A convenient workaround to that would be to take a screen capture of a formatted equation and then post it as a picture. On Windows it’s built-in called the “snipping tool”, and there’s a similar thing on Linux.
Cutting to the chase: how might this help us do significantly better than what we’re already doing? @Mzip has a working solution that seems to work pretty well, so I propose that as the baseline. I’ll leave it open as to what “better” means, but the usual trade-offs are among: more effective, faster, or cheaper.
As near as I could tell, this discussion was about whether RH% is the right signal for drying termination. What I apparently missed on my first reading was: if not that, what’s the better thing we should be using instead? Perhaps a concrete example would help illustrate. In it’s favor is that it seems to work for @Mzip, and it’s a lot more convenient than weighing the spool at several points in time and trying to judge when the weight has reached close enough to an asymptote. Again, I probably missed it, but what’s the proposed alternative, and what would make it better?
After a few days I have calmed down enough to make a last attempt to make a positive contribution to this topic. I suppose I wanted to give a bit of background first. I felt my solution needed the theory up-front in order to model the solution. As a research scientist, I rarely started with the solution and worked backwards to find out why it worked.
Anyway, I’ll summarize my approach and then MZip can tear it apart as he sees fit.
I’ve built a flexible heated transfer line out of PTFE tubing (the type used for Bowden tubes). I use a small air pump to pull or push a low flow of air through this heated tube through which the filament is being fed to the extruder. I monitor the ABSOLUTE HUMIDITY of the air exiting this tube and use this to close-loop control the pump speed to optimize the moisture extraction so that the pump lifetime is extended, and it makes the possible use of desiccants more practical.
My proof-of-concept testing has been very positive and I’m now moving onto putting together a practical implementation. I’ve found that the filament residence time in 1m to 2m of this heated tubing at 50°C is enough to dry PLA filament as it’s being fed to the extruder. Other filaments that can withstand higher temperatures could use shorter lengths. I need about 10 watts/meter to raise the temperature to 100°C and about 10 to 20mL/min of air to purge out the moisture. This is orders of magnitude less than what is needed to dry a whole spool of filament. It is looking to be very practical too - the heated tube is still thin and flexible and can be connected up the same way as regular PTFE tubing is used to feed filament to a printer.
At the moment I’m trying to determine the water vapor partition coefficients for some of the filaments I have to see if I can get away without using desiccants. Using the flow of air under reduced pressure is another option to investigate.
Welcome back to the discussion, @AndyT Glad you didn’t give up so easily, as I think you definitely have something to add to the discussion. As I pointed out in an earlier post, if temp is constant, then the relationship between AH and RH is fixed, right? So the whole argument about whether it is better to use AH or RH is kind of a red herring if you are using RH at a fixed temperature.
Anyway, I think your idea has merit. It will be interesting to see how your experiments turn out. How are you actually measuring AH? I think most of us just use RH because RH meters are cheap and easy to find. I personally don’t have much of an issue with PLA, but I do need to dry my PETG-HF to get good results. I have a couple of dryer boxes and it has always concerned me that there is really no place for the moisture to go other than the increased capacity of the heated air to hold water, although I have been able to get reasonable results. So I have been monitoring these discussions mostly for my scientific curiosity (an old computer engineer here) and enjoy the technical discussions.
Thank you for the kind words. in previous posts I think we’ve got rather hung-up on the distinction between RH and AH. In order to determine how successful a drying strategy is, I think it’s important to be able to quantify the rate of extraction - which needs some measure of the weight of water being extracted - hence I need the AH of the air eluting from the heated tube. In the system I’m putting together, I’m able to quantify the extracted water in real time and change the applied conditions accordingly.
Also because this air is moving around through pipes and pumps externally, it’s temperature will be changing as so will the RH even though the concentration hasn’t changed.
Note that if you are drying your PETG at 65°C, a 3°C change in temperature will cause a change of almost 11% in the RH reading but only 0.8% in the AH reading. I’d be surprised in the temperature gradients in your dryer will be as good as 3°C.
The AH can be calculated from the RH and the temperature using the equations I posted earlier and that’s what I do.
Yesterday, I improved the insulation around the heated tube and got to 160°C just using a 12V, 1A power supply. I might be able to run this system off batteries
@AndyT
Maybe post some photos? A few clicks or something whenever you feel your build has made forward progress. It would make it easier to understand.
That’s because your trust level is “new”. In order to become more trusted, message user discobot (that’s a bot that does a tutorial for discourse) and do discobot’s tutorials. That will up your trust level quite fast. (I’ve been to multiple discourse forums, it’s always the same way, just gotta up your trust level and it’s grand)
For anyone who might want to play around with using a vacuum, but who lacks a vacuum pump, I was earlier today reminded of a possible poor man’s way to create a fairly strong vacuum… And it’s something I think we probably all know, and yet we sometimes forget or it simply doesn’t bubble up to consciousness when discussing vacuums in some other context. For sure, anyone who has done canning would doubtless be aware of it.
In short, I had 1.1kg of silica gel desiccant that I thoroughly dried in my blast oven at 140C. I then transferred it to this gasket air-sealed metal container so that it could cool down to ambient while it remained dry and absorbed only the small amount of humidity in the sealed “Evolv” container:
As it turns out, my garage is close to freezing thanks to the currently active polar vortex, so cooling to ambient meant cooling down close to 0C. As a result of the cooling process, a strong vacuum was formed–so much so that it was impossible for me to remove the lid using only my hands.
I’ll be adding a small metal ball valve so that the next time this happens I can release the vacuum that way.
Maybe (?) there would be some value in this method as a low-effort polishing phase at the end of drying filament. since the filament wold already have been heated up during the bulk drying phase.
Further to my earlier post on possible temperature gradients in something like a Sunlu dryer causing erroneous RH readings. I carried out a few filament heating tests in my Sunlu V2 dryer. I used a new spool of SainSmart TPU for this test as I could easily push a temperature probe into the coil windings. Using freshly calibrated thermocouple probes I inserted one into the center of the filament coil from the side and put the other in the air just above the coil. The starting temperature inside the coil was 16°C. I powered up the dryer and set it to 65°C.
The results were a big surprise…
The Sunlu temperature readout reached 65°C after 2.5 hours.
The internal temperature slowly increased until it reached a stable reading of 59.6°C after 7 hours.
At this steady state the air temperature above the coil was 68°C
I checked the temperature of the air gap between the spool and the front wall of the dryer (where the humidity sensor is located) and read 54°C
I did the same for the gap at the back and read 51°C.
A 1kg spool of filament will have a high thermal capacity. I calculated that this would be around 95000 Joules. There’s other stuff to heat as well. And there will be large heat losses through the dryer walls. Finally, the thermal conductivity of TPU is low so it will take time for heat to pass into the interior of the coil. This is all difficult to accurately model but I’m not surprised that it took 7 hours to equilibrate. I feel that this means that just the outer coils of the filament coil will be heated to the set temperature and be dried. The bulk of the filament will not reach the set temperature (except for long prints) but will continue to release moisture which will inhibit drying of the surface filament.
The second concern is the very wide variation in the internal temperature inside the dryer - even when at a steady state. It appears that the humidity sensor is at least 10°C below the setpoint. This would represent a 60% increase in the apparent RH reading. These gradients also mean that most of the filament may never reach the temperature set point expect when reaching the surface.
All this indicates that these dryers may not be working quite as we expect.
The new parts about to arrive and I’ll continue building the inline dryer I’m proposing, which I hope, won’t suffer from any of these effects.
My dryer doesn’t have any fans. You’re right, a fan would help a lot. What’s really surprising is the amount of heat lost through the walls and the amount of heat (and time) needed to heat a whole spool of filament.
Yeah, that’s one of the things that got me interested in vacuum as well. What I recall from some survey reading was that the HVAC guys pull a strong vacuum of around, say, 20 microns, and then turn off the pump and watch their gauges. If it creeps up to around 200 microns (I’m forgetting the number, but something like that) and kinda stabilizes at that level, then it indicates there’s still water in the system that needs to be pumped out. In contrast to that, if the pressure just keeps rising and rising, then they know they have a leak somewhere.
[Edit: Only if the pressure remains low after turning off the pump and waiting awhile do they know that they’ve removed enough moisture, that the system doesn’t leak, and that the system is ready to receive refrigerant. So, I look at that and think: gee, that’s pretty cool that they’ve devised a way to assess moisture levels by looking at a vacuum measurement. Now, whether the same kind of approach would work for 3D printing filament, or not, I don’t know. Maybe? But I can also imagine that maybe it wouldn’t, especially depending on the filament. Maybe somebody reading this has already tried the experiment or otherwise knows the answer? If so, please post.]
Your right - fans and a higher wattage heater would certainly reduce the gradients in the air and speed up the heating rate of the air. However, I don’t think it will do much to speed up the time it takes to heat a 1kg spool of filament. There’s so much thermal mass and low thermal conductivity that will work against this.
As water emerges from the filament it will turn into vapor and so reduce the vacuum. 1kg filament containing 10g water would produce ~14 liters of water vapor at an elevated temperature. In a sealed container you may end up with a higher pressure than when you started.
The vacuum needs to be maintained - either by leaving the pump running or run it at frequent intervals.
I’m not sure what pressure units you’re using - normally the following is used Pa, kPa, psi, torr, atm, bar, mm Hg, in. Hg, in. water, etc. I’ve not heard of microns.
In the end, it doesn’t really matter what unit of measure you pick, because they’re all easily convertible back and forth into one another. In this context I think microns was picked to avoid using lots of decimals places or scientific notation, so it’s just a convenience, and perhaps the choice varies by trade.
A micron is just 1 millionth of a meter. You see it used alot in liquid filter ratings. 100 micron filter would filter everything larger than 100 microns.
But yeah, to pull water out you would need to maintain the vacuum. On cars, you just leave the pump going for a long time. Then you shut it off eventually and try and hold the vacuum.
The reason why I veered away from the vacuum approach was that some amount of the moisture you extract via vacuum finds its way into the vacuum pump oil. So, if doing that enough and for long enough, you eventually have to change the vacuum pump oil just to get rid of the moisture that accumulates in it. How often does that need to happen? Good question! I don’t know, and nothing in my reading suggested an answer. But changing vacuum pump oil seems like a lot more work than just changing and recharging desiccant, like in @Mzip’s filament dryer.
Maybe if it were possible to insert some kind of desiccant (like maybe molecular sieves or perhaps dryerite) between the vacuum pump and the filament vacuum chamber, then maybe the extracted moisture wouldn’t pollute the vacuum pump oil as quickly, or maybe at all? Well, that would be interesting, and perhaps compelling if it meant the filament could be dried much more quickly. But that’s a lot of if’s, and it’s more questions than I have answers. But maybe somebody reading this knows.
