I've been thinking of getting some roller inductors to play with next time I
visit a hamfest, and it occurred to me that a generic roller inductor
doesn't have a linear change in inductance with roller position... does it?
E.g., at the half-way point you should get 1/4 the total inductance, right?

Or is the form wound such that the turns are closer together at one end than
nothing, thereby linearizing the inductace vs. roller position to some
degree?

Thanks,
---Joel Kolstad

P.S. -- They never made small (say, size of an old Kodak film
canister)-sized roller inductors meant for QRP work, did they? Most seem to
be sized for hundreds if not thousands of watts.

Hi Joel,

On Tue, 08 May 2007 21:50:06 -0700, Joel Kolstad wrote:

I've been thinking of getting some roller inductors to play with next
time I visit a hamfest, and it occurred to me that a generic roller
inductor doesn't have a linear change in inductance with roller
position... does it? E.g., at the half-way point you should get 1/4 the
total inductance, right?

Right, unless it's wound linear inductance. Those can be found as surplus
from test equipment. The ones found surplus in receiver's are usually
linear frequency, rather than linear inductance. They are visibly quite
different.


P.S. -- They never made small (say, size of an old Kodak film
canister)-sized roller inductors meant for QRP work, did they? Most
seem to be sized for hundreds if not thousands of watts.

They exist. I could barely give a dual one away on Ebay last year. Dual
QRP roller too! Finally someone grabbed it for $5.

Cheers,

Gregg

It would be a shame to do this now -- but I used the roller inductor from a
BC-375 for a transmitter I made about 40 years ago.

Hi,

I've been thinking of getting some roller inductors to play with next time I
visit a hamfest, and it occurred to me that a generic roller inductor
doesn't have a linear change in inductance with roller position... does it?
E.g., at the half-way point you should get 1/4 the total inductance, right?

With a roller inductor the ratio of length to diameter is not
constant as the tapping point changes so the inductance is not
simply proportional to N^2. Using the standard handbook formula
(Nagaoka's), I get the inductance at the centre to be just under
half (44%) of that across the whole coil. Doubtless though, toward
the end, this formula becomes rather inaccurate.


Cheers - Joe G3LLV

On May 8, 8:50?pm, "Joel Kolstad"
wrote:
I've been thinking of getting some roller inductors to play with next time I
visit a hamfest, and it occurred to me that a generic roller inductor
doesn't have a linear change in inductance with roller position... does it?
E.g., at the half-way point you should get 1/4 the total inductance, right?

Or is the form wound such that the turns are closer together at one end than
nothing, thereby linearizing the inductace vs. roller position to some
degree?

In general, a linearly-wound inductor will follow the cylindrical
inductor equations. There might be some changes from that
from the influence of the "left-over" (unconnected) windings,
including fringing capacitance to ground.

In the few cases of linear-frequency-tuning, notably in Collins
PTO units and the "rack" assemblies in the R-390 family, the
windings were deliberately spaced to handle the powdered-
iron core position effect on inductance. I suspect that Collins
did a lot of cut-and-try to achieve the correct spacing changes
on those; very little quantitative information on it is available
in text or on the web. :-)

73, Len AF6AY

On May 9, 2:19 pm, AF6AY wrote:
In the few cases of linear-frequency-tuning, notably in Collins
PTO units and the "rack" assemblies in the R-390 family, the
windings were deliberately spaced to handle the powdered-
iron core position effect on inductance. I suspect that Collins
did a lot of cut-and-try to achieve the correct spacing changes
on those; very little quantitative information on it is available
in text or on the web. :-)

Note that the RF section slugs do not follow a linear-with-frequency
position; the while the shafts move linearly with frequency, the cams
introduce a necessary nonlinearity in the slug's linear position that
is necessary in the overall design.

And to get back to the OP's case of antenna tuners/pi matching
networks etc. it is not necessarily desirable to have the tuner's (or
pi-network's) inductances and capacitances vary linearly with knob
position. You actually want the curve of knob position vs value to be
logarithmic (look at the Hammarlund and Millen and National ads from
the 40's through the 60's to see all the various nonlinear variable
capacitor curves that are desirable in various uses) to make tuning
less critical on the high bands and more useful on the low bands.

Folks who grew up with pocket calculators and digital multimeters
might assume that everything should be linear. Those of us who learned
with slide rules know that in the real world, logarithmic is more
useful!

Tim.

On May 11, 4:51�am, Tim Shoppa wrote:
On May 9, 2:19 pm, AF6AY wrote:

In the few cases of linear-frequency-tuning, notably in Collins
PTO units and the "rack" assemblies in the R-390 family, the
windings were deliberately spaced to handle the powdered-
iron core position effect on inductance. *I suspect that Collins
did a lot of cut-and-try to achieve the correct spacing changes
on those; very little quantitative information on it is available
in text or on the web. *:-)

Note that the RF section slugs do not follow a linear-with-frequency
position; the while the shafts move linearly with frequency, the cams
introduce a necessary nonlinearity in the slug's linear position that
is necessary in the overall design.

I was referring specifically to the "PTO" (Permeability Tuned
Oscillator) that first saw wide use in Collins radios. That
tuning *IS* linear in frequency versus knob position. If you've
ever opened one to examine it, you will understand. Most of
the Collins PTOs also had a minor correction mechanism for
that tuning but, in essence, the winding pitch of the tuning
inductor varied according to some internal Collins manufacturing
rule.

I've also had my hands on many an R-391 and understand that
mighty mechanical monster still - even though it was four decades
ago - and would say that trying to find a correct winding pitch for
all those (broadband relative to tuning rate) inductors would have
made the production cost way too high. It wouldn't be
important because the actual Q of those variable inductors made
the front-end resonators broad enough so that "crude" cam
adjustments were okay for practical purposes.

And to get back to the OP's case of antenna tuners/pi matching
networks etc. it is not necessarily desirable to have the tuner's (or
pi-network's) inductances and capacitances vary linearly with knob
position. You actually want the curve of knob position vs value to be
logarithmic (look at the Hammarlund and Millen and National ads from
the 40's through the 60's to see all the various nonlinear variable
capacitor curves that are desirable in various uses) to make tuning
less critical on the high bands and more useful on the low bands.

I began actually handling of lots of high-power HF transmitters
in the early 1950s, notably the ones made by the Lewyt Vacuum
Cleaner Company! :-) No real relationship of capacitance or
inductance curves to tuning...wayyyy too many variables involved
in antennas-lines-etc. to pin down any "necessary curves."

Folks who grew up with pocket calculators and digital multimeters
might assume that everything should be linear. Those of us who learned
with slide rules know that in the real world, logarithmic is more
useful!

High disagreement there. Having learned and owned a slide
rule in high school of the late 1940s (and understood logs
and elementary transcendentals), I switched to sceientific
calculators (forever, I think) as soon as they were on the
market, never looked back. I severely dislike non-linear tuning
as it has NO physical-sensor-body correlation with signal
bandwidths or carrier positions.

I see Joel's roller inductor question connected with my
experience with Press-Wireless 15 KW transmitters and
their dual copper tubing final amplifier inductors. QSYs
on those PW-15s required unwrenching the shorting links
then rearranging the shorting links, remounting them,
again with a wrench. Having helped make better sets of
shorting links and doing some intial tuning (to set preset
tables for number of links), the unshorted links were
generally in the inductance values predictable by old-style
equations of inductors. There was some effect of the
shorted turns which was arrived at during the preset tune-
up tests. Those were feeding rhombics directly and there
were no fancy tuning-matching circuits involved in the
typical 4 to 18 MHz RF region.

73, Len AF6AY

On May 8, 9:50 pm, "Joel Kolstad"
wrote:
....
P.S. -- They never made small (say, size of an old Kodak film
canister)-sized roller inductors meant for QRP work, did they? Most seem to
be sized for hundreds if not thousands of watts.

An automatic tuner I worked on a long time ago had motor-driven
variable inductors only slightly larger than that. I guess it handled
about 200 watts. The design wasn't roller, but slider; it tuned very
fast for a mechanical system. There were one or two motor-driven
vacuum variables in there too.

Cheers,
Tom

"Joel Kolstad" wrote in message
...
I've been thinking of getting some roller inductors to play with next time
I
visit a hamfest, and it occurred to me that a generic roller inductor
doesn't have a linear change in inductance with roller position... does
it?
E.g., at the half-way point you should get 1/4 the total inductance,
right?

Or is the form wound such that the turns are closer together at one end
than
nothing, thereby linearizing the inductace vs. roller position to some
degree?

Thanks,
---Joel Kolstad

P.S. -- They never made small (say, size of an old Kodak film
canister)-sized roller inductors meant for QRP work, did they? Most seem
to
be sized for hundreds if not thousands of watts.


Interesting puzzler and as I've a number of these Roller Coasters built in
to various old military radios, I measured one.
This particular inductor is the "Aerial Tuning" inductor fitted inside the
casing of a 1946 vintage, UK "Wireless set number 62", a TR-RX running about
2 to 8MHz. The inductor is a Paxolin former 6.5" long by 2" diameter, wound
with what looks like 1mm dia' Silver wire. There's exactly 100 turns wound
evenly along it's length and the radio front panel has a little window
showing a "00" to "99" turns counter, operated by the tuning knob and a neat
Geneva mechanism. For measurement purposes, it's ideal.

TURNS. INDUCTANCE.
100 132uH
90 120uH
80 103uH
70 89uH
60 73uH
50 58uH
40 43uH
30 29uH
20 15.8uH
10 4.7uH
0 0.12uH

Values look not too non-linear and probably follow the solenoid formulas but
I can't be arsed doing the calcs .












--
Posted via a free Usenet account from http://www.teranews.com

Using Mathcad, the coil mentioned in this thread has the following
properties:

1) From 0 to 20 turns the inductance is 0.038 times n squared microhrenies
(square law).

2) From 20 to 100 turns the inductance is (1.49 n - 16) microhenries (very
linear).

For the first 20 turns the coil has the square law property because all of
the turns are tightly coupled to each other. This is not the case for the
other 80 turns of a long solenoid.

A PDF of the Mathcad calculation and graph are available by email from W0IYH
at QRZ.COM).

Bill W0IYH

TURNS. INDUCTANCE.
100 132uH
90 120uH
80 103uH
70 89uH
60 73uH
50 58uH
40 43uH
30 29uH
20 15.8uH
10 4.7uH
0 0.12uH