JPH063435A - Detecting method for pulse interval - Google Patents

Abstract

PURPOSE:To minimize a zero crossing error of a pulse waveform having noise overlapped to highly accurately measure a pulse interval by setting a fixed number of a high pass filter to a specific range of a half width of pulse. CONSTITUTION:In order to measure a time interval between two pulses having different amplitude and the same waveform, a high pass filter comprising C and R is used to differentiate pulse signals so that an interval of a time point where the signal crosses a zero level is a pulse interval. At this time if a time constant CR of the high pass filter is set within a range of 0.1 to 2.0 times a half width DELTA of pulse, an error of zero crossing can be minimized. Where a ratio CR/DELTA of the high pass filter time constant CR to the half width DELTA of pulse is on the axis of abscissas and a ratio (%) of the zero crossing error width to the half width DELTA is on the axis of ordinates, when CR/DELTA is less than 0.1, the zero crossing error increases with reduction of CR, while when CR/DELTA is more than 0.1, the zero crossing error increases with increase of CR. Therefore zero crossing error width narrows in a range of CR/DELTA of 1.0 to 2.0 and highly accurate measurement of the pulse interval is possible.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for accurately detecting time intervals of identical pulse waveforms having different amplitudes which appear in a light receiving element when distance measurement is performed using Er laser light. If two pulsed lights are received by the light receiving element and the pulse interval of the output signal is measured, the distance to the object can be found by multiplying this by 1/2 of the speed of light c. In a device that measures the distance to an object using the light of an Er pulse laser, the pulse waveform at the time of transmission is the same as the pulse waveform at the time when this is hit by the object and reflected back to be received.

However, the pulse for transmission and the pulse for reception have the same waveform but different amplitudes. However, the reflected light is weak and increases the sensitivity of the detection system consisting of the light receiving element and the amplifier, but if so, a lot of noise is included. In order to make accurate distance measurement, it is desirable to detect time intervals so as to minimize the influence of noise.

[0003]

2. Description of the Related Art If the time interval between two pulses having the same amplitude and the same waveform is measured, the comparator output is passed through a comparator whose output is inverted when a certain threshold V ₀ is exceeded. If the inversion timing is the start and end of timing, an accurate pulse time interval can be obtained. However, when measuring the time interval between two pulses having the same waveform but different amplitudes, if the timing of the timing is given by using a comparator having a constant threshold V ₀ , the same portion of the pulse does not correspond. The time measurement error becomes large. Here, the pulses having the same waveform but different amplitudes refer to a group of waveforms as shown in FIG. The pulse width and shape are the same, but the height is different.

When measuring pulse intervals having the same waveform but different amplitudes, the pulse intervals are the respective peaks.
It should be defined as the time interval from peak to peak. The pulse waveform also differs depending on the type of laser. Some are closer to Gaussian, and some are pulse waveforms closer to Lorentian. The Er laser is Er-doped in the medium, and the pulse waveform of this laser is similar to Gaussian, but has a tail in the latter half.

When it comes to the peak of a pulse, it must be the timing when the pulse is differentiated and this is zero. Therefore, in order to measure the time interval between two pulses having the same pulse waveform with different amplitudes, a high-pass filter (generally composed of a capacitor and a resistor) is used to pseudo-differentiate the pulse. Then, a method may be used in which the interval between the times t ₁ and t ₂ at which the differential waveform crosses the zero level is the pulse interval (t ₂ −t ₁ ).

Now, the point where the differential waveform crosses the zero level is called the zero point. Since the electrical differentiating circuit does not strictly correspond to the mathematical differentiation, the peak point of the original waveform and the zero point of the differential waveform do not coincide with each other, and they deviate from each other. However, if the waveforms are the same, the amount of the deviation is not affected by the amplitude, so (t ₂ −t ₁ ) can be used as the pulse interval. That is, the zero points t ₁ and t ₂ do not coincide with the peak point of the pulse, but (t ₂ −t ₁ ) theoretically coincides with the pulse peak interval. In addition, a comparator is generally used to detect the zero point.

It is generally said that a differentiating circuit consisting of a capacitor C and a resistor R needs to use C and R having a sufficiently small time constant with respect to the pulse half width Δ. Figure 3 makes this a problem. This shows several characteristics, and the gain becomes almost 1 at a frequency higher than the cutoff frequency f _c (1 / 2πRC). It does not function as a differentiation circuit in this area. It only works as a differentiating circuit at a frequency lower than f _c . In this case, the output gain is much less than unity.

[0008]

Since it is possible to obtain a result close to mathematical differentiation by using a differentiating circuit having a small time constant RC, it seems that there is no problem if a small CR is used. But it's not that easy. If CR is small, the amplitude of the waveform after passing through the filter will be small. This point will be explained with reference to FIG.

FIG. 4 shows the relationship between the cutoff frequency of the filter and the waveform after passing through the filter. It is assumed that the input waveform is a pulse like A. When the fundamental frequency f _{p of the} pulse is much smaller than the cutoff frequency f _c (1 / 2πRC) of the filter, that is, when RC is small, the output waveform B approaches a mathematically differentiated waveform, but the wave height becomes small. Here, the pulse fundamental frequency is defined as 1 / 2πΔ. That is, when RC / Δ is small, the output waveform is close to a mathematical derivative, but the wave height is small.

On the contrary, when the fundamental frequency f _p of the pulse is higher than the cutoff frequency f _{c of the} filter, that is, RC
When / Δ is large, the output waveform C does not become a differential waveform but becomes a linear combination of the original waveform and the differential waveform, and the larger the ratio is, the more the output waveform C becomes closer to the original waveform.

The explanation will be made more clearly. In the filter of FIG. 2B, assuming that the input voltage is v ₁ , the output voltage is v ₂ , and the output current is 0, v ₁ = v ₂ + ∫ (v ₂ / RC) dt ( The formula 1) holds. The second term on the right side is the voltage applied to the capacitor C, and the first term on the right side is the voltage applied to the resistor R. If RC is sufficiently large, the second term on the right side of (1) becomes smaller than the first term, and the output v ₂ has a waveform substantially equal to the input v ₁ . It corresponds to FIG. If RC is sufficiently small, the first term on the right side will be small and the output v ₂ will be close to the time derivative of the input v ₁ .

At this time, approximately v ₂ = RC (dv ₁ / dt) (2), so that if RC is small, the peak wave height becomes small.

When the differential wave height is low, the differential signal is embedded in the noise, and it is not possible to correctly detect the time point (that is, the peak point) where (dv ₁ / dt) = 0. The original pulse signal is accompanied by noise with a small wave height. Even if the wave height of the noise is small, it is possible that the noise is magnified by passing through the differentiating circuit and the point where the derivative becomes 0 becomes unclear. This is because noise has a small wave height but a high frequency. As described above, the comparator is used to detect the zero point of the differential waveform.
₁ = 0) is accompanied by a hysteresis circuit because the output does not fluctuate due to noise.

This is shown in FIG. Comparator - resistance to the inverting input R _2, R ₃ of motor is connected. The other end of R ₃ is grounded and the other end of R ₂ is connected to the output. It is assumed that the inverting input is P, the non-inverting input is S, and the output is Q. Since it is a comparator, the output takes two values. The low level is L and the high level is H. Hysteresis is provided by resistors R ₂ and R ₃ .

The width of hysteresis is arbitrarily determined by the resistance ratio. When the output is at high level H, the non-inverting input S is S ₂ = HR ₃ / (R ₂ + R ₃ ) (3). Since the inverting input P to which the differential waveform is to be input is v ₁ = 0 when there is no signal, v ₁ <S ₂ . If the resistance ratio is determined such that the maximum amplitude of noise is smaller than S ₂ , the output will not fluctuate due to noise and the output will always be H.

The output is inverted only when the input P exceeds S ₂ . The output goes low. At this time, the non-inverting input S drops, and S ₁ = LR ₃ / (R ₂ + R ₃ ) (4). S ₁ <S ₂ . The output maintains the L level as long as the input P is larger than S ₁ . The output is inverted and returns to the H level only when the signal input P becomes smaller than S ₁ . Since the output inversion thresholds are different in this way, the influence of noise when v ₁ = 0 can be blocked. Of course, in order to do so, the value of the resistor must be set appropriately according to the noise. That is, S ₂ > maximum value of noise may be determined. For example, the noise level is set to 3 mV, and this S ₂ is set to 9 mV, for example. Since the H and L levels of the output are determined by the voltage of the power supply, etc., S
_{When 2} is decided, the lower threshold S ₁ is also decided.

FIG. 8 shows waveforms of the output Q and the hysteresis S when the differential waveform P containing noise is input in the circuit having such a hysteresis. The input waveform contains noise, but as long as this is near 0 V, the output Q takes the H level, and therefore the hysteresis S has the upper threshold value S ₂ . Turn off the S ₂ in the stomach point input waveform P rises. At this time, the output is inverted and falls to the L level. At the same time, the hysteresis drops to the low threshold S ₁ . The input waveform rises further and reaches the maximum at point B. This time, it begins to decrease and at the point C the low threshold S ₁
Cut from top to bottom. The output is inverted and goes up to H level.
At the same time, the hysteresis returns to the high threshold value S ₂ . After that, the output Q and the hysteresis S do not change although they fall to the point C and return to the point E (zero level). After all, the point C where the input waveform falls below the zero level can be detected as the output rising timing.

With such a hysteresis circuit, fluctuations in the output of the comparator when the input is at zero level can be eliminated. However, the hysteresis circuit can only block the effect of noise when it is at zero level, but cannot eliminate the effect of noise when the input crosses zero level. S ₁ is cut off when the input, which is a differential waveform, falls, but since there is noise, a timing error proportional to the height of the noise occurs. This is the zero cross error.

This is shown in FIG. FIG. 2A shows an original pulse waveform. This is a waveform with one peak, but with a lot of noise. When this is passed through the high-pass filter of FIG. 2B, it becomes as shown in FIG. Since it is a pulse having one peak, it is differentiated into a waveform having positive and negative halves, but noise is also differentiated and accompanying it. Due to noise, this signal passes through the zero point many times.
At zero level, the hysteresis circuit described above prevents malfunction. However, there is uncertainty as to when the input rises and falls below which it crosses the zero level. Set the noise level to low. Is at zero level.

Consider a pure signal without noise in the input. This is called a true signal. Consider the input as the sum of the true signal and noise. What we want to find is when the true signal crosses G (zero level). However, when the downward noise happens to be superimposed, the input crosses the zero level at an earlier re-point and the output is inverted. On the other hand, when upward noise is superposed, the input crosses the zero level at the slower point than the null point. Even if the true signals are the same, the time to cross the zero level fluctuates between reels due to accidental noise. An error caused by this uncertainty is called a zero cross error. In addition, the time width of a portion that may cross the zero point is also referred to as a zero cross error.

The zero-crossing error is an error in time measurement in the timekeeping method in which the timing of the peak of the pulse is when the signal crosses the zero point. It is desirable that this is smaller. The zero-cross error increases when the noise amplitude is large and the signal level is low. As described above, when the time constant of the differentiating circuit is small, the amplitude of the signal becomes small, so that it is susceptible to noise. For this reason, the use of a high-pass filter having a small time constant CR undesirably increases the zero-cross error.

On the other hand, when the time constant CR is large, the action of the high-pass filter deviates from the mathematical differentiation and approaches the original signal, that is, the differential waveform crosses the zero level after the peak point of the original signal. . As will be described in detail later, if CR is large, the gradient of the differential waveform at the zero point becomes small, and the zero-cross error may become rather large. That is, if the time constant CR of the high-pass filter is too large or too small, the zero-crossing error increases, which is also not good. C
There should be an optimal range for R. The present invention aims to provide this.

[0023]

The present invention measures the time interval between two pulses having the same waveform but different amplitudes.
In a method of differentiating a pulse signal by a high-pass filter composed of C and R and setting the interval at the time when the signal crosses the zero level as the pulse interval, in order to minimize the zero cross error,
The time constant CR of the high-pass filter is set to 0.
It is characterized in that the range is set to 1 to 2.0 times.

[0024]

The zero-cross error can be minimized by setting the time constant CR of the high-pass filter to 0.1 to 2.0 times the pulse half width Δ. FIG. 1 shows an example of calculation of the high-pass filter time constant and zero-cross error. The horizontal axis is the ratio CR / Δ of the pulse half-width Δ of the high-pass filter time constant CR. The vertical axis represents the ratio of the zero-cross error width to the pulse half width Δ.

Here, the pulse waveform is a giant pulse when a Q switch is used for the solid-state laser. The noise is a sine wave having a frequency four times the pulse fundamental frequency f _p (up to 1 / 2πΔ), and the peak to peak of the noise amplitude is 1/50 of the pulse peak value V _p . When CR / Δ is less than 0.1, the zero crossing error increases with decreasing CR. In this case, the differentiated waveform approaches the mathematical derivative, but since the gain of the high-pass filter of the true signal containing no noise becomes low, the differential coefficient becomes low and the zero-cross error increases.

On the contrary, when CR / Δ is larger than 2.0, the zero-crossing error increases with increasing CR. This is because the differential waveform deviates from the mathematical differential and crosses the zero point at the tail of the peak, so that the differential coefficient becomes small. Therefore, the zero cross error increases.

Of course, the larger the noise amplitude, the larger the zero-crossing error. However, it is an upward translation of the graph of FIG. Therefore, the conclusion when the range is fixed by CR / Δ does not change. Strictly speaking, it also depends on the waveform of the original pulse.
However, in general, it can be said that the zero crossing error width becomes narrow in the range of CR / Δ between 0.1 and 2.0. The present invention sets CR / Δ in this range. It is thought that the present invention is the first to find the optimum range by recognizing such a mutual limiting relationship between the differentiating circuit and the pulse width.

The apparatus of the present invention was applied to a rangefinder using a laser. This is to detect the pulse by detecting the pulsed light of the laser from the collimator lens into a parallel beam, which is emitted, and which is reflected by the target object and the light reflected back by the lens. Difference between the times t ₁ and t ₂ (t
_2- t ₁ ) to measure the distance from the object.
The principle is well known. In this case, the intensity of the reflected light from the object varies depending on the distance, weather conditions, the reflectance of the object, and the like. The magnitude of reflected light (transmitted pulsed light) from a lens or the like is constant, but the magnitude of reflected light (received pulsed light) is not constant. However, the intensity changes of both pulsed lights, that is, the pulse waveforms are the same. Therefore, if the time interval from the peak to the peak is measured, the time required for the light to travel back and forth from the rangefinder to the target object can be known.

FIG. 5A shows a circuit for differentiating the output of the light receiving element and shaping the waveform. It consists of a CR differentiating circuit and a comparator. Differentiator circuit has a capacitor C and a resistor R
_{Consists of 1} . This connection point gives the differentiated waveform which is connected to the inverting input of the comparator. The non-inverting input is grounded by resistor R ₃ . The output of the comparator and the non-inverting input are connected by a resistor R ₂ . These resistors provide the aforementioned hysteresis. It is the same as that of FIG. When the input P is at the zero level, the output is H, the non-inverting input that gives the hysteresis is at the high threshold S ₂ , and the input P does not exceed this even if there is noise, so the output does not reverse. The high threshold S ₂ is given by equation (3).

When there is no pulsed light but only noise light,
Since the inverting input is higher than the non-inverting input S ₂ , the output of the comparator is H level. When the pulsed light enters the light receiving circuit, a pulse voltage waveform proportional to the pulsed light waveform is input to the differentiating circuit. This is v ₁ . A differentiating circuit differentiates this, and the result is v ₂ . The differentiating circuit produces a waveform as shown in FIG. The output of the comparator should be inverted when crossing the zero level line, but this is not the case due to noise, and the output is inverted when the true signal crosses from one to the other. . Then, a zero cross error occurs. The present invention is not intended to suppress the zero cross error, but is characterized in that the time constant of the differentiating circuit is selected so as to reduce the zero cross error.
The zero-crossing error never disappears, but it gets smaller.

It is correct to evaluate the zero crossing error starting from equation (1), but this is difficult. v ₁ = v ₂ + ∫ (v ₂ / RC) dt (5) With v ₁ known, the output v _{2 of the} differentiating circuit is

V ₂ = exp (−t / CR) ∫v ₁ ′ exp (τ / CR) dτ (6)

It becomes Although v ₁ and v ₂ are functions of time, the display of independent variables is omitted for simplicity. v ₁ ′ is the time derivative of v ₁ . τ is a variable of integration. Integration range is −∞
From time t. If the zero-crossing error is E, this can be evaluated as the time resulting from dividing the noise magnitude n by the time derivative of v ₂ = 0 (v ₂ cuts the zero level from top to bottom). So

E = n / (dv ₂ / dt) However, it can be defined as v ₂ = 0 (6 ′). However, since it is the value at v ₂ = 0,
From (1) this is

E = n / (dv ₁ / dt) However, it can be written that v ₂ = 0 (7). In step (6), v ₁ is given, v ₂ is analytically calculated, and the differential coefficient dv ₁ when it becomes 0
If / dt is obtained, the zero cross error can be found. The zero-crossing error can be obtained as E / Δ using the half-width Δ of the pulse as a parameter for the time constant CR.

The above is a general theory, and calculation cannot be performed unless the functional form of the input pulse v ₁ is known. Actually it is not Gaussian, but if the input is Gaussian, for example: v ₁ = exp (−t ² / δ ² ) (8)

However, the half widths Δ and δ have a relationship of δ = Δ (log _e 2) ^−1/2/2 (9).

V ₂ = v ₁ − (CR) ⁻¹ ∫ exp {−τ ² / δ ² − (t−τ) / CR} dτ (10)

The range of integration is from -∞ to time t. v ₂
We need to find the time when crosses the zero level from top to bottom. That is, v ₂ = 0.

1 = (CR) ⁻¹ ∫ exp {(t ² −τ ² ) / δ ² − (t−τ) / CR} dτ (11)

From the above, the time t at which v ₂ = 0 can be obtained. The variable transformation of η = t−τ (12) is performed,

CR = ∫exp {(− η ² + 2ηt) / δ ² −η / CR} dη (13)

It becomes The integration range is 0 to infinity. ^{CR = (δπ 1/2 / 2)} exp ^{[δ 2 {2tδ -2 - (CR} ) -1} 2/4 ] (14)

Thus, the time t at which v ₂ = 0 is obtained.
By substituting this value into the derivative v ₁ ′ = − (2t / δ ² ) exp (−t ² / δ ² ) (15) of (8), the zero-crossing error can be obtained by (7).

Although this calculation cannot be performed analytically, it can be calculated by determining the values of δ and CR. Although this example is Gaussian, the zero-cross error can be evaluated by the same calculation even if other waveforms are assumed. FIG. 1 shows the calculation result, not the actual measurement result. The present invention obtains the range of CR / Δ that minimizes the zero-crossing error by such calculation.

FIG. 5B shows the output of the light receiving element. Here, pulses having the same waveform but different wave heights are drawn. This has a waveform as shown in FIG. 2 (c) through the CR differentiating circuit, but since it passes through the comparator, it has a pulse waveform as shown in FIG. 5 (c). The rising time of this pulse substantially corresponds to the time of the peak of each pulse. Therefore, it is possible to measure the time interval of the peak of each pulse. As described repeatedly, the differential circuit is not a mathematical differential in a strict sense, so the zero point does not correspond to the peak point. However, the interval of the zero points corresponds to the interval of the peak points.

Although signal processing including such a differentiation circuit and a hysteresis circuit is commonplace, in the present invention, the time constant CR of the differentiation circuit is set to 0.1 to 2.0 of the pulse half width Δ.
Therefore, the zero-cross error is minimized. Therefore, according to the present invention, the pulse interval can be accurately detected as the interval when the digital pulse rises. The interval between two digital pulses can be easily measured using a commercially available counter.

[0048]

EXAMPLE The present invention was applied to a pulse laser range finder using an Er laser. First, the pulse waveform is obtained. The rate equation for the Er laser can be written as In this rate equation, only the two levels directly related to the laser oscillation are considered.

(DN ₁ / dt) = − (N ₁ −N ₂ ) BW−γ ₁ N ₁ + Φ ₁ (16) (dN ₂ / dt) = + (N ₁ −N ₂ ) BW−γ ₂ N ₁ + Φ ₂ (17) (dW / dt) = hf (N ₂ −N ₁ ) BW−2 kW (18)

Here, N ₁ is the density of atoms having electrons at one level, N ₂ is the density of atoms having electrons at two levels, W is the energy density of laser light, and B is Einstein's stimulated emission. It is a coefficient. γ ₁ and γ ₂ are relaxation coefficients of each level, Φ ₁ and Φ
₂ is the excitation rate to each level, K is the laser resonator,
Attenuation rate of light amplitude. h is Planck's constant and f is the frequency of light. As the physical property constant, the value of Er laser was used.

The intensity W (t) of light is obtained by solving this non-linear differential equation. Since it cannot be solved analytically, it was solved by a computer. This is a pulse waveform that resembles a Gaussian waveform but has a tail. This is a waveform peculiar to the laser.

Fourier transform of W (t) is performed. A pulse waveform H (f) in the frequency domain is obtained. H (f) = ∫W (t) exp (j2πft) dt = F {W} (19) The integration range is −∞ to + ∞. This calculation is also computational
FFT processing is performed by the computer.

Next, the time domain waveform g (t) after passing through the light receiving circuit and the differentiating circuit is obtained. As described above, a light receiving circuit made by using an operational amplifier or the like generally has a low pass filter characteristic. The low pass filter can be defined by the cutoff frequency. In this example, the cut-off frequency f _h = 1.5 / Δ is determined as the lowest frequency that can be received without deforming the laser pulse waveform.

The differentiating circuit is composed of a resistor R and a capacitor C. The cutoff frequency f _c of the differentiating circuit is f _c
= 1 / 2πRC. The present invention is an invention as to how to determine this RC. Low
The Fourier transform of a pass filter and the Fourier transform of a differentiating circuit are well known. Multiply this by the above H (f),
Obtain the Fourier transform of the waveform after the pulse has passed through the low pass filter and the differentiating circuit.

H (f) {1+ (jf / f _h )} ⁻¹ (jf / f _c ) {1+ (jf / f _c )} ⁻¹ (20)

When this is Fourier-inversely transformed, a time display g (t) of the signal waveform passing through the differentiating circuit is obtained. g (t) = F ⁻¹ [H (f) {1+ (jf / f _h )} ⁻¹ (jf / f _c ) {1+ (j f / f _c )} ⁻¹ ] (21)

The slope of g (t) at the zero crossing point g '(t
₀ ) {g (t ₀ ) = 0} is obtained.

The amplitude V _{noise pp} appearing at the output of the differential circuit of constant frequency and single frequency _noise is determined. For simplicity, it is assumed that the noise has a constant amplitude and a single frequency and is superimposed on the output of the light receiving circuit. Noise amplitude is S / N = 50
Assuming that It is assumed that the upper limit of the noise frequency is up to 3f _h and the lower limit is up to about f _c / 3. The former assumption is that the light receiving circuit substantially has a low-pass filter, so that noise having a frequency of three times the cutoff frequency or more does not appear. The latter assumption is that the differentiating circuit is a high-pass filter, so that frequencies below f _c hardly pass. With a margin, it is assumed that there is no noise with a frequency below f _c / 3.

From the above and above, the zero crossing error E is obtained by E = V _{noise pp} / g (t ₀ ) (22). Here, V _{noise pp} means the difference in voltage from peak to peak of the noise level.

The above operation was performed for various time constants CR to obtain the zero cross error. FIG. 1 shows the result. Originally, the graph of FIG. 1 should change depending on the relationship between the time constant of the differentiating circuit and the noise frequency.
However, this result did not change significantly even if the noise frequency was changed, so here, f _n = 4 / Δ = 8f _h
The result at the time of / 3 is shown. From the figure, CR / Δ = 0.5
In the vicinity, the value of the zero-crossing error with respect to the pulse half-width is the smallest. The minimum value is about 1 / (S / N) if the S / N value is large. The above-mentioned coefficient of performance can be hidden by S / N because V _{noise pp,} which is the numerator of equation (22), is proportional to the reciprocal of the S / N ratio.

As the optimum range of the time constant of the differentiating circuit, the ratio of the zero cross error to the pulse half width is 1.5.
/ (S / N) or less. This is 0.1 ≦ CR / Δ ≦ 2.0 (23). The above operation was performed for many CR / Δ, and some concrete waveforms are shown. This is because the waveform of the Er laser is specifically exemplified to make the present invention easier to understand.

[Example] Case of CR / Δ = 0.053 FIG. 9 shows waveforms at the input / output of the differential circuit in the case of CR / Δ = 0.053. The horizontal axis is time, and one unit is 25 ns. A waveform similar to the (channel 1) Gaussian shown below is the waveform v ₁ of the output of the light receiving element and the amplifier. That is, it corresponds to pulsed light. The pulse half width Δ is 57 ns. On the other hand, one scale on the vertical axis is 150 mV. The waveform having the zero point of (channel 2) shown above is the output v ₂ (g (t)) of the differentiating circuit. On the other hand, one scale on the vertical axis is 10 mV. v ₁ is one of the typical Er Les - has become a rise in is a The pulse waveform forward early, falling slow asymmetry of the waveform at the back.

FIG. 12 shows the configuration of the light receiving element, the amplifier and the differentiating circuit. v ₁ is the voltage at the output of the amplifier. v ₂ is the voltage at the connection point between the capacitor and the resistor. v ₂ and v ₁
Satisfies the constraint condition (1). The reason that v ₂ is smaller than v ₁ is that CR is small, so that
This is because the terms are superior. In this case v ₂ approaches the derivative of v ₁ , but it is still not a perfect derivative. 15 from the point where v ₁ is maximum to the point where v ₂ = 0
There is a delay of ns. A vertical broken line is drawn at the point where v ₂ = 0. When the derivative v ₂ ′ of this value is obtained, it is 1.6 mV / ns.

When the noise V _{noise pp} after differentiation is obtained, it is 8 mV _pp . Here, mV _pp is the difference in voltage between the upper peak and the lower peak expressed in mV. Is. From this, the zero cross error width is 5 ns.
The ratio of the zero-cross error to the pulse half width Δ (57 ns) is 0.088. The width of the zero cross error is large because the differentiated waveform v ₂ becomes a signal with a small amplitude and the zero cross error becomes large.

[Example] Case of CR / Δ = 0.39 FIG. 10 shows a waveform at the input / output of the differentiating circuit in the case of CR / Δ = 0.39. The horizontal axis is time, and one unit is 25 ns. A waveform similar to the (channel 1) Gaussian shown below is the output waveform v ₁ of the light receiving element. Pulse half width is 57
ns. The vertical axis is voltage, but one scale is 150 mV
Is. The waveform having the zero point shown above is the output v of the differentiating circuit.
₂ (corresponding to g (t) above). On the other hand, one scale on the vertical axis is 40 mV. There is a delay of 19 ns from the point where v ₁ takes a maximum value to the point where v ₂ = 0, which is larger than that of the previous example. However, the amplitude of v ₂ increases by that amount, and the slope of v ₂ at the point where v ₂ = 0 becomes 5.
It was 1 mV.

Peak-to-peak value V of noise after differentiation
_When asked for _{noise pp} , it was 10 mV _pp .
The zero-cross error is E = 2.0 ns obtained by dividing the latter by the former. The ratio of the zero crossing error to the pulse half width is
It is 0.035. Compared to the case of the example, it is reduced to about half. This is because the gradient v ₂ ′ at the point where v ₂ = 0 becomes large.

[Example] Case of CR / Δ = 8.2 FIG. 11 shows waveforms at the input / output of the differentiating circuit in the case of CR / Δ = 8.2. The horizontal axis is time, and one scale is 25 ns. The lower graph is the input v ₁ of the differentiating circuit. The pulse width Δ is 57 ns. On the vertical axis, one scale is 150 mV. The upper graph is the waveform of the output v ₂ of the differentiating circuit. One scale is 100 mV. v ₂ comes close to v ₁ in terms of waveform height. This is because v ₂ is far from the mathematical derivative and becomes almost equal to the original function v ₁ due to the large CR / Δ. Although v ₂ has a zero, this is far from the peak of v ₁ . This shift is 66 ns. The slope of v ₂ at the point of v ₂ = 0 is 3.3 mV / ns.

The noise after differentiation is 16 mV _pp . This is divided by the gradient value to obtain the zero crossing error, which is 4.8 ns.
And grows. This is because CR / Δ is too large, so v ₂ approximates v ₁ and v ₂ ′ at the point where v ₂ = 0.
Because it became smaller. This is because the noise after differentiation is large.

In Table 1, CR / Δ = 0.053, 0.39,
The value of the capacitor, the value of the resistance, the pulse half width, the input noise, the pulse peak value, and S for the three cases of 8.2.
/ N ratio, noise after differentiation, gradient at zero cross point, zero cross error width, and ratio of zero cross error width to pulse half width.

[0070]

[Table 1]

The zero cross error is determined by the gradient of v ₂ and the magnitude of noise when v ₂ = 0. Since the former has already been described, noise will be described here. The noise level on the input side of the differentiating circuit is constant at 16 mV _pp from example to example. The size changes as it passes through the differentiating circuit. This is 8 mV _pp when CR / Δ = 0.053. This is because the time constant of the differentiating circuit is small and the differential amplitude is small. CR /
When Δ = 0.39, the noise is 10 mV _pp . This reflects the fact that the time constant is slightly increased and the differential amplitude is increased. When CR / Δ = 8.2, the noise level is 16 mV _pp . This is because CR / Δ is large, so that v ₁ = v ₂ is almost satisfied in the equation (1).

[0072]

According to the present invention, the time constant CR of the high pass filter is 0.1 to 2.0 times the pulse half width Δ in the device for measuring the distance using the pulsed light of the Er laser. By selecting, it is possible to minimize the zero-crossing error of the pulse waveform on which noise is superimposed. Therefore, it is possible to measure the pulse interval with high accuracy. Therefore, it is effective for the time measurement of the pulse type laser range finder using the Er laser as the light source.

[Brief description of drawings]

FIG. 1 is a graph showing the relationship between a high-pass filter time constant and zero-cross error.

FIG. 2 is a pulse waveform diagram (a) on which noise is superimposed, a differential circuit diagram (b), and a differential waveform diagram (c).

FIG. 3 is a frequency characteristic diagram of a high-pass filter using CR.

FIG. 4 is a diagram showing a pulse signal and a waveform of the signal after passing the pulse signal through a CR differentiating circuit. (A) shows the case where the cutoff frequency f _c of the filter is much larger than the pulse fundamental frequency f _p , and (b) shows the case where f _c is smaller than or almost equal to f _p .

FIG. 5 is a signal processing circuit diagram (a) according to the embodiment, an input waveform diagram (b), and an output waveform diagram (c) showing an output after processing the signal by the signal processing circuit.

FIG. 6 is a diagram showing the same pulse waveform with different amplitudes.

FIG. 7 is a circuit diagram of a comparator having a hysteresis circuit.

8 is a waveform diagram of input, hysteresis, and output in the circuit of FIG.

FIG. 9 is a diagram showing pulse waveforms before and after receiving Er laser light and passing through a differentiating circuit when CR / Δ = 0.053.

FIG. 10 is a diagram showing waveforms of pulses before and after receiving Er laser light and passing through a differentiating circuit when CR / Δ = 0.39.

FIG. 11 is a diagram showing the waveforms of pulses before and after receiving the light of the Er laser and passing through a differentiating circuit in the case of CR / Δ = 8.2.

FIG. 12 is a measurement circuit diagram corresponding to the example.

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Junichi Asano 3-3-22 Nakanoshima, Kita-ku, Osaka City Kansai Electric Power Co., Inc. (72) Toshio Nagai 3-3-22 Nakanoshima, Kita-ku, Osaka City Kansai Electric Power Within the corporation

Claims (2)

[Claims] 1. A pulsed light of an Er laser is applied to an object,
To receive the transmitted pulsed light and the received pulsed light, obtain two pulses that have the same waveform including noise and the same pulse half width Δ, but different amplitudes, and measure the time interval between these two pulses. The pulse signal containing the noise is passed through a high-pass filter having a time constant of 0.1 to 2.0 times the full width at half maximum Δ to differentiate the pulse signal, and the time when the differential waveform crosses the zero level is obtained. A method for detecting a pulse interval, which is characterized by measuring 2. A pulsed light of an Er laser is applied to an object,
To receive the transmitted pulsed light and the received pulsed light, obtain two pulses that have the same waveform including noise and the same pulse half width Δ, but different amplitudes, and measure the time interval between these two pulses. The pulse signal containing the noise is passed through a high-pass filter having a time constant of 0.1 to 2.0 times the value width Δ to differentiate the output, and the output and the non-inverting input are connected to the resistor R
_The differential signal is input to the inverting input of the comparator which is connected by ₂ and the hysteresis is given by grounding the non-inverting input by the resistor R ₃ , and the time when the output of the comparator crosses the zero level is obtained. A method for detecting a pulse interval, which is characterized by measuring the pulse interval by the difference.