TheRadioBoard

i started this as Alpal and i started stealing this thread.
http://theradioboard.com/rb/viewtopic.php?t=1557

first some from that earlyer thread.




alpal asked a few more questions and gave comments but i felt that if he want them here he can introduce them, even including this felt almost like, uhm, theft?, kidnapping? (yeah I am weird, well aware of it).

i use the active probe I built in an Al box with schematic from here, to be honest i have built 4, first was not that good, got some oscillations, scrapped that and second worked.
http://www.crystal-radio.eu/enfetamp.htm

after a while built a third, modified it by using a larger "window" in front (15mm) , and by putting the first FET in another position (leaning slightly away from "bulkhead" to reduce stray capacitance, then built version 4 almost exactly same now ready to send to the person that paid for it (he lives just 40km away, I will NOT make one and mail, I feel they are to delicate for that)

My last 3 versions measures about 40Mohm input resistance, so quite a bit less of a load than your 10M.
also as the Al box is grounded to frame of cap via a crock clip on end of a 150mm long 1mm bare copper wire screwed to box I get almost no hum.
I thought it would increase load, but it turned out that it did little damage, seems the capacitive coupling to the surrounding is more lossy for me.

oh, and I use same type of lead for the input, only there I use a 0.6mm silver plated copper wire.

I started using the bare wires after Electronic Technician over on rap'ntap noticed that the PVC insulation could drag Q down over 100 units on a 1000Q coil.

i will include images later, Monica has the digital camera at her sister just now (they are makig a "my first year" book for her youngest daughter)

I have used a O-scope, multi times repaired probes, 3 DB load, and a old signal generator for years to design coils looking at the 3DB bandwidth. All are questionable as far as accuracy.
Comparing my notes with similar coils on the boards I get a 30% lower calculated Q reading on the solid wire wound coils. I suspect that the reading would be a lot farther off on higher Q coils.

yes i have fond that care needs to be taken to get repeatable results, +-5% is about my limit for same tank on different days.
I have a "test coil" o use as "reference" the median Q on that coil is well known together with my test cap.
That cap was selected among 23 candidates (a 3 day affair) and is my "reference", from time to time I test new candidates with that as reference, to date it has stayed on top (it came from a HP signal generator).

Hi Guys,
With all this Q talk, I decided to try some experiments. Before I got my
HP Q meter, I measured my coils using a signal generator, good capacitor
and my scope. I used what I thought was my best capacitor. This was one
of those ones that is very heavy, with good ceramic ball insulators, great
wiper arms and so on. It is pictured below in the first picture.

I can measure how "good" other capacitors are by measuring the Q of a coil,
with my HP Q meter with the internal capacitor and then adding a test
variable capacitor. I reduce the internal capacitor of the Q meter and
substitute the test capacitor. My measurements were at 1200 KHz and
used a 150 uH spider coil made with 165/46 litz wire.

The measured Q using only the internal capacitor is 610.

Using the first capacitor below, the Q was 530.

I then used a capacitor from an HP generator, with ceramic and three
wiper arms shown in the second picture. I was surprised that this also
measured a Q of 530.

I then grabbed an "XSS" type capacitor and did the same measurement.
This time I got a Q of 410. This was not surprising.

This tells me what many of you already know and that is if you want to
measure Q, use the best capacitor you can find.

Dave





\

_________________
Please visit my website at: http://www.makearadio.com

actually,my 3 best caps are in my competition set (the ATU is my test tank.-/ )

OErjan


Thanks for that pic & the advice.
I was going to build such a device at one stage but I didn't have the
time to do that & the other mods I needed to do on my Lyonodyne
Xtal set. Do you have trouble with 50 Hz mains getting into your
measurements? High impedance circuits pick up the mains easily.
I heard of another method just using a 100 M ohm resistor to
an oscilloscope input.
I wonder if a differential type amplifier design would be the best
to get rid of 50 Hz hum?

I also wonder if any variable capacitor in the world can really take
full advantage of 660/46 Litz wire?
46 gauge wire has a diameter of 1.57 thou or 0.00157 inches.
This means that it's circumference of exposed surface area is 4.9323 thou.
If we multiply this by 660 we get an equivalent area of 3.255 inches
as a straight line.
If we divide that by 2 we get an equivalent strip of thin copper
conducting both sides that is 1.6 inches wide.
The terminals on the capacitor where the Litz is soldered would have to
be 1.6 inches wide to not reduce the RF resistance at that point.
Also the wiper arm contact leading to the rotating part of the capacitor
would also need to have a contact length of 3.2 inches in a straight line.
No such capacitor exists but maybe an ordinary capacitor could be modified
to have these specs & a good Q meter would enable the
results to be confirmed from my calculations.
I have a feeling that the Q of air litz coil tanks circuits could climb to
over 2000 & maybe 3000 if a variable cap was modified.

Just a thought for future experiments!

It's fun to to the calculations, but where is all this leading to, may I ask?

I most definately understand that a high Q coil, capacitor and the whole tank circuit should be as high a Q as possible. But what good is a stand alone tank circuit with exceptionally high Q?

To use it, you have to couple energy to it and there goes all this extremely high Q due to the loading effects, so there you are back to where you started.

What am I not seeing here?
Curt

_________________
Curt, N7AH
CW forever
Connoisseur of the cold 807

Hi Curt. Yes, 'loaded Q' is the other side of the coin but the higher you start then the higher you'll wind up if care is taken in matching - both at the detector and the antenna side.

if I use high Q approach all through the set I end up with less losses, also if I match things I get less losses.

i will try to giva an descriptive example
Q is more or less what losses i have, the higher the Q the less losses.

If I load a tank that has unloaded Q of 2000 at 1700kHz with a good Diode followed by a good matching network from tank via diode to phones, I likely end up close to bandwidth we would get with unloaded coil of say 600 Q and still have a fairly sensitive AND selective radio.

also as Q rises I can decrease coupling to the antenna/ground network and further unload the set and get better bandwidth and still get decent voltages (antenna and ground are a fairly big part of the load to a crystal radio)

so simply striving for highest Q AND more or less as a Siamese twin higher impedance we get more and more efficient sets even if we load them down with Ant, ground, detectors, matching transformers...

OK, thanks. I was just wondering if you were approaching the point of diminishing returns by going to such extremes. I understand about starting off with the best, as everything you add to the circuit will take away from the Q of the tank circuit.

Does anybody know what is the highest Q ever obtained in a tank circuit for comparison purposes? Interesting topic of discussion.
Curt

_________________
Curt, N7AH
CW forever
Connoisseur of the cold 807

for, BCB I think the 2000+ obtained with the tri-filar coil linked above close to maximum Q made this far, certainly not many better have ever been made, if something commercial was ever made with higher Q they would be "holy grail" for all BCB DXers.
As frequency goes up to HF and above things change.

Its easy to get too sharp at the low end of the band. Even a halfway decent set can attain muffled audio. When we still had a station in the area on 690 I could find 'quiet' between a nearby 680 and 690. Being able to do that at the high end of the band is a real challenge.

In that regard a coil with 2000 Q at the low end of the band would be way overkill and would need some dumbing down to be usable.

I often make the comparison of a crystal set to a house of mirrors. EVERYTHING reflects back on everything else. As you improve one thing you have to improve everything else in concert. I figure a 2000 Q coil would need to be hung from the ceiling in the middle of the room since body effects would probably bollox it up.

Hi All,
I saw the interesting thread and wanted to mention I am the one who built the 2000+Q air-core coil that I posted on Rap'N'Tap a while back.

The coil is 6" dia x 8" long for the windings so it is not that much bigger than other air-core coils. It doesn't have to be suspended from the ceiling or anything special to work properly. When I measured it I had the coil elevated about a foot above the Q meter and there was some other equipment nearby so it is not overly sensitive to its surroundings. Just treat it like any other good coil and try to keep it away from metal and other lossy objects.

What I did notice was the volume on weak daytime stations was noticeably louder. The selectivity seemed very sharp on the low end of the band but not overly sharp and it all depends on how close you couple it to the antenna coil. If you want broader selectivity just bring the coils
closer together. A higher Q coil will always give you higher sensitivity and selectivity. You get louder volume for the same selectivity of a lower Q coil by just coupling the coils closer together.

As far as I'm concerned you can't have too much Q in a coil used for a crystal radio set. Q is the quality factor of the coil, so the higher the Q the higher the quality of the coil and the less losses in a circuit!

It is always worthwhile to strive for the highest Q coil you can build!

Also it is always worthwhile to strive for the lowest loss capacitor you can build or buy. I consider a low-loss capacitor just as important as a low loss coil. A typical air-variable with phenolic insulation can drag down the
Q of a high coil by several hundred Q.

I would only use 660/46 Litz wire for interconnects between a coil and capacitor. I have seen the effects of using tinned stranded wire for interconnects and it can drag down the Q a hundred points or more.

73 Todd WD4NGG








--------------------------------------------------------------------------------
New year...new news. Be the first to know what is making headlines.

The poor volume noted may be a result of overcoupling which gives a camel hump double peak waveform. I'm running into that subbing Litz equivalent coils for my solid wire coils. Adjustable coupling appears to be a necessity when using Litz.
Using Litz for interconnect connections has resulted in a net loss over large solid wire runs so far. I normally use 16 ga copper connected at the coil which also serves as a support.