Brain teaser - analog angular position sensor

[Turd Furguson]

Hello everyone!

Long time reader and subscriber of the EEVblog; first time poster. I’ve been scratching my head trying to recreate an analog angular position sensor from the late 1980s-early 1990s that interfaces with an IC. It is believed to be a 1k dual-gang 360-degree variable resistor opposed by 90 or 180 degrees given the popularity at the time. However, I cannot recreate the expected output readings via potentiometer testing on a breadboard.

Through testing and poking around with my oscilloscope I’ve determined the following:


Knowns – external direction sensor:

• Direction sensor has 4 wires (5v, gnd, sense1, sense2)
• When testing wiper readings directly on sense1 & sense2, readings are expected to oscillate sinusoidally; max reading 3.5-4.5v; min 0.5-1.5v.
• Expected dead-range within 0-0.5v & 4.5-5v.
• Note: I don’t have this part… objective of this mission is trying to recreate it.

Schematic:




Knowns: mainboard

• Sense1 & sense2 wires feed directly into an 8-channel analog multiplexor/switch (IC4 - CH2 & CH3)
  
• From here, signal is routed to an opAmp configured as an integrator to support analog-digital conversion via common pin 3.
• Capacitor C9 charges to a value equal to the integral of the input
• Channel 1 is selected on IC4 which discharges C9 until TR2 turns off
• TR2 Output to IC is a square wave and the discharge time is measured by the microprocessor.
• R15, R16 & R17 provide a reference voltage for the integrator and source voltage for the constant rate discharge.

Unknowns/thoughts:

• Degree separation & expected voltage differences of the two wires by the microprocessor.
  
• Opamp circuit seems to indicate 2.5ish volts input/output based on oscilloscope readings – not sure how 4.5v input is handled, and I don’t see any spikes based on my recreating efforts above.
• Recreating voltages of 0.5-4.5v using regular 10k pots on Sense1 & Sense2 generates no output behaviour
• This patent seems relevant as I see similar signal patterns in the circuitry, and this also supports rotational angular sensing: https://data.epo.org/publication-server/rest/v1.0/publication-dates/19870225/patents/ep0211477nwa2/document.html

Can anyone provide any guidance or direction? This seems like a simple circuit but it’s hurting my brain.
 

[duak]

Here's an interesting part: https://www.alps.com/prod/info/E/PDF/Sensor/Position/RDC80/RDC80.pdf   It accepts the full 360 deg rotation and has two linear outputs .

I was initially thinking of quadraure outputs (sin & cos) like a rotary transformer but it has both a CW and a CCW output which, when I think about it, is easier to use (unless you need sin and cos).

[Gribo]

Can it be an RVDT aka resolver?

[Benta]

Can it be an RVDT aka resolver?

I doubt it. A 360 deg potentiometer is more likely from the circuit shown.

[Gyro]

A sine/cosine pot?

[Turd Furguson]

Thanks to all for your inputs.

It's definitely not a sin/cos pot - I have one that tested unsuccessfully. That was my original theory before diving in this deep.

The RDC80 that duak linked may be what I'm looking for (image below). Based on the tolerances shown in the PDF (+/- 3%) it would explain my unsuccessful attempts to generate anything meaningful by manually adjustment of two regular pots.

Current thought is to replicate the output of the RDC80 with a digital potentiometer for greater accuracy.