from 1pps up to 700+ bps creating different effects.
bulletMeasurements are easier and more accurate.
bulletBang size is predictable and reproducible.
bulletEasily converted to A/C operation.
bulletDrawbacks of DC (as I currently see them) -
bulletMore points of failure. In particular the bridge rectifier as the resistor bank is unlikely to fail.
bulletReliability of the bridge (Further testing and adjustment of the resistance may make them more robust).
bulletFailure Analysis -
bulletLooks to me like I over did the max forward current. 15,000volts/2,500ohms=6amps My 1N5408 bridge should have been able to handle 3a@40kv and the H2412-22 bridge  600mA@24kv.
bulletAll legs tested OK at 120v and the signal looked the same as with good diodes. So much for my testing procedure using a 40w 120v bulb in series with the diode in question.
bulletMeasuring the resistance though told a different story. Each bridge had one bad leg (one was on the positive side and one was on the negative side).
bulletThe twenty 1N5408 diodes in the bad leg measured 10M-ohm in either direction. A good diode measures 10M one way and 300M-400M the other way.
bulletThe two H2612-22 diodes measured different. One was 10M-ohm in both directions and the other one was 14M one way and 25M the other way.
bulletNext Step -
bulletThe next step is to fix the bridges, increase the limiting resistance to 5K-ohms, then re run the full power tests.
bulletIt would also be useful to capture the charging current on the scope while performing the tests above.
bulletInvestigate ways of hardening the bridge including adding filters.

Questions and comments                Copyright 1997,2006 Brian D. Basura                This site was last updated 04/02/06