PID problems

Hello, I'm kind of a nubee (at this) I was a Allen Bradley certified systems integrator and the only person for service for the Bandit series CNC's in North West Ohio for 13 years. These little machines seem to have the same kinds of problems as the big ones..or is that the other way arround...... Anyways I have a home designed system very little pre-made(kit parts) Made with IKO rail ways, ballscrew Z feed. 8020 frame 10x10" table with X254mm x Y245 x Z265mm print area. Arduino Mega control, Erik-Marlin - V1.0.0.RC2-134-gb7fc14b firmware, hand wired "RAMPS" basically Ramps with my own pinout. Pololu drives. Latest Pronterface and Slic3r ( as of this week).Hot end is a MakerGear 50mmx6mm tube and a .35 BigHead nozzle.
Home design heat core, Brass bobbin with 5.55 ohms of #31 nichrome, a thin coat of Cotronics 989FS cement then wrapped with Cotronics ceramic tape. The thermistor is fit into a small piece of brass tubing soldered to the bobbin at the head end for the best thermal contact. OK so now the latest problem (other than basic print parameters) is the PID loop. If set for Bang-Bang the temperature oscillates around the set point (240 C) about +/- 6 C like expected. When set to PID the temp. rises to 239 C then drops to 224 C then back to 239 C to start all over again. No mater what I set the PID values to always the same result and range 224 C to 239 C even the same cycle rate.Yes I was watching out for that sneaky EEPROM it got me before. PID loops should not be that Honey Badger. A change from I=0 to I=10 should make a big difference. P500, I10, D455 should be very different than P5, I10,D455 or P100, I0,D0. Usualy when I tune a thermal PID I set I and D to "0" , set the gain till the temp gets close to set point , increase the I to get compliance then make big step changes and adjust D for 1 cycle of overshoot of about 5% or so.
Is there something I'm missing. I only found the one #Define for PID turn on. Also turning on the PID diagnostics only results in a no compile it did NOT like the forced byte to int conversion on the echo line. Found out about an hour ago that the Autotune requires a cold start to work, which would explain why no response when I tried to use it, will try tonight. Went and got the Rev3 candidate of Marlin it wouldn't compile at all with audrino 1.0.
A lot said in a little space so if misunderstood just ask, I don't insult very easy.
Would appreciate any thoughts.
Thanx
Todd

P.S. I actually think you could print on thin air if you could wipe it down first with ABS/Acetone monkey spooge. Had to use a small hammer to remove the last print from my bare Magnesium bed (printed at 90C - 75 C bed with no warping or pull away. I still cant get the one before off the Kapton tape. Pulled it right off the bed, WON'T come off the tape.

*** turn off verbose ***

Is the smoothing factor set to 1? It needs to be .95 or less, otherwise the PID never behaves like it should.

---

www.Fablicator.com

Yes set as defaulted at .95. Ran autotune last night from a cold start. It worked quite well. My hats off to the boys who wrote that software. The temperature controllers and the E-beam evaporator controller I use at work don't self tune that well! Once tuned it held temperature like a champ. I still don't understand why it was so unresponsive the first time. The values it arrived at weren't very far from where I started at, so some of the numbers I used should have made the loop radically unstable or so dampened that changes would have been deathly slow. It works now I guess I will leave some mysteries alone. Thank you. Now it looks like my biggest problem left is a filament feed problem. When I print something of a specific size every dimension is over sized (holes under) like the walls had a thickness of the nozzle added to the outside surface so externals are bigger by 2x nozzle and internals 2x nozzle undersized. also solid fills have a top surface that is 'fuzzy' not as if spider webs are being drug over the surface. more like pencil eraser sand. These harden quickly enough to deflect the head or even stall the drive. I have some pretty cool fails. one shifted so perfectly that it looks like it printed on thin air. At this time I think that I'm over extruding just a little too much so I will be tuning it down a little. I also find a little skipping after a filament retract. I would have thought after a retract the feeder would advance the same amount as the retract to refill the chamber to the same point as when it left off. Was planning to use the 'Extra Length On Restart' to compensate for it. It would probably only take a small re-feed. but not quite what it retracted. Will experiment with it. The biggest problem I have is the lack of documentation. What there is out there is totally net based, no down-loadable file. I have no internet at home so it leads to a lot of guessing, fun, but not productive.
Thanx again, I really do appreciate the response.

Todd