Filament Dehumidifier

[jono035 (Jono)]

@bitingmidge

the silica can’t both remove water and put it back at the same time (dry itself) … or can it?

No, not at the same time. The Silica gel just absorbs/releases water relative to its temperature and the partial pressure of water in the air around it.

When we use it as a dessicant we’re specifically exploiting the fact that it holds very little moisture when hot in order to dry it and then that it will absorb moisture from even very low RH air when it’s cool.

The systems that generate a constant supply of dry air via dessicants use multiple chambers so that one can be regenerating while another does the drying. That’s basically all we’re doing when we pull the dessicant out of the chamber to regenerate it manually, just that it’s in-line and automatic in those systems.

In this case of conventional filament dryers we can use a single chamber because the moisture absorption/removal process for the filament is so slow that we spending a couple of hours dumping moisture out of the dessicant into the airflow around the filament doesn’t do anything appreciable to the filament’s moisture content.

[jamiek (Jamie)]

@bitingmidge

silica can’t both remove water and put it back at the same time

You are completely correct.

My thought is that this would be two stages, one stage is drying desiccant like a food dehydrator. The other stage is using the desiccant.

@jamiek

image430×628 23.4 KB

(Yellow heat exchanger block is optional and might improve energy efficiency by retaining some of the heat when dumping the humid air.)

Then using the desiccant would be blowing air through the desiccant into one or more dry boxes:

@jamiek

In this desiccant mode, return hoses from the dry boxes would be much better than drawing in ambient air because the desiccant will last much longer before getting loaded with moisture.

Currently it would be a manual switch/reconfiguration between regenerating the desiccant and using it.

Everything else is probably overkill and more of an experimental concept to see if I can make ultra-low RH with multiple stages or something.

The picture above where I have 20% RH and 54% RH would be the usage stage, and shows that when you blow air through desiccant, the air becomes dry.

---

I tried using the heated bed of my printer to run the regeneration cycle. I have some old silica gel packets that are about 44% RH if I seal them in a bag with a sensor, and I thought I might try to dry them.

It appeared to be completely ineffective, and I think the reason was that the fan was too weak and the impedance of the bottom heater module was too high. The channels were long and thin, so there was not enough airflow and it didn’t do anything.

I redesigned the blower to use dual 5015 radial fans, and I made the bottom heater block with fatter, shorter channels.

I just turned it on again and I don’t know yet if it is working…

The green segment contains the old silica gel. The blower is the pink/purple segment with the wires coming out.

[jamiek (Jamie)]

Well, I think I am being forced to conclude that this is a failure.

I think there are still some good concepts, but specifically, going for 100% printed (in PLA to start with) and using the printer heat bed as heat source is not good. Either the bed is anemic or the heat transfer to the air is extremely poor, and the hot air source is not nearly high enough. Then on top of that, a huge amount of heat is being lost before it can do anything useful.

I let the system run for a couple hours today, and apart from a slightly lower ambient humidity, the result is essentially the same as the last picture.

The inlet is about 47% RH at 72F.
After passing the heat module it’s 18% RH at 103F.
Then after the silica gel it’s 32% RH at 88F.

The inlet is at A. After heating it’s at B and since it is being purely heated it should (and does) move horizontally. The vertical axis is absolute water content, which doesn’t change from A to B even though relative humidity does change.

From B to C the temperature decreases and the absolute humidity does increase slightly, meaning it is drying the silica gel a bit, or maybe it’s doing nothing and it’s within the margin of error of the sensors.

But either way, I feel like it is not nearly hot enough to be effective. Even when I remove the silica gel and run an empty plastic chamber, the temperature is barely any higher, so it seems the heat loss is really bad.

It doesn’t do nothing, and this old silica gel is now at 33% RH instead of 40 something, but I’m still saying it’s a failure.

On my previous iteration I cranked the heat bed temp up to 65 and it started to melt the PLA, so I kept the bed at 50C this time.

I think a dedicated heater is the way to go, and it needs a controller, and it can’t be made from PLA, which should have been obvious from the start.

I’ll have to rethink the consequences for the requirements. If the system is standalone with sensors and a controller, then maybe it can be fully automatic for continuous use with not much more investment. (Continuous use would mean redirecting airflow and cycling itself to regenerate the desiccant.)

[jono035 (Jono)]

I think you’re correct that the temperature swing needs to be higher. The missing part of that graph is time as it’s an equillibrium state over long term.

I also wouldn’t trust those humidity sensors to be better than +/- 10%, maybe worse at the low end. The sensors themselves are hideously non-linear and usually just roughly curve-fit to be ‘right-ish’ at the usual operating humidity ranges.

[Notnewton (Mark)]

Just to through out another train of thought(not to derail). Does anyone have a vacuum chamber for epoxy or something? Couldn’t you dehydtate just by leaving the filament under vacuum for a while? It should do the same as the dessicant packs/beads? Thoughts?

[MakerJim]

@Notnewton

Couldn’t you dehydtate just by leaving the filament under vacuum for a while?

Interesting thought. What else in the filament outgasses under a vacuum?
I bet a lot of the materials in the filament would not like being in vacuum.

Maybe pulling a vacuum on the desiccant store as part of a desiccant bead water release cycle would be useful.

[jamiek (Jamie)]

Vacuum is interesting. Based only on partial pressure, putting the filament into a perfect vacuum might be essentially equivalent to putting the filament into a chamber with perfectly dry air at 0.0% relative humidity. With the perfectly dry air, the filament will release some moisture into the air until it reaches a balance at whatever percent RH, and I would guess that a vacuum would release the same quantity of water and reach the same partial pressure of H2O, but just without the nitrogen and oxygen.

If you turn on the vacuum pump every now and then, perhaps you can pump out the water vapor that the filament has emitted, assuming the partial pressure of H2O is high enough and the pump is good enough. If the vacuum pump can’t pump down low enough, maybe you could let in a bit of air and pump it down again. A sort of gas rinse so to speak.

I think Jim might be on to something using the vacuum to regenerate desiccant, because you could extract a lot more water from a chamber full of silica gel than from a chamber full of filament. If your silica gel is at equilibrium with 30% RH, then that’s 150g H2O per kg silica gel, whereas 1 kg of filament will have only a few grams of water. It would be more effective to charge up silica gel in the vacuum and use the silica gel to dry the filament.

Something else that I’ve sorta been avoiding, because it’s complicated, is the effect of temperature. Yes, increased temperature lowers the RH for a given absolute quantity of water vapor, and that reduction in RH in turn helps to dry stuff. But I think for some materials there is an additional temperature effect beyond just the RH reduction, meaning you could leave the filament indefinitely in 0.0% RH (or in a perfect vacuum) and some water would not be released, but when heated, that water does get released. I think this happens with Nylon for example.

So far, my efforts have been pretending this heating effect does not exist, but a complete solution would have to address it.

[barry99705 (Barry)]

The best way to dry the filament would be get it hot, then pull a vacuum, but most home gamers aren’t going to have the equipment for that, so we just need to go with heat, then store in a dry box. From everything I’ve read, you can’t dry filament with desiccant alone.

[bitingmidge (Peter H)]

@jamiek

I’ll have to rethink the consequences for the requirements.

I’ve been doing that constantly for years, and look where it’s got me! I do like the principal though.

@jono035

I also wouldn’t trust those humidity sensors to be better than +/- 10%

I think that’s a reasonable margin for what we are doing. They are a guide. I am surprised that some “content creators” demonstrate terrible outcomes for “wet” filaments that I can only reproduce on the cheap stuff (PLA), (which I mostly don’t use). Even the PETG stored in my dryboxes at 50%+RH for six or eight months is still printing perfectly, having been correctly dried initially.

The PLA which was dried with inconsistent heat in the Eibos tends to go brittle on the outer winding where it was overheated.

I am still watching @jamiek’s progress with great interest.

@barry99705

From everything I’ve read, you can’t dry filament with desiccant alone.

I have read the same, and on the one hand it doesn’t make sense , given enough time surely the moisture will reach equilibrium, on the other neither does the need to heat filaments at different temperatures - my brain says if nylon can absorb moisture at room temperature, surely it can release it at room temperature too, but the science says differently apparently.

[MakerJim]

Does anyone know of a decent, yet affordable vacuum pump to do some experimentation with?

[jamiek (Jamie)]

@bitingmidge

@barry99705

From everything I’ve read, you can’t dry filament with desiccant alone.

I have read the same

Wut? I had not heard this. This sorta blows the premise of my intended design, unless it is just some materials that require heat to be dried, and other materials are okay simply with low RH.

If there is a material with a high equilibrium water content at room temperature, even at low RH, then I think that means you might have a relationship like this:

Where blue curve is at room temperature, and even very low RH doesn’t get fully dry, and red is at elevated temperature, where drying is possible. These curves are made up of course.

In this scenario you are going to have a very hard time storing the filament, and even dry desiccant at say 10% RH could make it worse. You pretty much always need to cook it if it has been a while. Maybe this is where the inline heater/dryer is needed to dry a few hundred mm of filament just before it goes into the extruder.

[bitingmidge (Peter H)]

My dry storage concept (not drying) has been to take an My “project” in my head goes something like this - (for storage, NOT for drying) has been:-

Ikea FABRIKÖR glass cabinet with the door appropriately sealed - I have thought of using insulated panels instead of the glass and cobbling a double glazed door, but in our climate I don’t think the heat loss is worth worrying about.

Then Stick in
El-cheapo compressor dehumidifier

• Rated Input Power: 220W (27°C 60%)
• Max. Rated Input Power: 270W (32°C 90%)
• Rated Input Current: 1.35A (27°C 60%)
• Max. Rated Input Current: 1.48A (32°C 90%)

Jamie’s concept has prompted me to think about a simple fan and pre-heater with a return

@jamiek

unless it is just some materials that require heat to be dried, and other materials are okay simply with low RH.

My present conclusions (right or wrong) are that if filament is stored dry ie after drying, then it should not require further treatment. Since “dry” for me is “outrageously humid” for others, this is the source of all of my vexation!

I’ve said it before but my “good” (Prusament) PLA and PETG filament is nowhere near as susceptable to humidity as others. Here’s yesterday’s example - these are tiny train game markers. All except the grey is Prusament PLA - stored in my “dry” boxes at up to 55% humidity, age up to at least three years. The grey is an elcheapo local brand PLA, well tuned, received and dried a week ago but left out in 85% plus.

I’ve noted that warming before printing often makes a noticable difference to print quality, so the pre-dryer might make some sense even if it’s not drying.

@jono035 did explain somewhere about polymers and water takeup, but I still don’t understand why for instance Nylon needs 5 hours @ 70° and why 20 hours at 40 won’t do the same job.

How many times can you heat-treat filament before it becomes tempered? Some of my PLA got brittle when it was over-exposed to heat from the Eibos, and the last six or so inches directly in the extruder and above it can also be brittle.

Again, a reminder for the kids at home - none of this is strictly necessary - even in extreme humidity things still work.

[jono035 (Jono)]

From memory our vacuum setup for resin degassing was pretty cheap so I imagine there are cheap 2nd hand options around. Looking at Amazon there are options with a vacuum pump in the ~$100 range.

I’ve seen things go past where people have used fridge compressors for vacuum bagging stuff.

A vacuum pump with an old fridge motor

We also briefly looked at venturi vacuum generators but the airflow requirements are pretty up there. I guess you could get around this with a small compressor by manually controlling the valves to drop the pressure slowly, allowing the compressor to recharge in between.

[bitingmidge (Peter H)]

It’s all been done before of course!

Vacuum VS Dehydrator! What's the best way to dry filament?

[Notnewton (Mark)]

Id bet you could minimally heat the ss vacuum chamber above as well by setting it on a hot plate or wrapping it with a heating pad.
I was wondering if the need for heat has to do with overcoming the energy that’s holding/ attracting water molecules in the 1st place. You’re not really evaporating the water with some of these chemicals, but overcoming the energy of attraction.

[jamiek (Jamie)]

@Notnewton

You’re not really evaporating the water with some of these chemicals, but overcoming the energy of attraction.

Yes, well put.

I worry that at even 10% RH, the energy of attraction for nylon for example could draw the water out of the air and leave the air below 10%. Then if you have silica gel at 10%, it could release moisture into the air and effectively the filament draws water out of the silica gel even if it is fairly dry.

[MakerJim]

@Notnewton

You’re not really evaporating the water with some of these chemicals, but overcoming the energy of attraction.

Yes, this. Pulling a vacuum also lowers all of the associated release points so it is a double benefit, except…

@Notnewton

Id bet you could minimally heat the ss vacuum chamber above as well by setting it on a hot plate or wrapping it with a heating pad.

Transferring heat in a vacuum is hard. (I put things up to 10 tonnes into vacuum on purpose to test them).

Conduction is the way, or alternately IR lamps as you can use radiation and conduction in a vacuum but not convection. (Obviously)

@jamiek

hen if you have silica gel at 10%, it could release moisture into the air and effectively the filament draws water out of the silica gel even if it is fairly dry.

This is likely a big part of how the materials are working against us.
Not at all as simple as it seems at first thought.

[bitingmidge (Peter H)]

So 137 posts later, I’m back more or less where I started I think, albeit with a little more thought on dry boxes and how I use them.

For DRYING - sometime in the next week or two (two) I’ll be printing the mods to the Eibos, which should give it much more even drying as long as the PC parts don’t melt and providing I can actually dry the PC well enough. I’ll use the food dehydrator for PLA and PETG because it’s much more gentle, and the the Eibos when I want to dry as I print.

For STORAGE Twenty days on, those two empty boxes are still holding 11 and 17% respectively, so there’s no doubt increasing the amount of silica gel to 200 grams minimum helped a lot. Now understanding that the cardboard spool centres holds as much moisture as it does - I am pretty much convinced that drying the filament before storage rather than before use is the most sensible approach, and as long as I can keep it below 20% only TPU, PC and ASA will need a “touch up” as I go. (I think)

For USE this is an ongoing thing, but PLA stored as above is unlikely to be a problem ever. I may convert my PETG boxes @Tokoloshe style to feed directly from them - and TPU, PC and ASA I will print from the appropriately heated Eibos.

BULK STORAGE my original idea was to take an IKEA Billy unit and modify the door to clamp it against a seal, then add a small compressor style dehumidifier into the cupboard, with an external water waste. Now I am wondering if air was to be recirculated via a fan and filtered through a few kg of silica before dehumidifying if that would even out the drying cycle time.

The only way to test is to test I guess, but for now, time is my enemy, and the cost of doing that will be similar to a (very small) commercial cabinet, so I’d kind of like to know how it’s going to work before I start. I may just buy the machine and run it in our under-stair storage area for a while to get a handle on what it can and can’t do. If I can hold humidity in the cupboard below 20% max, then I’d be on a winner I think.

In the meantime, I refreshed the dessicant in my camera gear box yesterday, and it’s slowly dropping from 82 to 41 and I expect in a day or two it’ll be around 18 - there are a lot of cloth lens bags in there that a smarter person would have dried thoroughly before packing them away

[barry99705 (Barry)]

My modified food dehydrator worked great, even when I was running it in my garage in Florida. I store my filament in a sealed plastic tub, but I only keep about a dozen spools at any one time.

[Crash_CNC]

I’m currently on the search for my holy grail of dry filament storage.

In the past 9 months or so, I’ve been experimenting with a bunch of cereal containers that each hold silica gel balls. They keep the humidity down for a few months then they need to be recharged which is a hassle. My guess is the air pressure variations push moisture past the seals and over-run the desiccant.

At the same time, I tried two 79 quart tubs. Inside these two tubs I have small bags of calcium chloride as the main desiccant but there is also some silica desiccant in there. These two tubs stay as dry as the meter can read even though they are opened more often.

The calcium chloride was $50 for 50 pounds and I don’t think I have used a single pound in the last 9 months. I store the excess in two 5 gallon buckets with watertight screw lids.

I’m currently mocking up a proper container that will allow the spools to hopefully feed a filament changer. It’s looking like I can fit 14 spools into the container if I mount one row on the bottom and the rear row about 3 inches higher.

One thing that am thinking about trying is to take a 5 gallon bucket and turn it into a super-dry-air reservoir with calcium chloride sitting on a mesh bed to allow for any liquid drainage. Near the top of the bucket, I would have a fish tank air pump. The air pump would supply dry air to the filament storage tubs based on the humidity sensor linked earlier in this thread. I was thinking of using pneumatic air fittings like we use on 3d printers for the supply and return lines or maybe just fish-tank tubing and hose barbs.

Thoughts?