Power Smart Heads

From BEAM Robotics Wiki

Jump to: navigation, search

We apologize for the need to display ads on this wiki. But somehow we must pay for hosting.


[edit] Original Version

Wilf's original Power Smart Head article BEAM Heads 101 contains a complete explanation of his original design, but he recently posted this explanation of its workings:


The PD1 / PD2 photobridge acts as a voltage divider with the midpoint at Vcc / 2 when the light on each PD is equal. The response is linear over a wide range of ambient light levels. Some PDs are very efficient and may need some light shielding (i.e. heat shrink) to reduce the photo current level in very bright light.

The hi / lo oscillator uses R1 / C1 to set the basic frequency. Resistor R3 connected from the photobridge midpoint to the R / C node of the oscillator is used to influence that frequency. That frequency depends on the voltage of the midpoint as well as the absolute resistance of the PDs. In general, the frequency is the lowest when the PD bridge center voltage is near the threshold of the 7AC240 inverter input.

With an R1 to R3 = 10 to 1 ratio, the frequency and the duty cycle varies when the PD output voltage is within 10% (+ / -) of the threshold. Voltage levels above and below that +/-10% range will cause the output of the oscillator to stop and will be steady high or low, inverted with respect to the input voltage. The band of oscillation can be made more or less by decreasing or increasing the ratio of R1 and R3 values (i.e. if R1=R3 it always oscillates). I use typical ratios of between 2 to 1 and 10 to 1. If the ratio is high, the sensitivity is greater but the tendency to wiggle is also greater. Remember a shaky head is not power efficient.

For some applications that have lots of built in damping (i.e. Submarine head, Mazola head) R3 can be 0 ohms ;).

The Nv / Nu driver (also called Nx driver) both differentiates and integrates the complementary outputs from the oscillator. When the oscillator oscillates, the Nx driver AC couples the output pulses noninverted as the Nx time constant is longer than the oscillator period. With AC coupled (non-inverted AC) pulses in phase across the motor, the PSHead is in the power save mode but with the brake on! That is important to be able to stop the motor rotation quickly to avoid mechanical overshoot when the head is aligned with the light source.

When the oscillator stops oscillating there are no pulses to AC couple and instead, the steady output is DC coupled (and inverted) to cause a differential voltage across the motor which then rotates.

The ratio of the R1 / C1 and R2 / C2 time constant should be about 1 to 10 for low standby power. Smaller ratios can be used for faster and more "varied" response but sometimes higher standby current. Since the R1 / C1 time constant is influenced by R3 and the photo bridge output voltage, that ratio will vary with light level. The ratios of those two RC time constant component values cause a variety of behaviour, sensitivity and efficiency.

For solar powered PS Heads, the time constant of R2 / C2 should (probably) be much shorter than the duration of the SE pop or some unpredictable (perhaps interesting) side effects may occur.

[edit] Solarized Power Smart Head

This PowerSmart head circuit was solarized, courtesy of Darrell Johnson, and dubbed the Solar Power Smart Head (SPSH). The SPSH circuit has been tweaked from time to time (each time improving its efficiency); the most recent update posted by Wilf is shown here:



[edit] 2 Degrees of Freedom Version

There's also a corresponding 2 degree of freedom (i.e., for a 2-axis head) version of this circuit:


[edit] Power Smart Head Version 3

Recently, Wilf posted version 3 of the SPSH -- this time with an LED flasher circuit. Here's Wilf's explanation:

I checked some RS FLEDs I have and the average current is about 2 ma. I designed a new 74HC240 LED flasher circuit which has an average current of about 100 uA. The LED flashes much brighter than the FLEDs. The circuit provides some voltage boost so it works down to 1.5V. The low power LED flasher comes on when the SE triggers and stays on until the SE resets. The rest of the SPSH layout has been changed somewhat to simplify connections between IC pins. A SPSH4 design is next that replaces the 1381 with two transistors and provides dual motor drivers.



[edit] Power Smart Head Version 5

Wilf posted version 5 of the SPSH in the BEAM Yahoo group on Feburary 18, 2006. Here's Wilf's explanation of the new circuit:

The new solar engine (Charge=Discharge or CD SE) is reminiscent of the PM1 SE but was derived from a LED flasher subsystem of a new Li- Ion charger that I will discuss later.

This CD SE has several nice features:

1) It pulses on/off even when powered from a battery ie no lockup. 2) The pulse width on/off ratio can be adjusted.

The function of a solar engine is to concentrate energy acquired from a low current source such as a solar cell, "compress" it into a capacitor or battery and release at a higher current rate or in a short powerful pulse to move a motor or light a LED.

The basic CD SE consists of the 1381, C2, R1 and R2.

When the voltage of the main storage cap C1 is high enough for R1 charges C2 to the 1381C trigger level, the 1381 output goes high at 2.4V, which connects C2 to R2 and causes C2 to discharge to the 1381 reset level of about 2.1V.

The active high 1381 output and the inverted active low output are used to cause single flash of the LED using a 3x voltage pump that can supply up to 7V across the LED. The LED must be a blue or white color that draws no current at 2.5V forward voltage. Alternatively two red or green LEDs in series can be used.

The inverted 1381 signal also enables the photobridge, the HLO comparator circuit and, after a delay to let the control circuit stabilize, enables the motor driver that rotate the motor if the photobridge is unbalanced.

After a short time, the CD SE resets and the cycle repeats.

The CD SE triggers slightly above the rated 1381C level or at about 2.4V. This includes the 50mV hysteresis as well as the 200mV drop of the 1381 supply current across R1 just before the 1381 triggers. The 1381 reset voltage is typically 2.15V which means that the voltage on C2 is a 250mV p/p sawtooth waveform oscillating between 2.15V and 2.4V. Once triggered, the SE on time dependents on the time constant of R1, R2 and C2 and not on the voltage of maon storage cap C1, which may be fully discharged in the mean time.

The delay circuit formed by R3 and C3 is kind of interesting as it delays the eneble of the B inverter motor drivers after the 1381 triggers but disables the motor drivers instantly when the 1381 resets. This ensures that there is no wasteful twitching of the motor and discharging of C1 during the SE turn-on and turn-off transients.

The tripler flasher causes both C4 and C5 to charge up to 2.4V while the SE is in the reset state. The voltage across the LED just prior to the 1381 trigger is the same as C1 which is also 2.4V. When the SE triggers, capacitors C1, C4 and C5 are effectively connected in series across the LED and can generate a pulse with a peak voltage of 7.2V maximum which cause a short but very bright flash.


[edit] Also See:

[edit] External References

[edit] Related Beam Yahoo Group Post's of Interest

This wiki is sponsored and hosted by Interactive Matter
Personal tools
Ads to finance this wiki