JPH0768456A - Method for corrective grinding - Google Patents

Abstract

PURPOSE:To provide a method for corrective grinding where the grinding and of a unsymmetrical, highly precise die, a non-spherical lens or the like can be automated with high precision. CONSTITUTION:A self-copying grinding device 1 and an automatic measuring instrument 2 are provided on a spindle head 4 of an NC machine tool to be controlled by a control device 3 to execute the measurement and the corrective grinding of the surface of a work, and during the measurement, the moving error and the thermal drift of the machine is corrected, and at the same time, the real grinding quantity per pass is monitored from the results of measurement, the following grinding quantity is presumed thereby, the number of the following grinding is set, and the correction factor is introduced to prevent the overshoot due to the variance of the grinding quantity.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a correction polishing method capable of automating and highly precise polishing of a precision metal mold, aspherical lens or the like having high precision which is not axially symmetric.

[0002]

2. Description of the Related Art Conventionally, an axially symmetric aspherical lens is at a technical level capable of being automatically processed by a spindle rotating type (lathe type) ultra-precision processing machine. However, in the case of an aspherical surface which is not axially symmetric, it cannot be processed by the above-mentioned processing machine, and therefore, the measurement by the three-dimensional measuring machine and the manual finish polishing based on the measurement result are performed. There is.

[0003]

In the polishing process of aspherical lenses which are not axially symmetric, the polishing amount itself (polishing depth) is not constant because the polishing process itself is dependent on manual work. The predetermined final shape is obtained for the first time by repeating the work and the polishing work by hand finishing many times. These series of steps are extremely complicated and patient work, which is an inefficient processing method, and automation has long been desired.

Therefore, it is an object of the present invention to provide a modified polishing method capable of automating and improving the precision of precision dies and aspherical lenses having high precision which are not axially symmetric.

[0005]

In order to achieve the above object, the present invention is to install an automatic measuring device in a self-profiling polishing device whose operation is controlled by a numerical control and a copying control device, and to set a spindle rotation speed, a feed speed, The polishing pressure, pick amount, polishing tool, abrasive grain type, polishing amount per pass, and target shape are stored in advance in the control device, and the polishing area on the surface of the workpiece is divided into grids, and each grid point The movement error (flatness) of the corresponding machine is measured in advance and a machine error map is created and stored in the control device. Prior to machining, each grid point is measured by the automatic measuring device, and each grid point is measured. The deviation from the target shape is corrected, and the deviation is corrected by the movement error of the machine corresponding to the grid point on the machine error map stored in advance, and
Each time measurement is performed, thermal drift correction is performed based on the difference between the measured values at arbitrary positions, the correction coefficient is added to this to make the next polishing amount, and this polishing amount is divided by the polishing amount per pass to obtain the number of polishing times. Obtain the (number of repeated passes), perform the corrective polishing process on the area including each grid point based on this polishing number, and after the completion of this process, measure again to find the deviation from the target shape, and move this deviation to the machine movement. Perform the error correction and the thermal drift correction, monitor the actual polishing amount per pass from the measurement result together with this, estimate the next polishing amount from this actual polishing amount, and calculate the number of polishing in the same manner as described above. This is performed and this is repeated until the deviation becomes equal to or less than a preset allowable value.

[0006]

According to the present invention, the machine movement error correction and the thermal drift correction can be automated, and the actual polishing amount per pass is monitored from the measurement result, and the next polishing number is set from this actual polishing amount. At the same time, by introducing the correction coefficient, it is possible to prevent overshoot due to excessive polishing.

[0007]

EXAMPLE FIG. 1 is an explanatory view of a surface polishing method for an aspherical lens, in which a target shape f (x, y) of the aspherical lens is theoretically calculated, and the polishing area is as shown in FIG. To xy
Pre-processed surface shape g ₀ (x, y) by dividing into a grid in the plane
The deviation h (x, y) between the target shape f (x, y) and the target shape f (x, y) is measured by an automatic measuring device for each grid point.

FIG. 2 is a schematic explanatory view of an essential part of an apparatus for carrying out the method of the present invention. An automatic measuring apparatus (2) is attached to a self-copy polishing apparatus (1) whose operation is controlled by a control apparatus (3). It is installed. The self-profiling polishing device (1) includes a uniaxial constant-pressure self-profiling polishing head (1c), which has a driving means such as an air motor built therein and rotatably supports a rotating spindle (1b) having a polishing tool (1a) mounted at its tip. It is mounted on the spindle head (4) of an NC machine tool (hereinafter simply referred to as a machine). It is desirable to use the self-copying polishing head (1c) described in Japanese Patent Publication No. 4-48578. The self-profiling polishing head disclosed in Japanese Patent Publication No. 4-48578 is
It is attached to the spindle head of the machine while the rotating spindle is supported so that it can be elastically displaced in the XYZ directions, the polishing tool is pressed against the surface of the workpiece with a constant pressure, and the spindle head of the machine is made to follow the movement. The amount of elastic displacement of the rotating spindle in the XYZ axis directions due to contact with the workpiece is detected, and the copying operation of the spindle head of the machine is controlled so that the combined value of the detected displacements becomes a constant value, and the surface of the workpiece is ground under constant pressure. To do.

In the automatic measuring device (2), a measuring head (2a) with a retract function (up and down retracting operation) is attached to the spindle head (4), and measurement is automatically performed by a control device (3) such as a personal computer. Done. That is, the measurement operation of the automatic measuring device (2) expands / contracts in two stages as shown in FIG. 2, and the measurement head (2a) is first lowered to position the measurement head (2a) in the measurement process. The probe (2b) is lowered by the retract cylinder at each position, and the surface position of the workpiece is measured by an appropriate displacement sensor. According to this configuration, since the machine does not move during measurement, the measurement error can be minimized. The retract stroke is, for example, 100 mm.

FIG. 3 shows a correction polishing processing flow of the present invention. As a control method of the polishing amount by the correction polishing,
This can be achieved by controlling the number of polishing operations, controlling the residence time (scanning speed), varying the polishing pressure, and changing the polishing conditions (setting the pick amount). In the present invention, however, a method of controlling the number of polishing operations is used to make the surface properties uniform. It is adopted. What is characteristic here is that (a) the movement accuracy of the machine and the thermal drift correction at the time of measurement are automated, and (b) 1 from the measurement results.
It can be mentioned that the actual polishing amount per pass is monitored and the next polishing number is set, and at the same time, the correction coefficient α is introduced in order to prevent overshoot due to variation in the polishing amount.

A detailed description will be given below according to the flow of FIG. On-the-machine measurement accuracy is directly affected by machine movement accuracy. In the present invention, the movement error (flatness) of the machine is measured in advance corresponding to each grid point in FIG. 1, and the movement error map of the machine is stored in the controller (3) based on this. At the same time, a correction function by linear interpolation is added to the thermal drift due to the temperature fluctuation during measurement.

Raw measurement data from the measurement displacement sensor may cause a problem in measurement accuracy. That is, it is necessary to consider the following three items as error factors that affect the measurement accuracy. (1) Measuring accuracy of measuring instrument (2) Moving accuracy of machine (geometrical error): g (x) (3) Thermal displacement of machine (thermal drift): t (x)

Since the present invention uses the digital linear scale (resolution 0.1 μm) as the displacement sensor, the measurement error of the measuring instrument of (1) can be ignored,
The measurement method that cancels the errors in the items (2) and (3) is adopted.

In FIG. 4, for the true shape f (x),
The actually measured value is the sum of the geometric error g (x) and the thermal drift t (x), that is, {f
(X) + g (x) + t (x)}. Therefore, the geometrical error g (x) is measured in advance, stored as an error map in the control device (3), and the error corresponding to g (x) is corrected at the time of measurement.

Regarding the thermal drift t (x), the difference between the start point and the end point at the time of measurement, actually, the start point → the end point → the start point, is sequentially measured for each grid point, and the first and second same measurement points ( It becomes possible to perform correction by performing trend correction (correction by linear interpolation) from the difference between the measured values at the (start point) (corresponding to a kind of mastering function). Note that the thermal drift t
It has been confirmed that (x) can be approximated by a substantially straight line as shown in the actually measured data on the upper side of FIG. In addition, the following measures are taken to improve reliability. (1) Data is captured twice at the same measurement point. If the difference between the two measured values is within 2 μm, for example, it is determined to be OK and the next measurement is performed. (2) Over / Under Displacement Setting Set the upper and lower limits of the measured value, and if it is over / under measured value, judge as abnormal.

FIG. 5 shows the results of measuring the movement error (flatness) and thermal drift of the machine on the workpiece mounting surface.
The in-plane flatness error is about 14 μm, and a thermal drift of about 17 μm occurs at a temperature fluctuation width of 1.8 ° C. In contrast,
By performing the mechanical accuracy correction and the thermal drift correction, FIG.
As shown in the lower side of the figure, it can be seen that the correction accuracy of 0.6 μm is obtained. Since it has been clarified that sufficient accuracy of measurement can be guaranteed by the above method, the correction polishing method of the present invention will be described below.

Once the deviation from the ideal shape is obtained,
In principle, it is possible to finish into a predetermined shape by pushing in the amount of error corresponding to the deviation by polishing. However, in reality, it is empirically known that the polishing characteristics (polishing amount) of the polishing tool (1a) change from moment to moment depending on the polishing conditions and tool life. Therefore, when the polishing amount is always treated as a constant value, it does not converge to the ideal shape forever.
Alternatively, a problem that convergence takes time occurs. Therefore, a method of measuring the polishing amount immediately before polishing, estimating the next polishing amount from this, and correcting each time to a polishing amount that is closer to the actual polishing amount (corresponding to a kind of learning control)
Has been adopted.

The polishing amount for one pass is determined by the spindle speed, feed rate,
Although it varies depending on the polishing pressure, the pick amount, the polishing tool, and the type of abrasive grains, the polishing amount d per pass under a certain condition is obtained in advance by experiments and is stored as data in the control device (3). It is assumed that

Therefore, assuming that a predetermined shape is finished by repeating measurement and correction polishing n times (see FIG. 1), the first, second, third, ..., Nth times (however, generally, n ≧
The number Ni of polishing passes and the polishing amount di per pass in the polishing region (square or circular region) centered on any point P (x, y) up to 3) are as follows. The first _{N 1 (x, y) =} {g 0 (x, y) -f (x, y)} / nd 1 per pass polishing amount _{d 1 = {g 0 (x} , y) -g 1 ( x, y)}
/ N ₁ 2nd time N ₂ (x, y) = {g ₁ (x, y) −f (x, y)} / nd
₁ polishing amount per one pass _{d 2 = {g 1 (x} , y) -g 2 (x, y)}
/ N ₂ 3rd time N ₃ (x, y) = {g ₂ (x, y) −f (x, y)} / nd
Polishing amount per ₂ 1 pass _{d 3 = {g 2 (x} , y) -g 3 (x, y)}
/ N ₃ ······· Polishing amount per pass d _n-1 = {g _n-2 (x, y) −g _n-1
(x, y)} / N _n-1 nth time N _n (x, y) = {g _n-1 (x, y)
-F (x, y)} / nd _n-1

That is, instead of calculating the number of polishing passes N _i as a constant until the end by using the previously determined polishing amount d per pass, the previous actual polishing amount d _i-1 is constantly monitored. A method of correcting the polishing pass number N _i based on this is adopted. According to this method, fluctuations in the polishing amount due to subtle changes in the polishing tool (1a) can be eliminated, and the polishing amount d obtained in the experiment can be a rough set value, which is extremely convenient. When the measurement results satisfy the following conditions: ┃gi (x, y) -f (x, y) ┃ ≦ ε That is, if the preset allowable value ε or less, it is judged that the machining is completed and the measurement is performed. Skip point polishing. When all the measurement points satisfy the above conditions, all processing is completed.

In order to improve the polishing efficiency, the total polishing amount is not divided into n equal parts, but the polishing amount per time is ₁ as N ₁ , N ₂ ... N _n increases. / 2, 1/4, 1/8
It is a good idea from the viewpoint of machining efficiency and prevention of over-polishing to reduce it in geometric progression. That is, if the error amount corresponding to the deviation amount is driven in 100% as it is, there is a risk of overshoot due to overpolishing. Therefore, the correction coefficient α is introduced.

An example of changes in the shape deviation by the above method is shown in FIG.
Is shown in. The cross-sectional shape became closer to the target shape with the number of times of polishing, and reached the allowable value after 7 times of polishing, but overcorrection due to overshoot occurred in the lower right portion because the correction coefficient was set to α = 1. FIG. 7 shows the residual sum of squares of the shape error and the change in the polishing amount per time when the correction coefficient is changed. From FIG. 7, the polishing amount changes with time. It is automatically corrected by learning the previous polishing amount. Note that it is dangerous to keep the shape deviation amount of 100% as it is, and it is effective to avoid this with a correction coefficient. As a matter of course, the larger the correction coefficient value, the faster the convergence, and thus the polishing function can be used in the shortest time by using the learning function. Further, it is possible to control the shape accuracy in units of 0.1 μm by reducing the polishing amount per pass, that is, by increasing the minimum resolution.

[0023]

EFFECTS OF THE INVENTION Conventionally, in the polishing processing of high precision precision molds and aspherical lenses which are not axially symmetric, a spindle rotating type super precision processing machine or the like cannot be used. According to the present invention, an automatic measuring device is installed in a self-copying polishing device, and a series of processes including the above-mentioned measurement and correction polishing process are performed. The work can be automated and unmanned, and this type of polishing process can be automated and highly accurate.

[Brief description of drawings]

FIG. 1 is an explanatory diagram of a surface polishing method for an aspherical lens.

FIG. 2 is a schematic explanatory view of a main part of an apparatus for carrying out the method of the present invention.

FIG. 3 is a flow chart of a correction polishing processing method of the present invention.

FIG. 4 is an explanatory diagram of measurement error.

FIG. 5 is an explanatory diagram of surface shape measurement values of a workpiece at different room temperature X-direction positions and mechanical accuracy correction and thermal drift correction.

FIG. 6 is an explanatory view showing changes in the number of times of polishing and the polishing cross-sectional shape of a workpiece.

FIG. 7 is an explanatory diagram showing changes in the residual sum of squares and the polishing amount per pass for various correction coefficients.

[Explanation of symbols]

 1 Self-copying polishing device 2 Automatic measuring device 3 Control device 4 NC machine tool spindle head 5 Indicator

Claims (1)

[Claims] 1. An automatic measuring device is installed in a self-copying polishing device whose operation is controlled by a numerical control and a copying control device, and the spindle rotation speed, feed rate, polishing pressure, pick amount, polishing tool, abrasive grain type, 1 The amount of polishing per pass and the target shape are stored in advance in the controller, the polishing area on the surface of the workpiece is divided into a grid, and the movement error (flatness) of the machine corresponding to each grid point is measured in advance. A machine error map is created and stored in the control device, each grid point is measured with an automatic measuring device prior to machining, the deviation from the target shape for each grid point is calculated, and this deviation is stored in advance. It is corrected by the movement error of the machine corresponding to the lattice point on a certain machine error map, and
Each time measurement is performed, thermal drift correction is performed based on the difference between the measured values at arbitrary positions, the correction coefficient is added to this to make the next polishing amount, and this polishing amount is divided by the polishing amount per pass to obtain the number of polishing times. Obtain the (number of repeated passes), perform the corrective polishing process on the area including each grid point based on this polishing number, and after the completion of this process, measure again to find the deviation from the target shape, and move this deviation to the machine movement. Perform the error correction and the thermal drift correction, monitor the actual polishing amount per pass from the measurement result together with this, estimate the next polishing amount from this actual polishing amount, and calculate the number of polishing in the same manner as described above. A method of correcting polishing, characterized in that it is carried out and is repeated until the deviation becomes equal to or less than a preset allowable value.