QRPp journal Vol. IV No. 3
Cheers,
Mark.
|
Kurt is onto something with "no solder".
Tin-lead solder has 10x the resistivity of copper, tin-silver is not much better:
? ??
More of a difference than I would have expected.
I'm sure most amateur built mag loops are using solder.
Currents are so large in a high Q loop that even
a small fraction of an ohm makes a difference.
Would be interesting to evaluate solder vs no solder construction?of a mag-loop for efficiency.
There are plenty of competing naratives about how effective a mag loop is.
Things like the use of solder could definitely contribute to the disparity.
And a dielectric such as teflon that is usually very good might well be marginal in these conditions.
On the other hand, a thin film of solder at an overlap between two sheets of copper
would present a very short path for current flow through solder.
The shorter that path is (and the larger the area of overlap), the less impact it would have on resistance.
Even with 10x the inherent resistivity, the effect of such a junction should be much less than that of
the resistivity of the much longer path through the copper loop.
Jerry
|
No solder...rule number one. ?
Rule number two...no metal other than the loop and coupler.
As far as stability. ?Great point and I can¡¯t yet speak to power more than qrp. ?It works fine for qrp. ??
As far as higher power and stability? ?Looking forward to that exploration. ?If there is a stability issue at high power, it is likely a control problem that now days can be solved with some feedback systems, a $20 microcontroller, and some code. ?Active control. ?They fly bricks that shouldn¡¯t in air now days that way...I¡¯m just trying to make some QSOs from my kitchen table.
73,
Kurt?
toggle quoted message
Show quoted text
On Dec 21, 2019, at 8:57 PM, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:
?Looks like teflon has a sufficiently low loss tangent, is being used successfully as a dielectric on mag loops. I'm sure a vacuum for the dielectric is preferred, but nothing costs way too much. Could be compression, book, or trombone style, but having the cap be one piece with the copper loop should be a win, assuming it can somehow be mechanically stable enough to stay tuned within a few khz. Yes, it would be tough to get this to work well, currents are tens of amps, several kV across the cap. But I don't yet see what is inherently wrong or inefficient with overlapping two strips of copper with a teflon dielectric. ? No soldering between loop and cap? That rules out most of the mag loop construction techniques I've seen. And an argument to try overlapping the two ends of a copper strip to form a cap. Jerry, KE7ER On Sat, Dec 21, 2019 at 01:29 PM, Kirk Kleinschmidt wrote:
Those are all great ideas...until you try them. :)
?
Mag loops don't work (well) unless we build them in ways that only mag loops require. Overlapping strip capacitors, as you describe, while they will "work" after a fashion, have huge losses in mag loop applications, and they don't stay stable with temperature and humidity, etc.
?
Loop conductors should be continuous and NEVER soldered. Capacitors should be vacuum variable or butterfly types with no mechanical wipers/contacts (huge losses). Mag loops in no way work like the antennas we're all used to, and different (weird?) construction techniques are required to make a good one.
?
Read this from Leigh, a VK5 expert in mag loop design and the underlying science. See:
?
There a lots of junky mag loops on Youtube, etc. They work -- as do 6-inch whip antennas -- just not very well.
?
There are some mag loop groups here on .io
?
Properly built mag loops are awesome. The opposite is also true :)
?
Good luck,
?
--Kirk, NT0Z
? Rochester, MN
?
My book, "Stealth Amateur Radio," is now available from www.stealthamateur.com and on the Amazon Kindle (soon)
?
|
I understand that their used to be a "Elmer 101" box available.? My little Builders group is going to be building SWxx's soon and would like to get a hold of this book for the builders.? Anyone have any ideas as to where I can get some of these or even one of these??? I have found all the Elmer 101 lessons on the net but if i can I would like to get the book.
Thanks
Steve NU0P
|
Looks like teflon has a sufficiently low loss tangent, is being used successfully as a dielectric on mag loops.
I'm sure a vacuum for the dielectric is preferred, but nothing costs way too much.
Could be compression, book, or trombone style, but having the cap be one piece with the copper loop
should be a win, assuming it can somehow be mechanically stable enough to stay tuned within a few khz.
Yes, it would be tough to get this to work well, currents are tens of amps, several kV across the cap.
But I don't yet see what is inherently wrong or inefficient with overlapping two strips of copper with a teflon dielectric.
?
No soldering between loop and cap?
That rules out most of the mag loop construction techniques I've seen.
And an argument to try overlapping the two ends of a copper strip to form a cap.
Jerry, KE7ER
toggle quoted message
Show quoted text
On Sat, Dec 21, 2019 at 01:29 PM, Kirk Kleinschmidt wrote:
Those are all great ideas...until you try them. :)
?
Mag loops don't work (well) unless we build them in ways that only mag loops require. Overlapping strip capacitors, as you describe, while they will "work" after a fashion, have huge losses in mag loop applications, and they don't stay stable with temperature and humidity, etc.
?
Loop conductors should be continuous and NEVER soldered. Capacitors should be vacuum variable or butterfly types with no mechanical wipers/contacts (huge losses). Mag loops in no way work like the antennas we're all used to, and different (weird?) construction techniques are required to make a good one.
?
Read this from Leigh, a VK5 expert in mag loop design and the underlying science. See:
?
There a lots of junky mag loops on Youtube, etc. They work -- as do 6-inch whip antennas -- just not very well.
?
There are some mag loop groups here on .io
?
Properly built mag loops are awesome. The opposite is also true :)
?
Good luck,
?
--Kirk, NT0Z
? Rochester, MN
?
My book, "Stealth Amateur Radio," is now available from www.stealthamateur.com and on the Amazon Kindle (soon)
?
|
Hello Jerry,
There is more to this capacitor than meets the eye...but at its basics it is indeed controlled separation ?as a tuning mechanism...but dialectic choice is very important. ?Your formica dialectic wouldn¡¯t work too well, I¡¯m afraid. ?I once tried to make a loop something like this using a playing card as a dialectic and it didn¡¯t work out too well. ? I learned about dialectic loss tangent after than experience.
One way you can tell that you have big losses in a mag loop is that you will have trouble getting a coupling loop to give you decent SWR without it being much larger than the typical 1/5 circumstance of the main loop...or you will need to make it a squished down oval close to the main loop. ? This antenna I show does have some loss, more than a several hundred dollar vacuum variable would give, but it is not too bad for my needs or most people¡¯s needs I suspect. ?This antenna has a bandwidth of less than 40 kHz at 20 meters. ?Teflon is your friend. There is no soldering allowed in my design. ? ?The dirty secret of copper tubes is that everything inside the skin depth is a waste of money. ?My strip of copper here has more surface area than what is typically used by folks making it out of tubes...and it is easy to collapse.
The cap I show is actually breaking my long term rule for this loop. ?In my loop no metal other than the main loop and coupler is allowed above the feed point, but I am using metal push to fit connectors for my Teflon tuning drive in this prototype. ?I will 3D print these in the next prototype.?
The map I show of my WSPR receipts were from inside my house with the loop sitting on my kitchen table. ?But yeah...I¡¯m sure others could do better.?
Kurt
toggle quoted message
Show quoted text
On Dec 21, 2019, at 3:12 PM, Jerry Gaffke via Groups.Io <jgaffke@...> wrote: ?I don't see anything recognizable as a cap in that photo, am curious how you are going about it.
Best guess is it's the metalic assembly above the PC board with coax in, a white wire coming out.
And the black thing under it is the motor drive?
I've never messed with a magnetic loop, but have considered it.
Was gluing formica type laminate to kitchen countertops here a few years ago,
part of it involved cutting long inch wide strips for the vertical edges with a table saw.
Occurred to me then I could glue copper foil to one side of such a laminate strip for a magloop.
And if the laminate proved to be a suitable dielectric for a capacitor, perhaps just overlap the ends,
compress that overlap to tune the resultant cap.? If not, then perhaps attach teflon (somehow)
to one end of the laminate strip for the capacitor dielectric.? The laminate bends easily enough,
a long strip could be wound into something a half meter in diameter for transport or storage.
Jerry, KE7ER
|
Hey, Jerry,
Those are all great ideas...until you try them. :)
Mag loops don't work (well) unless we build them in ways that only mag loops require. Overlapping strip capacitors, as you describe, while they will "work" after a fashion, have huge losses in mag loop applications, and they don't stay stable with temperature and humidity, etc.
Loop conductors should be continuous and NEVER soldered. Capacitors should be vacuum variable or butterfly types with no mechanical wipers/contacts (huge losses). Mag loops in no way work like the antennas we're all used to, and different (weird?) construction techniques are required to make a good one.
Read this from Leigh, a VK5 expert in mag loop design and the underlying science. See:
There a lots of junky mag loops on Youtube, etc. They work -- as do 6-inch whip antennas -- just not very well.
There are some mag loop groups here on .io
Properly built mag loops are awesome. The opposite is also true :)
Good luck,
--Kirk, NT0Z ? Rochester, MN
My book, "Stealth Amateur Radio," is now available from
www.stealthamateur.com and on the Amazon Kindle (soon)
On Saturday, December 21, 2019, 3:12:13 PM CST, Jerry Gaffke via Groups.Io <jgaffke@...> wrote:
I don't see anything recognizable as a cap in that photo, am curious how you are going about it.
Best guess is it's the metalic assembly above the PC board with coax in, a white wire coming out.
And the black thing under it is the motor drive?
I've never messed with a magnetic loop, but have considered it.
Was gluing formica type laminate to kitchen countertops here a few years ago,
part of it involved cutting long inch wide strips for the vertical edges with a table saw.
Occurred to me then I could glue copper foil to one side of such a laminate strip for a magloop.
And if the laminate proved to be a suitable dielectric for a capacitor, perhaps just overlap the ends,
compress that overlap to tune the resultant cap.? If not, then perhaps attach teflon (somehow)
to one end of the laminate strip for the capacitor dielectric.? The laminate bends easily enough,
a long strip could be wound into something a half meter in diameter for transport or storage.
Jerry, KE7ER
|
I don't see anything recognizable as a cap in that photo, am curious how you are going about it.
Best guess is it's the metalic assembly above the PC board with coax in, a white wire coming out.
And the black thing under it is the motor drive?
I've never messed with a magnetic loop, but have considered it.
Was gluing formica type laminate to kitchen countertops here a few years ago,
part of it involved cutting long inch wide strips for the vertical edges with a table saw.
Occurred to me then I could glue copper foil to one side of such a laminate strip for a magloop.
And if the laminate proved to be a suitable dielectric for a capacitor, perhaps just overlap the ends,
compress that overlap to tune the resultant cap.? If not, then perhaps attach teflon (somehow)
to one end of the laminate strip for the capacitor dielectric.? The laminate bends easily enough,
a long strip could be wound into something a half meter in diameter for transport or storage.
Jerry, KE7ER
|
Thanks to all your replies. There is a lot of valuable information. This is my first toroid and I want to make sure I have it right.
I'm winding T3 which is an FT37-61 with 23 turns of #28 and 6 turns of #26.
Looking at:
... suggests that the 23 turns should give me 29uH and I'm measuring 28.4uH on my L/C meter. The calculator on that page assumes #28 wire. I haven't added the secondary yet. I just wanted to make sure I was on the right track and that the 0.6uH is within tolerance.
I'm using one of these:
Thanks to all.
73,
Paul VA3ZC
|
Hello all,
I¡¯m working on a prototype for a low cost magloop. ?Cost of ALL parts including WiFi/low energy Bluetooth ¡°arduino-based¡± electronics and mechanical assembly for driving the capacitor is about $100 (no joke). ?Attached is a pic of the current prototype. ?It is mounted on some form board as a temporary support for testing, but I have plans to make this thing portable with a quick deployment system. ?The loop itself is 10mil 1 inch wide copper. ?Still working on the software, but It will self-tune and be manually controllable from an iPhone.
The capacitor is 3D printed from my own CAD design and still a work in progress. ?My current testing is confined to 20m, but I should be able to modify this thing to cover 10-30m at least...maybe 40...but I would rather have a 2nd loop for 40 thru 80.
?
The capacitor dialectic should handle at least 100W, but since I don't have a non-qrp radio in the house, I don¡¯t know how stable it would be at that power...I plan to stress test this thing in the near future with the help of a non-qrp ham friend. ?It works well at qrp power.
I put it on 0.2 watts of WSPR today. ?My results:
Anyways...thought I would share. ?
73,
Kurt-AE6UJ
?
?
|
Re: Varactor Diode C Measurement
Correction, at 9v the 1n4001 graph shows 10pf, not 30pf:
? wrong:? ?Using the midrange 9v:30pf pair to compute? ?K = 10*sqrt(9) = 30,
? right:? ? ?Using the midrange 9v:10pf pair to compute? ?K = 10*sqrt(9) = 30,
Just in case anybody was trying to make sense of this stuff.
> Perhaps performance at high frequencies, where the 1n4001 might show a bit of PIN behavior?
I usually blame poor performance at high frequencies on capacitance, but not the case here.
So how would a 1SV322 be better than a 1n4001?
To show "PIN behavior" it would have to occasionally be forward biased, that's not the case here.
But is there a similar behavior that might cause trouble at high frequencies?
No diode used as a variable capacitor will have a linear response to voltage, but some are better than others.
This is significant when trying to tune a VFO, don't want all the action to be happening at the bottom end of pot rotation.
How close to linear is not a characteristic of varactor diodes I've ever seen discussed.
Most analysis is a quick look at min and max capacitance.
I think I'm done here, will go back to fiddling with si5351's.
We now return you to your regularly scheduled programming.
Jerry, KE7ER
|
Sorry, this time copied correctly.
|
On Sat, Dec 21, 2019 at 04:22 AM, n3fel@... wrote:
Al Dutcher, EDN April 12, 2007
https://www.edn.com/.../ test... measurement/.../ Cheap-and-easy- inductance- tester-uses-few-components Howard, n3fel
|
It's a good idea to wind and measure the inductance of the toriod before installation.? In my experience not all cores are alike especially the five for a dime variety found on Ebay.? Testing of magnetic material is steeped in theory and the art is well practiced.? Check the litergy on test methods.?
I used a multivibrator design by Al Dutcher, EDN April 12, 2007 that formed a charge/ discharge oscilator and ran at a few Mhz which I viewed on a scope and measured on a frequency counter.? The frequency of oscillation related to the inductance.? A general coverage receiver quite loosely coupled to the large signal of the multivibrator might also work if the oscillation is within range of its rx coverage.? I keep an ancient RS DX302 general coverage receiver on the test bench just for that purpose
Buy the cores from Kits and Parts.? You'll get what you want within a few days of your order rather than waiting 6 weeks or more.? ?Howard, n3fel
|
Re: Varactor Diode C Measurement
The 1n4001 seems to follow the? ?Cj = K/sqrt(V)? ?formula rather closely.
From the graph of fig 4 on page 2 of the 1n4001 datasheet at? ?
? ??
I read the following sample of values:??1v:30pf? 9v:10pf? 30v:5pf? 80v:3pf
Using the midrange 9v:30pf pair to compute? ?K = 10*sqrt(9) = 30,
we get the following results from Cj = K/sqrt(V) = 30/sqrt(V):
? ? 1v: 30.0pf? ? 9v 10.0pf? ? 30v: 5.48pf? ?80v: 3.35pf
The figures computed from? Cj=K/sqrt(V)? follow the graph quite well.
Some diodes don't follow our formula so closely.
Take a look at the figure on the left side of page 2 for the 1SV322:
? ??
Using that figure, I read:? ? 6v:5pf 4.5v:6pf 2.8v:10pf 1.6v:20pf? 0.5v:38pf
Using the 2.8v:10pf center value pair, I find that K=16.73? and Cj=K/sqrt(V) gives:
? ?6v: 6.83pf? ? 4.5v: 7.88pf? ? 2.8v: 10.0pf? ? 1.6v 13.23pf? ? 0.5v: 23.66pf
Note that values from the figure have higher capacitance at low voltages than our computed values.
So this specially designed 1SV322 varactor diode is even more non-linear than the 1n4001.
Do I have something wrong here?
Why would we not use a 1n4001 instead?
Perhaps performance at high frequencies, where the 1n4001 might show a bit of PIN behavior?
Curious.
Jerry, KE7ER
?
|
Hi
The phase noise numbers for the RFzero was recorded subject to a 50 ohms load, i.e. no additional 74xx buffer stage:
100 Hz -103 dBc/Hz
1 kHz -128 dBc/Hz
10 kHz -136 dBc/Hz
100 kHz -138 dBc/Hz
1 MHz -144 dBc/Hz
at 9 MHz. From
Denis, G0OLX, uses an RFzero generating ~100 MHz driving the GB3SEE 24 GHz beacon. The frequency shifting, not the CW-FSK, comes from Denis walking around the lab (Doppler), and the fact that the FT-817 is xtal controlled. Given the x240 multiplication the phase noise is not bad. OK not tuning to the sides, but still good audio. Like any other RF circuits also the Si5351A needs a good home to do its best.
Bo
|
Re: Varactor Diode C Measurement
A minor error in my previous post where I said:? "assume our 10v power supply has 0.1v of pk-pk ripple on it" My subsequent calculations for the variation in capacitance assume noise on the power supply of 0.2v pk-pk, not 0.1v pk-pk. Chuck wrote: > **I did this stuff about 60 years ago.? Kinda hazy now. My college days were back in the 1970's. So I've got a 15 year advantage on you. It's scary to revisit this stuff and realize how much I forgot. For those who forgot high school algebra, this might be confusing:
If we lump all those constants into a single value I'll call K, we get:
Cj = K/sqrt(V)
Assume we have a varactor diode that measures 100pf when we put 1 volt across it.
What is the capacitance when we put 10 volts across it?
At one volt: K = Cj*sqrt(v) = 100*sqrt(1) = 100
At 10 volts: Cj = 100/sqrt(10) = 31.62 pf
We could break it down a bit further:
Cj = K/sqrt(V)? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?# Forget the nasty formula.? All we care about is that capacitance is proportional to? 1/sqrt(V)
Cj * sqrt(V) = K/sqrt(V) * sqrt(V)? ? ? ? ? ? ?# multiply both sides by sqrt(V)
Cj * sqrt(V) = K * (1/sqrt(V)? * sqrt(V)? ? ?# regroup Cj * sqrt(V) = K? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?#? sqrt(V) divided by itself is equal to 1 K = Cj * sqrt(V)? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?#? switch sides So if we know the capacitance Cj is 100pf at a bias of 1v, we can calculate the value of K: K = Cj * sqrt(V)? = 100 * sqrt(1) =? 100? The two different instances of "100" are an unfortunate coincidence, and thus a source of confusion. The first 100 is 100pf, the second 100 is the computed value of the constant K for this particular diode. Now that we know the value of K for this diode, we can calculate the capacitance Cj for a bias of 10v: Cj = K/sqrt(V)? = 100/sqrt(10) = 31.62 pf Jerry,? ?KE7ER
On Fri, Dec 20, 2019 at 07:55 AM, Jerry Gaffke wrote:
Getting back to our original conversation, assume our 10v power supply has 0.1v of pk-pk ripple on it. How much variation in capacitance do we get with this particular diode for a bias at 1v versus a bias at 10v?
|
I have built a lot of QRP radios over the years, both kits and from scratch.? Including the NC40A from the old group's project.? Have probably wound 100s of toroids in the process and have never measured the inductance let alone the frequency response of any of them.? Don't have the test equipment for the latter anyway.
All of the radios including the NC40A have worked well.? Most, including the NC40A are still in use here? The only problem ever encountered due to toroids was getting 43 and 61 mix cores "mixed up" in couple of instances.? Worst case is sometimes especially in PA output networks the toroid windings need to be spread out or bunched a little to hit the sweet spot.
Not criticizing measurement for those with the equipment and inclination. Personally haven't found it necessary to building a working project.? Recently, I find using a cheap multitester to check for counterfeited parts a worthwhile effort.? Avoided installing a lot of bad capacitors and active components recently, all provided in radio kits.
Curt KB5JO
|
Re: SMT Varactor vs Thru Hole
I think the curves on Diz's page are in the best
format for varactor capacitance. We normally consider the
voltage input linearly. If I'm graphing one, I automatically
mark off the V (x) axis in evenly spaced volt increments. But
I'm normally looking for something different on the output.
Marking the C (y) axis linearly makes no sense. We are rarely
looking at the linear response of a varactor because we aren't
really concerned with capacitance. We're normally concerned with
the resulting resonant frequency response in a tuned circuit.
Log/linear can help us with that.
Most of the time, I want a frequency readout to
be linear. A varactor capacitance with a non-linear response to
voltage is going to give me that. On Diz's graphs, I see a
linear voltage input. If I see a fairly straight curve for C,
I'm probably going to get a fairly linear frequency response. So
log/linear makes sense. If he used linear/linear, you would see
a deceptively "hooked" curve that truly shows how C responds,
but it wouldn't be as useful for predicting frequency response.
Log/linear provides more insight that you can use.
Some of the newer hams may not have noticed the
shape of the plates on larger "tuning" capacitors. They weren't
symmetrical. They were very oddly shaped to make the dial
calibration more linear.
There. Marketing guys owe me one.
BTW, I used to follow your website pretty closely
in the Rock Mite days. Fun stuff.
Eric KE6US
On 12/20/2019 6:54 AM, Chuck Carpenter
wrote:
toggle quoted message
Show quoted text
Looking closer
at the chart posted previously?from Diz's?diodes page
SMD/SMT Varactor 1SV322.?would be a good choice.??
The V/C curve is about the same as the thru-hole?MV-209 or?BB910.
Note that the curves for these two are plotted differently:? SMT
part is linear/log and the Thru part is log/linear.? It gives the?
plot a different look even though they are much the same.? Depends
on what the marketing folks want to emphasize...8^)
Plot them linear/linear if you're curious?about?the "real" shape
of the?V/C?curve.
--
Chuck, W5USJ (ex K2OFN)
Point, Rains Co, TX? EM22cv
|
Re: Varactor Diode C Measurement
Chuck, It took me awhile to figure out how to make sense of it. Could be errors, but this is what I am currently thinking:
Here's that equation again, a power of (1/2) means we take the square root:
Cj = A/2 * ((2*q*e/(V0-V))*(Nd*Na/(Nd+Na)))**(1/2)
V0 is very close to zero volts so we can ignore it, and V is a negative voltage.
A is the cross sectional area of the cap, Nd and Na have to do with doping levels.
Those all get determined when the diode is fabricated, and won't change for us.
e (epsilon) is the permitivity, should be about the same for all silicon diodes.
q is the charge of an electron, determined back when the universe was born.
If we lump all those constants into a single value I'll call K, we get:
Cj = K/sqrt(V)
Assume we have a varactor diode that measures 100pf when we put 1 volt across it.
What is the capacitance when we put 10 volts across it?
At one volt: K = Cj*sqrt(v) = 100*sqrt(1) = 100
At 10 volts: Cj = 100/sqrt(10) = 31.62 pf
Getting back to our original conversation, assume our 10v power supply
has 0.1v of pk-pk ripple on it. How much variation in capacitance do we get
with this particular diode for a bias at 1v versus a bias at 10v?
Here's the difference between max and min capacitance at a nominal 10v of bias:
delta-C = 100/sqrt(10-0.1) - 100/sqrt(10+0.1) = 0.316 pf
When we adjust the pot to divide down the power supply to give 1v of bias,
we also divide down the noise by a factor of ten.
Here's the difference between max and min capacitance at a nominal 1v of bias:
delta-C = 100/sqrt(1-0.01) - 100/sqrt(1+0.01) = 1.000 pf
Conclusion:
Our oscillator will be significantly more sensitive to power supply noise
at low bias voltages across the varactor diode. This is in spite of the fact
that the voltage fluctuation on a bias of 1 volt is 1/10 what it would be for a 10 volt bias.
?
On Fri, Dec 20, 2019 at 02:12 AM, Chuck Carpenter wrote:
Jerry,? That's a fun equation!
The half-power Sqrt sez is not linear. I can't quite visualize** how the parallel capacitor calculation works out.? I may just have to get really curious and check out the constants and variables and work some examples from the back of the book.
In the meantime, I'll just use my measuring gadget...8^)
Fun Stuff
**I did this stuff about 60 years ago.? Kinda hazy now.
--
Chuck, W5USJ (ex K2OFN)
Point, Rains Co, TX? EM22cv
|
Kirk Kleinschmidt, NT0Z
Bo, OZ2M