Blobbing due to small pauses. [help me]

I recently put together a Mendel Isaac, using Gen6 electronics and an Arcol hot end. I absolutely love the setup except for the fact that anywhere a perimeter starts i get small blobs.

Using jitter in skeinforge makes it very easy to tell that regardless of my retract settings the blobs are staying about the same size. The retract settings i have work fantastic to completely eliminate any oozing during a move between parts. The ooze ONLY shows up at the beginning of a perimeter line. The only thing i can see that might be causing this is that between when the extruder moves back after a retract there is about 1/4 to 1/2 a second between that and when it starts moving the print head around to draw the perimeter line. during that time is when the blobs show up.

I'm printing with green PLA from ultimachine. I've tried every temperature from 170 to 200 and had no changes to the oozing with temperature changes

I also took the advice of someone over on the bfb forums and setup replace.csv to take out the extra commands that aren't needed by a stepper extruder, hoping that would help, but still nothing.

I'm really pulling my hair out trying to figure this out, any help would be greatly appreciated

I print 4043D PLA at 210 degrees C with excellent results.

---

Bob Morrison
Wörth am Rhein, Germany
"Luke, use the source!"
BLOG - PHOTOS - Thingiverse

Quote

The only thing i can see that might be causing this is that between when the extruder moves back after a retract there is about 1/4 to 1/2 a second between that and when it starts moving the print head around to draw the perimeter line. during that time is when the blobs show up.

Well that is sure to cause a blob. It should move off within a few milliseconds of fast forwarding the extruder.

---

[www.hydraraptor.blogspot.com]

i'm honestly not entirely sure how long the pause is, but it is under 1 second, but still very noticeable.

I'm really getting immensely frustrated trying to figure this out. Could it be some setting in repsnapper that I'm missing maybe?

Everywhere a perimeter starts? Or simply at the start of each layer? A slow Z move could certainly result in a blob per layer, and would take a half second or so.

it could be the z speed.. honestly I have no idea how to set the z speed in skeinforge. And I haven't noticed any blobs in the infill, so it could just be happening on the outer perimeter

at one point I did try doing the infill first to try to get the blobs to stay on the interior, but that did not help at all. That's why I thought that it was happening at the start of each line it's extruding.
Also I believe that when you are using retractions on the extruder it doesn't fast forward till after the Z move. So that would eliminate the ooze potential from moving up a layer.

I get a blob every time a thread starts. I am very interested in finding a solution as well.

Could it be due to the fact there is no accelleration? The head does not move at the specified speed immediately, or perhaps some form of backlash... although I doubt backlash because it is at the start of every thread.

@AgeingHippy - I had this problem when the PTFE thermal insulation on Adrian's old hot end design with nichrome wire was getting too hot and making too soft. Blowing cold air fixed the problem in this case.

I've also had pauses when printing when I set my extruder stepper to microstepping using Gen 3 Techzone remix. I found that I have to limit myself to 1/2 step on the motor or the microcontroller/firmware can't deal with the fast retract/restart speeds.

Hmm... that is interesting information brnrd. I will try lowering my temperature and see what happens.

If you lower the tip temperature too much, then PLA becomes hard to push and it builds up more pressure. In this case, I've found that I can't get retract/restart to stop the ooze. The only way for me was to cool down the thermal insulation with air. Try this experiment. Let your extruder cool down to ambient temperature. Then turn on the heater. As soon as you reach temperature, try extruding with retract and see if it stops the ooze. In my case, the retract worked in this situation. As you allow for more time for the thermal insulation to heat up, the retract won't work anymore and you will get blob during restarts. This is what i observed and what led me to the conclusion that it's the thermal insulation.

If you have an aquarium, you can try borrowing the air pump to blow air over your insulator.

Cheers

I did not read your post with full attention.

I am planning on getting one of those small fans tomorrow so will try cooling down the thermal break and see what happens.

You can also switch to ABS. It's not as sensitive to the thermal break getting too hot.

good information brnrd, I don't think that solves my problem as the filament is only getting heated a couple mm up from the actual heat zone. And wouldn't explain why there's oozing on every layer including the first few. With your theory there wouldn't be any oozing at first and would get worse as it goes. Whereas I'm getting consistent ooze every layer.

Could it be an issue with the gen6 electronics? Anyone else out there using gen6 electronics run into this problem?

I'm really out of ideas at this point, and at a loss as to where to continue trying to solve this issue.

Here's another idea.

Are you microstepping the extruder stepper motor? With the Techzone Gen 3 electronics, I found that I get a pause whenever it does the retract/restart with microstepping so I set the Pololu to 1/2 step for the extrude. Others have told me that the Five D firmware (reprap) does it's calculations in floating point so it slows it down. This problem gets worse as you increase the retract rate and distance.

What kind of hot end are you using? I'm amazed that the heat is only limited to 2 mm beyond the hot zone. Without any cooling, heat will travel up tip even through the thermal insulation and soften PLA.

[blog.arcol.hu]

that is the hot end i'm using, as he states in there, it melts around 10mm of the filament, which is only a few mm higher than the nozzle itself i believe (i haven't actually measured how long the nozzle itself is) but i can let the filament sit with the hot end heated all the way up to around 200C and just start extruding without the filament bending or being a problem at all. Nor do I get any problems with the filament buckling during extrusion.

The gen6 electronics do use microstepping, but to be completely honest i have no idea where to set that in the firmware, or if i even can.

I'll try playing around with the retract rate and see if increasing speeds/distances makes the problem worse. that would be a good litmus test to see if that is where the problem is.

I guess all that heatsink in the cold side helps. it would probably be longer with PLA than with ABS. I've found ABS to be easier to work with that PLA.

I have fixed this same issue on my bot. (Gen6 as well) It's not the Z speed, or Jitter, or Lash.

It's CLIP!

Activate Clip and start witht the default 0.50 clip and increase untill the blob is gone. Got my major blob to go away around 0.56 clip.

Here's a video of my issue-Corner Blob

Now that same print goes counter-clockwise until the layer is done and goes up quickly with no blob. I hope this works for you. Good luck!

Edited 1 time(s). Last edit at 04/14/2011 12:52PM by MrJohn.

With a clip that large, you will usually get big gaps in the perimeters. The correct way to fix the problem is to get the retract/restart to work right. You would need to play around with the extruder temperature. Too hot and you will increase blob. Too cold and you will get ooze or it will not extrude. PLA seems to be more picky with temperature than ABS.

well i just tried with your settings for clip mrjohn, and it totally worked, blobs reduced in size by about 90%, I'll up it by another .01 or .02 and that should get rid of it entirely.

Thank you so much.

This (the Clip idea) sounds interesting.

I do think the temperature issue does have merit though.

I have reduced my print temperature from 210 to 195 (first layer 210) and the blobs seem somewhat smaller...

yea not pushing the temperature too high helps. I'm now using retract (about .75mm works well for me), printing at 200, using clip, and printing infill>loops>perimiter that seems to get around the problem the most.

it doesn't completely get rid of the problem of pausing at the start of a thread, but with printing in that order, clip will move mostly seamlessly from one line to the next. Since it is trying to bridge lines that end close to where the next line starts.

So i still get a blob now and then, but they are MUCH smaller now, and definitely not enough to ruin a print at all.

I was having the same issue. Using Gen 6 board with mendel-parts V5 hot end modified to fit wade's extruder. Firmware is Klimentkip. Setting retract to 30mm/s really helped. I have my temp set to 205 degrees and the blobbing is very minimal, but before getting my temp set right it was pretty bad. I plan on switching to RAMPS here very soon becuase the gen 6 board is pretty limiting If it fixes my issue just with the swap I'll let ya know!

let us know how that works out inh, it would be nice to be certain of what is actually causing the problem instead of just finding workarounds.

I can't really advise on any of the above comments, but you can determine if it is a problem in your extruder by measuring how much material comes out of your nozzle in ten seconds after the motor stops. Figure that out, and post your results. Hoefully some here will reciprocate with their times.

Will give Clip a tweak as I'm getting similar at the moment

---

__________________________________________________________________________
Experimenting in 3D in New Zealand

I've seen similar things with my Gen6 at the start of a thread. I've watched the communication log and the pauses (and hence the blobs) always seem to coincide with a communication error. I suspect that the processor is running out of oomph at a critical moment, like during retraction or a Z-Move and misses a command. The delay for the resend causes the small pause and the blob.

I haven't really spent much time digging in the firmware trying to figure out what is causing the packet to be lost.

JB

Probably the high step rate during the retraction takes up all the CPU time.

---

[www.hydraraptor.blogspot.com]

let us know what you find out jim, that sounds very promising towards figuring out an actual solution to this instead of just workarounds.

From what I can gather, getting e steps per mm right for the extruder, finding the right temps and retract, using CLIP and JITTER, seem to eliminate the blobs on one corner. It's also due to the size of that test piece being so small. If I could a fix out of thin air it would be Z up in flight. Meaning the Z would rise while the x or y is still moving.

I ran a few experiments this weekend. The first was to remove retraction and reduce the Z-axis speed to verrrry slow. This was to test my theory that these two events were causing the comms errors due to lack of processor resources. With this configuration the comms errors went away (of course the blobs got much worse!).

Then I started digging around in the firmware. From what I was able to determine (without using a profiler) the retraction and z-axis moves require so many steps in such quick succession that the ISR for the UART (which is lower priority than the timer interrupts) cannot be serviced. Watching a trace of the UART transmissions from the Gen6 during these periods shows the characters coming through in 1 character increments rather than in big blocks - also suggesting that UART routines are having problems.

Looking at the echo of the commands that are sent, during the periods where the UART is slowing down, the command that causes the buffer flush echo's back as just the first character. This essentially means that the Gen6 received the first character (an 'N') then missed all the rest until it got the carriage return ('\n') indicating the end of the packet. Due to the poor flow control, this causes the buffers to be flushed and all the data resent.

There is another problem at play, Repsnapper does not wait for the 'ok' to be sent from the Gen6 before it sends the next packet. It sends the next packet the moment it receives the first byte from Gen6. Since the 'ok' occurs at the end of the period when the UART is slowing, waiting for the 'ok' might be beneficial.

To try to fix this I wrote my own simple host tools to send the GCodes to the Gen6 to wait for the 'ok's before sending the packets. This did help with the pauses at the end of strings, but it was a minor improvement and not worth writing a whole new tool chain for. But fixing Repsnapper might be worthwhile.

I then decided to look at the retractions. There are a lot of retractions that are unnecessary, they cause more of a blob due to the pause then they would if the head just moved straight away. So I wrote a script that removed any retractions where the next move was less than 1mm away. This also helped noticeably with the blobs. Removing those at the end of strings, but it didn't help with blobs at the end of a layer, which are the most noticeable ones since they are on the outside surfaces.

Increasing the Z-axis speed does help reduce these blobs. But, these are all band-aid patches - the real issue is the processor still can't service the UART interrupts when it is making fast movements. Either the code needs to be restructured to allow the UART some processor time or a faster processor is required or a better flow control system that prevents the host sending characters when the processor can't handle them would help resolve the problem completely.

JB