The I-Sense Indicator Page

A "Motion" Indicator for the Diamond K9000 (and other motorized devices)

Updated 10/12/2020

This page describes the I-Sense Indicator Project

This is a small circuit that provides an LED indication when a certain current level is detected flowing in a 13.8Vdc power connection. While the circuit has application to many different use-cases, the primary need came about as the result of my purchase of a Diamond K9000 motorized fold-over antenna mount for my vehicle. The controller (a pair of push buttons, one for "UP" and one for "DOWN") has an LED that indicates when one of the buttons are pressed, but there was no indication provided for when the mount reached the full "UP" or full "DOWN" position. This circuit provides an LED indication that the motor for the K9000 was "moving" which, by inference, suggests that when it stops moving, it has reached the up or down position.

The photo (left) illustrates the K9000 control module with the current sensor in series with the power lead just below.

Penumbra: The What and Why...

The K9000 mount's motor (see Figure 1) is controlled by reversing the voltage (approximately 13.8Vdc) on the leads to the motorized mount. Limit switches inside the mount control the stop-action of the motor and diodes across these switches allow motion in the opposite direction from that which activated the limit switch. This configuration allows for a simple "open-loop" controller that requires just two wires and no complex circuits. The starting current for the motor is in excess of 1 A, but the running current is something like 400 mA (give or take). The sample resistor can drop a fair amount of voltage while still allowing the motor to move (albeit a bit more slowly). I ultimately settled on 0.5 ohms as a middle ground between I-V gain and I-R loss.

Figure 1. The K9000 Motorized Mount.

But first, there were a couple of incarnations of the circuit. Originally, I used half of a dual op-amp to amplify the voltage across a high-side, 0.8 ohm, sample resisor and the other half to perform a comparator detection on the result. This worked, somewhat, but I wanted a simpler circuit. The TL331 is a small comparator that can operate at up to 36V with a 20mA drive capability. I found that it was possible to fashion a hysteresis circuit to allow for a wide hysteresis range with reasonable adjustment range on the thresholds. This both simplified the circuit and improved its operation.

For those not familiar with hysteresis (see Figure 2 for an example of a hysteresis plot), it is simply a condition whereby two different thresholds are used based on the output state of the circuit (hysteresis can also apply to other phenomena, but here, I am only discussing it as it pertains to comparators). The idea is to move the threshold away (higher or lower, depending on the nature of the circuit) such that it becomes "more difficult" to return to the previous state. For example, Imagine that there are two thresholds, one at 1V and another at 2V. When the output is at a low state, the 2V threshold is selected. As the positive input increases past the 2V point, the output changes state and the 1V threshold is then selected. Now, the positive input must fall below 1V before the circuit returns to the original state. This greatly improves the noise immunity of the comparator and is crucial to prevent "chatter" on the output.

Figure 2. An example of a hysteresis "loop"

A comparator is really just a differential amplifier what is optimized to compare two inputs. They generally feature high (open-loop) gain and an open-collector or open-drain output. Thus, the output of a typical comparator generally only has two valid states: OPEN and GND (the TL331 used herein is no exception to this "rule"). However, like a traditional differential amplifier, the comparator drives it's output the same way: it's output tries to drive the "-" input towards the voltage seen at the "+" input. Thus, if the "-" input voltage is less-than the "+" input, the output is OPEN (which, for most comparators, is assumed to be a high voltage because a pull-up resistor is placed there). If the "-" input voltage is greater-than the "+" input, the output is GND.

Hysteresis is a positive-feedback phenomena. Whereas most linear amplifiers employ negative feedback to achieve stable gain, comparator circuits generally utilize positive feedback to "trick" the circuit into moving the positive input voltage based on the state of the output. The easiest way I've found to design and analyze hysteresis is to write down two different schematic models of the circuit, one for the output = open state and one for the output = GND state. Figure 3 illustrates a simple comprator circuit broken into it's two states.

Figure 3. Comparator "OPEN" and "GND" equivalent circuits

As can be seen by inspection, the "+" input will be different in each state because the pull-up resistor is grounded-out when the output is GND. Also note that the "Vref" and "Vin" signal names can be exchanged such that the output polarity of the comparator matches the desired operation. The Kirchhoff's Voltage Law (KVL) equations are as follows:

V(+)open = (Ri * (Vcc - Vref)/(Ri + Rf + Rpu))) + Vref

V(+)gnd = Vref * ( 1 - (Rf/(Rf + Ri)))

As can be readily seen, there are a lot more than 2 unknowns. Whenever there are more unknowns than equations, a closed-form solution is not possible. Fortunately, some of these "unknowns" are constrained for us, while others generally yield to a rough guess. Vcc, for example, is generally dictated by what is available or by the circuit down-stream from the comparator. Other unknowns may be constrained by "rule-of-thumb" choices. Rpu is generally chosen to be on the order of 1K to 50K ohms with 10K as the nominal choice. Rf should be as large as practical with values in the 100K to 1M ohms range being typical (a high value for Rf helps keep the gain of the compaarator high). The threshold values (V(+)open and V(+)gnd) are generally dictated by the application. With these ideas set forth, it is usually a simple matter to get the math down to two equations with two unkowns which can then be easily solved.

While spice simulation tools are readily available, I generally prefer to use a spreadsheet to simulate the comparator operation. I enter the equations for each version of the circuit (typically, side-by-side) and provide "hacking" cells with the component and circuit values. I'll then manually adjust the respective cells and observe the result. In short order, I can usually narrow in on the resistor values I need to get the desired thresholds to function as required. It is also possible to graph a solution across a range of input values (my spreadsheet, linked at the bottom of this page, features both methods).

This circuit cannot be placed in the leads to the antenna mount, as those leads see a polarity reversal depending on the direction of the motor. In the main power lead, there is some overhead current in the LED and possible leakage sources, so the comparator needs to have a threshold higher than 0V to get it back into the "Vout = OPEN" state. As well, the current level required to place the comparator into the "Vout = GND" state needs to be something like half the nominal load current while the motor is moving so that the comparator LED is not falsly activated by noise. I chose 20mA for the low threshold to account for the Diamond Control Module LED, and about 200mA as the "motor on" threshold. Using the spreadsheet I created, I was able to determine values for the feedback resistors using a 2W, 0.5 ohm sense resistor (a Vishay WSR2R5000FEB SMD resistor), a lower limit of 20 mA and an upper limit of 200 mA (see the schematic at the end of this page). Actual testing gave values of between 26 and 37 mA for the lower threshold, and about 240 mA for the upper threshold...close enough.

High-Side vs. Low-Side (sensing)

Where the sense resistor lives is an important decision. High-side sensing places the sense resistor in series with the positive power lead to the load while "low-side" sensing places the resistor in the power return path. High-side is generally desireable because it reduces the restrictions on the grounding of the load. If you perform low-side sensing, the load can have only ONE return path for power. Any other GND connection will defeat the sense resistor rendering the sensing technique useless.

However, high-side sensing is not without it's pitfalls. For one, the sensing device (in this case, the comparator) must operate at or above the supply voltage connected to the sensing resistor. In addition, common-mode effects can introduce significant errors to the differential inputs of the device. Mitigating these issues is possible, but can become onerous in short order. For the K9000, the load is isolated from ground so low-side sensing can be accomodated. Furthermore, the TL331 inputs must operate no higher than Vcc-1.5V, so this also points to low-side sensing.

For those situations where high-side sensing is required, Analog Devices offers several current sense amps (such as the INA180A or INA186A series) that can be used to capture the voltage across the sense resistor. This voltage can then be applied to the comparator circuit of choice. The cost, and there is always a cost with comprimise, is that you need an additional part. Actually, possibly two or three additional parts as the sensing amplifiers can not usually operate with a Vcc of greater than 5V (although they can connect to a high-side resistor that is operated at up to 40V). So, if a supply of between 3.3-5V is not available already, you will have to add a voltage regulator. Also, for these situations, you generally want a stable voltage reference for the comparator, which is another I.C. and some resistors and a capacitor or two. So, for high-side applications, the circuit complexity is generally higher. The same process for designing the comparator network still applies, however.

The HOW...

Small proto PCBs (mostly from China) are cheap and easy to come by as of this writing. However, I still will cobble together a hand-wired prototype on occasion, if just for practice. In truth, I can usually have a simple prototype operational within a couple of hours. Unless I'm gunna make a lot of a thing ("a lot" is generally anything more than 1), this is how I'll make it happen. The images below illustrate the assembly of the circuit. Note that these images anre not necissarilly from the "final" version and are offered for illustrative effect.

First, I pencil-sketch the pattern I want to cut. I put pencil direct to the copper using actual components in most cases. Once I have all of the cuts penciled down, I will start slicing the copper with an Exacto knife. I can reliably carve with 10mil traces and 10 mil spaces. Anything less than that is also possible, but gets to be more difficult. Copper is very soft and plyable. As such, it will distort as the result of the carving. This leaves burrs and "threads" which are cleaned up by the blunt edge of the Exactor blade. Then, I use water-clean flux to solder-tin the surface. This image is after the carving, deburr, and tinning have taken place:

Next, the components are soldered in place. This technique generally leaves a lot more copper on the board than traditionally etched designs might. I follow the rule: if it doesn't need to be removed, leave it. This photo shows the completed assembly:

The LED is at the top of the image, just to the right of center. The "input" GND (from the supply) is at the lower-right, while the "output" GND (to the load) is at the lower left. The power connection is at the upper left. It took about 30 to 45 minutes to carve, tin, and solder the parts to the board.

The resulting board is small enough (about 0.8" x 0.7") to fit inside the Diamond control module. However, the Diamond controller doesn't have enough vertical clearance. 3D printing a new back-cover would address this, but then what to do about the LED? Besides, I don't have a 3D printer and am fine with the board as a wired-in dongle right next to the control module (as shown in the image at the top of this page). A piece of 1/2" clear heat-shrink tubing keeps the assembly from making sparks, and allows me to see the LED. This is a schematic of the final circuit:

Motor Current Sense Circuit.

The values illustrated above give threshold limits corresponding to current values of Im < 18mA (GND->OPEN) and Im > 200 mA (OPEN->GND). Of course, in between these two current values, the comparator will maintain its previous state. The LED illuminates when the motor is running (comparator output = GND). For Vcc = 13.8V, Iq is approximately 20uA (LED = off), and approximately 10.5 mA (LED = on). Assuming no unforseen calamity, the circuit will draw only about 500uA (this includes the TL331 current) unless the target device (e.g., motor) is operating. Assuming a 100 AHr vehicle battery, this would deplete the battery charge after about 22 years. Not surprisingly, a typical automobile battery won't last more than a few months without periodic charging, so this level of idle current poses no issue for the typical automobile installation.

Umbra...

This circuit does a great job of detecting the operating current for an isolated load and requires a minimum of components. It is good to finally have an indication of what is happening with the antenna mount on the roof of my vehicle.

Some related links:

TL331 Datasheet

Diamond Motorized Antenna Mount

Analog Devices LTSpice tool

Comparator Component Calculator (Excel file)