20 Revolution Position Sensor?

Thread Starter

MikeML

Joined Oct 2, 2009
5,444
(Also posted to ETO).

I am building an Arduino-controlled remote LC antenna tuning network with a motorized inductor and capacitor on a common shaft. I have a 99uH roller-inductor that takes 20 revolutions of a 0.25" shaft to adjust the inductance from ~1uH to 99uH. Total travel of the shaft: 20rev*360°/rev=7200°. The DC gear motor that will drive the shaft has no internal position sensor.

I need an absolute shaft position sensor that would tell me the present position of the roller tap in the inductor. Absolute means that I can read it on power-up to determine the position of the shaft. I need sufficient resolution to determine which revolution (1 of 20) and at least a few steps within that revolution. Quantizing the position with a 10bit AD would be sufficient.

A 20Turn sensor could couple directly to the end of the rotating shaft. I could also use 2:1 gears or cog belt to drive a 10Turn sensor.

Any ideas and USA sources of supply of the sensor and/or the drive train (gears/belt) .?
 
Last edited:

GopherT

Joined Nov 23, 2012
8,009
I am building an Arduino-controlled remote LC antenna tuning network with a motorized inductor and capacitor on a common shaft. I have a 99uH roller-inductor that takes 20 revolutions of a 0.25" shaft to adjust the inductance from ~1uH to 99uH. Total travel of the shaft: 20rev*360°/rev=7200°. The DC gear motor that will drive the shaft has no internal position sensor.

I need an absolute shaft position sensor that would tell me the present position of the roller tap in the inductor. Absolute means that I can read it on power-up to determine the position of the shaft. I need sufficient resolution to determine which revolution (1 of 20) and at least a few steps within that revolution. Quantizing the position with a 10bit AD would be sufficient.

A 20Turn sensor could couple directly to the end of the rotating shaft. I could also use 2:1 gears or cog belt to drive a 10Turn sensor.

Any ideas and USA sources of supply of the sensor and/or the drive train (gears/belt) .?
Could you assume the position by measuring the inductance?
 

Thread Starter

MikeML

Joined Oct 2, 2009
5,444
Could you assume the position by measuring the inductance?
Yes, but it complicates things because the tuner is being driven with a few Watts of RF at the operating frequency while being tuned. The Arduino will have the ability to measure the frequency of the excitation frequency (1.8 to 18MHz), and will slew the tuner to the (previously determined) approximate position as fast as possible (continuously reading the sensor), and then switch to a "search" mode to find the final angular position (jog the shaft) based on reflected power on the feedline.

I thought about switching the inductor into an oscillator whose frequency is a function of the inductance, but this would take time, and I would have to turn off the excitation in the middle of the scan...
 

GopherT

Joined Nov 23, 2012
8,009
@MikeML
I would use a jack-shaft with 2:1 ratio (pulleys/belt or gear/chain) and a multi-turn pot coupled to the bottom of the jack-shaft, add 5v across the pot and use the ADC on the arduino to determine the voltage on the wiper. Arduino has onboard 10-bit ADC.

Even easier if you find a 20 turn pot (no jack-shaft needed). Some trimmers get a bit funky at the limits so you might want to keep the range away from the extremes.
 
Last edited:

AnalogKid

Joined Aug 1, 2013
10,987
A change of 1 uH would be 1% of the total and 1/5 of a revolution. If this is good-enough minimum resolution, a precision pot, like those used in industrial controls a few decades ago, has better than 1% repeatability. With the proper belts, gears, and other connecting goop, the pot would tell you the absolute position by reading the voltage at the wiper. These pots sometimes had 10 turn counting dials external, and sometimes had planetary gearing internally so 10 turns of the shaft equaled one revolution of the wiper.

Gopher beat me by seconds.

ak
 

GopherT

Joined Nov 23, 2012
8,009
A change of 1 uH would be 1% of the total and 1/5 of a revolution. If this is good-enough minimum resolution, a precision pot, like those used in industrial controls a few decades ago, has better than 1% repeatability. With the proper belts, gears, and other connecting goop, the pot would tell you the absolute position by reading the voltage at the wiper. These pots sometimes had 10 turn counting dials external, and sometimes had planetary gearing internally so 10 turns of the shaft equaled one revolution of the wiper.

Gopher beat me by seconds.

ak

I happen to have a small collection of those 10-turn industrial and instrumentation pots. Spectrol and Backmann "hiPot" were the brands of choice. 1/4" shaft.

Date code is 44/69 but, even after 46 years, they are spot on. I just tested on with a 100 mV excitation. I was able to read to the 0.01 mV with smooth control across the entire range. Using a very steady hand and a large knob, I got 0.01 mV resolution. Amazing for the age (my hands and this component). 0.01 mV over 0 to 100 mV is just over 13-bits of resolution.

@MikeML

Let me know if you need a pot.

image.jpg
 
Top