@user_3026326371 So as not to dilute this thread, I created a fork:
Let’s continue it over there.
Thanks, @NeverDie but I’m going to let this one die since it got so long and start another when I post the design for questions/issues.
All good. I’d really like to get folks an answer for these filament issues. Summer in the northern hemisphere seems to cause lots of problems.
After I post some very fresh (and good) results I’ll let this die, that is… From pump on it was about two minutes to the output hygrometer pegging low at 10%. Output pressure is well more than 4.5 inches of water column with all the flow anyone could ask for. It’s not so obvious in that image but it was almost a solid stream of air. Of note to those just joining - 10% is the lowest the hygrometers indicate. The actual output humidity is somewhere between 0 (not likely) and 10%.
Next is an actual filament drying to see how it does with a comparison to room air, and then it’s just an assembly line of drying my existing rolls to put them in cereal containers and see how long before output humidity starts indicating above 10%.
Really amazed at all the volume it has now. Feed this into a filament dryer and I think it will dry much faster and better but I’ll have weights in and out to know. The jade white will be a good comparison between ambient air drying and dry air drying.
2 Likes
So far in 1 hour the humidity in the Sunlu S2 has dropped from 44% to 32% RH at 55C. The spool of jade white basic PLA has lost 0.30g of water so far with 39% ambient humidity into the air dryer with air out less than 10% RH. The jade white was fresh from its bag, shipped as a 2-pack. Throttle valve on the aquarium pump about 2/3 open. The lid is not propped open. I don’t know how that compares to other’s results. Feels like good numbers though.
I didn’t think to weigh the desiccant before starting and need to rework the end of tube fitting that keeps gel beads out of the plumbing. It’s too hard to push into the beads which is why once it was on and I remembered I needed a tare on them, it was too much of a pain to take back apart. It doesn’t need to be a big diameter anyway. Probably helps some but it’s just collecting air after it passes through the beads. So a minor change necessary to make changing beads easier. But as for data uses, the big thing I’m interested in is practical - how many spools of filament can I dry before changing beads?
Edit - Starting on the 3rd hour drying now. Humidity in the dryer just dropped a point to 31% RH but was 32%RH still when I weighed the spool at the 2 hour weighing. In the first two hours it had lost 0.79g water total.
A different, but easy way to test for benefit–especially if you had only one spool of a particular filament-- would be:
- dry the spool to asymptote loss of weight using regular ambient make-up air, as you might do before you had your air dryer.
- Afterward, activate your air dryer to supply the dry make-up air and see how much more moisture you are able to extract. All that extra extracted moisture is bonus compared to what you were getting prior to your device, and this method quantifies what the incremental benefit is.
1 Like
That’s a good point and sidesteps the issue of only being able to test two different but very similar rolls that could and probably will still be a little different.
I was going to dry the second roll conventionally and compare but can do this for added information.
Good idea!
This is new PLA right from the shipping bag from the box that I’ve had on hand (with the other spool) for a few months.
I’m looking forward to that other test, @NeverDie. That will be interesting both for comparing to today’s run but also if the dry air picks up more water over what the dryer alone can do.
Don’t know starting water concentration in the filament but at least under these conditions definitely losing some weight. Sunlu dryer reporting 22% RH inside at 6 hours in - down significantly from 44% RH starting humidity. Ambient humidity at the 6 hour mark (and during the whole test) is 43% RH.
6 hours in total weight loss is now 1.60g with Sunlu chamber at 22% RH.
Y axis is water loss in grams. X axis is minutes of drying.
I had to cut yesterday’s test short. The spool was still losing water at a fairly steady rate give or take given the noise. The graph is weight loss in grams vs minutes of drying. The top curve is with the dry air. The bottom curve is no dry air with the door propped open using the brace. Ambient humidity is 43-43%
I was expecting more difference in 6 hours but there is just a hint of the door open curve starting to roll off. I put the dry air curve spool immediately into a plastic cereal box and sealed it up (no desiccant) so may be able to pick up where I left off if the weight doesn’t change too much from when I took it out.
The different slopes show dry air drying is faster than ambient air drying at 44% RH or so with the door propped open. But they also show pretty good weight loss with the Sunlu propped open. We’ve been having showers again and humidity is highish now. If the door open curve is rolling off that will also be important so going to get data as long as I can tonight. Once it rolls off enough, I was going to add dry air but might have to put this spool in a cereal box until tomorrow too to complete with dry air.
1 Like
Yes, it would be nice to see the performance when ambient humidity is high. Outdoors here it’s regularly 50-70%RH ambient. My garage is a bit less, but it tends to track that. My garage is where I’d prefer to do my printing. Less convenient, but nicely isolated.
1 Like
With dry air purge the performance should always be the same for drying. The difference will be how many rolls of filament until the beads need regenerating.
I don’t plan on any more open door runs at various humidities, though. I’ve really invested way too much time in this.
Humidity is about 10 points higher today. I stored the open door spool in a sealed poly cereal box overnight same as I did the dry air roll now from a couple of days ago. When I weighed the propped-open door spool this morning to restart testing it had picked up a very little weight (could also be noise in the weight measurements) so it looks like it’s probably ok to start and stop testing for long runs as long as you store the filament in a sealed up impermeable container.
But, with the higher humidity today, door open, and the roll and chamber starting cold, during the first hour the spool increased in weight 0.09g. I double checked the weight and had done multiple measurements at the start so it should be real. The spool had been drying all day yesterday and losing weight so the filament would have a higher affinity for water than when it first went in the dryer. With it sitting in moist air while the dryer was heating up, the relatively dry filament got wetter. I actually saw it happen. Not quite fair though - weights are dropping again. But it shows it can happen especially if the filament is fairly dry and humidity is high.
But the weight is dropping again now that the whole roll has warmed some and the last measurement was just barely heavier than the starting weight. So three measurements over two hours and next measurement should be back to or past where I ended the testing yesterday.
It looks like with the door propped open the stock Sunlu can do ok with the above caveats on really humid days with filament that is already fairly dry. Purging with dry air or recirculating over desiccant looks like the ticket for faster and more consistent drying though.
I’ve now got some performance comparison numbers but still don’t know how long the beads will last in practice. And it’s not a thorough test, no replication, and the first dry air run had an issue that cut into how well it performed.
With dry air I dried a spool of jade white to a Sunlu S2+ chamber RH% of 22% in 6 hours for a 1.6g weight loss. This run had a slight leak at the hygrometer in a custom/replacement Sunlu base so possibly could have been better. Ambient humidity was around 39-40%. Door was closed. Aquarium pump was set at about 3/4 flow setting. This run had to be terminated due dinner with friends but weight was still dropping and this roll has been saved in a sealed poly cereal box without desiccant so should be able to continue drying tomorrow for comparison with…
The second spool in the two pack got the “stock” Sunlu S2+ with no dry air and the door propped open using blecheimer’s door prop. Ambient humidity ranged between 44% and 61% RH and it took 17 hours to lose the same 1.6g that the dry air removed in 6 hours. That was also where the open door weight loss leveled out and stopped decreasing. That’s give or take because while continuing to do measurements to make sure it bottomed out, there was variability up and down in the weights that could be noise. Measurements varied by 0.04g was all.
Once the open door spool had its weight loss level off, I hooked up the dry air to see how far and how fast it would continue to drop. So far it’s lost an additional 0.64g in 5 hours and the Sunlu RH% is showing at 19% with an ambient humidity that’s ranged between 52 and 62% RH.
So at least for these rolls and so far, dry air being pumped into the Sunlu removes the same amount of water in 6 hours that the Sunlu can by itself with the door propped open in 17 hours. The Sunlu by itself with door propped open was able to remove 1.61g of water in that 17 hours and that was as far as it could go. Adding dry air has removed another 0.64g in another 5 hours.
Neither is fast. To get to this same point on just dry air is looking like 11 hours running but the filament just keeps getting dryer still. It hasn’t bottomed out but probably will get where it’s taking way too long. Some stop at an indicated humidity of 20% in the Sunlu and it hit that between 9 and 10 hours in.
So a little more to do to verify but that is the comparison. Dry air gets filament more dry and does it faster than a propped open stock dryer. The lowest humidity the Sunlu reported with the door propped open was 32%. It never hit 20% Continuing the test a few more hours but that’s pretty much it.
Let me throw this out there as a point of comparison.
I have a true blast oven, and I’m drying for far longer than you are. For a spool of creality Hyper ABS (mounted on a cardboard spool), it shakes out like this from a recent drying session on a brand new package delivered from amazon:
After 4 hours, about 6-7g lost.
After another 20 hours, about another 6-7g lost.
After another 48 hours, about another 1.5g lost.
This is drying it at a very reliable and stable 70C, but in the garage where the high humidity (50-70%RH) of the ambient make-up air is no doubt a factor. Still, it’s a fairly lengthy process. I’d say these numbers are fairly typical for almost any filament mounted on a cardboard spool, so I’d hazard a guess that a lot of the moisture is from the cardboard… I probably could have stopped after the first 24 hours, but at 70C the filament doesn’t seem bothered by leaving it in longer. For this particular filament, at 80C the inner windings start melting together, resulting in a useless spool. At 70C, it seems like it can go indefinitely.
1 Like
From the tests I’ve done which admittedly aren’t that many, the run with the door propped open seemed to follow the ambient humidity. With ambient air there is only the heat to lower RH as you know and it can only go so far.
I was surprised at how well the Sunlu did with the door propped open, though. I kept waiting for weight to stabilize and it kept dropping even a little bit. But it hit a floor because of the water in the air and as you suggested, turning on the dry air brought it down farther.
I stopped it at 18% RH showing on the Sunlu chamber humidity and put it away. The dry air spool is in now finishing up and I’ll be rolling through all my spools to see how long this can go before the output air hygrometer starts indicating above 10%.
I’m guessing I should get 10 or so spools before the beads start losing the ability to dry the air enough but that’s a wag. Time will tell.
Added - And about how fast it’s using up silica gel beads - I now have no way to know other than watching the exit hygrometer. With the plastic bottles I was using before to hold desiccant on the test stand, they were nice and clear in the IR. I could see how much of the beads were being used by the heat they released using the IR viewer. The new glass reservoir is totally opaque to those IR wavelengths and now I just see the outer glass which doesn’t show the heating effect. Unfortunate because I could know early on how much of the gel beads were grabbing water.
Added some more - on the 5th roll now and all things nominal. Output air is still below 10% RH. Because of a big void volume in the new base for the Sunlu (to be fixed), things go much faster if I don’t shut down like overnight between spools. Doing spools back to back it looks like I can do two per day or almost two per day. Storing the spools in the cereal poly containers works great. The spools grab a few tens of milligrams of water weight but lose it once it heats up again next day and RH drops from new dry air. It picks right up where it left off but that’s about two hours or so to get back to temperature. As the spool warms RH in the Sunlu spikes and then trails back down. It’s working like clockwork. Will see how many spools before output RH starts rising but depending on how much that affects drying times, might be able to go further than when the output RH starts indicating. But output RH is still somewhere below 10% just no idea where.
I’m on the 7th roll of filament going through the dryer now. Each has lost around 2g of water and completed in around 8 hours give or take with the flow setting at about 3/4, around 50% RH, and drying to an indicated humidity of 19% in the filament drying chamber. I also found more leaks and ways to optimize that I’m trying in another print to put this to bed.
As it has been running, I’m now seeing an effect I wasn’t counting on - the indicator is starting to change color to show how much silica gel is being used up. I had thought that the way the silica gel would work is humidity in the silica gel wouldn’t get high enough to cause the color change. I thought I’d see water breakthrough first and the only indication would be the exit hygrometer coming up off the 10% minimum reading it can do. I was wrong.
At least running these spools back to back and only turning it off overnight, the humidity gets high enough to start seeing the color change where the air enters the reservoir. This is really good news because the silica gel is holding more water than I had planned for. Above that region the humidity is still too low to change the color so it still polishes the air to low RH%. Will be interesting to see just how long the silia gel can hold the RH% low but the color change is low in the reservoir just like the heat was confined before. It may be saying I’ve got water holding capacity to dry more than the 10 rolls I thought I might be able to do before regenerating the beads.
Here it is this morning cooking away. You can see the desiccant darkening up at the reservoir bottom where the air first enters the column. Decreased the dryer itself to 50C just to see what the dryings look like there. 55C is enough to affect the filament so a little cooler would probably be better.
1 Like
Great work! If you would, please mind me what the “3/4” flow setting means?
The aquarium pump I’m using has a throttle valve built in and that’s what I have set to 3/4 open.
I’m almost done drying the 7th spool now and if the color change can be trusted, it might be a number of more spools before the water breaks through enough to start seeing the exit hygrometer start increasing its reading. I wish I knew the actual RH of the dry air but sometime in the next 5-10 spools I bet it starts climbing. Will update whenever I see it.
It has got a lot of time in so far just pumping away and the water carrying ability of those beads is surprising. I don’t have a starting weight on them though and really wish I did. The fitting I had on the end of the drinking straw was too big and flat and was difficult to push through the beads so I only did it once and redesigned it to better stab through the beads. But it had the equivalent running time now of 2-3 days solid and still going strong.
I’ve got all of this rounds mods in the print and it should be the last one. Have a new drinking bottle on the way to build a whole ‘nuther unit but think it’s the one I can post.
It does take a little soldering to put pigtails on the hygrometers, you have to paint the whole inside and outside with cyanoacrylate to seal it up (have yet to print one that is air tight), there’s some drilling in the water bottle lid, and be able to apply a good gasket-like bead of silicone seal to set the top of the water bottle in place. It’s a semi-complicated build.
But it is working like a champ. Way better than I had thought. The aquarium pump was key to making this work (IIRC that was your idea?) since it also flushes out the moist air in the dryer itself. Not super efficient but efficient enough. And silica gel regenerates easier than alumina or Drierite, etc.
1 Like
That was @islandBill. He and the others on that thread had all sorts of good ideas.
It’s great that it’s working as well as it apparently is on PLA. Do you think you might ever have any ambitions for printing with Nylon, TPU, or whatever else is to known to be a tricky moisture sponge? I’d be real interested in just how dry the air through your desiccant column really is, and whether it can get to the -40C Dewpoint that Visioneer specified. Equally interesting is whether that -40C spec is truly needed for any of the filaments we can actually print on a Bambu printer, or whether it is actually only of relevance to even fancier filaments that are outside the reach of our Bambu printers. I’m starting to suspect it’s the latter case, which would be fine by me, but then it simply shifts the question as to what the true dewpoint spec should be.
I’m just thinking out loud here as I sharpen my own thoughts on the matter… Anyone reading this please feel free to jump in.
1 Like
I don’t have a way to measure the humidity below 10% that the hygrometers can do. I saw water in the Drierite when I was running the test stand so there was some breakthrough even though the silica gel exit hygrometer was pegged at 10%. Maybe it starts out somewhere near a -40 dewpoint, but the exit RH% will be climbing as the pump runs. You can use any desiccant in it, though. Drierite dries to -100 F so you’d get that for a while but you’d want to know where you were at especially if you need such dry air. I’m lucky with PLA with people just drying to 20% RH in the chamber. If you find a way to measure that low you’re golden.
No plans to use those other filaments for now but maybe in the future?
1 Like
You’ve clearly got “the knack.” Have you dabbled at all in Arduino? If not, you’d surely pick it up quickly. If so, though, it’s a fairly simple matter of hooking one up to a high quality TH sensor. Very often there’s a corresponding demo library on github, so development time is minutes. In terms of high quality but inexpensive, I think SHT45 seems inexpensive but still “good”. Not the best, but not bad. You can buy a breakout board with the sensor already mounted from Aliexpress for $3.80:
https://www.aliexpress.us/item/3256805824443252.html?spm=a2g0o.productlist.main.3.7a774771bByjXU&algo_pvid=5f42b9c8-2c40-46dc-ac5e-728a5fe33f00&algo_exp_id=5f42b9c8-2c40-46dc-ac5e-728a5fe33f00-1&pdp_npi=4%40dis!USD!3.80!3.80!!!3.80!3.80!%402103011017237213066494912e0a63!12000035308426138!sea!US!726672273!X&curPageLogUid=m3ZMwD4oDgaz&utparam-url=scene%3Asearch|query_from%3A
Buy 3 and you’re shipping would be free.
Here’s the datasheet:
https://www.mouser.com/datasheet/2/682/HT_DS_Datasheet_SHT4x-3454169.pdf
It can measure the full RH range, from 0% to 100%.
As you can see from Figure 5:
from 20C to 70C ambient, typical RH% accuracy is plus or minus 1%, and maximum error is plus or minus 2%.
Well, actually, that’s not entirely true. Actual accuracy depends on both temperature and actual humidity:
i.e. typical accuracy is plus/minus 1.5%RH across the entire range.
Would that be to your liking? If not, TI makes a $4.70 chip that’s accurate to about plus or minus 0.5%RH:
It’s been a while since I last checked, but last I checked aliexpress doesn’t have a breakout board for it, but digikey does… for a bit more coin, but not a lot more.
I already have the aliexpress SHT45 breakout board. I’ll probably eventually order the TI breakout board as well.
1 Like
I am familiar with Arduino and have a few doing different simple things around the house. Very cool little devices for sure!
I don’t think I’ll be looking deeper into the low humidity aspects, though there would be value to getting accurate numbers below 10% RH. Exit humidity will be affected by flow rate and water weight in the desiccant from use. When water starts breaking through it will also be affected by ambient humidity. It’s not a good reason but it’s already pretty complicated and I just want to be done with it. My wife is a bit tired of my setup on the end of the dining table as I test to get an idea of capacity.
It would be good information but if you load it with fresh Drierite the air coming out will be around -100F dew point at least for a while. If it behaves like it’s doing with silica gel, you would probably get at least 10 spools dried before having to worry. Possibly more since Drierite is a chemical reaction and has a pretty high capacity for water. For my purposes with PLA, silica gel as desiccant and just knowing the exit humidity is below 10% is plenty. And redesigning to incorporate an Arduino, sensor, and making sure it’s all leak-free (way bigger issue than I thought it would be) would be a big job.
But a standalone measuring system that could be added inline would be great and easy to use for those who need to know what the humidity is or who want to experiment with other desiccants. And it could provide very useful information. I’ve got my hands full right now but would be happy to help or test if you want to create one.
1 Like