Strange infill gaps

Hi,
I have a strange infill issue on large prints. For the first 4 layers or so, 100% infill looks good. However afterward, there are pretty large gaps in the infill (it looks like 95% for all intents and purposes), which means small holes in the end model as well as reduced part strength. Small parts however print just fine, with no gaps between the top infill lines. Hints at underextrusion, but the lines are still smooth and continuous, and there is no indication of filament grinding (I can usually hear it pretty well when it does grind).

I tried a number of things:
Increased the infill multiplier by over 20%, still gaps in later layers. Possibly smaller, but I can't tell easily.
Tested the thickness of the problematic infill layer, it appears to be .20mm as it should be, so it's not z axis moving too much.
Made sure z axis is moving correct amount of mm overall (calipers and calibration model, spot on)

I can increase the extrusion multiplier further, but I'm worried at this point my bottom layers will turn out badly, it's already 1.1. My e steps are well-calibrated (recently too). Anyone experienced something like that or has any ideas?
Thanks

Prusa i3 rework
Printing with natural PETG at 255 C

This has been an issue with PETG for several users. The only cure, I can think of is printing really slow.
Or try to use a different pattern for the infill. In my experience long straight lines are the worst, so try out Honeycomb or Hilbert curve for a change.

I had good results with 238°C and 30mm/s.

Edited 1 time(s). Last edit at 01/10/2017 02:43AM by o_lampe.

I'm printing at 20mm/s... Much slower and it'll be printing through the night.

This is a bizarre issue, are there no clues as to what might be causing it?

I've never tried honeycomb, does that print 100% infill? Can't imagine what that would look like. I can also try lowering my temps, but I seriously can't see a higher temp producing this problem.
Thanks.

Confirming partial improvement if print is paused at problematic layer, then nozzle is raised and some filament is extruded in air, then print resumed. The improvement lasts only a layer or two though. Nozzle flow reduction is obvious, but the reason (or permanent fix) is unclear.

I will next try different printing speeds and report how these turn out.

Biggest mystery though is that I've successfully printed this exact model before with a twin roll of eSun PETG, at the same exact settings. Problem seems to have started when I switched rolls, but no guarantee that that's the cause.

Edit: confirming gaps even at 10mm/s, as well as 30mm/s. Doesn't seem to get better with low speed (or higher speed).

Edited 1 time(s). Last edit at 01/11/2017 05:32AM by anvoice.

My first roll of PETG was natural and I found it very difficult to print well (filaments.ca brand). I recently tried a roll of MG chemicals red PETG and found it much easier, so variation between rolls, in my case maybe due to brand or colour is a definite possibility.

With softer/lower temp filaments I've had problems later in a print where the filament was heating up as it passed by the hot stepper motor on the way in to the extruder which caused symptoms like you describe, with gradual under-extrusion developing. However, I haven't seen that with high temp filaments like PETG. (I solved the problem by using a piece of PTFE tube to insulate the filament from the motor for about 5 cm before entering into the direct drive extruder.)

Edited 1 time(s). Last edit at 01/11/2017 01:26PM by JamesK.

I dug out an old roll of near-finished MatterHackers natural PETG, and managed to print without gaps it seems, except for the ridges from overextrusion still feel like the gaps did (hard to see with clear filament so I usually went by touch to check for gaps). So perhaps I have a bad roll of filament, although it measures pretty much spot on at 1.75mm diameter.

Unfortunately I don't have other colors to test at the moment, only 2 more rolls of natural eSun in the mail. I sincerely hope I can get this to print, as I prefer natural, and have already printed numerous good parts with PETG (mostly with the MatterHackers filament).

I'll report back if I manage to solve this problem for the eSun PETG or if the new rolls print better. In the meantime, still gladly taking suggestions.

Quote

hard to see with clear filament so I usually went by touch to check for gaps

That brings back memories. When I started I assumed that the natural filaments would be the easiest to print with, so that's all I ordered. Then I discovered how hard it is to judge over or under extrusion with clear filament!

Too true. Maybe I'll get some color as well just for testing purposes, though I don't see using it much for actual parts. Or maybe just get one of those small digital microscopes, that will clear things up with clear filament.

Confirming no gaps using new roll of eSun natural PETG at 255 C.

Now suspecting moisture in previous roll. Or just bad filament. At some point will dry the old filament and try again.