Ultrasonic Alignment tool for Bed Leveling+zeroing of stage + Filament monitor

[Mjolinor]

I think that temperature variation will kill it anyway unless you employ some very clever compensation techniques. You have minimum 10% variation in velocity with temperature and velocity is the main factor in any timing that you must use to measure distance. That temperature variation cannot be accounted for because it may go from 100 C at the head through a cold patch at (I guess) around 30C then back up to 60C as the signal approaches the bed. Once you had set the thing with a warm head and bed then the results would be way off with a cold head and bed so it would need setting for every different temperature you use and for every location on the bed as temperatures vary depending on where you are then once you have opened a door or walked round the room the air moves and your readings are wrong again.

You also have a small variation in the velocity depending on frequency so any frequency sweep method would need more compensation tables. If you use eg. 40 kHz then you have lambda of around 8mm. No way on earth can you distinguish how many wavelengths you are away without moving the bed up and down significant amounts and without knowing if the head will crash how can you know if you are one wavelength away or more.

I am watching with interest to see how you overcome some of these (and significant other) problems. I looked at doing this a couple of years ago as I have done significant design work in ultrasonic sensing, I abandoned the idea because the further in you go the more problems arise.

[ShadowX]

I agree. Its not that it can't be done, but at what cost. Its interesting that they claim they can compensate for environmental factors with doppler, phase difference, etc. It is a very challenging environment especially on a 3D printer with a heated bed. The hot end near the sensor may be at 240C and the bed is at 110C and the air surrounding the bed is at 25C with a fan blowing over it. I would be shocked if the air density remains constant between the probe and the bed unless its a very short distance. At least with a physical mechanical probe, you have a set known distance that shouldn't change. Here, you are dependent on the sound waves travelling through air onto a hot surface and bouncing back. It might work in a relatively short distance from the sensor, I just don't know. They have to prove that it works and its reliable.

The 10F to 160F (-12C to 71C) is actually a very reasonable temperature swing that would accelerate the life cycle testing of the electronics. Most high reliability electronics can handle those temperature swings within a limited amount of cycles (ie: 30 cycles). You have to temp cycle it through the range and measure with a fixed calibrated height. Its the only way you know what the measurements are doing in those environments. At least you can do a plot and get an idea of the delta in readings and plot the change over time/temperature. You may find out that its only accurate at a specific temperature range. The accuracy and precision may even drift with the temperature swings. If they claim something is accurate or precise to a certain level, the entire device, including sensor and the electronic components has to output a signal within those spec limits. If the electronic components drift over temperature, even though the sensor is stable, the output would also drift. You may end up having to source high precision components with low thermal drift to maintain the requirements.

I hope others are asking similar questions and getting honest answers. I don't mind if someone doesn't know an answer, but if they give me a BS answer, I will be all over them in a short time. They have three PhD graduates or candidates working on the project, so they should be able to answer some of these technical questions easily.

I think that temperature variation will kill it anyway unless you employ some very clever compensation techniques. You have minimum 10% variation in velocity with temperature and velocity is the main factor in any timing that you must use to measure distance. That temperature variation cannot be accounted for because it may go from 100 C at the head through a cold patch at (I guess) around 30C then back up to 60C as the signal approaches the bed. Once you had set the thing with a warm head and bed then the results would be way off with a cold head and bed so it would need setting for every different temperature you use and for every location on the bed as temperatures vary depending on where you are then once you have opened a door or walked round the room the air moves and your readings are wrong again.

You also have a small variation in the velocity depending on frequency so any frequency sweep method would need more compensation tables. If you use eg. 40 kHz then you have lambda of around 8mm. No way on earth can you distinguish how many wavelengths you are away without moving the bed up and down significant amounts and without knowing if the head will crash how can you know if you are one wavelength away or more.

I am watching with interest to see how you overcome some of these (and significant other) problems. I looked at doing this a couple of years ago as I have done significant design work in ultrasonic sensing, I abandoned the idea because the further in you go the more problems arise.

[dsayan]

I agree. Its not that it can't be done, but at what cost. Its interesting that they claim they can compensate for environmental factors with doppler, phase difference, etc. It is a very challenging environment especially on a 3D printer with a heated bed. The hot end near the sensor may be at 240C and the bed is at 110C and the air surrounding the bed is at 25C with a fan blowing over it. I would be shocked if the air density remains constant between the probe and the bed unless its a very short distance. At least with a physical mechanical probe, you have a set known distance that shouldn't change. Here, you are dependent on the sound waves travelling through air onto a hot surface and bouncing back. It might work in a relatively short distance from the sensor, I just don't know. They have to prove that it works and its reliable.

The 10F to 160F (-12C to 71C) is actually a very reasonable temperature swing that would accelerate the life cycle testing of the electronics. Most high reliability electronics can handle those temperature swings within a limited amount of cycles (ie: 30 cycles). You have to temp cycle it through the range and measure with a fixed calibrated height. Its the only way you know what the measurements are doing in those environments. At least you can do a plot and get an idea of the delta in readings and plot the change over time/temperature. You may find out that its only accurate at a specific temperature range. The accuracy and precision may even drift with the temperature swings. If they claim something is accurate or precise to a certain level, the entire device, including sensor and the electronic components has to output a signal within those spec limits. If the electronic components drift over temperature, even though the sensor is stable, the output would also drift. You may end up having to source high precision components with low thermal drift to maintain the requirements.


I hope others are asking similar questions and getting honest answers. I don't mind if someone doesn't know an answer, but if they give me a BS answer, I will be all over them in a short time. They have three PhD graduates or candidates working on the project, so they should be able to answer some of these technical questions easily.

”I agree. Its not that it can't be done, but at what cost. Its interesting that they claim they can compensate for environmental factors with doppler, phase difference, etc. It is a very challenging environment especially on a 3D printer with a heated bed. The hot end near the sensor may be at 240C and the bed is at 110C and the air surrounding the bed is at 25C with a fan blowing over it. I would be shocked if the air density remains constant between the probe and the bed unless its a very short distance. At least with a physical mechanical probe, you have a set known distance that shouldn't change. Here, you are dependent on the sound waves travelling through air onto a hot surface and bouncing back. It might work in a relatively short distance from the sensor, I just don't know. They have to prove that it works and its reliable.
The 10F to 160F (-12C to 71C) is actually a very reasonable temperature swing that would accelerate the life cycle testing of the electronics. Most high reliability electronics can handle those temperature swings within a limited amount of cycles (ie: 30 cycles). You have to temp cycle it through the range and measure with a fixed calibrated height. Its the only way you know what the measurements are doing in those environments. At least you can do a plot and get an idea of the delta in readings and plot the change over time/temperature. You may find out that its only accurate at a specific temperature range. The accuracy and precision may even drift with the temperature swings. If they claim something is accurate or precise to a certain level, the entire device, including sensor and the electronic components has to output a signal within those spec limits. If the electronic components drift over temperature, even though the sensor is stable, the output would also drift. You may end up having to source high precision components with low thermal drift to maintain the requirements.
I hope others are asking similar questions and getting honest answers. I don't mind if someone doesn't know an answer, but if they give me a BS answer, I will be all over them in a short time. They have three PhD graduates or candidates working on the project, so they should be able to answer some of these technical questions easily. “
alignG compensates for the temperature gradient. It does not mean that there is no effect; but, it is minimized. There are many ways to do that. But, please don’t make it that much bold. In the worst condition, after compensation, we may have less than 20% change in the accuracy. In the other hand, it is a systematic error in the measurement; if you always do the measurement on a bed with 120C, you will get an approximately consistent result; so, it is not really matter. But if you plan to always play with the bed temperature during the bed leveling, that’s a different deal. You are not also have to do the measurement when the extruder and bed are hot, if you are that much concerned.
The parameters that you mentioned, directly affect the transit time based ultrasonic sensors. We are not doing a transit time measurement. For Phase/Doppler measurement, there is a different story. Our measurement is not as simple as sending a pulse and measuring the time that it returns. There are many techniques employed by many researchers during past decades to minimize the effect of sound speed on the Doppler/Phase measurements. A simple search in google or IEEEXplore website will show you many practiced techniques including multi-frequency approach, T&R orientation, …
Any measurement instrument is sensitive to the temperature. We never claimed that we have an ideal sensor that solves all the problems for everybody on any condition. What are the sensors that you can use for auto bed leveling today? Please let me know the available choices. Which of them are not dependent to the temperature? Is an inductive proximity sensor or IR sensor kept several millimeters away from the hot bed not affected by its temp? You can keep alignG 10cm away from the bed and do the measurement.

[dsayan]

I think that temperature variation will kill it anyway unless you employ some very clever compensation techniques. You have minimum 10% variation in velocity with temperature and velocity is the main factor in any timing that you must use to measure distance. That temperature variation cannot be accounted for because it may go from 100 C at the head through a cold patch at (I guess) around 30C then back up to 60C as the signal approaches the bed. Once you had set the thing with a warm head and bed then the results would be way off with a cold head and bed so it would need setting for every different temperature you use and for every location on the bed as temperatures vary depending on where you are then once you have opened a door or walked round the room the air moves and your readings are wrong again.

You also have a small variation in the velocity depending on frequency so any frequency sweep method would need more compensation tables. If you use eg. 40 kHz then you have lambda of around 8mm. No way on earth can you distinguish how many wavelengths you are away without moving the bed up and down significant amounts and without knowing if the head will crash how can you know if you are one wavelength away or more.

I am watching with interest to see how you overcome some of these (and significant other) problems. I looked at doing this a couple of years ago as I have done significant design work in ultrasonic sensing, I abandoned the idea because the further in you go the more problems arise.

I think that temperature variation will kill it anyway unless you employ some very clever compensation techniques. You have minimum 10% variation in velocity with temperature and velocity is the main factor in any timing that you must use to measure distance. That temperature variation cannot be accounted for because it may go from 100 C at the head through a cold patch at (I guess) around 30C then back up to 60C as the signal approaches the bed. Once you had set the thing with a warm head and bed then the results would be way off with a cold head and bed so it would need setting for every different temperature you use and for every location on the bed as temperatures vary depending on where you are then once you have opened a door or walked round the room the air moves and your readings are wrong again.”
alignG compensates for the temperature gradient. It does not mean that there isnt any effect; but, it is minimized. There are many ways to do that. But, please don’t make it that much bold. In the worst condition, after compensation, we may have less than 20% change in the accuracy. In the other hand, it is a systematic error in the measurement; if you always do the measurement on a bed with 120C, you will get an approximately consistent result; so, it does not really matter. But if you plan always to play with the bed temperature during the bed leveling, that’s a different deal. You do not also have to do the measurement when the extruder and bed are hot if you are that much concerned.
The parameters that you mentioned, directly affect the transit time based ultrasonic sensors. We are not doing a transit time measurement. For Phase/Doppler measurement, there is a different story. Our measurement is not as simple as sending a pulse and measuring the time that it returns. There are many techniques employed by many researchers during past decades to minimize the effect of sound speed on the Doppler/Phase measurements. A simple search in google or IEEEXplore website will show you many practiced techniques including multi-frequency approach, T&R orientation, …
Any measurement instrument is sensitive to the temperature. We never claimed that we have an ideal sensor that solves all the problems for everybody on any condition. What are the sensors that you can use for auto bed leveling today? Please let me know the available choices. Which of them are not dependent to the temperature? Is an inductive proximity sensor or IR sensor kept several millimeters away from the hot bed not affected by its temp? You can keep alignG 10cm away from the bed and do the measurement.

“You also have a small variation in the velocity depending on frequency so any frequency sweep method would need more compensation tables. If you use eg. 40 kHz then you have lambda of around 8mm. No way on earth can you distinguish how many wavelengths you are away without moving the bed up and down significant amounts and without knowing if the head will crash how can you know if you are one wavelength away or more.
I am watching with interest to see how you overcome some of these (and significant other) problems. I looked at doing this a couple of years ago as I have done significant design work in ultrasonic sensing, I abandoned the idea because the further in you go the more problems arise.”
Does the sound velocity depend on its frequency?
No way on earth? Seriously?