fullsquare@awful.systems 1 week ago
The dielectric between the plates in this case is 0.4mm of ABS plastics (+ a bit of air in the 3d print layer lines).
in terms of losses, PP or PE is a bit better than ABS, teflon or FEP is a bit better than PP, but air is superior to either (this is part of the reason why foam coax is a thing). not sure which ones are printable, or whether it’s practical at this size, but try to introduce as many voids as possible (perhaps requires larger thickness of dielectric). it doesn’t matter much in your case, because of low power (warping of plastic because of excessive heat is probably not a problem). if your coax has solid dielectric, then by introducing enough air in 3d-print your variable might become less lossy than that
The Capacitors allows my 80cm diameter loop to tune from 20Mhz to 37Mhz. Sweeping the whole range is a bit slow due to the low RPM of the motor and takes about 6min. But that is kinda nice when fine adjusting to a frequency.
you have probably noticed that position vs resonant frequency relationship is rather nonlinear. you can get higher sweep speeds at lower end without losing much accuracy at higher end by tapering end of side plates into a triangle shape (it will get longer overall). it doesn’t matter much in your case, because it’s all approx monoband, but if you want to go multiband with this, then it’ll be a nice enhancement. similar effect happens when air variable capacitors have moving plates shaped in such a way that one end is longer than the other, and external edge has shape roughly like a section of logarithmic spiral. precise movement of variables like this is done by use of worm drive with large wheel
I am not sure what is causing this, but i assume it could be due to increase of dielectric losses in the capacitor getting bigger when more of the plates overlap because then the electric field has to flow thru a bigger area of dielectric, increasing the potential for losses.
loss tangent of dielectric is material property, that is ratio of equivalent loss resistance to capacitance should remain constant at given frequency. so i guess that losses should remain roughly the same, if dielectric is to blame, but at any rate lossy capacitor should make bandwidth broader and SWR lower. my guess would be that it’s a matter of coupling loop becoming wrong-sized or wrong-positioned at some point with change in frequency (try moving it up or down? there’s gotta be some optimum position for your entire range of interest)
einfach_orangensaft@sh.itjust.works 1 week ago
ok, so the next version gona have a air gap then
great idea, gona add that to the next version too
thought about that too, the capacitor kinda has long leads on it, they could act as circumference extensions for the loop, i tryed shortening the loop a bit and it improved the SWR a bit, i now managed to get it down to 1.46 at 27mhz (before loop shortening it was 2.6). sadly that seams to be the limit of what shortening can do. Now got a SWR of 1.05 at 31,58Mhz…hope the next iteration of the capacitor moves that down to 27Mhz
fullsquare@awful.systems 1 week ago
i was thinking more like, thin external plastic shell and empty cells inside, perhaps with another thin plastic shell on inside, and internal metal shell (on plastic support?) fitting in snugly, for mechanical stability, idk 3dprinting
keeping leads short and nonmagnetic (dramatic reduction in skin layer depth) would be a good thing because of losses, but the longest object in capacitor would be just capacitor plates, and either way in wavelength terms it’s rather small. more precisely you can model it as open transmission line stub with some weird and low impedance, but it’s so small that you don’t have to