[Amps] 3-1000Z Filament Voltage Question

[David Kirkby]

Densis, I hope you don't mind me putting my reply to your private email 
on the mailing list, but there is nothing confidential in it and your 
mail server is blocking my IP address. I had this happen only a week or 
so ago - I think spamcop has decided onetel is a source of spam, so is 
blocking onetel domains.

I got this back:

4.186.192.25 does not like recipient.
Remote host said: 554 Service unavailable; Client host [212.67.120.104] blocked using bl.spamcop.net; Blocked - see http://www.spamcop.net/bl.shtml?212.67.120.104
Giving up on 64.186.192.25.

Now to my reply.

Dennis Sokol wrote:

> Dr. Dave:
> Thanks for the input. You are absolutely correct. One measurement does 
> not tell the story. I've sampled the plate voltage now over several 
> days. I was wrong in my original post. HV is 2825 to 2880, averages
> 2850 key down with 750 ma plate current.

Those figures give a regulation of +/- 0.9% about the mean. Unless I'm 
very much mistaken, that is not going to be realistic.

> Getting about 59% efficiency in Class B which is a bit lower than 
> expected. I still think that low filament voltage contributes to this 
> so my next effort is to move the tap on the fil xformer and insert a 
> dropping resisitor in the primary to get the voltage right. After I do 
> that, I'll report my results on the reflector.
>
I think moving the tap and adding a resistor is sensible, but I would 
check the mains voltage a bit more often. Unless the mains voltage where 
you are is one hell of a lot better than here in the UK, the variations 
will be a lot more than than what you quote.

One option, which is not 100% accurate, as it assumes the source 
impedance of the mains does not change, is to measure just the mains 
voltage (not on the amp). Sure it will drop when the key is down, but if 
you make the assumption that the impedance of the mains does not change, 
you can ignore that, as it will just scale. A decent DVM will be able 
to  measure min/max/average etc just of the raw mains. I have one here 
that would be easy to leave on for two week.

If you get a +/-5% change in mains voltage, it is hard to see how you 
can expect to get less of a change in HT (and heater) voltage, unless 
the impedance of the wiring changes to compensate, which is a bit 
unlikely. 

I'm not convinced reduced heater voltage will lead to lower efficiency 
either. Low heater voltage will of course limit the cathode current, but 
will just give less power, not lower efficiency. However, there may be 
other effects that My guess (and its only a guess) is that
* your heater voltage is not affecting your efficiency or power
* if the heater voltage was very low, where it did limit cathode 
current, the tube would tend to edge towards class C, as it would be an 
on/off device, with the on-current limited by the  cathode emmision. 
That being the case, I might expect efficiency to improve!

BTW, what are you using to measure the RF power with? DC voltage and 
current can be read to better than 1% with a half-decent DVM, but RF 
power can not *easily* be measured to such accuracy. You state the 
efficiency is 59%, but if you are using a Bird through line to measure 
the RF power, I would say they are no better than +/-10% of FSD. If you 
can only measure the power to +/- 10%, you efficiency calculation is not 
going to be any better.

The spec of 5% of FSD for Bird 43's is not true - I have bought a new 
element and it was out, and tested lots of them as part of a sandwich 
degree placement at the MOD. We could measure RF power quite accurately 
(can't recall how accurate), but I can assure you the Birds don't meet 
the spec that is claimed of them. If you want to measure power 
accurately, use a calibrated attenuator and decent RF power meter from 
HP/Agilent. If you want to do it even better, then at high power you can 
use a water calorimeter.

-- 
Dr. David Kirkby, 
G8WRB

Please check out http://www.g8wrb.org/ 
of if you live in Essex http://www.southminster-branch-line.org.uk/