The crossflow fan approach

[lars86 42]

Love the concept guys!

I ordered in the FX121 fan for myself.

What have you guys done to power it? Off the UM board? External?

On a side note, is there somewhere I can pick up premade extensions for the stepper motor cables?

 

[nick-foley 5]

FX121 can be powered directly from the stock fan connection. The only drawback is that it doesn't do PWM, so it's all-or-nothing with cooling. Great for PLA, not so much for ABS or PET (unless you have a heated chamber?).

Also been looking for premade stepper extensions... haven't found much. Would really like ones with locking connectors, and twisted pairs would probably be good, though I'm not an EE so I don't know how much difference it would really make.

 

[lars86 42]

Is there a second fan power source on the board? I'm using the primary fan circuit to drive a small actively cooled print head.

 

[duesentrieb 0]

There is no second fan power source on the board. Maybe there are some free Arduino pins so you could add a second transistor and control this fan also with PWM. But I don't know how to implement this in Software.

Another posibility is to add an extern fan control, like mnis did with his modification. You can get these cheap on ebay and control up to 5 fans with them.

Philip

[duesentrieb 0]

Cleaned up all the provisonal wiring, printed out the moulding for MadOverlords Hot End Thermal Isolation Sock ( http://www.thingiverse.com/thing:165750)%20and%20made%20one.%20The%20tip%20useing%20corn%20starch%20for%20faster%20cureing%20of%20the%20silicone%20was%20great.%20Installed%20the%20thermal%20sock%20and%20its%20the%20ideal%20match%20for%20the%20cross%20flow%20fan.

Can run the fan even at full speed now without any effect on the hot end temperature. Before I had to reduce the RPM of the fan to around 20%, because on higher speeds it reduced the hot end temperature so much I got underextrusion, even if I had isolated it with kapton and a wind shield.

Doing a test print with brand new Colorfabb PLA, which arrived today, and the fan at 30%. Looks very promising. Still like the cross flow fan setup very much. Will add better looking side panels and stick with this setup.

 

[jonnybischof 60]

Has anyone ever though about 3D-printing a crossflow fan?

It "should" need no more than a simple DC motor (cheaper than a crossflow fan) and some ball bearing on the other side to hold the rotating cylinder. The rest can be 3D printed to fit the exact dimensions of the printer.

Go, or no go?

I know it's not "that" easy to create a crossflow fan, but I'm sure there's someone around here who can take a look at a design and improve it..

Maybe someone can simulate the airflow? I don't have access to such software... :(

/edit:

Maybe taking a standard fan apart and using its motor could work. I know the crossflow fan is a much bigger load than the normal fan blades, but it might work...

 

[Dim3nsioneer 557]

The price depends on the product you buy. There are cheap crossflow fans, even with AC motor (cheaper but not as efficient as modern brushless dc motors).

Then there are the other products where you get additional functuality such as a pwm line for setting the speed, tacho signal, alarm signal etc.

The blade geometry of cross-flow fans is usually not very high-sophisticated compared to the blade geometry of e.g. axial or radial fans nowadays. So, I think no CFD is needed.

But actual printing of the blades might be quite tricky as they are long and curved.

 

[tommyph1208 55]

Surely it is possible to make something that blows air... (that's the essential part right?)

But its definitely not going to be very silent :/

On another note... I just ordered a DC type crossflow fan, what was the conclusion on how/where to connect them?

Can you drive a fan like that dirctly from the UM board, or do you need a separate circuit with eg. a relay driven by the fan output?

 

[jonnybischof 60]

It should be possible to use the standard fan output (PWM).

But (!!) this is a 19V output (if you use the standard power supply). So make sure your crossflow fan doesn't have a problem wit 19V. Using a 12V fan should work (it also works with the standard fan...), but you are running the fan out of it's spec...

Depending on how much current your crossflow fan draws, you might overheat the PWM transistor (Q4) on the board. Theoretically, it supports up to 4A. But that depends on the cooling, so make sure it doesn't get hot.

I don't know how well the crossflow fan handles the PWM signal. You'll probably have to push the speed setting to 100% to make it start, but once it's spinning it should be possible to lower the speed. Good fan controls "kick start" the fan by delivering 100% for a few seconds until they start regulating the speed.

About a homegrown-crossflow fan:

It's actually not that easy to find a suitable motor. Brushed DC motors aren't useful because they're usually noisy. Brushless DC motors with integrated commutation-logic are rare and expensive.

I have some Tiger brushless motors (multicopter motors) laying around. Guess I'll try one of these. The problem is, I'll need an ESC to control them -.-

Theoretically, one could use a stepper motor, but only if used in closed-loop mode (which can't be done with the Pololu stepper drivers) which again needs external control logic.

Maybe it is actually best to just dismantle a really strong 120mm fan and use it's motor. It would be the cheapest solution.

It should be possible to print the whole fan rotor in one part, with the blades standing vertically. I'm printing with Colorfabb XT at the moment and I believe it wouldn't be a problem to do that print with it.

Most crossflow fans' rotors are divided into several sections. The printed version could as well be split up into sections which can then be stacked on top of each other. It would be very easy to adjust the fan's overall width for whatever uses it may be needed.

/edit:

http://turbomachinery.asmedigitalcollection.asme.org/article.aspx?articleid=1485164

This article about optimisation of fan blade geometry seems very interesting. I don't have a credit card though -.-

Maybe I can get it through my company..

/edit2:

The most important information can already be seen from the previews.. I think I'll give it a try and design something soon. I want to improve the UM1 frame, why not integrate the crossflow fan directly into the frame..

 

[nick-foley 5]

Is anyone using a crossflow fan that can be PWM'd? The Silverstone ones I've been using for the past few months are fantastic at constant full blast for PLA and XT, but I would like to be able to turn them down for using materials which have worse layer adhesion like ABS and PET+.

Suggestions?

 

[Dim3nsioneer 557]

There are crossflow fans from ebmpapst which can be PWM'd (separate PWM input signal). But they are quite pricy.

 

[nick-foley 5]

Hmmm, yeah, mouser has their smallest one at $88/unit... probably a little above what it is worth to me...

Maybe will try a large axial fan...

 

[Dim3nsioneer 557]

You will hardly get a flow as laminar as from a cross flow fan with an axial fan alone. You might try to build a 'laminarizer', i.e. something with a lot of straight channels where the air will flow through (if the axial fan is strong enough).

 

[jonnybischof 60]

Has anyone worked with an ebmpapst QG030 crossflow fan before?

I've just ordered one from Mouser (QG030-198/12) which should be ideally sized (198mm fan width).

They didn't have the 24V version stocked, meh :(

I guess this model doesn't have a PWM input? There is no useful information on the manufacturer's page, and the datasheet doesn't mention anything about the connection of the fan (why bother with such details, anyways? -.-).

It's pricey, but it will probably outlive me so I'm fine with it.

 

[Dim3nsioneer 557]

Nope, the QL series has a PWM input. Check page 26 of the http://www.ebmpapst.com/media/content/info-center/downloads_10/catalogs/epL_Katalog2009_DE.pdffor more details about the QG030.

Life span is 30'000 hours. Should be ok for a start (more than 3 years of continous printing).

 

[IRobertI 520]

You will hardly get a flow as laminar as from a cross flow fan with an axial fan alone. You might try to build a 'laminarizer', i.e. something with a lot of straight channels where the air will flow through (if the axial fan is strong enough).

Like I did on page 2?

http://umforum.ultimaker.com/index.php?/topic/3890-the-crossflow-fan-approach/page-2

 

[Dim3nsioneer 557]

Yep. Like you did a veeeeeery long time ago in January...

 

[jonnybischof 60]

GRRR

I hate it when they use different part number styles in the catalogue than online! I've been on this catalogue, and I've searched it for "QG030" (actual part number) which returned nothing. The catalogue lists it as QG 030.

Thank you for finding it

Sadly, it's only the same half-hearted info as is provided in the datasheet.

The technical drawing is ridiculous. The most important measures are not there. Where is the axis, and where exactly is the air outlet??? People, learn how to make drawings... :(

I can see how to mount the fan in this picture, but I don't see WHERE I have to position it in order to get my airstream where I need it...

I guess I'll also have to make a custom regulation for this fan. Using PWM on non-PWM fans can shorten their lifespan a lot.

 

[Dim3nsioneer 557]

GRRR

I hate it when they use different part number styles in the catalogue than online! I've been on this catalogue, and I've searched it for "QG030" (actual part number) which returned nothing. The catalogue lists it as QG 030.

Thank you for finding it

Sadly, it's only the same half-hearted info as is provided in the datasheet.

The technical drawing is ridiculous. The most important measures are not there. Where is the axis, and where exactly is the air outlet??? People, learn how to make drawings... :(

I can see how to mount the fan in this picture, but I don't see WHERE I have to position it in order to get my airstream where I need it...

I guess I'll also have to make a custom regulation for this fan. Using PWM on non-PWM fans can shorten their lifespan a lot.

 

Maybe page 11 of the same document helps...

Shouldn't be too difficult for you to provide a 8-14V voltage out of the PWM...

You can also go below the 8V. ebmpapst will just no guarantee that the fan runs properly all the time, but if it works, it's fine. As the motor has an overheat-protection you should not be able to destroy it by running too low.

 

[jonnybischof 60]

8V "minimum" is probably the usual "minimum voltage where the fan can startup properly" value. Just kickstart it, then regulate back. Usually, fans work at very low voltages that way.

I've once had a 12V PC fan that actually ran at 2.5V. You could see the blades - it was that slow. But it kept turning and didn't make the slightest noise

I'll have to think about how to regulate it. Sadly, there's not even an RPM feedback on this fan which would've been very useful. I'll see about that. First I need to get it mounted...

/edit:

The "good" ones from ebmpapst cost 180 CHF. That's definitely not an option...

 

[jonnybischof 60]

Just received my QG030

This thing is powerful!!

It generates a lot more static pressure than any computer fan I've seen (and I've seen quite some...).

Also, it starts at 3.3V and is probably more than powerful enough at 5V. 12V makes it roar like hell, that would probably blow the plastic right off the build platform

Anyways, I will try to make it work with PWM but I'll add a filter in order to get a linear voltage to the fan. Will report what happens... (If you don't hear from me again, it blew up...)

 

[Dim3nsioneer 557]

It generates a lot more static pressure than any computer fan I've seen (and I've seen quite some...).

 

What you feel is the large flow, not static pressure. Cross flow fans are known for producing a large flow but with very limited static pressure. Highest static pressure is reached with radial fans followed by diagonal fans and axial fans.

Of course you feel the large flow as long as it can blow more or less freely but if you block e.g. the outlet the output will go down very quickly.

 

[jonnybischof 60]

Isn't a crossflow fan just a "widened" radial fan? The fan geometry certainly is the same...

But yes, put an obstacle at the wrong place and there is no more flow.

/edit:

ok, the air enters a radial fan differently than a crossflow fan...

Well anyways, I'm very happy with the fan so far. Now I just have to put it to use and see how well it plays with the printer..

 

[jonnybischof 60]

I'm in the process of integrating the QG030-198 crossflow fan now, and just wanted to note something:

You can actually take the fan apart and switch the motor from the right to the left side, if necessary. The relevant parts are mirrored, so you just have to pull out some screws, flip the motor and put the screws back.

Might be good to know when selecting a crossflow fan.

I'd also strongly recommend to take measures for vibration dampening, as the fan does create quite some noise & vibrations.

 

[duesentrieb 0]

Just deinstalled my cross flow fan after half a year of useage. For some parts and materials it was great, for others not. I think I will change the wiring later so that I can switch between cross flow and small head mounted fans. For PLA the cross flow fan works. With ABS I had the problem that it cooled down the heated bed too much so parts became loose and the whole print was ruined. Right now I need a lot of ABS parts, so I am not useing it. With vibration and noise I did not have any problems because it offered a lot more cooling than needed so most of the time I ran it with the lowest possible setting to keep it still constantly spinning, around 18 PWM and at that speed it is very quiet.

The downside is, you have a good side and a bad side of a print. The good one is the one faceing the fan, the bad the other one. For some prints it doesn´t matter, for other especially with overhangs it does. It is a good concept but in my opinion not the non plus ultra. Maybe it´s also only my exact setup which needs optimization or higher requirements I have now... don´t know.

I am curious which experiences you will make with this setup, Jonny.