Added support for MCP3008 and MCP3004 ADC for temperature read #205
Conversation
temp.c
Outdated
| do { | ||
| uint8_t j, table_num; | ||
| //Read current temperature | ||
| temp = analog_read(i); |
There was a problem hiding this comment.
This was a bug. Should be temp = analog_read(temp_sensors[i].temp_pin) because we want to read ADC pin of i-th sensor rather then ADC pin i. Fixed.
There was a problem hiding this comment.
Analog read is index based. So we don't read the temp_pin. We read the ADC of index i, which is configured in analog_xxx.c.
There was a problem hiding this comment.
Yes. You're right. I read analog part and somehow I thought that it reads pin it gets in argument directly. Reverted to analog_read(i)
There was a problem hiding this comment.
And then there is this thread: http://forums.reprap.org/read.php?1,647767,648074#msg-648074 Is this the reason why the existing code was suspect to you?
Other than that: nice patch! I'll apply it before too long.
There was a problem hiding this comment.
I haven't read this thread till now. The reason why I need separate ADC is that I have 5 temperature sensors in a print bed and 4 heaters. Print bed is 40x40cm with four 20x20 heaters. One sensor sits in the middle of each heater and the fifth one is in the middle of a test bed. You may see it in the video I tweeted some time ago: https://twitter.com/WitekSowa/status/713456028860473344
This number of sensors and heaters plus LCD, jog dial, SD and at least 5 endstops (max in Z axis, min/max in X/Y) requires much more pins than ATMega have so I run ADC/SD via SPI and LCD/Jog dial via I2C port multiplexer (MCP23017).
|
Good. Applied and squashed both commits to experimental with some whitespace editing and rearrangements (I'm a nitpicker :-) ). Regression tests pass. Please test wether the code on experimental works for you, then we can close this pull request. |
|
Will check in 2-3 weeks. I'm. On holiday now |
|
I have some issues to read that code. The MCP is an ADC over SPI. So you have a CS, which you can choose in the configtool. This is all ok. But for the channel you can only select some normal ADC-pins from the MCU not the channels from the MCP?!? |
|
@Traumflug I've checked experimental branch. MCP3008 support works fine but I got into problems with pre-computed slope in thermistor table. See more details here: #208 @Wurstnase Yes, I reuse AIOx names for denoting ADC channels. AIOx translates just to x so it works fine. And I find it more user friendly: AIOx read as analog input x may refer to MCU channel or to ADC channel. There's no need to name them differently if user selects a sensor type as well. |
I'm building custom 3d printer with custom electronics. Machine is based on ultimaker design while electronics on Gen7. I use MCP3008 8-channel 10 bit SPI ADC for temperature read instead of ADC build in AVR (I need to save pins). With this pull request I add support for this ADC to teacup. I hope someone else will make use of it.
Most changes are related to adding support for MCP3008 to python config tool rather than to actual teacup code so I suggest to start reviewing with C code and then go to python and configs