Capacitive droppers at switch-on.

[ENGLISH VICTOR]

Well let's have another look at the theory of capacitive mains droppers with diode protection. I beg to differ from Paul's opinion on a couple of points first it might help to consider what happens when a diode is connected in parallel with the heater chain with a resistor as the dropping element when the mains polarity reverse biases the diode the circuit funtions as if no diode existed, on the conducting half cycle however the heater chain receives only 0.6 volts, lets say nothing. Therefore the heater chain is only receiving half the power it should and would require the voltage to be increased by 1.414 to return the power to the correct value, the value of the dropping resistor would need to be reduced to bring this about., wasting power. With a capacitive dropper the situation is quite different in that no power is wasted during the diode conduction period, the capacitor just charges, it's source and load impedences being virtualy zero, as the current leads the voltage by almost 90 deg we have the classic wattless current that the supply authorities so dislike. Again the value of the cap needs to be reduced to such a value that it's reactance (the reciprocal of (6.82*50 for 50Hz. mains) supplies the correct power to the heater chain.Paul is correct in thinking that the diode will do nothing on the reverse cycle but remember it reaches it's peak in only 10mS from the zero datum thereafter it is a down hill current untill the mains phase begins to reverse, it only needs to rise to 600mV before the diode conducts, being to all intent and purposes a S\C and rupturing the fuse. The 10mS of applied mains does not impart enough energy to blow the heater chain normaly.
A variation to keep the capacitor the normal value would be to disregard the diode I suggest and use instead a pair of back to back power zenner diodes who's voltages are equal but exceed the peak heater chain voltage by say 15%. Finaly regarding Colin's comments I feel the idea of using a resistor to minimise the S\C damage excellent, I would have used a lower value only to maximise the wattless current other wise we are back to a dropper.I have not thought about the need to reduce the value of the capacitor as in Colin's case, it may be a function of the 220 Ohm resistor. The reatance of a 4.0 uF capacitor should be correct for a basic 300mA heater chain using just a few EF80s say six of them from 240V A.C.
For those who don't believe the wattless curent theory carefully check the temperature of the dropping capacitor, it will indicate the the POWER being lost to supply the heater chain. BTW using a parallel diode as originaly suggested does limit the available power to the heater chain, the zenner approach is fine for higher voltage chains the single diode is fine for say a 40 volt chain in a radio using 300mA valves.
Personaly I have used all methods and prefer to use either just one good qulity cap or better still a mains transformer.
Victor.

[ENGLISH VICTOR]

Sorry Colin I read the value of your surge resistor as being 220 Ohms not 22 Ohms as you in fact wrote. 22 Ohms is around the value I had in mind, an excellent compromise between efficiency and safety. BTW I never considered the cap diode to operate in the same manner as the diode resistor system, in the latter there are no reactive components to consider.
Regards Victor.

[dave_n_t]

Originally Posted by ukcol

This would involve using a fixed high wattage resistor in place of the chain (for example 1070 ohms for a 107 volt 100mA chain) and measuring the voltage developed across it by a given value of capacitance using a true RMS meter.

A trick used by G.A. French (of Suggested Circuits fame, in the Radio Constructor magazines) was to make a 'true RMS meter' by using a filament lamp and a photocell to measure its brilliance. I think he used an l.d.r. and ohmmeter at the time, but a phototransistor-based circuit might now be easier to resource. The lamp/ldr combination was calibrated using DC (whose rms value is somewhat trivial to measure).

[The reason he was using this 'true rms' technique was to verify the calculations behind silicon rectifier/resistor heater droppers .... ]

dave

[ukcol]

Any chance of a copy of the article Dave?

[paulsherwin]

Originally Posted by ukcol

Well actually you can protect a heater chain from the capacitor going short circuit using this method, I have done it myself and talked about it on this forum at the time.

I stand corrected Colin and Victor. What I *meant* to say was a protective diode used in this way would require a modified dropper cap value, and I didn't know how to calculate that. You two have obviously cracked this issue.

Paul

[dave_n_t]

Originally Posted by ukcol

Any chance of a copy of the article Dave?

I'll have a dig thru' my mag. store.

dave

[ukcol]

Thanks Dave.

[ukcol]

Just a couple of points to mention arising out of Victors posts.

The usual arrangement for the diode/resistor dropper is for the diode to be in series with the whole chain. During the half cycle that the diode is not conducing no power is dissipated in either the valve heaters or the dropper resistor. By placing the diode across the chain you are (as Victor says) wasting power in the dropper during the half-cycle that the diode conducts. The former method is the only one I have ever seen in manufactured equipment.

I am sure Victor is aware of this and was using the parallel example only for comparison in the discussion but I wanted to make it clear that the series type of resistor/diode dropper is the only type to use.

On the subject of fault current limiting; open circuit failure of power silicon diodes is almost, but not completely, unknown - with one exception. In a TV set that uses four discrete diodes to form a mains bridge rectifier a very large current can flow through a s/c diode and one of the remaining good ones. If there is not sufficient surge limiting in the design this current can be large enough to blow both diodes in half before the fuse has completed it rupture. In the cap/diode dropper if the diode blows apart and the fuse remains intact it spells disaster for the heater chain; hence the inclusion of a resistor.


Paul - I wish we did have it cracked, we still need a mathematical model.


Note on spelling.

I aways use a spell checker because my spelling is not the best. In this post I mis-spelt "rectifier" and after a spell check I had that well known device the "bridge ratifier".

Sorry that's taking us off topic, back to the subject.

[ukcol]

I did an experiment today based on the original observation (see 1st post in this thread).

I set up a "race" between a capacitive dropper and a resistive dropper to see if the capacitive type has indeed the slower start time.

The circuit is shown in the thumbnail attached below.

L1 and L2 are 240volt 200-watt tungsten lamps. C1 is a 14uF capacitor (a motor run type) and R1 is a high wattage 100-ohm wire-wound resistor.

At the time of the experiment the mains supply was measure at 233 volts. When the circuit is running there is 166 volts across L1 and a calculation suggested R1 should be about 100 ohms to give a similar voltage across L2. Using a 100-ohm resistor the measured voltage came out at 161 volts.

With cold lamps, when S1 is closed L1 takes about half a second to get to its target brightness, but L2 reaches target brightness much more quickly, in about a quarter to a third of the time. (Please note that all these times are not measured but judged subjectively). Judging short time periods is difficult, but by setting this experiment up as a race showed clearly that the capacitive dropper had a better "slow start performance" and by implication a better anti-surge performance. It is also important to note that this is only showing this result for a particular ratio of lamp to dropper impedance and may not be so for significantly different ratios.

I did a further test. With the circuit running the switch was turned to the off position and then turned on again almost immediately. This was done to show what would happen if the lamp filaments were warmer at switch on. When this was done the difference in start up times of the two lamps was less but still significant. I think this result is to be expected given the way the slow start mechanism works.

Originally Posted by dave_n_t

A trick used by G.A. French (of Suggested Circuits fame, in the Radio Constructor magazines) was to make a 'true RMS meter' by using a filament lamp and a photocell to measure its brilliance. I think he used an l.d.r. and ohmmeter at the time, but a phototransistor-based circuit might now be easier to resource. The lamp/ldr combination was calibrated using DC (whose rms value is somewhat trivial to measure).

Dave sent me a copy of this article and the technique use to measure True RMS may prove useful to analysing the capacitor/diode dropper. When I get round to trying it I shall let you know how I get on.

[dominicbeesley]

Would it also be possible to eliminate, or at least reduce, the initial cap charging surge with a small choke. I reckon that its only in the case where the circuit is switched on at the "wrong" point in the wave form so something that is reactive to a high freq but is negligible at low frequencies?

Dom

[ENGLISH VICTOR]

Hi all,
you are quite correct Collin in your assumption regarding comparison of circuits. Diode, inductive and capacitive droppers can so easily open a can of worms but at the same time in forum such as this can promote excellent and informative discussion.
Thankyou Paul for your kind words.
Victor.

[ENGLISH VICTOR]

I should have written in my thoughts on cap mains droppers have written: the value of the capacitor should be increased" not decreased as I wrote, I was probably thinking of the decreased reactance needed It does in fact become obvious by further reading. Just goes to show I should read my own copy before posting.
Victor.

[kibble]

I have used capacitors as droppers for several sets successfully. I never add a surge limiter resistor as it seams to be pointless, the valves always takes longer to heat up than when using a resistive dropper. In my simple way of thinking if it's OK to put a capacitor of several microfarads across the mains unprotected such as for PF correction in strip lights without blowing a fuse, then it's unlikely to blow a heater chain on switch on. As for protection against the capacitor going short, I usually put a VDR across the heater chain. They are small & cheap but surprisingly effective at blowing fuses when an over voltage occurs.

[paulsherwin]

Originally Posted by kibble

I have used capacitors as droppers for several sets successfully. I never add a surge limiter resistor as it seams to be pointless

I think 'surge limiter' is probably a misnomer, certainly after following the discussion in this thread. A small resistor is handy to fine tune the voltage though. as it can be difficult to get this exactly right without paralleling lots of small X2 caps.

Paul

[kibble]

From my experience I have found difficulty in making adjustments to the heater current by adding a resistor. It appears to me that the current remains near constant for a given value of capacitance. The resistance & voltage of the heater chain whether 6 or 120 volts makes little difference to the value of C, only the current seams to be important. With the heater chain shorted maximum current flows through the capacitor, this value does not reduce much when the heater chain is put back in circuit. Not what I would expect but what I observe. I find it easier to add a second lower value capacitor to trim the current than adjust resistance.

[kalee20]

Kibble's observation ties up fairly well with what theory would suggest.

Operating from 240V 50Hz, if 6V appears across the (resistive) heaters, then 239.92V appears across the capacitors (voltages adding in quadrature). So, for say 100mA current flow, the capacitor must be 1.326uF.

If you add a lot of extra valves to raise the heater voltage to 120V, then you still have 207.8V across the capacitor, so the current flow is still going to be 86.6mA - not a massive drop from 100mA.

It's not until the heater voltage drop exceeds 71% of the mains voltage, that the heater chain resistance really starts to make itself felt on the current.

[kibble]

Thanks kalee20.
I felt a little uncomfortable describing my experiance without being able to provide an explanation. Ohm's law is fine but j notation