A couple of other things of possible interest.
First, the most expensive drybox I’ve ever seen:
Similarly, the Prusa recommended drying oven is noteworthy for how expensive it is:
Thanks. I had no idea the bottom one existed.
Little cheaper here. 300c is nutty. Must be good for curing. or broiling a steak lol
or annealing. 
Up to 300C is pretty common for blast ovens.
1 Like
Interesting to read about Peltier devices, vacuum pumps and desiccants to dry air. One of the simplest ways to dry air is to pressurize it. Get an old compressor that will go up to 150psi, then the maximum RH of the air coming out will be 10% (AH = 1.7mg/L). This is why water traps are needed on compressors. Raise this air to 60°C and the RH comes to about 1.3% - even though you could have started with 100% RH.
PS Work on the inline dryer is proceeding and I’m starting to see some encouraging results.
Thanks for this information. I was rather hoping that there was a specific maximum water content to target for each filament rather than seeing if it worked afterwards.
Yep, that would be nice but I was wanting to use the printer for actual stuff.
It’s super easy to just dry to a value you know will eliminate water effects though if you can cut it closer you’d definitely be more efficient. I hit the level I dry to in 8-12 hours but cutting it closer would shorten that up.
If someone is in a production environment where that kind of thing would make significant difference, it’s no trick for them to do those experiments. For home use, a few hours probably doesn’t make a lot of difference.
Another thing is the closer you cut it to where you start seeing water effects or they exceed some level, the more water creeping back in will be able to cause problems or force touch-up drying.
You are right, though.
I started out with an approach similar to what you describe, expecting results similar to what you describe, but what I found is that if I compressed the air and then bled it off only very slowly, my water trap didn’t collect any water. None. Zero. Not even a fogging of the container. Maybe the water traps work better if you’re blowing out air at a fast enough rate that there’s a significant reduction in temperature, and it’s because of that chilling effect that the water condenses out and gets collected in the water trap?
In my case, the other complicating factor was that the act of compressing the air raised the temperature of it enough that it was still hot on exit. This is because I went with a small air brush compressor. Maybe if I had gone big with an 20 gallon shop compressor or something it would have cooled off enough between duty cycles, and the larger tank would have had more surface area to radiate heat away from the compressed air.
I abandoned this approach largely because even the airbrush compressor was already fairly loud, and a really big air compressor would have been even louder. I suppose I might have simply attached a much larger tank to the airbrush compressor, but I decided to change course instead given the disappointing early results.
Nonetheless, if you already have the equipment and could make it work, it would be interesting to hear the details of it.
Compressed air as a method to generate dry air is a non-starter for me due to the noise. That’s another reason I am thinking Peltiers.
There are some cheap Peltier dehumidifiers on Amazon meant to operate inside a room. I think Adam Savage has one in his dry box. 
I wonder if one of those could be retrofitted for this somehow.
Wait… hmm… big dry box which contains a room dehumidifier and an S4? Maybe put a small aquarium pump in to cycle air between the S4 and the dry box? Yes… yes… this may be the overkill I’m looking for…
The airbrush setup wasn’t a total failure. I suppose from just the compression part alone it managed to reduce the humidity from around 60rh on intake to around 30ish on exit. Maybe lower, as I don’t exactly recall now. However, it wasn’t low enough to be interesting for filament drying. That said, it only went to around 80 or 90psi, so well short of the 150psi or higher. I’m sure it could be made to work probably well enough if someone wanted to pursue it and didn’t mind the noise, and so I wouldn’t want to discourage anyone from trying it. I’m simply documenting here the unexpected hurdles I ran across. Because it was in the garage I didn’t think the ongoing occasional noise would bother me or my family, but it turns out it did. YMMV.
@Lou I genuinely think Mzip’s design is an excellent starting point if this is something you’re interested in. You may decide it’s all you really need.
It is, but I have an S4 which is much larger… maybe an aquarium pump wouldn’t keep up.
This is a test part printing from the silver silk PLA Bambu recently discontinued for the new silk+. Dried to 19% using a surprisingly old silica gel column that’s getting very dark but still pegging the output hygrometer at 10%. 0.2mm layers.
3 Likes
The seam must stand out on this pretty bad? Really nice looking print.
There is a flat side on the back where it mounts to a plate and the seam prints at the back along the line where flat meets curve so the mounting hides it. Seam is set to print at the back so positioned the flat at the back on the build plate to keep the seam hidden. It works great.
2 Likes
You’re right, the air in the compressor tank will be very hot initially and will need to cool before extracting the air. This may be impractical if a high flow of air is needed.
Here’s a new ay for deciding whether or not your filament is too wet for printing:
2 Likes
That’s beautiful! Makes total sense and is a relatively quick test for moisture.
It’s a bit more work to load a spool, heat the extruder, and perform the test, but it looks like a practical test that should provide accurate results since it’s based on the exact thing that messes up prints.
Also, that length scale would probably be unique for each filament type so good chance users would need to generate their own calibrations to find what lengths print well and where they get too long and prints fail. Long strands would probably be obvious fails, but the shorter lengths might be more iffy on how long is too long.
But the test seems sound and should work well I think. Neat video!
1 Like
I agree with all your points.
Seems like some kind of pinless moisture meter could be rigged for detecting moisture, and if it worked that would be the ideal gizmo, even if you had to calibrate it for different filament types. Afterward it would be quick and efficient.
Alas, more potential avenues to explore than I have available time. 
1 Like
I do want to put in a few words for those who have moisture issues, though. If you dry your filament effectively (this is key) and store it in dry containers, testing isn’t so much an issue.
If you put dried spools into some container with low water permeability and keep desiccant fresh in those containers, it stays dry. Put hygrometers in there and you have a way to keep tabs on how dry your filament in storage is.
Truly dry filament pays big dividends - first just in quality of prints by removing moisture effects, but desiccant in storage with dried filament stays “dry” for longer, desiccant in the AMS lasts longer, etc. Filament that isn’t “dry” carries more water into storage and into AMS units which loads up desiccant and forces early replacement or regeneration.
I leave spools in their shipping boxes until I’m planning to use them, then dry them, and put them in cereal boxen with desiccant and a hygrometer until it’s time to use them. They are dry and I don’t need to use a hot extruder to know how they will print, don’t need to weigh before and after drying, etc. Drying is incorporated into my filament handling chain for every incoming spool and that sets everything else up to work extremely well.
And I know there’s issues with the hygrometers in filament dryers and also the hygrometers I use in the cereal boxes. They don’t affect my process. I don’t really care so much about the true humidity values. It’s enough that I’m in the ballpark even though the hygrometers I use turn out to be pretty uniform in response. I dry a bit more than I apparently need to but everything else down the line benefits.
It might not be appropriate for all. The cereal boxes increase the filament storage requirements a bit, and it means drying every spool instead of skating on some. But not dragging water into storage and the AMS units saves hassle there too in having desiccants last long times. I’m still on the first packets of silica gel in my storage boxes from when I started this process last summer and they still all read the 10% minimum my hygrometers can read. I just replaced my AMS desiccant packs after 4-5 months of service when the humidities climbed to “2”.
I don’t weigh spools any more. Not doing any “how much moisture is allowable tests”. It’s just a procedure that so far has worked fine every time so it’s just an SOP now.
Not everyone needs to dry filament or store it with desiccant but turns out I did. Everyone’s MMV though.
1 Like