profile | register | search | main site
|Forums | Crystal Radio Think Tank ** Advanced Forum | Double slider question.||Post Reply||Send Topic To a Friend|
|golfguru||Posted - 19 February 2013 20:0 |
Supposing there are *actual* (measureable) losses involved with "dead-end" turns, if a double slider were adjusted, such that there are always ~equal numbers of "dead-end" turns at both ends of the coil (for the required inductance), would that eliminate the problem? (ie. "phasing" and all that).
(NOTE: As in the case where the coil would be used as an inductor only, not as a tappable auto transformer.)
Edited by - golfguru on 2/19/2013 8:05:43 PM
|homebrew||Posted - 20 February 2013 7:8 |
Dead end turns do not seem to be a problem in the broadcast band. In longwave circuits where low stray capacitance is desired the early expensive radios used dead end shorting switches.
Years ago the question was raised here about shorting out the dead ends. At least one experimenter got lower Q readings with the dead ends shorted out. Just the opposite of what I expected.
At the time I was using a commercial long wave loading coil on my antenna tapped every five turns. In the broadcast band I could tell no difference in tuning and volume with the dead ends shorted. Normally I had the dead ends at ground potential to minimize hand capacitance and inter component coupling problems. That would probably work best on a double slider set.
|Garry Nichols||Posted - 20 February 2013 8:50 |
Are we talking about leaving the unused end turns open circuit vs shorting across them?
Or are we talking about either of the above vs the hypothetical case of unused turns removed from the coil?
I vaguely recall from a booklet on ham radio mobile antennas that unused turns had a less degrading effect if left on the ground end of the coil and shorted across.
I'm interested in the opinions of the more learned on this.
My guess is that for coils of ordinary construction used in the BCB it may be hard to tell the difference of any of the above three alternatives (open, shorted, removed).
If the coil is of ordinary construction, two could be built (one with turns removed) and all three combinations tested. It would be easy to test shorted vs open on one coil.
|_J_||Posted - 20 February 2013 14:42 |
There are actual measurable losses in dead-end turns. However, dead end turns are sometimes not very significant compared to the rest of the coil losses, so may not be so bad if the rest of the coil is not very efficient.
To your question, reducing the currents / voltages in the dead end turns will reduce the losses. I don’t think dividing the end turns on each end can cancel out any currents in them though. So cancellation would not seem to help losses. However, it may still reduce losses from another mechanism; it will reduce the voltages at the ends to maybe half. This will reduce the losses a little. Here is my reasoning: since power loss is proportional to V*V, when you half the voltage = v, then v*v loss at each end still totals less than V*V.
Does that make sense?
|golfguru||Posted - 20 February 2013 16:27 |
>>>>> Does that make sense?
Yes. Maybe 4 times less?
>>>>>>>>>> Are we talking about leaving the unused end turns open circuit vs shorting across them?
Edited by - golfguru on 2/20/2013 5:05:11 PM
|Garry Nichols||Posted - 21 February 2013 10:2 |
I was suggesting removing turns as an experiment to see how it compared to unused turns. I was thinking unwrap the coil until a test station is tuned in. Laborious, I know! But you would get into the ballpark by seeing where the slider tuned in the same station.
Just thinking out loud . . . .
|golfguru||Posted - 21 February 2013 16:0 |
It was just a passing thought. I guess I am not worried about it enough to actually do any experimenting :-)
|Nyle Steiner||Posted - 22 February 2013 5:17 |
I experimented with this a couple of times in the past and I found that there was not much difference between leaving unused turns of an LC tuned circuit open and having unused turns shorted. The tuned circuit seemed to work quite well under these two conditions. I did notice however that putting a resistor across the unused turns made the LC tuned circuit operate very poorly.
|Richard||Posted - 22 February 2013 9:33 |
I've measured slightly higher Q readings in my NBS-121 set's detector coil when unused turns were left open vs. shorted.
In an NBS type set, unused antenna and detector coil turns should be confined to the 'cold' end of each coil with the hot ends facing each other for closer coupling, especially needed as one tunes beyond one Mhz and into SW frequencies.
I've also experimented with disconnecting unused turns from the active part of the detector coil (using switches) but it really doesn't seem to make an audible difference with solid wire coils.
The differences might be greater with litz wire coils but I haven't yet performed that experiment.
Edited by - Richard on 2/22/2013 9:36:07 AM
|_J_||Posted - 26 February 2013 13:51 |
Seems like, in theory (FWIW, never tried it) in a very low impedance radio leaving them open should work best; in a very high impedance radio, shorting them should work best.
|Bob Weaver||Posted - 26 February 2013 18:13 |
IMO, shorted turns are never a good thing. You are trading a small amount of self-capacitance (a non-loss reactance) for circulating currents and resistive I^2R losses.
However, the effect of shorted turns varies depending on how closely coupled the overall coil windings are. For example, if you wind a coil on a ferrite toroid where the turns have extremely high coupling, and then short out just a single turn, the inductance will drop drastically, and the Q will go nearly to zero.
On the other hand, if you have a space wound air core coil, and you short out a turn or two at one end, the shorted turns will only negatively affect one or two adjacent turns. So, the overall effect won't be as bad.
|_J_||Posted - 26 February 2013 21:56 |
But open turns are a problem too. Open end turns on a hi Z radio develop substantial voltages, acting like a Tesla coil. Can't that cause RF dielectric losses and V2/R losses in the insulation that exceed what the I2R losses caused by a tiny current of the shorted turns? Shorting a Tesla coil almost completely unloads the driver circuit.
|homebrew||Posted - 26 February 2013 22:40 |
You see dead end shorting switches in early longwave sets and loose couplers that cover LW to 7 MC or so.
These were solid wire cylinder coils used with tube circuits mainly. I suspect that one main design reason was to reduce unwanted intercomponent coupling from the dead ends acting as antenas in the compact sets.
|Bob Weaver||Posted - 27 February 2013 19:47 |
But open turns are a problem too. Open end turns on a hi Z radio develop substantial voltages, acting like a Tesla coil. Can't that cause RF dielectric losses and V2/R losses in the insulation that exceed what the I2R losses caused by a tiny current of the shorted turns?
It depends on whether we are talking about transmitters or receivers. In the output tank of a transmitter, then yes the situation can arise where there are very large resonant voltages generated on open unused turns, and for that reason, these are often shorted out. These are always space wound coils so that the shorted turns have less effect on Q.
I can't imagine any situation in a crystal set or regen where you could generate the kind of voltages that would be a problem, unless the insulation/core material is incredibly lossy. The unused portion of the coil would have to be considerably more than 50% of the whole winding, and if you shorted out that much of a winding, then I would expect it to have a serious effect on Q.
I guess the best way to find out is to do some experiments.
|Richard||Posted - 27 February 2013 21:8 |
"I guess the best way to find out is to do some experiments."
I agree. Speculation settles nothing. Experiments do.
|homebrew||Posted - 27 February 2013 22:30 |
Experiments have been done. Reread the posts. It appears that dead ends are not a problem in the BCB. At least as far as hearing a difference.
Experiments would have to be across a broad range of frequencies and set Q's, designed to ID small differences.
And some definition of how much dead end to use. Maybe 25% of the coil length? Or just add 25, 50, and 75% to the experiments?
Q meter comparisons, local boomer, fading weak stations, and signal generator tests?
Click Here To Close Thread, Administrators & Moderators Only.
Show All Forums | Post Reply