[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [TCML] mmc cap protection by spark-gaps-any ideas?



Ray -
   
  Perhaps some confusion here. Most of your comments regarding bleeder resistors are 100% applicable to high voltage DC power supplies used in radio transmitters, radar systems, etc. In these circuits, the bleeder resistors perform a critical role in voltage regulation, and a secondary role as safety dischagre resistors. However, the capacitors (and their bleeder resistors) used in a Tesla coil tank circuit operate under very different conditions, and fulfull very different functions.
   
  Generally speaking, the resonant frequency of conventional spark-gap Tesla coils falls between 80 and 400 KHz. This is the frequency of the alternating current that the capacitors in the MMC are subjected to. Lets look at a medium sized TC with a resonant frequency of 200KHz, using an MMC type tank capacitor assembled with series strings of Cornell Dubilier .15uF/2KV film capacitors. Each individual capacitor has a 10 meg ohm resistor connected in parallel.
   
  The function of the bleeder resistors is only to discharge the individual capacitors when power is removed from the system. Even though the MMC capacitor is effectively "shorted" by the secondary of the HV supply transformer, and the net voltage across the capacitor's end terminals is zero, it is possible for individual capacitors within the string to maintain an "isolated charge" IF each individual capacitor does not have its own bleeder resistor. I have built MMCs without bleeder resistors, and measured between 200 and 800 volts on individual capacitors a few seconds after power was cut. Only a few capacitors in the string would have residual voltage, and different capacitors would have a retained charge every time the system was power cycled.
   
  As for providing "voltage equalization" across the individual capacitors in a string, this can be easily disproven by looking at the impedance of the capacitor at the operating frequency vs the resistance of the bleeder. Lets look at the impedance of an individual .15uF capacitor from our MMC operating at 200KHz. Xc = 1/2 * pi * frequency (in Hz) * capacitance (in Farads), or about 5.3 Ohms. For the purpose of "voltage equalization", a ten million ohm resistor in parallel with a 5.3 Ohm impedance just doesn't have much effect.
   
  Incidentally, for the last 30 years it has been fairly common practice to use series strings of large electrolytic capacitors for the filter capacitors in HV DC power supplies for radio transmitters (check any ARRL handbook for details). The non-polarized Pyranol capacitors you mention date from the 1940's & early 1950's.  
   
  None of this really should have any influence on safety practices around any sort of equipment utilizing high voltage. Always assume that EVERY large HV capacitor is perpetually charged to maximum operating voltage until you have personally discharged it and connected a temporary jumper across its terminals. Realistically, this is probably not practical with an MMC, which is why I recommend that every capacitor in an MMC be equipped with a bleeder resistor of the appropriate voltage rating, and that the entire MMC be "shielded"  with some sort of protective enclosure to prevent inadvertant contact even when the system is powered off.
   
  Regards,
  Herr Zapp
   
  

Ray von Postel <vonpostel@xxxxxxxxxxx> wrote:
  Gary:

On Aug 19, 2008, at 10:31 AM, Lau, Gary wrote:

> The resistors you discuss in DC power supplies are used for an 
> entirely different reason than why they're used in MMC's. In MMC's, 
> the power is AC, and the voltage division is determined solely by 
> the ratio if the capacitances. The bleeder resistor values are far 
> higher than the AC capacitive reactance, so the resistors have no 
> bearing on voltage division ratios.
>
When you connect a series string of resistors across a voltage source, 
a.c. or d.c. the voltage drop across each resistor depends upon the 
current through the string and since each individual resistor is in 
parallel with a capacitor, the a.c. or d.c. voltage drop will be the 
same across both capacitor and resistor.

> The wire wound resistors used in the radio PS need to be high power 
> (which makes them more failure-prone) because the leakage current of 
> electrolytic is extremely variable over time and temperature, so the 
> resistor must be chosen to be significantly lower in resistance to 
> guarantee that IT determines the division ratio and not the variable 
> leakage resistance. That's just not the case in MMC bleeder 
> resistors.
The bleeder resistor across the output of a power supply is chosen as 
part of the filter system and in many cases the filter system 
consists of nothing but resistors and capacitors in a typical a.c./dc 
cheap plastic radio. I don't think you noticed that the voltage 
involved in the accident was 6 kv. That in itself should indicate 
that in so far as the accident as concerned it did not involve 
electrolytic capacitors. The capacitors involved were General 
Electric, Pyranol, (oil filled) 1 mfd. 10 kv. working. The bleeder 
resistors were there to help the regulation of the power supply. Just 
what do the resistors do in an MMC? So far as I can see the MMC is 
the equivalent of a single capacitor with low dielectric resistance.
>
> If it might diminish your horror, consider that the likelihood of 
> _ALL_ of the many bleeder resistors in an MMC failing open is very 
> remote. The consequence of just one failing isn't that great. And 
> even if they did, the discharge path formed by the NST or other 
> transformer would ALSO have to fail for there to be a genuine hazard.

I have no horror of getting killed. As a matter of fact, I don't know 
you and you don't know me so neither of us should have any horror of 
the other getting killed. Of course if enough of us get killed 
playing with Tesla coils we can expect Big Daddy to step in and outlaw 
them as toys. That would be a disaster for a lot of nice people. So 
out of common courtesy we should observe Safety First.

Any electrical shock any time and of any magnitude has the potential 
of being a hazard. What can go wrong will go wrong.
>
> It's not clear what you are proposing as an alternative. If you are 
> suggesting that folks not use bleeder resistors and instead use a 
> manual shorting stick every time they turn off their coil, I would 
> counter that:
> 1) Shorting just the ends of the MMC won't guarantee that the 
> individual caps are all at zero charge.
> 2) I would bet any money that users are more likely to neglect to 
> use the shorting stick, than there would be massive failures of all 
> bleeder resistors.

I am not proposing any circuit changes. The use of resistors in a MMC 
is wholly justified but for good electronic reasons and NOT FOR 
SAFETY. The MMC itself is a work around to keep the cost of the hobby 
down. Nothing wrong with that.

> I believe that it's far safer to have a 2nd tier safety mechanism 
> (bleeders) that is always present (the 1st tier is the presence of 
> the transformer winding), than to have a 2nd tier mechanism that 
> requires considerable diligence and is likely to be ignored.

My statement is that bleeder resistors across capacitors are not there 
for reasons of personnel safety. There are many readers of this list 
who have little or no experience with electronics in general nor high 
voltage in particular. To hold out the concept that it is safe to 
turn off the power to a Tesla coil and then get your fingers in it
strikes me as being just plain bad practice. Your are of course 
right that safety devices are often ignored. Some people have been 
known to park a car on a hill with out setting the emergency brake nor 
putting the car in park. I doubt you are advising people not to 
provide a shorting or grounding stick because they might not use it. 
There is no doubt that bleeder resistors do provide some safety, but 
they are not designed to be either fail safe nor infallible.

It should be pointed out that the retained charge in a capacitor is 
d.c. They do not hold an a.c. charge (if there were such a thing). As 
to a Tesla coil operating at a "high" frequency having anything to do 
with it, I would remind you that "High" in the case of the Tesla coil 
is only relevant to the 50/60 Hz. line voltage. Most operate within 
the low or medium frequency range.

The mention was made in another post of the fact that capacitors with 
a high voltage rating are often shipped with their terminals shorted. 
Who ever shipped them was looking out for the next fellow. Capacitors 
can frequently be charged by static electricity while in storage or 
transit.

Frankly, I am wondering why we are having this discussion. How many 
times have people on this list commented on the fact they have been 
"bit" by a turned off coil and they should have waited longer. Any 
thing we can do to keep some one from being hurt is worth doing.

I still want an explanation as to what you think the resistors do in 
an MMC capacitor bank.

Safe coiling
Ray

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla

_______________________________________________
Tesla mailing list
Tesla@xxxxxxxxxxxxxx
http://www.pupman.com/mailman/listinfo/tesla