Hi Since I seem to have accidentally hit myself into the face while I was asleep, due to a nightmare, and now have a kind of swollen eye, I've put off recording my batch load video and went deeper into the research for example clips instead. During that research I realized, that the most common method for loading and unloading has shifted from rubber bands to full blown robot arms, and I was thinking, we could actually also use robot arms. I could actually 3D print that robot arm and control it with an Arduino: https://www.youtube.com/watch?v=_B3gWd3A_SI&ab_channel=HowToMechatronics
Opinions?
-lev
Hello David.
The best wafer handling systems I know using the Bernoulli's principle https://en.wikipedia.org/wiki/Bernoulli%27s_principle and the manipulator arm has similarities to a forklift.
So I looked at youtube and found this clip :-) https://www.youtube.com/watch?v=BxkEfLpB3RU
Well, being cost-effective and reliable I'd suggest to use components similar to CD or DVD players. Their drawers mechanic is in principle all we need when the cassette always could stand vertically. Open up the drawer and slide into the cassette, grep the wafer onto drawer, slide out of the cassette while closing the drawer back into our instrument..
BTW, I would transport the wafer batches as seen many times with rail-mounted cages on ceiling.
Regards, Hagen.
On 1/24/22 9:05 PM, David Lanzendörfer wrote:
Hi Since I seem to have accidentally hit myself into the face while I was asleep, due to a nightmare, and now have a kind of swollen eye, I've put off recording my batch load video and went deeper into the research for example clips instead. During that research I realized, that the most common method for loading and unloading has shifted from rubber bands to full blown robot arms, and I was thinking, we could actually also use robot arms. I could actually 3D print that robot arm and control it with an Arduino: https://www.youtube.com/watch?v=_B3gWd3A_SI&ab_channel=HowToMechatronics
Opinions?
-lev
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi I was thinking more along the lines of just having one robot arm, with a simplified low cost version of the wafer transport grab shown in the video you've linked. We could mount a small low cost CMOS sensor at the endpoint as well, and then use OpenCV to detect the presense or absense of wafers and making sure, the graber doesn't damage any of the wafers.
I'm going to buy a 3D printer soon (gotta hack some code for my day job first, in order to buy a proper one), and there's already simple robot arm examples on the web which I could simply print out and then modify them as needed. The grabbing part might however require being made from metal, which will require me to use one of those metal part prototyping services here in Portugal, which will probably increase the cost, so I'll try a 3D printed plastic variant first.
The advantage of a robo arm with a fork lift system is, that I only need one cassette, because the robot can just work slot by slot and put the wafer back after processing.
I could put another inexpensive CMOS sensor directly over the loading stage and use OpenCV to make sure that the wafer is properly placed on the stage. In general, I think OpenCV and cameras is a fantastic way to drastically reduce the amount of sensors required in the system, so we should make as much use of it as possible.
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
Opinion?
Cheers -lev
PS: Today I got my Liberland passport. The greenscreen/chromakey background and the LCOS projector modules are still on their way.
On Tuesday, January 25, 2022 6:14:29 AM WET Hagen SANKOWSKI wrote:
Hello David.
The best wafer handling systems I know using the Bernoulli's principle https://en.wikipedia.org/wiki/Bernoulli%27s_principle and the manipulator arm has similarities to a forklift.
So I looked at youtube and found this clip :-) https://www.youtube.com/watch?v=BxkEfLpB3RU
Well, being cost-effective and reliable I'd suggest to use components similar to CD or DVD players. Their drawers mechanic is in principle all we need when the cassette always could stand vertically. Open up the drawer and slide into the cassette, grep the wafer onto drawer, slide out of the cassette while closing the drawer back into our instrument..
BTW, I would transport the wafer batches as seen many times with rail-mounted cages on ceiling.
Regards, Hagen.
On 1/24/22 9:05 PM, David Lanzendörfer wrote:
Hi Since I seem to have accidentally hit myself into the face while I was asleep, due to a nightmare, and now have a kind of swollen eye, I've put off recording my batch load video and went deeper into the research for example clips instead. During that research I realized, that the most common method for loading and unloading has shifted from rubber bands to full blown robot arms, and I was thinking, we could actually also use robot arms. I could actually 3D print that robot arm and control it with an Arduino: https://www.youtube.com/watch?v=_B3gWd3A_SI&ab_channel=HowToMechatronics
Opinions?
-lev
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hello David.
Regarding the Mini-Computer, please look for flex-flat-cable MIPI CSI-2 connectors.
https://en.wikipedia.org/wiki/Camera_Serial_Interface makes it incredible easy to connect cheap cameras to them.
Nowadays always all small smartphone cameras using this (proprietary) Low-Voltage Differential Signaling (LVDS).
The Rasperry Pi was the first palm-size computer which got a MIPI CSI-2. I am sure, a couple of China Clones has this interface too.
Regards, Hagen.
On 1/25/22 12:55 PM, David Lanzendörfer wrote:
Hi I was thinking more along the lines of just having one robot arm, with a simplified low cost version of the wafer transport grab shown in the video you've linked. We could mount a small low cost CMOS sensor at the endpoint as well, and then use OpenCV to detect the presense or absense of wafers and making sure, the graber doesn't damage any of the wafers.
I'm going to buy a 3D printer soon (gotta hack some code for my day job first, in order to buy a proper one), and there's already simple robot arm examples on the web which I could simply print out and then modify them as needed. The grabbing part might however require being made from metal, which will require me to use one of those metal part prototyping services here in Portugal, which will probably increase the cost, so I'll try a 3D printed plastic variant first.
The advantage of a robo arm with a fork lift system is, that I only need one cassette, because the robot can just work slot by slot and put the wafer back after processing.
I could put another inexpensive CMOS sensor directly over the loading stage and use OpenCV to make sure that the wafer is properly placed on the stage. In general, I think OpenCV and cameras is a fantastic way to drastically reduce the amount of sensors required in the system, so we should make as much use of it as possible.
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
Opinion?
Cheers -lev
PS: Today I got my Liberland passport. The greenscreen/chromakey background and the LCOS projector modules are still on their way.
On Tuesday, January 25, 2022 6:14:29 AM WET Hagen SANKOWSKI wrote:
Hello David.
The best wafer handling systems I know using the Bernoulli's principle https://en.wikipedia.org/wiki/Bernoulli%27s_principle and the manipulator arm has similarities to a forklift.
So I looked at youtube and found this clip :-) https://www.youtube.com/watch?v=BxkEfLpB3RU
Well, being cost-effective and reliable I'd suggest to use components similar to CD or DVD players. Their drawers mechanic is in principle all we need when the cassette always could stand vertically. Open up the drawer and slide into the cassette, grep the wafer onto drawer, slide out of the cassette while closing the drawer back into our instrument..
BTW, I would transport the wafer batches as seen many times with rail-mounted cages on ceiling.
Regards, Hagen.
On 1/24/22 9:05 PM, David Lanzendörfer wrote:
Hi Since I seem to have accidentally hit myself into the face while I was asleep, due to a nightmare, and now have a kind of swollen eye, I've put off recording my batch load video and went deeper into the research for example clips instead. During that research I realized, that the most common method for loading and unloading has shifted from rubber bands to full blown robot arms, and I was thinking, we could actually also use robot arms. I could actually 3D print that robot arm and control it with an Arduino: https://www.youtube.com/watch?v=_B3gWd3A_SI&ab_channel=HowToMechatronics
Opinions?
-lev
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi Hagen
I'll need to design an adapter with multiple such connectors on it anyway, considering, that there will be multiple cams, so I won't need a computer, which actually has those on board, because I can wire it over USB, to the machine running the software. I'll also have to add the controller for the LCOS chip on there as well, so it will be just one or two USB plugs, and then probably a CentOS with the software. RedHat/CentOS is used in the industry, whenever Linux is required, because of the support contracts which are possible, so it will be easier to do marketing, when providing CentOS/RedHat images ready to be installed on any desktop machine you want.
Cheers -lev
On Tuesday, January 25, 2022 1:28:43 PM WET Hagen SANKOWSKI wrote:
Hello David.
Regarding the Mini-Computer, please look for flex-flat-cable MIPI CSI-2 connectors.
https://en.wikipedia.org/wiki/Camera_Serial_Interface makes it incredible easy to connect cheap cameras to them.
Nowadays always all small smartphone cameras using this (proprietary) Low-Voltage Differential Signaling (LVDS).
The Rasperry Pi was the first palm-size computer which got a MIPI CSI-2. I am sure, a couple of China Clones has this interface too.
Regards, Hagen.
On 1/25/22 12:55 PM, David Lanzendörfer wrote:
Hi I was thinking more along the lines of just having one robot arm, with a simplified low cost version of the wafer transport grab shown in the video you've linked. We could mount a small low cost CMOS sensor at the endpoint as well, and then use OpenCV to detect the presense or absense of wafers and making sure, the graber doesn't damage any of the wafers.
I'm going to buy a 3D printer soon (gotta hack some code for my day job first, in order to buy a proper one), and there's already simple robot arm examples on the web which I could simply print out and then modify them as needed. The grabbing part might however require being made from metal, which will require me to use one of those metal part prototyping services here in Portugal, which will probably increase the cost, so I'll try a 3D printed plastic variant first.
The advantage of a robo arm with a fork lift system is, that I only need one cassette, because the robot can just work slot by slot and put the wafer back after processing.
I could put another inexpensive CMOS sensor directly over the loading stage and use OpenCV to make sure that the wafer is properly placed on the stage. In general, I think OpenCV and cameras is a fantastic way to drastically reduce the amount of sensors required in the system, so we should make as much use of it as possible.
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
Opinion?
Cheers -lev
PS: Today I got my Liberland passport. The greenscreen/chromakey background and the LCOS projector modules are still on their way.
On Tuesday, January 25, 2022 6:14:29 AM WET Hagen SANKOWSKI wrote:
Hello David.
The best wafer handling systems I know using the Bernoulli's principle https://en.wikipedia.org/wiki/Bernoulli%27s_principle and the manipulator arm has similarities to a forklift.
So I looked at youtube and found this clip :-) https://www.youtube.com/watch?v=BxkEfLpB3RU
Well, being cost-effective and reliable I'd suggest to use components similar to CD or DVD players. Their drawers mechanic is in principle all we need when the cassette always could stand vertically. Open up the drawer and slide into the cassette, grep the wafer onto drawer, slide out of the cassette while closing the drawer back into our instrument..
BTW, I would transport the wafer batches as seen many times with rail-mounted cages on ceiling.
Regards, Hagen.
On 1/24/22 9:05 PM, David Lanzendörfer wrote:
Hi Since I seem to have accidentally hit myself into the face while I was asleep, due to a nightmare, and now have a kind of swollen eye, I've put off recording my batch load video and went deeper into the research for example clips instead. During that research I realized, that the most common method for loading and unloading has shifted from rubber bands to full blown robot arms, and I was thinking, we could actually also use robot arms. I could actually 3D print that robot arm and control it with an Arduino: https://www.youtube.com/watch?v=_B3gWd3A_SI&ab_channel=HowToMechatronics
Opinions?
-lev
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi I guess those plus a USB hub should do :-) https://pt.aliexpress.com/item/1005001499097554.html
Cheers -lev
On Tuesday, January 25, 2022 2:20:27 PM WET David Lanzendörfer wrote:
Hi Hagen
I'll need to design an adapter with multiple such connectors on it anyway, considering, that there will be multiple cams, so I won't need a computer, which actually has those on board, because I can wire it over USB, to the machine running the software. I'll also have to add the controller for the LCOS chip on there as well, so it will be just one or two USB plugs, and then probably a CentOS with the software. RedHat/CentOS is used in the industry, whenever Linux is required, because of the support contracts which are possible, so it will be easier to do marketing, when providing CentOS/RedHat images ready to be installed on any desktop machine you want.
Cheers -lev
On Tuesday, January 25, 2022 1:28:43 PM WET Hagen SANKOWSKI wrote:
Hello David.
Regarding the Mini-Computer, please look for flex-flat-cable MIPI CSI-2 connectors.
https://en.wikipedia.org/wiki/Camera_Serial_Interface makes it incredible easy to connect cheap cameras to them.
Nowadays always all small smartphone cameras using this (proprietary) Low-Voltage Differential Signaling (LVDS).
The Rasperry Pi was the first palm-size computer which got a MIPI CSI-2. I am sure, a couple of China Clones has this interface too.
Regards, Hagen.
On 1/25/22 12:55 PM, David Lanzendörfer wrote:
Hi I was thinking more along the lines of just having one robot arm, with a simplified low cost version of the wafer transport grab shown in the video you've linked. We could mount a small low cost CMOS sensor at the endpoint as well, and then use OpenCV to detect the presense or absense of wafers and making sure, the graber doesn't damage any of the wafers.
I'm going to buy a 3D printer soon (gotta hack some code for my day job first, in order to buy a proper one), and there's already simple robot arm examples on the web which I could simply print out and then modify them as needed. The grabbing part might however require being made from metal, which will require me to use one of those metal part prototyping services here in Portugal, which will probably increase the cost, so I'll try a 3D printed plastic variant first.
The advantage of a robo arm with a fork lift system is, that I only need one cassette, because the robot can just work slot by slot and put the wafer back after processing.
I could put another inexpensive CMOS sensor directly over the loading stage and use OpenCV to make sure that the wafer is properly placed on the stage. In general, I think OpenCV and cameras is a fantastic way to drastically reduce the amount of sensors required in the system, so we should make as much use of it as possible.
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
Opinion?
Cheers -lev
PS: Today I got my Liberland passport. The greenscreen/chromakey background and the LCOS projector modules are still on their way.
On Tuesday, January 25, 2022 6:14:29 AM WET Hagen SANKOWSKI wrote:
Hello David.
The best wafer handling systems I know using the Bernoulli's principle https://en.wikipedia.org/wiki/Bernoulli%27s_principle and the manipulator arm has similarities to a forklift.
So I looked at youtube and found this clip :-) https://www.youtube.com/watch?v=BxkEfLpB3RU
Well, being cost-effective and reliable I'd suggest to use components similar to CD or DVD players. Their drawers mechanic is in principle all we need when the cassette always could stand vertically. Open up the drawer and slide into the cassette, grep the wafer onto drawer, slide out of the cassette while closing the drawer back into our instrument..
BTW, I would transport the wafer batches as seen many times with rail-mounted cages on ceiling.
Regards, Hagen.
On 1/24/22 9:05 PM, David Lanzendörfer wrote:
Hi Since I seem to have accidentally hit myself into the face while I was asleep, due to a nightmare, and now have a kind of swollen eye, I've put off recording my batch load video and went deeper into the research for example clips instead. During that research I realized, that the most common method for loading and unloading has shifted from rubber bands to full blown robot arms, and I was thinking, we could actually also use robot arms. I could actually 3D print that robot arm and control it with an Arduino: https://www.youtube.com/watch?v=_B3gWd3A_SI&ab_channel=HowToMechatroni cs
Opinions?
-lev
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
On January 25, 2022 11:55:13 AM UTC, "David Lanzendörfer" leviathan@libresilicon.com wrote:
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
do not for god's sake use RAMPS for a 3D printer controller.
consider getting a board designed by think3dprint3d adrian bowyer
they use the superb Trinamic 2.8A stepper drivers which have high accuracy and high frequency interpolation, and Adrian properly designed the PCB to actually take heat away.
you might think you are doing yourself a favour by getting a cheap china knockoff but it is a huge risk
l.
Hi So I kindly asked in Chinese for the schematics of those LCOS PCBs I've ordered and they were very forthcoming. I now have the schematics. When it comes to stepper electronics, I hadn't made up my mind yet, but I was going to check for driver boards which exactly have high accuracy, considering that I'm shooting for a lambda of 500nm, and those stepper motors are supposed to... well step over the wafer without too much offset... I've decided to not move the wafer, but instead moving the optics on top of the wafer, because the thing I've ordered will be so lightweight, that there won't be any vibrations, especially, considering that I'm anyway stepping VERY slow... I'll cross the bridge of what drivers to use, as soon as I'm there :-)
Cheers -lev
On Tuesday, January 25, 2022 2:49:45 PM WET lkcl wrote:
On January 25, 2022 11:55:13 AM UTC, "David Lanzendörfer"
leviathan@libresilicon.com wrote:
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
do not for god's sake use RAMPS for a 3D printer controller.
consider getting a board designed by think3dprint3d adrian bowyer
they use the superb Trinamic 2.8A stepper drivers which have high accuracy and high frequency interpolation, and Adrian properly designed the PCB to actually take heat away.
you might think you are doing yourself a favour by getting a cheap china knockoff but it is a huge risk
l. _______________________________________________ Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
On January 25, 2022 3:47:39 PM UTC, "David Lanzendörfer" leviathan@libresilicon.com wrote:
that there won't be any vibrations, especially, considering that I'm anyway stepping VERY slow... I'll cross the bridge of what drivers to use, as soon as I'm there :-)
Trinamic ICs have 1/256 phase interpolation and generate proper pure sine waves, where ordinary ones have 1/8 if you are lucky and shove out domething closer to a sawtooth than a sine wave. the difference is that Trinamic use 32bit internal microcontrollers whereas your average driver is likely an 8051.
consequently the average drivers have some serious interpolation artefacts that you really do not want to have the hassle of finding out how bad they are *and how PCB designers failed to read the frickin datasheet*
l.
Just a fun-fact:
rumours told me, in quite new designs Trinamic uses a RISC-V core for this "32-bit internal microcontrollers" :-)
Ups, there is already a PR-Statement about this. :-o
https://www.trinamic.com/company/news/news-detail/trinamic-introduces-worlds...
Isn't it a strong buy recommendation??
On 1/25/22 7:14 PM, lkcl wrote:
Trinamic ICs have 1/256 phase interpolation and generate proper pure sine waves, where ordinary ones have 1/8 if you are lucky and shove out domething closer to a sawtooth than a sine wave. the difference is that Trinamic use 32bit internal microcontrollers whereas your average driver is likely an 8051.
On Tue, Jan 25, 2022 at 6:27 PM Hagen SANKOWSKI hsank@posteo.de wrote:
Just a fun-fact:
rumours told me, in quite new designs Trinamic uses a RISC-V core for this "32-bit internal microcontrollers" :-)
yes, they were one of the very first adopters, even before the standard was ratified, because it saved them a hell of a lot of money and it's entirely closed-off and proprietary "in-chip".
l.
On Tue, Jan 25, 2022, 10:16 lkcl luke.leighton@gmail.com wrote:
Trinamic ICs have 1/256 phase interpolation and generate proper pure sine waves,
it may not be all that fancy, but i wrote some code many years ago that does 1:2048 micro-stepping with dithering so the output is pretty smooth (it's a set of 8 256-microstep ramps, with bitreverse-based dithering (kinda like class-D amplifier output)). it uses an ATTiny85 (8-pin AVR). Feel free to try it out if you like -- it just repeatedly slowly accelerates/decelerates to a pretty high speed (to the point where my motor couldn't keep up at like >500 rpm), so you'll have to adapt the code if you want a different drive pattern. you need to continuously rapidly call SetMotorPosition even if you aren't changing the motor position because it uses sw to do the dithering. i remember it being almost silent at low speeds thanks to the microstepping.
Contact me if you want a different foss license, GPLv2 was just whatever i had lying around at the moment.
https://github.com/programmerjake/stepper-motor
Jacob
Hello Everyone,
"I was going to check for driver boards which exactly have high accuracy, considering that I'm shooting for a lambda of 500nm, and those stepper motors are supposed to... well step over the wafer without too much offset..."
I think it will be impossible to find or construct a stepper-driven screw with 500nm accuracy. This problem is usually solved in the way that is used in CD/DVD drives: a screw-shaft with a stepper motor provides wide-range but coarse and imprecise actuation over the entire range, while a short-range magnetic actuator (essentially a solenoid facing a permanent magnet suspended on springs) provides fine movement (resolution dependent only on the ADC, if any, driving the solenoid amplifier) over 2-3 steps distance.
In an optical drive, the analog tracking error signal is fed into the solenoid via an analog feedback loop, keeping the lens over the track. When the solenoid control signal approaches the end of its range in either direction, a comparator senses it and causes the stepper to step one forward or backward (while the movement of the lens caused by the stepping is canceled out by the feedback loop), bringing the solenoid back to the center of its range of movement.
Regards,
Ferenc
On 1/25/22 16:47, David Lanzendörfer wrote:
Hi So I kindly asked in Chinese for the schematics of those LCOS PCBs I've ordered and they were very forthcoming. I now have the schematics. When it comes to stepper electronics, I hadn't made up my mind yet, but I was going to check for driver boards which exactly have high accuracy, considering that I'm shooting for a lambda of 500nm, and those stepper motors are supposed to... well step over the wafer without too much offset... I've decided to not move the wafer, but instead moving the optics on top of the wafer, because the thing I've ordered will be so lightweight, that there won't be any vibrations, especially, considering that I'm anyway stepping VERY slow... I'll cross the bridge of what drivers to use, as soon as I'm there :-)
Cheers -lev
On Tuesday, January 25, 2022 2:49:45 PM WET lkcl wrote:
On January 25, 2022 11:55:13 AM UTC, "David Lanzendörfer"
leviathan@libresilicon.com wrote:
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
do not for god's sake use RAMPS for a 3D printer controller.
consider getting a board designed by think3dprint3d adrian bowyer
they use the superb Trinamic 2.8A stepper drivers which have high accuracy and high frequency interpolation, and Adrian properly designed the PCB to actually take heat away.
you might think you are doing yourself a favour by getting a cheap china knockoff but it is a huge risk
l. _______________________________________________ Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi Ferenc Good point with the precision issue... Do you know of any good off-the-shelf laser distance measuring sensor which has a sufficient resolution for determining the exact distance (with an acceptable tolerance), so that we can do the fine adjustment?
Cheers -lev
On Tuesday, January 25, 2022 6:22:25 PM WET Ferenc Éger wrote:
Hello Everyone,
"I was going to check for driver boards which exactly have high accuracy, considering that I'm shooting for a lambda of 500nm, and those stepper motors are supposed to... well step over the wafer without too much offset..."
I think it will be impossible to find or construct a stepper-driven screw with 500nm accuracy. This problem is usually solved in the way that is used in CD/DVD drives: a screw-shaft with a stepper motor provides wide-range but coarse and imprecise actuation over the entire range, while a short-range magnetic actuator (essentially a solenoid facing a permanent magnet suspended on springs) provides fine movement (resolution dependent only on the ADC, if any, driving the solenoid amplifier) over 2-3 steps distance.
In an optical drive, the analog tracking error signal is fed into the solenoid via an analog feedback loop, keeping the lens over the track. When the solenoid control signal approaches the end of its range in either direction, a comparator senses it and causes the stepper to step one forward or backward (while the movement of the lens caused by the stepping is canceled out by the feedback loop), bringing the solenoid back to the center of its range of movement.
Regards,
Ferenc
On 1/25/22 16:47, David Lanzendörfer wrote:
Hi So I kindly asked in Chinese for the schematics of those LCOS PCBs I've ordered and they were very forthcoming. I now have the schematics. When it comes to stepper electronics, I hadn't made up my mind yet, but I was going to check for driver boards which exactly have high accuracy, considering that I'm shooting for a lambda of 500nm, and those stepper motors are supposed to... well step over the wafer without too much offset... I've decided to not move the wafer, but instead moving the optics on top of the wafer, because the thing I've ordered will be so lightweight, that there won't be any vibrations, especially, considering that I'm anyway stepping VERY slow... I'll cross the bridge of what drivers to use, as soon as I'm there :-)
Cheers -lev
On Tuesday, January 25, 2022 2:49:45 PM WET lkcl wrote:
On January 25, 2022 11:55:13 AM UTC, "David Lanzendörfer"
leviathan@libresilicon.com wrote:
I'd be using some Mini PC like this one, for hooking up all the "webcams" (CMOS sensors) to, and then run all the software headless on there, as well as the Arduino(s) for the motion control: https://www.aliexpress.com/item/1005003020832871.html
do not for god's sake use RAMPS for a 3D printer controller.
consider getting a board designed by think3dprint3d adrian bowyer
they use the superb Trinamic 2.8A stepper drivers which have high accuracy and high frequency interpolation, and Adrian properly designed the PCB to actually take heat away.
you might think you are doing yourself a favour by getting a cheap china knockoff but it is a huge risk
l. _______________________________________________ Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
On Tue, Jan 25, 2022, 12:05 David Lanzendörfer leviathan@libresilicon.com wrote:
Hi Ferenc Good point with the precision issue...
Do you know of any good off-the-shelf laser distance measuring sensor
which has a sufficient resolution for determining the exact distance (with an acceptable tolerance), so that we can do the fine adjustment?
it isn't off-the-shelf but you could likely easily build one based off of laser-diode self-mixing: https://youtu.be/MUdro-6u2Zg
Jacob
Hello List,
I just found a laser diode with a built-in photodetector in my home. I will play with it and report in the upcoming days. Since these are used in optical drives, they are supposed to be atom-cheap. What is evident right now is that the diode's temperature has to be controlled to keep the wavelength constant.
Regards,
Ferenc
On 1/25/22 21:42, Jacob Lifshay wrote:
On Tue, Jan 25, 2022, 12:05 David Lanzendörfer leviathan@libresilicon.com wrote:
Hi Ferenc Good point with the precision issue... Do you know of any good off-the-shelf laser distance measuring sensor which has a sufficient resolution for determining the exact distance (with an acceptable tolerance), so that we can do the fine adjustment?it isn't off-the-shelf but you could likely easily build one based off of laser-diode self-mixing: https://youtu.be/MUdro-6u2Zg
Jacob
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi Ferenc Awesome! I'm looking forward to hearing your results! I wonder how we can get a somewhat reproducible result and how we can calibrate the distance measuring setup... Any idea?
Cheers -lev
On Tuesday, January 25, 2022 10:05:33 PM WET Ferenc Éger wrote:
Hello List,
I just found a laser diode with a built-in photodetector in my home. I will play with it and report in the upcoming days. Since these are used in optical drives, they are supposed to be atom-cheap. What is evident right now is that the diode's temperature has to be controlled to keep the wavelength constant.
Regards,
Ferenc
On 1/25/22 21:42, Jacob Lifshay wrote:
On Tue, Jan 25, 2022, 12:05 David Lanzendörfer
leviathan@libresilicon.com wrote: Hi Ferenc Good point with the precision issue...
Do you know of any good off-the-shelf laser distance measuring sensor which has a sufficient resolution for determining the exact distance (with an acceptable tolerance), so that we can do the fine adjustment?it isn't off-the-shelf but you could likely easily build one based off of laser-diode self-mixing: https://youtu.be/MUdro-6u2Zg
Jacob
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi
If we have a wafer that is already patterned, we need to localize the fiducials on on it at first. Their positions, as reported in interferometer increments, will define an arbitrary-unit scale across the wafer that depends on wavelength. In that case, not knowing lambda is not a problem as long as it remains the same during stepping that wafer. For a new wafer, I would recommend to expose and etch a "zero-layer" on it (like in HKUST) with a conventional mask and contact alignment.
Regards,
Ferenc
On 1/25/22 23:07, David Lanzendörfer wrote:
Hi Ferenc Awesome! I'm looking forward to hearing your results! I wonder how we can get a somewhat reproducible result and how we can calibrate the distance measuring setup... Any idea?
Cheers -lev
On Tuesday, January 25, 2022 10:05:33 PM WET Ferenc Éger wrote:
Hello List,
I just found a laser diode with a built-in photodetector in my home. I will play with it and report in the upcoming days. Since these are used in optical drives, they are supposed to be atom-cheap. What is evident right now is that the diode's temperature has to be controlled to keep the wavelength constant.
Regards,
Ferenc
On 1/25/22 21:42, Jacob Lifshay wrote:
On Tue, Jan 25, 2022, 12:05 David Lanzendörfer
leviathan@libresilicon.com wrote: Hi Ferenc Good point with the precision issue...
Do you know of any good off-the-shelf laser distance measuring sensor which has a sufficient resolution for determining the exact distance (with an acceptable tolerance), so that we can do the fine adjustment?it isn't off-the-shelf but you could likely easily build one based off of laser-diode self-mixing: https://youtu.be/MUdro-6u2Zg
Jacob
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Hi Yes. I intend to etch alignment markers. And thanks! That solves the problem!
Cheers -lev
On Tuesday, January 25, 2022 10:19:19 PM WET Ferenc Éger wrote:
Hi
If we have a wafer that is already patterned, we need to localize the fiducials on on it at first. Their positions, as reported in interferometer increments, will define an arbitrary-unit scale across the wafer that depends on wavelength. In that case, not knowing lambda is not a problem as long as it remains the same during stepping that wafer. For a new wafer, I would recommend to expose and etch a "zero-layer" on it (like in HKUST) with a conventional mask and contact alignment.
Regards,
Ferenc
On 1/25/22 23:07, David Lanzendörfer wrote:
Hi Ferenc Awesome! I'm looking forward to hearing your results! I wonder how we can get a somewhat reproducible result and how we can calibrate the distance measuring setup... Any idea?
Cheers -lev
On Tuesday, January 25, 2022 10:05:33 PM WET Ferenc Éger wrote:
Hello List,
I just found a laser diode with a built-in photodetector in my home. I will play with it and report in the upcoming days. Since these are used in optical drives, they are supposed to be atom-cheap. What is evident right now is that the diode's temperature has to be controlled to keep the wavelength constant.
Regards,
Ferenc
On 1/25/22 21:42, Jacob Lifshay wrote:
On Tue, Jan 25, 2022, 12:05 David Lanzendörfer
leviathan@libresilicon.com wrote: Hi Ferenc Good point with the precision issue...
Do you know of any good off-the-shelf laser distance measuring sensor which has a sufficient resolution for determining the exact distance (with an acceptable tolerance), so that we can do the fine adjustment?it isn't off-the-shelf but you could likely easily build one based off of laser-diode self-mixing: https://youtu.be/MUdro-6u2Zg
Jacob
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
Libresilicon-developers mailing list Libresilicon-developers@list.libresilicon.com https://list.libresilicon.com/mailman/listinfo/libresilicon-developers
--- crowd-funded eco-conscious hardware: https://www.crowdsupply.com/eoma68
On Tue, Jan 25, 2022 at 6:22 PM Ferenc Éger eegerferenc@gmail.com wrote:
In an optical drive, the analog tracking error signal is fed into the solenoid via an analog feedback loop, keeping the lens over the track. When the solenoid control signal approaches the end of its range in either direction, a comparator senses it and causes the stepper to step one forward or backward (while the movement of the lens caused by the stepping is canceled out by the feedback loop), bringing the solenoid back to the center of its range of movement.
ohh, don't they use some sort of 2D interference pattern for focussing? you get 4 spots and it's easy to measure the intensity of them. something like that, i can't remember where i heard about this, details are fuzzy in my mind.
l.
libresilicon-developers@list.libresilicon.com
-
David Lanzendörfer -
Ferenc Éger -
Hagen SANKOWSKI -
Jacob Lifshay -
lkcl -
Luke Kenneth Casson Leighton