JPH05226433A - Voltage measuring device - Google Patents

Abstract

PURPOSE:To stabilize the focussing actuation of analyzed voltage for improving the measurement precision by a method wherein the voltage of the finite difference between the first mean voltage and the second mean voltage whereto the first voltage wave form is added is to be assumed as the analyzed voltage. CONSTITUTION:The first specific gain (gamma) is applied to the deviation of the output of an analyzer (c) from a target voltage 1 to compute the first voltage 3 corresponding to the first voltage wave form 2 and the mean voltage. Besides, the second gain (alpha) larger than or equal to the first gain (gamma) is applied to said deviation to compute the second mean voltage 5 corresponding to the second voltage wave form 4 and the mean voltage. At this time, the voltage of the finite difference between the first mean voltage 3 and the second mean voltage 5 whereto the first voltage wave form 2 is added is to be assumed as the analyzed voltage. Furthermore, the mean wave form 4 at every measurement is outputted as the measurement results corresponding to the measured voltage. Through these procedures, the fluctuation in the analyzed voltage can be restrained to avoid the divergence.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention relates to a voltage measuring device such as a packaged LSI or an unsealed LS.
The present invention relates to an improved technique of a voltage measuring device suitable for measuring the terminal voltage of I.

[0002]

2. Description of the Related Art FIG. 7 shows an example of a conventional voltage measuring device.
This example shows an opening 1 of a vacuum chamber 1 (also called a lens barrel).
Attach the socket 3 to the atmosphere side of the substrate 2 that closes a,
LSI product 4 detachably attached to the socket 3
(Hereinafter, simply LSI) each package lead 4a (hereinafter
Bottom, simply leads) and formed on the vacuum side of the substrate 2
Electrical connection between each measuring terminal 2a (hereinafter referred to as terminal)
And place the analysis grid 5 in the immediate vicinity of the terminal 2a.
Constitute. Incidentally, 6 is an electron gun for generating an electron beam, and 7
Is an electron lens that focuses the electron beam, and 8 is the electron beam
Pulse gate for making pulse (so-called EB pulse gate
G), 9 is a focused and pulsed electron beam B _PWear
Deflection coil 10 for deflection on the eye terminal 2a is an analysis grid.
Voltage applied to the terminal 5 (hereinafter referred to as analysis voltage V_r) Occurs
Analysis voltage generator, 11 is analysis voltage V_rAnalysis in quantities related to
Secondary electron B from the target terminal 2a passing through the grid 5_R
Analyzer for detecting, 12 is a deflection for driving the deflection coil 9
The driver 13 is a gate driver for driving the pulse gate 8.
Ibar 14 is an exhaust for keeping the inside of the vacuum chamber 1 in a vacuum state.
Air pump, 15 is an example required to actually operate LSI4
For example, the reference phase data, strobe signal and power supply voltage are generated.
LSI drive circuit, 16 is a predetermined inspection sequence
Therefore, various data and signals necessary for each part of the device are appropriately
It is a control circuit that is generated.

Here, the control circuit 16 outputs two types of voltage data (hereinafter referred to as terminal voltage data) when the device is installed or when the system is adjusted. The LSI drive circuit 15 generates a first reference voltage V _s-low and a second reference voltage V _s-high in response to these two data, and outputs these reference voltages to a predetermined terminal 2a of interest. Apply with staggered time. The first reference voltage V _s-low is, for example, the lowest voltage that can be taken by the terminal 2a, and the second reference voltage V _{s- high} is, for example, the highest voltage that can be taken by the terminal 2a. However, it is necessary that both have known voltages.

Now, V _s-low (or V
_s-high ), the electron beam B _P
Of the analysis voltage V _r applied to the analysis grid 5
When is swept from the minimum value to the maximum value, the amount of secondary electrons B _R passing through the analysis grid 5 and captured by the analyzer 11 is
It changes into an almost S-shape with the above sweep operation. That is, an S-shaped curve is drawn such that after a stable transition at a predetermined minimum value, a step of rapidly increasing while drawing a function curve, and then reaching a predetermined maximum value and stabilizing.
FIG. 8 is a diagram showing an ideal S-shaped curve (hereinafter, S curve), the horizontal axis represents the magnitude of the analysis voltage V _r , and the vertical axis represents the amount of secondary electrons. S _L is a slice level (hereinafter referred to as a target voltage) set substantially in the center of the S curve, and the target voltage and the output of the analyzer 11 (voltage value corresponding to the secondary electron signal amount) are analyzed so that they match each other. Feedback control the voltage.

For example, the analysis voltage is V_R0When S₀Become
If the secondary electron signal amount is obtained, follow the S curve
"Α (S₀-S_L) ”Only new analysis shifted to the right
Voltage V _R1And further, this V_R1Obtained in response to
S₁Using "α (S₁-S_L) ”Is just shifted to the right
New analysis voltage V_R2become. That is, in this example,
In the region where the slope of the S curve is small, the analysis voltage is increased (however,
Then S₀<S_LThen S₀> S_LOn the decrease side)
In the area where the slope of the S curve is large while making a large change
Is for increasing the analysis voltage (however, S₁<S_LThen S₁>
S_LIf it is, decrease gradually to
Target voltage S_LHas converged to.

The determination of convergence is made based on whether the change width of the analysis voltage is within a predetermined value or not. The measurement of the analysis voltage in the converged state is repeated a number of times, and the averaging thereof is performed as a measurement result (the voltage of the terminal 2a of interest). Equivalent) to take out. Here, the above α is a feedback gain (so-called convergence coefficient), and is given by the reciprocal of the slope β of the straight line contacting the intersection of the S curve and the target voltage S _L (however, sign inversion), as shown in FIG. ..

[0007]

However, in such a conventional voltage measuring device, since a constant feedback gain α given by the slope of the S curve is used, for example, the S of the secondary electron signal is changed. When the / N ratio is small, there is a problem that the analysis voltage largely deviates from the target voltage S _L and diverges.

FIG. 10 is a diagram showing an actual S curve.
The range between the two virtual lines 20 and 21 is the range in which the secondary electron signal varies due to the influence of noise, and the solid line 22
Represents the range. The main cause of variation is extremely short pulse width (eg 100 ps)
This is because the S / N ratio of the secondary electron signal deteriorates due to the use of the stroboscopic electron beam.

Here, the actual S-curve shown by the solid line 22 is not an S-shape but a mountain shape having one peak, as can be seen from the figure. This is because as the analysis voltage is increased, the efficiency of trapping secondary electrons in the analyzer 11 decreases above a certain voltage. Therefore, if the analysis voltage is located at the peak position or on the left side of the peak, such as V _R0 , V _R1, or V _R2 , the analysis voltage can be controlled to increase or decrease so as to match the target voltage S _L. However, for example, _VR3 or _VR4
As in the case which is located on the right side of the peak, V _R3
Since <S _L and V _R4 <S _L , the analysis voltage after change is controlled to the increasing side, and eventually the analysis voltage is controlled in the direction away from the control intention, that is, the direction in which the secondary electron detection amount decreases. There is a problem that it will be done. By the way,
FIG. 11 shows an unfavorable measurement result. In contrast to the stable waveform shown in FIG. 8A, the defective waveform shows a stable waveform in FIG.
As shown in, large fluctuations appear in places.

Therefore, an object of the present invention is to optimize the feedback gain to stabilize the convergence operation of the analysis voltage and improve the measurement accuracy.

[0011]

In order to achieve the above-mentioned object, the present invention irradiates an electron beam to a predetermined terminal to which a voltage to be measured is applied as shown in the principle diagram of FIG. An analysis voltage is applied to an analysis grid arranged in the vicinity of, and the amount of secondary electrons from the terminal taken out in association with the analysis voltage is detected by the analyzer, and the output of the analyzer becomes a predetermined target voltage. A voltage measuring device for increasing / decreasing the analysis voltage as described above, which corresponds to a first voltage waveform and its average voltage by applying a predetermined first gain to the deviation between the output of the analyzer and the target voltage. The first average voltage is obtained, and the deviation between the output of the analyzer and the target voltage is calculated as the first average voltage.
A second gain corresponding to the second voltage waveform and its average voltage by applying a second gain that is greater than or equal to
Is obtained, and the voltage obtained by adding the first voltage waveform to the difference between the first average voltage and the second average voltage is set as the analysis voltage, and the second voltage waveform for each measurement The average value of the integrated values of is output as a measurement result corresponding to the measured voltage.

[0012]

According to the present invention, if the second gain is made larger than the first gain, the first voltage waveform with less noise (but including the offset) and its average voltage are generated, and the offset is generated. A second voltage waveform that does not include (but has a lot of noise) and its average voltage are generated. Then, from these voltages, an optimum analysis voltage that has less noise and does not include offset is created.

Therefore, the convergence coefficient of the analysis voltage is mainly given by the first gain, the fluctuation of the analysis voltage is suppressed, and the divergence is prevented.

[0014]

Embodiments of the present invention will be described below with reference to the drawings. 2 to 6 show a voltage measuring device according to the present invention as an LSI.
It is an example applied to a test apparatus. The parts corresponding to those in the conventional example (FIG. 7) are designated by the same reference numerals, and the duplicate description thereof will be omitted.

In FIG. 2, the analyzer 11 outputs a secondary electron signal having a magnitude related to the voltage applied to the analysis grid 5 (analysis voltage). In response to the signal from the counter circuit 30 that counts the number of measured phase point data, this signal is integrated by the secondary electronic signal arithmetic mean circuit 31 for each measured phase point, and the arithmetic mean value (secondary electronic signal amount) Are transferred to the control computer 32. On the other hand, the analysis voltage generator 10
Is supplied with probe (terminal) voltage data from the control computer 32, and the analysis voltage generator 10 generates a voltage (analysis voltage) according to this data and applies it to the analysis grid 5.

Reference numeral 33 is a delay unit (corresponding to the gate driver 13 in FIG. 7), and 34 is a clock generator. The control computer 32 executes a predetermined measurement program which will be described later. For example, the feedback voltage storage buffer 32a, the voltage waveform addition buffer 32b,
It has a storage space such as the probe voltage storage buffer 32c and the measurement voltage storage buffer 32d, and updates the contents of these buffers sequentially while executing the measurement program.

FIG. 3 is a flow chart of the main part of the measurement program. In this flowchart, first, when the secondary electrons are acquired in step 40, the convergence coefficient α corresponding to the second gain is calculated based on the deviation between the secondary electron amount and a predetermined target voltage (see the slice level S _{L in} FIG. 8). To obtain the second voltage waveform and its average voltage (second average voltage),
Next, in step 41, the convergence coefficient γ corresponding to the first gain is applied to the deviation to obtain the first voltage waveform and its average voltage (first average voltage).

Here, the convergence coefficient α corresponding to the second gain is the reciprocal of the slope β of the line tangent to the intersection of the S curve and the target voltage S _L (however, the sign is the same as in the conventional example (see FIG. 9)). However, the convergence coefficient γ corresponding to the first gain is given by a value obtained by multiplying the above α by a predetermined coefficient k (k is a number that exceeds 0 and is 1 or less). That is, γ = α · k
(However, 0 <k ≦ 1), and γ is always smaller than α (when k is less than 1) or equivalent (when k is 1).

FIG. 4A is a diagram showing a second voltage waveform obtained by applying the second gain (α) and its average voltage (second average voltage), and FIG. b) is the first
It is a figure which shows the 1st voltage waveform and its average voltage (1st average voltage) acquired by applying the gain ((gamma)) of. Second
The characteristic of the voltage waveform of is that a large amount of noise is included and the amount of offset is not included.
The characteristic of the voltage waveform of is that the offset component is included, and the noise component is not included (or only a small noise component is included). That is, the first voltage waveform and the second voltage waveform
The voltage waveforms of have opposite characteristics.

Next, in step 43, the difference between the first average voltage and the second average voltage is obtained, and the offset of the first voltage waveform is corrected by this difference to generate a new analysis voltage (probe voltage). To do. Then, after the second voltage waveform of this time and the second voltage waveform of the past are added and averaged in step 44 to generate a measurement result, step 40 and subsequent steps are repeated until the end of measurement is determined in step 45.

As described above, in this embodiment, the first gain "γ" and the second gain "α" are set, and these gains are used to calculate the first voltage waveform and its average voltage (first Average voltage) and the second voltage waveform and its average voltage (second average voltage) are generated. Since new analysis voltage and measurement result are obtained from these generated voltages, the offset component is corrected. When a correct analysis voltage can be obtained and, for example, α> γ (that is, 0 <k <1), a small gain (γ) can suppress fluctuations in the analysis voltage.
It is possible to prevent the divergence of the feedback loop.

Even when α = γ (that is, k = 1), the offset of the voltage waveform can be corrected.
Sufficient measurement accuracy can be obtained even from a waveform with a poor / N ratio. FIG. 5 is a histogram analysis diagram of the analysis voltage subjected to feedback control, and is a graph showing the appearance frequency of the fluctuation amount (that is, drift voltage) of the analysis voltage at each measurement timing (however, if 3σ or more is an unstable timing. Detected as). According to this figure, a Gaussian distribution can be obtained that spreads laterally evenly around the average value. A region that greatly deviates from the Gaussian distribution is a feedback region that is extremely unstable and is likely to diverge, and when a feedback voltage exists in such a region, the first region described above is used.
It is desirable to make the gain (γ) of 1 as small as possible (k close to 0). By doing this, timing 1 in FIG.
The protrusion in the positive direction up to 4 can be stabilized. The protrusion of the timing 5 in the negative direction in FIG. 6 is basically stable because the signal amount of the S curve monotonously decreases.

[0023]

According to the present invention, since the feedback gain is optimized, the convergence operation of the analysis voltage can be stabilized and the measurement accuracy can be improved.

[Brief description of drawings]

FIG. 1 is a principle diagram of the present invention.

FIG. 2 is a configuration of an embodiment.

FIG. 3 is a flowchart of an embodiment.

FIG. 4 is a first voltage waveform (and its average voltage) of one embodiment.
FIG. 7 is a diagram of a second voltage waveform (and its average voltage).

FIG. 5 is a histogram analysis diagram of feedback-controlled analysis voltage according to an embodiment.

6 is an unstable timing diagram detected by the histogram analysis of FIG.

FIG. 7 is a configuration diagram of a conventional example.

FIG. 8 is an ideal S curve diagram.

FIG. 9 is a conceptual diagram of a convergence coefficient α.

FIG. 10 is an actual S curve diagram.

FIG. 11 is an unfavorable measurement waveform diagram shown in comparison with a stable waveform.

[Explanation of symbols]

B _P : electron beam S _L : slice level (target voltage) γ: first gain α: second gain 2a: terminal 5: analysis grid 11: analyzer

Claims (1)

[Claims] 1. A terminal which is irradiated with an electron beam to a predetermined terminal to which a voltage to be measured is applied, gives an analysis voltage to an analysis grid arranged in the vicinity of the terminal, and is taken out in association with the analysis voltage. A voltage measuring device for detecting an amount of secondary electrons from the analyzer by an analyzer and increasing / decreasing the analysis voltage so that the output of the analyzer reaches a predetermined target voltage. A predetermined first gain is applied to the deviation to obtain a first voltage waveform and a first average voltage corresponding to the average voltage thereof, and the deviation between the output of the analyzer and the target voltage is the first voltage. A second gain larger than or equal to the gain is applied to obtain a second voltage waveform and a second average voltage corresponding to the average voltage, and the first average voltage and the second average voltage are equal to each other. The voltage obtained by adding the first voltage waveform to the difference Was said analyzing voltage, and a voltage measuring device according to claim average value of the integrated value of the second voltage waveform for each measurement count to output as a measurement result corresponding to the voltage to be measured.