my PID Autotuning Figures are wildly erratic

[analog_banana]

I've been trying to get my Rostock Max to print for 8 months,
One of the things I have noticed is that when I perform the PID autotune process...
the figures I'm getting back are wildly different.

Here's an example of 2 tests I ran, both starting from room temp going up to 215 degrees.

=TEST 1 at 21:12pm====
-------------------------
21:12:49.279 : Clasic PID
21:12:49.279 : Kp: 69.67
21:12:49.279 : Ki: 12.08
21:12:49.279 : Kd: 100.45

21:13:11.350 : Clasic PID
21:13:11.350 : Kp: 35.71
21:13:11.350 : Ki: 3.24
21:13:11.350 : Kd: 98.30

21:13:23.056 : Clasic PID
21:13:23.056 : Kp: 56.84
21:13:23.060 : Ki: 9.74
21:13:23.060 : Kd: 82.88

=TEST 2 at 21:43pm====
-------------------------
21:43:17.170 : Clasic PID
21:43:17.170 : Kp: 39.81
21:43:17.170 : Ki: 6.68
21:43:17.170 : Kd: 59.35

21:43:31.228 : Clasic PID
21:43:31.228 : Kp: 30.67
21:43:31.231 : Ki: 4.37
21:43:31.231 : Kd: 53.77

21:43:50.946 : Clasic PID
21:43:50.946 : Kp: 47.50
21:43:50.946 : Ki: 4.83
21:43:50.950 : Kd: 116.73
-------------------------

In GeneB's manual the figures seem a lot more consistent,
Has anyone with experience seen this kind of thing?

Also Does anyone here know the correct settings for the "TEMP SENSOR"s in the Marlin Firmware.
I beleive the thermistors I am using are the standard ones supplied with Rostock Max Kits in March/April 2013

This is waht the firmware looks like to set the thermistor type:
#define TEMP_SENSOR_0 1
#define TEMP_SENSOR_1 0
#define TEMP_SENSOR_2 0
#define TEMP_SENSOR_BED 1

Thanks in advance for any guidance!

[analog_banana]

I talked to Mhackney, and these figures are within the bounds of normality so I'm dismissing PID tuning as being a problem with my machine.

[mhackney]

PID tuning is a black art. The parameters are dependent on a lot pif different factors and I don't think there is one correct "answer". The goal is to maintain a steady temperature without a lot of fluctuation and to reach that temperature relatively quickly. If you tune for an endpoint of 80°C you will get different settings than if you tune for an endpoint of 90°C. But the same settings will work for both (and quite well), they just won't be the "optimal" settings.

[analog_banana]

Great - I'm going to set it and forget it for the time being. Turns out my problem was more to do with filament choice and layer height.
I'm finally getting successful prints!
Examples on my build log here:
http://forum.seemecnc.com/viewtopic.php ... 800#p26800