datasheets.com EBN.com EDN.com EETimes.com Embedded.com PlanetAnalog.com TechOnline.com  
Events
UBM Tech
UBM Tech

Design Article

Comment


anonymous user

3/30/2004 12:02 PM EST



The polarity markings on the ...

More...



anonymous user

3/4/2004 6:28 PM EST



Yes, another error on the ...

More...

1-kV power supply produces a continuous arc

Robert Sheehan

2/5/2004 3:00 AM EST

Designing a high-voltage switching power supply that can produce a sustained arc can be challenging. This compact and efficient design delivers 1 kV at 20W and can withstand a continuous arcing, or short-circuit, condition (Figure 1). It uses standard, commercially available components. R1 sets the LTC1871 switching-regulator controller for a nominal operating frequency of 120 kHz. The circuit operates as a discontinuous flyback structure, producing 333V across C1. The diode/capacitor charge-pump multiplier triples this voltage to create 1000V at the output. Figure 2 shows the switching waveforms. When the primary switch, Q1, is on, the output rectifiers are reverse-biased, and energy is stored in the transformer, T1. When Q1 turns off, energy transfers to the secondary winding, and C2 and C3 pump up the output voltage through the rectifiers. The primary voltage goes high and is clamped through the transformer and rectifier, D1, by the voltage across C1. The transformer is well-coupled, so the leakage inductance creates little voltage spike. A small RC snubber across the primary winding damps the ringing and reduces EMI (electromagnetic interference).

For current-limit protection, the circuit in Figure 1 contains two active circuits and one passive element. The voltage across the current-sense resistor, R2, limits peak primary current to 7.5A. Q2 provides secondary-side current limit. Notice the bump on the leading edge of the current ramp of Trace 2 in Figure 2. This bump coincides with the positive excursion of the voltage across R3 in Trace 4, which is the refresh current for C2 and C3. When the circuit is overloaded, this slug of current becomes high enough to enhance Q2, folding back the load current (Figure 3). A hard short circuit results in relatively low power dissipation. Omitting Q2 for the secondary-side current limit results in substantially increased short-circuit current and internal power dissipation, resulting in failure of the primary switch Q1. R4 provides a load impedance for the power supply.

This load helps to limit the peak-current stress in the multiplier capacitors and diodes. Don't skimp on the power rating for R4, because dissipation during a continuous arc can be substantial. Should R4 fail open, the feedback circuit forces a full duty cycle with catastrophic results. Too low a value for R4 can result in charred circuits and hours of debugging. (Yes, a hearty explosion elicits a round of applause from the lab crew.) Arcing is the most stressful condition, and the output capacitor constantly charges and discharges (Figure 4) . As a final figure of merit, the circuit is efficient (Figure 5). The efficiency reaches 87.3% at 12V input and a full load of 20W and increases to 87.7% with an overload of 24W.

So what is this circuit good for? A battery-operated bug zapper, perhaps. And, like raking a live wire across a grounded file, this is a great tool for befuddling the AM-radio listeners on the production floor. The circuit probably doesn't deliver enough energy for use as an ion generator for a plasma cutter, though one engineer I knew was willing to give it a try. A previous version of the circuit used a monolithic switcher, and with the right materials for banana jack and plug, created a bright orange glow and enough heat to raise thoughts about the fire extinguisher (plenty of ozone, too). I'd stay away from using this circuit as a cat trainer or an electric fence. The circuit does generate a lethal voltage potential, and lawsuits can be quite costly. Prototype this circuit at your own risk.

Check out our Best of Design Ideas section!





Michael.Kiefer

2/5/2004 2:46 PM EST



Your dwg. in figure 1 shows the 1 kv output tied back to Vin. The connection should be to the 2 4.99 meg/ 12.4 k divider network to gnd.



Sign in to Reply



Michael.Kiefer

2/5/2004 2:46 PM EST



Your dwg. in figure 1 shows the 1 kv output tied back to Vin. The connection should be to the 2 4.99 meg/ 12.4 k divider network to gnd.



Sign in to Reply



anonymous user

2/5/2004 2:49 PM EST



I found an error in the 1kV generator.
Remove the dot between high-voltage feedback and +12V supply.
Otherwise it will blow it self upp (I guess).



Sign in to Reply



anonymous user

2/6/2004 8:44 AM EST



The 1-kV power supply feedback is connected to the low voltage
input, It should be connected only to the first 4.99M ohm resistor
in the feedback circuit.



Sign in to Reply



Milant

2/6/2004 2:46 PM EST



Re 1kV power supply: are we really sure that we want a connection from the HV output to the low-voltage input?



Sign in to Reply



anonymous user

2/10/2004 2:02 AM EST



Received My Feb. 5th issue of EDN today and noticed the high
voltage feedback hook up error to the Vin line. That would be
very interesting for a short time. Also, those two 4.99M resisters
should be located in the High voltage isolation area.



Sign in to Reply



anonymous user

2/17/2004 7:05 PM EST



Figure 1 seems to show the 1kV output tied back directly to the input voltage. Bad idea.



Sign in to Reply



anonymous user

2/18/2004 5:06 AM EST



Hi:
Feedback path from 1 kV output is tied directly to DC input... sure it makes a loud explosion. Guess there is an extra junction in the drawing.



Sign in to Reply



anonymous user

3/3/2004 5:23 PM EST



Thanks for all the comments, and thanks to EDN for fixing the schematic. I dusted off the circuit and shorted the output to the input. It just sits there happily chugging away with the output voltage clamped to the 12V input. It barely gets warm.



Sign in to Reply



anonymous user

3/3/2004 5:34 PM EST



There is one other error on the schematic, T1: Copper. This should be Cooper Electronics (formerly Coiltronics). Transformer by copper had us chuckling though.



Sign in to Reply



anonymous user

3/4/2004 11:16 AM EST



Nit picking is the delight (or curse?) of old engineers! The schematic polarity dots on T1, or else the scope waveform relative polarities for Ch1 and Ch3, are inverted. Then the text describes voltage tripling. That's not accurate. The output "voltage multiplier" is actually a peak rectifier followed by a two stage series multiplier. The unloaded output is positive peak plus double the peak-to-peak voltage delivered at the transformer secondary.



Sign in to Reply



anonymous user

3/4/2004 6:28 PM EST



Yes, another error on the scematic, this time on the dot notation for T1. It truly is a flyback design. My prototype has pin 7 going to the anode of D1, and pin 2 connected to R11.



Sign in to Reply



anonymous user

3/30/2004 12:02 PM EST



The polarity markings on the transformer appear to be reversed in the published article. Either the primary or the secondary need to be reversed.
Brad



Sign in to Reply



Please sign in to post comment

Navigate to related information

Datasheets.com Parts Search

185 million searchable parts
(please enter a part number or hit search to begin)