A Steps-per-Unit Change Mystery

Hello all. Something is happening with my Reprap, and I can't explain it. Can you?

The specs: Mendel i2 with Ramps 1.4 running Marlin.

The background: I used Slic3r to get gcode that produced this cool gear bearing. (Thing 53451.) Slic3r produced gcode that delivered a high quality print in about 125 minutes.

Then, I replaced my printed pulleys and T5 belts with machined pulleys and GT2 belts. I updated my DEFAULT_AXIS_STEPS_PER_UNIT for x and y from 64 to 80.192. Using *exactly the same gcode* that was used above, I'm printing another gear bearing, but the print is going just ok, at best. The small perimeters (i.e. any gear teeth) appear to be printing as good as ever. The large circles around the outside, however, look a little sparse, as if there wasn't enough extrusion happening. I also have a gut feeling that these circles are going more quickly than they used to. That suspicion is buttressed by pronterface's shown time remaining. It's going to come in somewhat under 2:05. The size of the print is identical to the one I already have, so I think my steps per unit value is correct.

The mystery (to me, anyway): why should the steps per unit change have any impact whatsoever on the speed of the print, which was limited by various mm/sec and mm/sec^2 values in Slic3r and Marlin? Shouldn't that change be perfectly offset by my new pulley size?

Any advice or clues welcome.

Pics to follow.

Edited 1 time(s). Last edit at 05/19/2013 10:26PM by Karmavore.

Just thought id pitch in with my 2 cents

Based on micro stepping we see steppers doing 200 steps per revolution (1.8 degrees) divided by 16 (micro steps) 0.1125 of a degree.
The issue i've come across is with your value of 80.192 steps. More to the point how your firmware and hardware deal with an input of 0.192 of a step. I can seem to find a definitive answer on how firmware/hardware deals with this. mechanically its impossible or appears to be. a micro step is a micro step. the hard/soft ware cant do a fraction of a step.
There are quite a few discussions on this forum, mostly relating to e steps and layer heights

I suppose the easiest way to try this would be by setting your X & Y steps to 82 and see if there's any improvement/ difference.
Please post back your findings if you do

Good luck

First, Marlin stores steps/mm in EEPROM, too, so your new settings might be ignored, even after re-uploading the firmware.

Then, are you running a very high steps/mm on the extruder? Marlin on a 16 MHz electronics is limited to about 12'000 steps/second, AFAIK, and exceeding that at high speeds isn't difficult.

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY

Thanks, bigfilsing and Traumflug. I'm confident that the new settings are being used, fractions and all, because the size of my post-GT2 parts are exactly the same as the pre-GT2 parts. If not, I'd expect that the part would be larger in the x and y directions by a factor of 80.192/64.

My E steps/mm is exactly (3200*(47/9))/(3.1415926535*6.86), which is about 775. I would need to try to extrude 15 mm of filament per second to get into 12,000 step territory, assuming I'm doing that math right. I don't think I'm extruding that fast.

In fact, I fear I may have sent you both on a wild goose chase. The second print finished in 123+ minutes, which isn't all that different. And later layers on the part did look better. I wonder if I was the victim of a segment of filament with slightly lower diameter or a higher melting point. Increasing the temperature of the hotend helped my next print.

Another suspect I have is the trimpot on E. I did futz with it before I discovered that I partially lost connection on wires to the E stepper.

Again, thanks for the help. Even when I post (it seems) faulty questions, I still learn something from the answers.

OK, now I *really* know what was wrong. My hinge on the Greg's Wade Reloaded just failed. I can only assume that it was on the verge of failure when I started noticing that I wasn't getting the extrusion I wanted. The hinge wasn't holding tight against the hobbed bolt, and the consequence was visible in my print.

It's nice to know that I'm not going crazy, even if I did spend a long time barking up the wrong tree. And: it turns out this was a pretty good mystery, even if it had a twist at the end that really couldn't be predicted.

Making good on my pics promise:

Glad to see it works now

---

Generation 7 Electronics	Teacup Firmware	RepRap DIY