Electromechanical damping stabilizes analog-meter readings
Electronics damps meter-needle swings.
Alexander Bell, Infosoft International Inc, Rego Park, NY; Edited by Brad Thompson and Fran Granville -- EDN, September 29, 2005
Before shipping moving-coil meters, manufacturers may short-circuit the meters' terminals with a length of wire, which provides effective electromagnetic damping and results in better immunity to external mechanical vibration and shocks that can occur during transportation. This Design Idea applies essentially the same principle to analog meters under normal operating conditions. Connecting a meter to a voltage source with low internal resistance applies electromagnetic damping and makes the meter's readings more stable. Increased immunity to external vibration and shock takes on importance in mobile- or portable-system applications and especially in automotive devices.
For example, suppose that your application requires measurement of a 0 to 10V power supply (Figure 1). You have available a typical electromechanical meter that presents a full-scale voltage rating, VFS, of 50 mV and a full-scale current rating of 1 mA. To obtain the 10V full-scale voltage range, you add a series resistance, RS. First, calculate the meter's internal resistance, RCOIL:
Next, calculate the multiplier resistor, RS, as follows:
The resistance of RS typically greatly exceeds that of RCOIL and therefore significantly reduces the electromagnetic damping action on the meter movement. Although you can improve damping by shunting the meter with a capacitor, this approach also increases the meter's settling time.
Figure 2 illustrates a better approach, in which a moving-coil meter connects to the output of an operational amplifier, IC1, embedded in a deep negative-voltage-feedback loop. Because the op amp presents an extremely low equivalent output resistance, the meter's terminals are "virtually shorted," providing effective electromechanical damping that results in more stable readings and increased shock and vibration resistance. In Figure 2, the resistive voltage divider comprising R1 and R2 connected to the op amp's noninverting input determines the meter's full-scale reading. You can add RF and CF to form an optional highpass filter to further improve the meter's settling time. Transistors Q1 and Q2 are also optional and added as overvoltage protection. Note that, for normal operation, the transistors' forward base-emitter voltage, VBE, should be several times larger than the meter's full-scale voltage, VFS, which is typically 50 to 100 mV.
A rail-to-rail-capable, single-supply micropower op amp makes a good choice for this application. If the input voltage, VIN, exceeds the op amp's minimum power-supply-voltage requirement, you can connect the op amp's VCC pin directly to the input terminal, as the dashed line in Figure 2 shows. In effect, the circuit combines the advantages of meter buffering and improved shock and vibration resistance with a traditional moving-coil meter's advantage of requiring no external power supply. You can choose from among many commercially available off-the-shelf, rail-to-rail output-micropower op amps that draw supply currents well below the full-scale current drain, IFS, of typical moving-coil meters. For example, Maxim's MAX4289 requires as little as 1V and 9 µA of power, and the MAX4470 requires a minimum of 1.8V but only 750 nA of supply current.
Although this Design Idea has so far related only to dc-voltage measurements, you can expand the circuit to include ac- and dc-voltage measurements (Figure 3). In this approach, you add a precision diodeless, full-wave-rectifier stage based on a single rail-to-rail operational amplifier and resistors R3, R4, and R5 (Reference 1). Resistors R1 and R2 determine the full-scale reading. This circuit requires an external dc power supply to drive op amps IC1 and IC2; voltage-limiting transistors Q1 and Q2 are optional.
| Reference |
Click here for more Design Ideas! |
-
Thanks for your comments, folks.
To Daniel P.
In regards to your suggestion containing resistive divider/caps:
1). I would assume that multi-uF capacitor (even mF-range could be needed for proper damping) most likely of electrolytic type, connected in parallel as you have suggested, will add its own leakage current to the meter and could degrade the accuracy of measurement in a much greater extent than caused by a copper TempCo, which, as you have specified, is pretty much predictable and as such - relatively easy to compensate. The capacitor leakage current as a function of temperature is far less predictable than copper coil resistance.
2). For the reason stated above I would doubt that any accurate x10 Probe would contain any electrolytic capacitors of mF-range; such probes mostly are designed to operate on a high frequencies, where ceramic caps took place. Still, there might be some overlapping area where both methods are equally applicable; as usual in engineering practice - actual numbers do matter.
To James H.
3). You can increase damping with parallel resistor (you have mentioned 50 Ohm), but that resistor will dramatically increase the overall power consumption and still be worse that OpAmp damping method; it’s because the OpAmp’s equivalent DYNAMIC output resistance could be well below 1Ohm, thus providing far better damping..
King Regards.
Alex B.
Alex Bell - 2006-12-12 13:36:00 PST -
Driving a moving-coil with zero added resistance leaves its voltage sensitivity 100% dependent on the temperature coefficient of its coil resistance. Copper resistivity increases by about 4000ppm/Celsius: this translates into a 4% sensitivity decrease with every 10 Celsius increase.
A better approach to consider is to keep the external series resistor, shunt the coil with many microfarads to short the coil back emf, and restore the needle speed by shunting the external R with another capacitor. This is the topology of any x10 compensated scope probe. Just make sure the op amp is able to drive this capacitive load with no unstability.
If the external R is big enough, it will also protect the meter from overload with no additional components.
Daniel Perez - 2005-21-10 09:45:00 PDT -
Couldn't you get damping by paralleling the meter with a resister? By using a 50 ohm resister in parallel with the meter the series resister drops to 4975 ohms. Another idea would be to split the series resister and place a zener shunt in to limit overvoltage situations of both polarities.
James Hutton - 2005-20-10 16:52:00 PDT
