Duet 0.8.5 with E3D PT100 sensor+amplifier ?

Looking at upgrading my arduino mega + ramps 1.4 to a Duet 0.8.5

I am currently using a single E3D PT100B sensor and E3D PT100 Amplifier board.

E3D Documentation
[wiki.e3d-online.com]

I want to use the Duet 0.8.5 for dual extrusion and would then need to use 2 of the PT100B sensors.

Can the sensors be used with Duet 0.8.5? I simply just can't make sense of the documentation to be sure.

I have mailed a bit with think3dprint3d but that didn't enlighten me really.

I currently use a dedicated AC to DC 5v psu to provide stable 5v, but it would be nice if the 5v could come from the Duet 0.8.5 instead. Would that be possible?

I do not have the knowledge to generate my own temperature tables or any other such custom Projects (I can configure marlin firmware just fine though)

I looked into this a couple of weeks ago.

First problem is that the E3D docs say nothing about whether it can be run from 3.3V and still give accurate results. Having looked at the schematic, I think it should be OK; but it would be nice to have it confirmed. If you run the board from 5V, then the ADC in the Duet will no longer be working ratiometrically and accuracy will suffer. Also an amplifier powered from 5V might produce more than 3.3V output e.g. when the PT100 is disconnected, and that is likely to damage the ADC. This applies to other 32-bit electronics boards too. They need to qualify the board for use with 3.3V electronics to keep up with the times.

Second problem is that the thermistor inputs on the Duet have the usual thermistor series resistor to +3.3V. Depending on the tolerance of the output series resistor of the PT100 amplifier board (which is not specified in the published schematic) and the output drive capability of the instrumentation amplifier, it may be possible to leave this in place and account for its effect in the firmware. Otherwise, either the series resistor must be removed, or another analog input pin used to connect the PT100 amplifier. Unfortunately there is only 1 uncommitted analog input when the DueX4 expansion board is also being used, although that can be increased to 2 if support for some types of Z probe is disabled. So you could do dual extrusion with PT100 sensors, but no more than that.

I bought a MAX31855 thermocouple board several months ago with the intention of adding thermocouple support to RepRapFirmware. In the end Dan Newman beat me to it, and his thermocouple support will be in my fork soon. The thermocouple board is nicer to work with because it has a digital interface to the Duet. This means that the only source of errors you get will be the thermocouple itself and the MAX31855 chip, which has a quoted accuracy of +/-2C. I believe Dan's firmware supports up to 4 of them. With a PT100, sources of error include the resistance of the leads and the connections to the PT100 (unless you use a 4-wire PT100, but I don't think the amplifier supports those), offset voltages in the amplifier, tolerances in the amplifier components, ground noise in the connection to the Duet or other electronics, and ADC offset and linearity errors in the microcontroller on the Duet, which are not insignificant.

However, the PT100 is easier to mount than a thermocouple because it comes in a small metal can. Thermocouples usually come either bare (so they need to be insulated) or in probes that are much too large to fit a 3D printer head. If someone gives me a PT100 and amplifier, I may find time to add support for it in the firmware. Or I might design an alternative amplifier board that carries not only the instrumentation amplifier, but also a small microcontroller that generates a MAX31855-compatible digital output, to feed back to the Duet or other electronics.

Edited 5 time(s). Last edit at 10/27/2015 04:48PM 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].

Hi,

I used one of these digital breakout boards for another project I needed PT100 breakout capability for:

[www.playingwithfusion.com]

It works with 4 wire RTD sensors, which improves accuracy quite a bit.... this board will also work with 2 and 3 wire RTDs which I also did for comparison (also compared to a 4 wire RTD temperature standard sensor...). The 4 wire RTD sensors were the most accurate sensors I tested, but they took longer to respond to temperature changes than small Thermistors and small Thermocouples which can affect their function in quickly changing temperature environments.

Quote
dc42
I looked into this a couple of weeks ago.

First problem is that the E3D docs say nothing about whether it can be run from 3.3V and still give accurate results.

The output table is strictly for 5v. It changes dramatically if there are variances in voltage input. This is the main reason I run it from a dedicated small 5v psu instead of from AUX power on RAMPS, as it wasn't really stable.
I have looked at the table and the temperatures needs to hit 500c to output 3v and 600c to output 3.33c - Means I should be safe in regards to the board outputting more than 3,3v
Output Characteristics:
[wiki.e3d-online.com]
Can Duet 0.8.5 provide stable 5v?

Quote
dc42
Second problem is that the thermistor inputs on the Duet have the usual thermistor series resistor to +3.3V

All this is over my head. I should not begin doing stuff like that I am not going to be using the expansion board.
The PT100B sensors from E3D are either with 2 wires OR 3 wires, where 2 of them are put together. - My sensor is not from E3D and I havn't been able to make it out for sure on the E3D part.
My sensor is a 3 wire version, where I have put 2 of them together (red wires both).

Quote
dc42
I bought a MAX31855 thermocouple board several months ago .

Does this mean I can use PT100B sensor with the Duet 0.8.5 board if I buy a MAX31855 thermocouple board? You said lot of Things that I really didn't grasp

Quote
dc42
However, the PT100 is easier to mount than a thermocouple because it comes in a small metal can.

Yes. It should really have the exact specifications as the E3D sensors, which is incidentailly (or not so much) the same as Ultimaker 2 PT100B sensors. I'm currently Building an UM2 clone, including the hotend using the 3mm PT100B sensors and 4mm cartridges.
The E3D PT100B hotends are also using these sizes, so I'd rather just stick with that. Especially since their sensors really aren't more expensive than Chinese bought stuff. (and I'd rather buy from E3D)
I know I can use some bead-sensors, but I'd really rather use the PT100B due to accuracy and mounting issues.

Quote
patrickrio
Hi,

I used one of these digital breakout boards for another project I needed PT100 breakout capability for:

[www.playingwithfusion.com]

It works with 4 wire RTD sensors, which improves accuracy quite a bit.... this board will also work with 2 and 3 wire RTDs which I also did for comparison (also compared to a 4 wire RTD temperature standard sensor...). The 4 wire RTD sensors were the most accurate sensors I tested, but they took longer to respond to temperature changes than small Thermistors and small Thermocouples which can affect their function in quickly changing temperature environments.

How do you power it and does it Work fine with Duet 0.8.5 without massive firmware modifications? I really need something "plug-n-play" - firmware setup is going to be needed of course, but I can not do a lot of modifications like add extra support for hardware and such.

Edited 1 time(s). Last edit at 10/28/2015 02:00AM by dintid.

Quote
patrickrio
Hi,

I used one of these digital breakout boards for another project I needed PT100 breakout capability for:

[www.playingwithfusion.com]

It works with 4 wire RTD sensors, which improves accuracy quite a bit.... this board will also work with 2 and 3 wire RTDs which I also did for comparison (also compared to a 4 wire RTD temperature standard sensor...). The 4 wire RTD sensors were the most accurate sensors I tested, but they took longer to respond to temperature changes than small Thermistors and small Thermocouples which can affect their function in quickly changing temperature environments.

Thanks, I wasn't aware of the MAX31865. That would be an excellent option for use with the Duet. That breakout board is relatively expensive, but the chip itself isn't. So I may manufacture a lower cost board that doesn't include the voltage regulator or the level shifters, because they are not needed when you are using 3.3V electronics.

Quote
dinted
The output table is strictly for 5v. It changes dramatically if there are variances in voltage input. This is the main reason I run it from a dedicated small 5v psu instead of from AUX power on RAMPS, as it wasn't really stable.
I have looked at the table and the temperatures needs to hit 500c to output 3v and 600c to output 3.33c - Means I should be safe in regards to the board outputting more than 3,3v
Output Characteristics:
[wiki.e3d-online.com]
Can Duet 0.8.5 provide stable 5v?

Yes, but that's irrelevant. Barring inaccuracies in the amplifier, the output voltage will scale with supply voltage (i.e. it's ratiometric). The important thing is to make the supply voltage the same as the ADC reference voltage, so that it doesn't matter exactly what the supply voltage is (because the ADC is also ratiometric). With 5v electronics, this means running the amplifer from 5V. With 3.3V electronics, it means running the amplifier from 3.3V. This assumes that the ADC reference voltage on the electronics is the supply voltage, which is practice is how 3D printer electronics are wired.

---

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
dc42

Quote
dinted
The output table is strictly for 5v. It changes dramatically if there are variances in voltage input. This is the main reason I run it from a dedicated small 5v psu instead of from AUX power on RAMPS, as it wasn't really stable.
I have looked at the table and the temperatures needs to hit 500c to output 3v and 600c to output 3.33c - Means I should be safe in regards to the board outputting more than 3,3v
Output Characteristics:
[wiki.e3d-online.com]
Can Duet 0.8.5 provide stable 5v?

Yes, but that's irrelevant. Barring inaccuracies in the amplifier, the output voltage will scale with supply voltage (i.e. it's ratiometric). The important thing is to make the supply voltage the same as the ADC reference voltage, so that it doesn't matter exactly what the supply voltage is (because the ADC is also ratiometric). With 5v electronics, this means running the amplifer from 5V. With 3.3V electronics, it means running the amplifier from 3.3V. This assumes that the ADC reference voltage on the electronics is the supply voltage, which is practice is how 3D printer electronics are wired.

I really don't understand any of that. Maybe you can tell it in plain speak? I don't know what ADC, ratiometric or any other of those terms means or what it pertains to.
I really need the laymans way with a yes or no and then explains what it takes - non-technically.

I do know that the output table needs to be modified if I provide 3.3v instead of 5v and I can't do that, which makes stable 5v input important.

I just need to know if Duet 0.8.5 can Work with the E3D amplifier without me having to come up with a brand new table.

In short: can it Work just like normal marlin firmware and if it can, must the 5v come from external source or can the Duet 0.8.5 provide it?

edits: trying to clarify

Edited 5 time(s). Last edit at 10/28/2015 05:12AM by dintid.

RepRapFirmware does not use thermistor tables, you tell it (via the config.g file) the R25 and B values of the thermistor you are using. A PT100 thermistor + amplifier does not produce an output like a thermistor - in particular, the amplifier output voltage increases with temperature, where as the voltage across a thermistor decreases. So the firmware would need code changes to support a PT100, regardless of what voltage you use to power the amplifier. Supporting a PT100 using the MAX31865 breakout board might be easier because it connects in the same way as the MAX31855 thermocouple board, but it would still need a firmware modification.

Edited 1 time(s). Last edit at 10/28/2015 08:06AM 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].

Quote
dc42
RepRapFirmware does not use thermistor tables, you tell it (via the config.g file) the R25 and B values of the thermistor you are using. A PT100 thermistor + amplifier does not produce an output like a thermistor - in particular, the amplifier output voltage increases with temperature, where as the voltage across a thermistor decreases. So the firmware would need code changes to support a PT100, regardless of what voltage you use to power the amplifier. Supporting a PT100 using the MAX31865 breakout board might be easier because it connects in the same way as the MAX31855 thermocouple board, but it would still need a firmware modification.

You totally do not understand my question as you overcomplicate everything (which I then do not folllow).

PT100B Works 100% in normal Marlin firmware for my Arduino + ramps 1.4. I really just wanted to know if it worked on Marlin for Duet. Guess it does not?

I do not understand what you are trying to tell me. Maybe you just say that it does not Work on duet?

Edited 3 time(s). Last edit at 10/28/2015 08:26AM by dintid.

Marlin has not been ported to the Duet. The Duet runs RepRapFirmware instead.

---

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
dc42
Marlin has not been ported to the Duet. The Duet runs RepRapFirmware instead.

ohh, I thought the firmware for Duet was your Marlin build.

It would be really nice if there was some documenation as to what sensors could be used and how - step-by-step.

A why would be nice, but the how is the important part. Often "guides" are made up of "whys" and then assume people can figure out the how based on the techincal whys. Which just leaves many people stranded. Myself included - or at least slowed severely Down.

I agree, the Duet needs better documentation. I have made a start by documenting the configuration files on the RepRapWiki, and listing known issues of the Duet 0.6 (I was tempted to add a known issues list to the RAMPS wiki entry too but I wasn't sure how long it would last).

Other people are having conversations about setting up a dedicated Wiki for the Duet that would cover the hardware too, and I hope this happens soon.

Edited 1 time(s). Last edit at 10/28/2015 12:43PM 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].

When I said I was using that breakout board for another project, I REALLY meant other project.

I used the PT100 RTD and the MAX38165 in an induction heated coffee roaster prototype.... Not a 3D printer so it was used with a Teensy and the standard Arduino PID Library and Tuner. Power was right from the appropriate pins on the Teensy which Is also available on the Duet... Should be really easy to wire it up, and not too hard to include the needed lines of code for it's operation. Most of the code you need is probably already in the links, so mostly cut and paste. For me it was purely cut and paste (which is about the only way I can program anyway as I am not skilled like dc42...) Roaster Prototype worked, but needs software refinement. Currently talking to a programmer so we will see how that goes.

DC42, by way of information, US built PT100 sensors and Europe built PT100 sensors have a slightly different platinum wire alloy, so they have a slight difference in resistance value. You will want to specify an exact sensor, or make sure that you can choose between the two settings. probably only matters if you use the better 4 wire sensors though.

Edited 2 time(s). Last edit at 10/28/2015 02:37PM by patrickrio.

Slightly off the OP topic and sorry if I missed this in another post, but since it seems that MAX31855 support might potentially already be available in a current version of the Dc42 RepRapFirmware - are there any instructions available on how to enable and test this on a Duet 0.85? I am more curious about how this would be enabled in config.g (or elsewhere) than on how to connect the MAX31855 to the Duet.

Quote
cglass
Slightly off the OP topic and sorry if I missed this in another post, but since it seems that MAX31855 support might potentially already be available in a current version of the Dc42 RepRapFirmware - are there any instructions available on how to enable and test this on a Duet 0.85? I am more curious about how this would be enabled in config.g (or elsewhere) than on how to connect the MAX31855 to the Duet.

MAX31855 support is implemented in Dan Newman's fork, available here [github.com]. it is based on my fork, and I will merge the MAX31855 support into my fork when I have finished the new PanelDue firmware. Dan and I discussed how to enable MAX31855 support in config.g, and I think he implemented my suggestion of using the X parameter in the M305 command, with channels 100 upwards being the MAX31855 channels (whereas channels 0-7 are the normal thermistor channels).

Edited 1 time(s). Last edit at 11/10/2015 03:55PM 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 look forward to your release.

After a brief look through dnewman's fork, it seems to mainly focus on radds - which (without any research) probably has a different pinout to the duet. In his release folder, he has a 'RepRapFirmware-1.09k-dc42.bin' - does this have the MAX31855 implemented? I suppose I might still need to recompile a new version using my preferred pins (unless there are already 5 recommended pins). This is a great starting point - Thanks!

After updating the firmware, I am still only able to read the thermistor values and not the MAX31855 values.

I suspect that my next step is to change the config.g file, or since I am using dnewman's fork it might need to be changed in config.h?

Any examples on exactly what to modify would be a nice help - after almost a week trying to get eclipse to actually produce out a valid bin file, I would really like to save a little time on getting this all finished and working!

If Dan followed my suggestion, then you will need to use the X parameter in the M305 P1 command in config.g to redirect the first hot end to use one of the thermocouple channels. The thermistor channels are X0 to X6 and the thermocouple channels are X100 to X103. The SPI pins to use on the expansion connector are the ones labelled MOSI0, MISO0, SPCK0. For the chip selects use NPCS0 for the first thermocouple and NPCS1 for the second. I think a further 2 are supported. Hopefully the Pins.h file for the Duet in Dan's fork has these set up already.

I'll be merging Dan's thermocouple support into my fork of RepRapFirmware soon.

Edited 1 time(s). Last edit at 11/17/2015 04:34AM 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].

Thank you! With everything working, I feel like Victor Frankenstein now!

I have 'M305 P1 X100' setup and the PanelDue is displaying the correct thermocouple readings.

The problem now is that the heater is not heating up.

Any suggestions as to what could be wrong / what to look at next?

1. Can you confirm that you have defined a tool using heater 1 with the M305 command, and that you have it selected?

2. Does the LED next to the E0 heater output illuminate when it is supposed to be heating?

Edited 1 time(s). Last edit at 11/21/2015 01:36PM 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].

The LED is not illuminating. I have tried both 'M563 P1 D0 H1' and 'M563 P0 D0 H1' (in case) without success. I also have 'G10 P1 S0 R0 X0 Y0' / 'G10 P0 S0 R0 X0 Y0' below M563.

So are you selecting the tool with T0 (if you used M563 P0) or T1 (if you used M563 P1) before you try to set the temperature?

If you switch to the Console tab of the web interface, are there any error messages displayed there?

---

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 am selecting the tool before trying to set the temperature. No error messages displayed in the Console initially, but when I 'home all' axes, something strange is happening.

I have two MAX31855's breakouts connected and have tested them connected to 'SPI0_NPCS0 (aka PA28)' and 'SPI0_NPCS1 (aka PA28)' and then to variations of PA10 - PA13 as a sanity check. I am using two Type K 'unground' thermocouples (I have tested that they are not ground and properly insulated). Initially the readings are correct, but after homing the axes I get the following error:

'Temperature fault on heater 1, sensor is shorted to the voltage rail, T=2000.0'

This error happens every time I home the axes and does not change even when I swop the pins used / MAX31855's / thermocouples?

I suspect the problem might be related to: '#define PERIPHERAL_CHANNEL_CS_PIN 78 // NPCS0' in MAX31855.cpp - since '78'(aka X15, aka PB23) is not NPCS0 (aka PA28) - but I will not know if this makes a difference until later this week when I get a chance to recompile the firmware again to test it.

Hopefully someone might have another potential cause / solution. I am pretty much out of ideas after this!

I think that's an issue for Dan to look at until I merge in the thermocouple code, which will hopefully be a day or two from now.

---

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 will try to look at this later today (my time; California). There may be a startup situation in which initially there is a fault reported and the heater disabled but the thermocouple then is operating correctly and reporting correct temperatures. I'll look more closely at this this evening. You may be able to clear the fault with

M562 Pnn ; nn is heater index

Again, I'll look at this this evening. The issue, I suspect, lies with the MAX31855's quirk of not reading reliably the first few reads and also when power is still stabilizing. The code presently is aware that the first read is sometimes all 1-bits and knows to ignore that. However, it may be necessary to just toss entirely the first several reads, especially if power isn't fully stable yet.

I played some with this on my Duet 0.6 CoreXY bot and cannot reproduce what you are seeing. (Majority of my testing has been on my RADDS + PanelDue + RRF + MAX31855s CoreXY bot.) I did make the code even more pedantic about detecting bad reads and declaring them an I/O error (temporary error) rather than, possibly, a permanent error (reacted too immediately by disabling the heater and declaring a fault). I also discovered a bug whereby a short to vcc was reported as a short to ground and vice versa. So if the short to vcc error you saw was from a valid data read from the MAX38155, then it actually was a short to ground and not a short to vcc. I've checked my changes in to github. You can pull and try them. Unfortunately, I again have to run. Sorry for the terse message.

Edited 1 time(s). Last edit at 11/23/2015 02:25AM by dnewman.

DC42, I have a duet 0.85 and some of the E3D sensors in question. I'll see about running them at 3.3v within a few days and report back.

Quote
dc42
If Dan followed my suggestion, then you will need to use the X parameter in the M305 P1 command in config.g to redirect the first hot end to use one of the thermocouple channels. The thermistor channels are X0 to X6 and the thermocouple channels are X100 to X103. The SPI pins to use on the expansion connector are the ones labelled MOSI0, MISO0, SPCK0. For the chip selects use NPCS0 for the first thermocouple and NPCS1 for the second. I think a further 2 are supported. Hopefully the Pins.h file for the Duet in Dan's fork has these set up already.

I'll be merging Dan's thermocouple support into my fork of RepRapFirmware soon.

Yes, it's X100 to X103 as per dc42's suggestion.

MISO0 tied to the MAX31855's data out
SPCK0 tied to the MAX31855's clock
No need to use MOSI; data is not transmitted from the Duet to the MAX31855. The Duet only reads data from the MAX31855.

The chip selects were based upon what was spare on the Duet expansion header. Since NPCS1 and 2 were already used (ZDIR, ZPROBE_MOD_PIN), and possibly even CS3, I instead used a block of 4 adjacent and, at the time, unused pins on the expansion header: RXD1, RXD2, TXD1, TXD0 (pins 11 - 14 on the 0.8.5 header). See

http://reprap.org/wiki/Duex4_pinout

Those assignments for the chip select pins are not yet firm; they are subject to change until such time that final choices are made when the work is rolled up into the dc42 repo. There's a good argument for using NPCS0 as it is, I believe available. Namely, if most people will only use a single thermocouple, then it may seem "natural" to use it. If people use more than one thermocouple, then that other block might be argued to make more sense. I personally was concerned that if I used NPCS0, then someone with two thermocouples may fail to carefully read the documentation and attempt to use NPCS1 as well as NPCS0. They'd then have strange interactions with, I believe, their Z stepper driver.

Running the E3d board at 3.29v I measured 0.80v (multimeter, no pull-up) on the output in a ~22C room. Multiplying the voltage offset and interpolating with E3D's wiki table, I get 23.75C. Not bad! Unfortunately I doubt this relationship is linear, so I'll be hooking this up to a stable 5v supply and plan on keeping my heaters below 600C!

Quote
Crispy001
Unfortunately I doubt this relationship is linear, so I'll be hooking this up to a stable 5v supply and plan on keeping my heaters below 600C!

Bad idea. If you do that, then the reading you get will vary with the ratio between the actual voltage on the 5V supply and the actual voltage on the 3.3V supply. If you power the amplifier from 3.3V instead, then it will not be sensitive to the exact supply voltage, because both the amplifier and the ADC are ratiometric and will be using the 3.3V supply as the reference.

---

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].