|
|
project snapshot: |
|
|
|
|
|
|
|
The first intrinsically motion sensitive capacitive proximity sensor |
|
|
|
|
|
|
|
|
|
|
PROBLEM |
Our client required a sensor to control the operation of a touch-free, battery operated, automatic dispenser of antibacterial hand soap, which he was developing for use in a variety of environments ranging from hospitals to restaurants. The performance and format requirements were:
o activation by presentation of a hand within a practical distance range below the dispenser so that individuals would not typically risk accidentally touching the housing; o extreme functional reliability; o zero or minimal calibration requirements; o low power requirements for long battery life; o small size; low cost. Additionally, it was also desirable that the sensor could be mounted within the housing, without the need to provide external access ports in the base of the dispenser. In this way the housing cover could be as smooth and streamlined as possible, which was also considered a desirable feature for ease of cleaning and minimizing the chances of cross-contamination. A prototype housing had already been designed and rendered as an SLA. The sensor hardware had to fit into this. |
One of the pre-production sensor boards. Designed to clip directly onto the sense electrodes. |
|
|
|
|
|
|
|
|
|
BACKGROUND |
The client manufactured the soap, which was a proprietary and successful product. He wanted to extend his business into devices. He had a sensor development effort in progress with another supplier and it was bogged down with problems. We were asked to perform a feasibility study of that particular technology – as a ‘sanity check’. We concluded that the approach had fundamental problems. Three months later the client returned to us having finally determined the same result. The client also had a broad appreciation for the business of soap dispensers and the limitations of current commercial designs. As a result, he specifically did not want to utilize an optical, (IR), proximity sensor. |
|
|
|
|
|
|
|
|
|
SOLUTION |
The human body is largely composed of water. Thereby, human hands have a high dielectric constant and would be detectable as a change of capacitance of a simple electrode arrangement which could be comfortably housed within the base of the dispenser. However, there were several foreseeable potential problems associated with standard capacitive sensing technology that detect changes in capacitance. The first is that the dispenser was intended for use across a variety of application environments and potential mounting locations. As such the set-up capacitance would tend to vary due to unit-unit variations in the environment, posing either an awkward problem for consistent range performance without calibration, or restrictions on installation. Additionally, very often, capacitive sensors can be prone to stray electrical interferences which could potentially lead to false activations. |
||
|
|
|
|
Frank is the master of noise reduction (in this case phase jitter) and shared the patent. |
Our solution was to create a capacitive sensor which was instead directly sensitive to the time rate of change of capacitance, (dC/dt), and not at all responsive to the absolute value of capacitance change. Our thinking here was that it is physically impossible for an object, (in this case a person’s hand), to enter detection range without actually moving. The lead electronics engineer on this project, Frank, then came up with the idea of using a phase-locked-loop, (PLL), circuit, and this idea worked and led to a highly customized PLL design. The result was a sensor which was adjusted to a sensitivity of about 50 fF/s, which was found by testing to be the best match to the application… anything lower and the sensor would ignore it and automatically adjust, anything greater than about 50 fF/s and the sensor would generate a single trigger signal until the rate subsided. |
|
|
|
|
The central advantage of having a sensor which is sensitive to the rate of change of capacitance instead of an incremental change, is that it will therefore automatically adjust to slow changes in capacitance due to environmental effects, (such as changes in humidity), slight differences between units due to different installation environments, and also small variations due to a host of component manufacturing and assembly variations. However, although a simple differentiating circuit tacked onto a more standard capacitive sensor design would also provide a sensor which would be sensitive to dC/dt, such an approach would necessarily result in a noisy and less reliable device. This design is instead intrinsically sensitive to dC/dt. For this application, the sensor was designed to trigger on positive dC/dt’s only, so that a soap pump activation signal would be generated to provide one squirt of soap only when a hand was approaching the base of the dispenser where the sense electrodes were installed, instead of when it was being withdrawn from that region. No calibration was required. |
US Patent 6,731,209 Issued May 4, 2004 Control system with capacitive detector Wadlow, et al. 24 claims Assigned to client.
CLAIM 1: A capacitive sensor system for controlling operation of a device, the system comprising:
sense electrodes for enabling establishment of an electric field for intercepting motion of a proximate object;
and an electronic circuit for providing a control output signal in response to a rate of change in capacitance of the sense electrodes due to motion of the proximate object within the field without intermediate electronic differentiation of signals related to a change in capacitance.
[Claim 1 is particularly broad in that it covers all alternative means of intrinsically detecting the time rate of change.] |
|
|
|
Due to the inherent properties of the circuit, it also has a relatively high degree of intrinsic immunity to interference from stray EM fields. We became a bit obsessive about delivering on this particular requirement because our client was determined to deliver a product that would be very reliable, especially with respect to false activation. For instance, the sensor circuit board also includes a low battery detection circuit which locks out the sensor in order to avoid erratic operation when the batteries get depleted. In view of the food service application, one of our first practical EMI susceptibility tests was to place an early prototype right next to an old kitchen microwave oven, whereupon it promptly activated! A few design iterations later had that particular problem solved, along with the elimination of a small degree of sensitivity to cell phone signals. In 1000’s of subsequent reported tests by our client using our pre-production sensor boards he reported a perfect, zero incidence of false activation and 100% reliable operation. |
||
|
|
|
|
|
|
|
|
|
SRC team |
Myself: Sensor specialist. Electrostatics experience. Frank: Lead electronics engineer; M.S.; 30 years experience; specialist in low noise circuit design. Ralph: external EE for design review; contributed design ideas; brainstorming sounding board. PCB layout designer. |
|
|
|
|
|
|
|
|
|
|
|
Copyright © 2007, Sensors Research Consulting, Inc. All rights reserved. |
|