Differential IR sensor and BuildTak

I am happy to report that BuildTak sent me some samples for testing with the differential IR sensor on my hot end board, and I have just done some tests. I ran tests on a single-nozzle board in my test rig, and on the dual-nozzle board on my Ormerod.

In both cases, the sensor worked on a BuildTak surface when it was operating in normal (differential) mode. That's the mode activated when you use M558 P1, or when the 4th wire to the sensor connector is not present. You can tell that the sensor is operating in this node because the sensor gives a reading of about 465 at a height slightly greater than the trigger height.

In simple modulated mode, the sensor does not work above BuildTak. This is not surprising, because BuildTak is very dark, and simple modulated mode really needs white targets just like the standard Ormerod 2 modulated sensor.

Just because both of my sensors worked above BuildTak doesn't mean that all sensors I have shipped will. Where the trigger height is higher than average (i.e. above 3mm over plain glass), it is possible that the signal level of the reflected IR may not meet the threshold in the firmware. When I make the new batch later this week, I will test them above BuildTak, and if necessary adjust the threshold in the firmware to get reliable operation with BuildTak as a target.

If anyone cannot get one of my differential IR sensors to work above BuildTak, then if you return the board to me I will test it and if necessary adjust the firmware. Please note that this applies only to the differential IR board, not to my earlier IR/ultrasonic board (but I would expect the ultrasonic mode to work with BuildTak anyway). I will charge you just the return postage.

Anyone wishing to order one of my differential IR sensor boards can find details at [forums.reprap.org].

Edited 1 time(s). Last edit at 03/22/2015 01:47PM by dc42.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I can confirm that my single nozzle board works fine with BuildTak. I have noticed though that the trigger hight might not be exactly the same with my kapton covered bed.

Best,
Achilles

Quote
Spaceskater
I can confirm that my single nozzle board works fine with BuildTak. I have noticed though that the trigger hight might not be exactly the same with my kapton covered bed.

Thanks for confirming this. The trigger height is somewhat affected by reflections from the underside of the glass, which is why the trigger height for transparent and semi-transparent surfaces such as glass and Kapton on glass is not quite the same as for an opaque surface such as BuildTak. A piece of thin black plastic or black paper placed between the glass and the aluminium heat spreader reduces the reflection.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

While I don't know what the case is with the BuildTak material, it is worth bearing in mind that colours that look dark or even jet-black under visible light may reflect IR well. I once had a monochrome camera that automatically switched to IR below a certain visible light level, and I recall seeing black clothing magically changing to white as the IR function cut in - it looked very strange because nothing else in the image changed appreciably. The camera could also see through black Perspex as if it were 100% clear. The Perspex was so opaque that you could look directly at the Sun through it quite comfortably, which brought home the point that using such ad-hoc material to view the Sun (or welding arcs) is dangerous as your retina would be hit with the full amount of the Sun's IR output without you realising it, which I should think would cause a lot of damage in a pretty short time.

Dave

Quote
dmould
While I don't know what the case is with the BuildTak material, it is worth bearing in mind that colours that look dark or even jet-black under visible light may reflect IR well.

That's a good point. However, in my tests I observed that the range over which the sensor detects the surface is quite small. In other words, when the sensor head is a little closer to the BuildTak than the trigger height, the trigger LED goes off again. That suggests to me that the IR signal from the far LED is below the threshold, which (along with the fact that at least one user can't get his sensor to work with BuildTak) makes me think the reflectivity is low.

When I've built the new batch of boards, I'll be testing them on BuildTak, and if I have any that don't work on that surface I will check the signals levels with a scope.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I guess I am the one user that could not get his sensor to work.

Well there is hope. My printer is at the moment in the shape of a dissected squid but I have the Duet up and running on the USB port and last night I did some test on sensing the BuildTak. By turning the Z-axis by hand I actually got the red LED on the senor board to lit up over the BuildTak.

What has happened is that the whole printer head has been disassembled (2-color upgrade) and during this I ever so slightly adjusted the angle of the IR diodes and the sensor. My theory is that due to the limited IR reflection of the BuildTak the angle of of the IR diodes and the IR sensor is even more crucial to get right compared to a glass bed.

My PSU died so until that is sorted out I cannot give a 'real life' confirmation that it is working. But as I said; there is hope!

If your PSU is dead, perhaps now would be a good time to return your hot end board to me for testing with BuildTak and reprogramming to increase the sensitivity?

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Quote
dmould
While I don't know what the case is with the BuildTak material, it is worth bearing in mind that colours that look dark or even jet-black under visible light may reflect IR well. I once had a monochrome camera that automatically switched to IR below a certain visible light level, and I recall seeing black clothing magically changing to white as the IR function cut in - it looked very strange because nothing else in the image changed appreciably. The camera could also see through black Perspex as if it were 100% clear. The Perspex was so opaque that you could look directly at the Sun through it quite comfortably, which brought home the point that using such ad-hoc material to view the Sun (or welding arcs) is dangerous as your retina would be hit with the full amount of the Sun's IR output without you realising it, which I should think would cause a lot of damage in a pretty short time.

Dave

Black objects are generally good emitters of IR as well as great absorbers, so you are not only seeing reflected IR, but emitted IR produced after absorbance of IR and other wavelengths. (as far as I know ;-) )

James

Quote
jamesf
Black objects are generally good emitters of IR as well as great absorbers, so you are not only seeing reflected IR, but emitted IR produced after absorbance of IR and other wavelengths. (as far as I know ;-) )

James

When the bed is hot, the dark surface is indeed likely to emit IR. My board uses a 3-way modulation scheme so that it only responds to the IR generated by the two on-board IR LEDs. In practice, reflected sunlight is a much bigger source of ambient IR at these wavelengths than emission by the bed itself.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Well I already have the new PSU on the bench it is just that life came in between. When I get everything together who knows if I fry something then I will have to order a new Z-Probe.

I have a feeling the Z-Probe will work just fine on BuildTak.... I hope...

A hot bed will indeed emit IR, but its wavelength is way below the wavelength detectable by the probe photodiodes (or a normal CCD camera). I haven't looked up the specs of those diodes, but they almost certainly detect radiation in the near IR region centred around 0.75um (400THz), and objects do not begin emitting IR at that frequency until they are almost red-hot (500 deg C at least). Thermal sensors (used in PIR sensors, IR thermometers and thermal imaging devices) operate in the far-infrared region, which has a wavelength about 10 times longer than that detectable by the probe diodes (centred around 7.5um or 40THz).

Dave

I have done some more detailed measurements on a single-nozzle board. In particular, I have measured the modulated IR signal level at the trigger height, for various surfaces. This is what I found:

Plain glass: 300mV
Kapton on glass: 400mV
BuildTak: 50mV

This confirms that BuildTak is less reflective than plain glass. The firmware has a threshold of 32mV, so if it sees a signal lower than that, it will not trigger. As this is quite close to the 50mV that I measured, I can well believe that some IR sensors will not trigger on BuildTak. I could reduce the firmware threshold, but accuracy may already suffer at low signal levels because the quantisation of the ADC is 3.2mV.

For the dual nozzle boards I will be making as soon as the PCBs arrive, I intend to double the value of the phototransistor load resistor, so as to double the sensitivity. The only disadvantage is that it will be easier to saturate the sensor with strong ambient IR. However, I have found it much harder to saturate my differential IR sensor than the original RepRappro sensor, in fact I have only seen it saturate once, and that was when there was bright sunlight reflecting directly off the spot directly below the sensor on the plain glass bed.

Tomas, if you have a fine-tipped soldering iron, then you could try modifying your board. Remove resistor R1 (value 2.7K) and replace it by 5.6K or higher (e.g. 10K) 0603 resistor.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

I now have tested to probe the BuildTak and it works!! I have even tested a 5-point bed compensation

With white paper I had contact at 1.0 mm. I removed the paper and lowered the nozzle until it got contact again at that was at Z 0.75. So as said in this thread the reflection is lower and subsequently the probe has to go lower to get a reading.

What has changed since I initially tried to use the Z-probe on BuildTak is that I have adjusted the angle of the IR diods and the sensor. I think they are more perpendicular to probe board now then before.

The only strange thing is that before adjusting the IR diods I had contact 2.7 mm above the bed using white paper. After adjusting it is at 1.0 mm. I seems to work fine so I will not bend the legs of the diods any ore then necessary.

I'm glad you got it working. The normal trigger height is with the edge of the sensor board between 2.5 and 3.5mm above the surface. The corresponding nozzle height will depend on how high the edge of the sensor board is relative to the nozzle.

---

Large delta printer [miscsolutions.wordpress.com], E3D tool changer, Robotdigg SCARA printer, Crane Quad and Ormerod

Disclosure: I design Duet electronics and work on RepRapFirmware, [duet3d.com].

Oh, of course. Sometimes I am a bit to thick, more than alowed.

It is nice to have it working!