Jim's EBAY Account wrote:
>To the collective intellect abroad at "AMPS"... <}:<}
>
>Regarding the use of surge limiting resistors in high voltage plate
>supplies:
>
>Once, somewhere, I read that the surge limiting resistor should be of a
>value
>that will drop at least 25% of the plate voltage at 5 times the normal "max"
>current of the supply. And, the wattage rating should be exceeded by a
>factor of 10 or more at that current (so it will fail as needed - like a
>fuse).
Don't know where you might have read that, but it doesn't look like good
advice, for two reasons...
[This information comes from Eimac Bulletin #17 on 'Fault Protection',
downloadable at: www.ifwtech.co.uk/g3sek/misc/bull17.pdf (about 65KB)]
1. The kinds of surge the resistors are protecting against are likely to
be low-resistance arcs, either inside or outside the tube. These are
pretty close to a crowbar short from B+ to ground, so current will not
stop at 5 times normal! It is limited only by the internal resistance
and inductance of the reservoir capacitor... and by the surge limiting
resistor.
Eimac recommend limiting the current to 40A for tubes up to 1500W anode
dissipation, so the value of the surge resistor should be the B+ voltage
divided by 40A (lower values down to even 10 ohms seem to protect the
tubes OK, but Eimac stay on the safe side).
2. The purpose of the glitch resistor is NOT to blow like a fuse! It is
to hang in there and limit the current until SOMETHING ELSE switches or
blows to disconnect the whole system from the mains.
[End of information taken directly from Bulletin #17]
>
>Let's do some math.
>
>If this is correct, a typical application might be:
>
>50 ohm, 100 watt ceramic power resistor in negative lead of a 5000
>vdc 1.0 amp CCS Power Supply
>
>...where,
>
>UNDER NORMAL CONDITIONS (under 1 ampere load):
>
>Plate B+ "Normal" Volts dropped Watts used at
>Voltage max Iplate at "normal" Iplate "normal" Iplate
>
>5000 vdc 1.0 amp 50 volts 50 watts
> (1/2
>the resistor's rating!)
>
That seems good...
>UNDER SURGE CONDITIONS (5 times normal current, or 5 amp surge):
>
Way too optimistic
[...]
>
>It is fairly obvious that the resistor will probably fail before a surge
>reaches, say,
>5 amps because 5 amps takes the resistor to over 62 times its power rating.
>
No, that resistor shouldn't ever fail, because there MUST be something
else designed to switch or blow first.
>Here's the question...
>
>What is the best "type" of power resistor for this? I read somewhere that
>normal
>ceramic power resistors can stand about 500 volts of "drop" per inch, but
>that
>they tend to be very "forgiving" with short duration surges, meaning they
>fail slowly.
>
The "volts per inch" rating is important, because the surge resistor
must NEVER arc along its length. Physically long resistors are obviously
best, and many people have had good experience with the long, thin
vitreous enamel coated wirewound resistors of 15/17W rating.
The glass coating over the wire prevents surface arcing, and a string
made up of typically 3 or 4 10-ohm resistors will handle the normal
power dissipation in a 1.5kW amp.
In surge conditions those resistors will comfortably handle at least
1-1.5kV each for a few milliseconds - which is as long as they should
need to hang in there, before something else shuts down the entire amp.
For higher voltage and power ratings, connect more resistors of lower
values in series.
>Many commercial HV power supplies I have repaired have large 'carborundum'
>type resistors - (expensive and hard to find surplus in the right values)
The carborundum type resistors are highly regarded, because they
generally have a very good volts/inch rating. On the other hand, with a
chain of the smaller wirewound types you can keep the volts/inch well
down, and also fold up the chain to save space.
> in
>series
>with special HV fuses (also expensive and hard to find).
>
Some people use HV fuses, others use a mains fuse or circuit breaker, or
a mains contactor triggered by a circuit that continuously monitors
anode current.
>Would the "dale" type power resistors that bolt to a chassis work better or
>worse?
Don't use them - they are only rated at about 2kV from case to interior,
and the insulation tends to break down with use. Many people won't even
use them for mains voltages. They also have very poor overload
performance.
>( I have attached a picture..! - Virus-free! Promise!)
>
(This is a plain-text-only list, so the list server removed the
attachment automatically.)
>What are the pros and cons of each type? It seems the "dale" bolt-down
>would
>tend to minimize 'shrapnel' in the power supply (if it blows) due to its
>basic design.
>
On the other hand, the metal case acts like a gun barrel...
>What are the "rules of thumb" with regards to joules (watt-seconds) ratings
>in such
>a case? My present project application will produce 5000 VDC at 1.0 amp,
>and I want to be such the tube is protected. I am currently considering 50
>ohms
>at 100 watts.
>
If you intend to run a supply of that size at full ratings, you need the
best protection you can get. A nice loooooong carborundum resistor is
the way to go.
--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek
|