RAMBO v1.2g, E3Dv6 hotend, TEMP_SENSOR_0=5... still wrong temperature

[update: argh sorry this should all say "RAMBO" not RUMBA!]

hi some advice appreciated: i'm doing a corexy folding printer, and using a RAMBO v1.2g with an E3Dv6 hotend. the configuration page for E3Dv6 says use TEMP_SENSOR_0=5, which i'm doing... and the temperature reads 40 degrees when it should be 25. with the thermistor disconnected the temperature reads 37.5 or thereabouts. also, the NTC EPCOS 100k printbed resistor (TEMP_SENSOR_BED=1) is the same deal: 37.5 ish when disconnected and a 40 degree reading when it should be 25.

this is with the following commit from https :// github.com / MarlinFirmware / Marlin.git

commit 122bdd187f9df02a6d9b59495630e40da307e82c
Merge: 9045ea3 f394c06
Author: Scott Lahteine 
Date:   Fri Mar 25 01:15:38 2016 -0700

    Merge branch 'RCBugFix' into RC

anyone any clues?

many thanks.

Edited 1 time(s). Last edit at 03/30/2016 12:22PM by lkcl.

i've just compiled and uploaded latest stable marlin firmware (1.0.x) directly from the marlinfirmware repository and exactly the same thing is happening: both sensors are reading 40C when both resistors measure approx 100kOhms and the actual temperature in the room is around 23C.

i have a mendel90 as well with pretty much exactly the same firmware, running on a Melzi, and that's reporting the correct temperatures. any clues where to begin investigating greatly appreciated.

ahhh... according to this post: [www.reddit.com] wrong readings can be caused by wrong pull-up resistors. so i measured the resistors R31 and R32 shown here [reprap.org] and sure enough, they're both measuring 4.0k not 4.7k. i haven't measured the others yet, the board's still installed so is very awkward to get to. i happen to have some 4k7 0402 resistors so i will see what happens after replacing them, but.... argh...

ok so i removed one of the resistors and measured it stand-alone: it's reading 4.7k which is the *correct* value. which is very strange. next possible investigation, measure VCC and also the voltage across the terminals without the thermistor installed, see what that reads...

update: VCC 4.98V, thermistor terminal to VCC: 0.19V.

Edited 1 time(s). Last edit at 03/30/2016 10:01AM by lkcl.

ok so a little more investigation, the voltage reading from TP20 is 4.59V and from TP22 is 4.58V at room temperature (appx 22 to 25C). the Semitech 104-GT2 thermistor for the E3Dv6 hotend is reading 116k, the NTC EPCOS B57621C104J62 is reading 106k - according to a datasheet for these thermistors if they're reading 40C they should be around 52k.... and they're not.

any ideas on what might be wrong greatly appreciated. if anyone has a Rumba V1.2, a reading of the voltages from the test pads TP20 and TP22 at room temperature would be greatly appreciated.

okaaay, so to solve the problem with RAMBOv1.2g boards not working properly i've bypassed the two voltage divider circuits for heater and hotend, and created equivalents from components that i had immediately to hand:



the dog's dinner on a 3x2 header with four resistors and two capacitors is a replacement pair of voltage dividers plugged into X-40. i've modified pins.h so that it uses ADC3 and ADC4, and i'm getting temperature readings of 20.3 and 20.4C as a result.

what we can conclude from this - especially given that all four thermistor circuits are wrong - is that the most likely fault on the RAMBOv1.2g board is the OVREF circuit. however as i have a solution - messy or not - i'm calling "time" on this one as i have so much else to get done. if anyone else encounters incorrect temperature readings with a RAMBOv1.2g please do speak up.