Hi, David,
Regarding the FCP 160-80 meter dual banding project and circuit component
sizing particulars ... I completely understand the urge to downsize.
However, the devil is in the blasted details.
This question faithfully recurs from time to time: Can we use smaller wire,
use a smaller core, smaller relays, etc? That is why I have replied to the
list instead of off-list. You have some company in your inquiry, and likely
will have more. I'll also likely turn this into an FAQ entry.
The question originates in the understandable but incorrect idea that the
only or essential reason for our using larger components is to prevent
burning them up while running QRO.
Back in pre-publish days Jack and I were locally famous for burning stuff
up and setting utility buildings on fire. And we DID work to have our
creations not catch on fire. But ultimately we realized that sneaky
accumulated miscellaneous losses were more sinister than fire risks. With a
main and ultimate goal of helping people to get out and make QSO's, we had
to refocus our thinking on eliminating loss to continue with new discovery
and progress.
In the FCP, and related projects in the loss list, the central design aim
is to REDUCE LOSSES. Avoiding burning up at QRO power levels IS one, but
ONLY ONE benefit from using larger components. And it is NOT the essential
or controlling issue. Nearly all of the sizing decisions on k2av.com have
to do with reducing dB loss, considerably more critical to results at QRP
than it is at QRO.
Loss reduces performance, and loss will affect QSO's far more often at QRP
power levels. QRP operators are most likely to be in the noise at the
distant end, even when QRO's extra 25 dB puts that QRO signal well out in
the clear and armchair copy. When one's TX signal is in the noise at the
far end on difficult propagation paths, all those recovered tenths of dB's
and possibly a few entire dB's added together can enable copy and QSO at
the far end.
The subject of toroid core sizing can be difficult to understand and
explain, but downsizing DOES have costs. In toroids with high coverage
windings, all else held equal, eddy current loss in watts increases with
the square of the inverse of the toroid volume, thus putting smaller
toroids at a disadvantage. The smaller inside circumference of smaller
toroids may force the use of smaller wire, with less insulation and more
copper loss in the winding, to keep the same number of turns in the
winding. These are the two main problems with using the T200 powdered iron
cores for FCP isolation transformers. Again, this hurts results at QRP more
than at QRO.
Separately there are the issues of breakdown voltages and capacitor RF
current handling. We can be tricked if we don't pay attention to some
sneaky square law subtleties.
Sizing of relays in this project involves breakdown voltage, not power
dissipation. Unfortunately for expense scrimpers (count me in that group),
voltages reduce only with the SQUARE ROOT of the reduction in power.
At 1500 watts NEC4 shows the RF voltage at the FCP shorting relay to be 8
kV peak-to-peak. Adding a 50% safety margin for situations not in the model
sets the required voltage rating at 12 kV. Dropping power from 1500 to 100
watts lowers voltage approximately by a factor of 4, to 3 kV which includes
the safety margin. Dropping from USA 1500 to British 400 watts only cuts
voltage numbers in half.
On the bright side, I have seen Jennings RD5B vacuum relays, well-suited
for the FCP shorting application, go for as little as 35 USD on eBay. I
paid a surplus house 90 USD for the one I am using.
Choice of a non-vacuum capacitor for the circuit does involve breakdown
voltage. BUT far more importantly for solid capacitors, the choice requires
knowing their internal resistance and/or their current carrying capacity.
Here, as with voltage, the current reduces only with the square root of
power reduction.
Regarding vacuum vs. solid capacitors, there is a large gap between the
current handling capacity of various film type capacitors and those using
special material such as HEC-50 series transmitting capacitors. Three HEC
170 pF in parallel for a fixed 510 pF for the 80m network cap can
reasonably handle 12 amps RF circulating current in the circuit.
Other types of caps seemed physically large enough, but had no published
manufacturer's ratings for RF current. They began overheating in only three
minutes at QRO, easily seen in quickly increasing SWR following the
heat-driven change in capacitance. In minutes, the caps were too hot to
touch. At my location at 1500 watts, only the HEC's and vacuum caps have
survived. More important, caps with runaway heat at QRO would have the same
dB LOSS at QRP, but offering no loss warning from smoke and wild SWR's when
TX.
Alas, there seems no cheap or junkbox way to construct without breakdown or
loss.
Using the Deltrol relays with increased contact spacing seems to clearly
cover voltage from only 100 or 200 watts, or possibly "no cheating" 400
watt limits. With a hi-pot meter that measured up to 15 kV we measured
breakdown of a Deltrol modified for two 80 mil (2 mm) gaps
series-connected. The relay started to arc across all four contacts at only
6000 VDC.
Scientifically, getting only 6000 VDC breakdown from the Deltrol's two gaps
in series, instead of double the single gap measurement of 4000, seems to
involve something about reduced gap resistance once an arc is in progress,
thus not allowing the two ratings to add in series. This *measured* DC
breakdown result tells us that a single modified Deltrol relay with
in-series gaps enlarged to 2 mm is at its properly engineered limit at 400
watts and higher power levels are eating the safety margin.
I note, as one correspondent suggested, that using TWO modified Deltrol
relays with their four 2 mm gaps in series would seem to handle QRO. But
the second relay only adds 2 kV per additional gap or 4 kV additional for
10 kV. This at least exceeds the expected QRO RF voltage of 8 kV RF
peak-to-peak at the FCP shorting point. But it reduces the safety margin to
25%. Maybe not so good an idea for a non-sealed relay used outdoors with
connections through the weather enclosure.
At this writing we only have good reports using non-vacuum relays with the
modified Deltrol (or similarly modified equivalent relay) at stations
operated at 100 watts.
Note that the ubiquitous RJ1A genre of vacuum relay types, with their
typical 3 kVDC ratings also DO NOT rate at QRO or even 400 w in this
application, retaining a safety margin only at 100 watts. The modified
Deltrol handles more at a fraction of the cost.
Personally I am using Jennings vacuum relays RD5B (SPST NO) and R3B (DPDT)
in my 160-80 setup. The R3B is a sentimental keepsake from the estate of
W0UCE, used on his 160-80 switching setup, and now on mine. Aside from
dropping it on a concrete floor or adjusting it with a sledge hammer, the
R3B seems indestructible. I can assure you Jack never gave it any rest.
Again, David, I do understand the urge to downsize, but would respectfully
suggest you not do it.
Personally, being able to do 160 very well, AND using the 160 wires to ALSO
get the top-drawer performance of an 80 meter end-fed half-wave L, **in my
only remaining place on the property adequate to erect such a wire**, would
be worth a lot of money for relays, vacuum caps and whatever. I must also
admit I did throw up some cheep stuff that didn't survive, just to see what
happened. I really didn't expect them to survive, and they didn't.
If you do try the cheep route, it would be useful to know for the record
what you tried at what power, and which of them arced, smoked, incinerated,
melted, etc, and which did not. Something like "A fellow from the UK
already tried that and it went up in flames," is a very useful tidbit for
some correspondents. (I don't give out source call signs unless I have
permission.)
Just remember that if it heats up, melts or breaks out in flames at QRO,
that the now-demonstrated large dB loss will also be present running QRP
and LP.
73 and good luck,
Guy K2AV
On Tue, Nov 13, 2018 at 5:25 PM David Cutter via Topband <
topband@contesting.com> wrote:
> I am planning an inverted L over a folded counterpoise as K2AV.
> My power levels will likely be about one third of the design levels for the
> K2AV design. I can imagine during tune up or adjustments that I might hit
> half the power level for short periods, so my component values need not be
> so extreme. I'm thinking mainly of the vacuum variables and other caps. I
> intend to use the 80m option, so relays will also have a relaxed spec.
>
> Has anyone made a lower power version or can advise thereon?
>
> David G3UNA/G6CP
>
> _________________
> Searchable Archives: http://www.contesting.com/_topband - Topband
> Reflector
>
_________________
Searchable Archives: http://www.contesting.com/_topband - Topband Reflector
|