Extruder fan specification ?  

Re: Extruder fan specification ?

@Paul
Thanks for the data Paul, I finally dug up my Multi-meter to test brake temps but your method makes the point well enough. Can you clarify your data a little though? You have 5 data sets, I am hazy on a couple things tested.
1. Can I assume you are testing a Noctua NF-A40x10 FLX 12V?
Yes I am sorry it was late and I was just trying to get the formatting to work lol
2. E3D Fan, is it the stock 30x10 fan?
Yes the stock fan that came with the printer
3. When you say Stator Vanes, can I assume you mounted this to the Noctua 40x10?
Yes I did it as the video recommended with the fan pulling in the ambient air and blowing it towards the heatsink (in the video it was bags and bits of paper) The video made it seem like it made a large difference but it was also an 80mm fan they demonstrated. What surprised me was that it was actually the worse of all the cooling solutions I tried. In my feeling of the air I think the vanes are just too big and too much air is actually pushed back out the back of the fan.
4. Can you give rough estimate of ambient temperature in °C?
It usually bounced between 23 and 24 C Since I do not have an enclosure and my heating is through hot water baseboards that is about as consistent as I can get. These are not scientific tests but it gives us a good overall performance.
5. When cooling air only, do you have strong underwear to hold your brass balls?
Well I needed a base performance of the heater. It was actually the last test I did because I was curious about something which was it seemed to me at 250 going to 225 that we are running things so hot that the overall environment seems to do more for the cooling then the Noctua Fan which seems to be the case. Since the fan is more for keeping temperatures consistent at a particular point (If you want to print at 225 the heater turns on and off based on the registered temperature and the fans job is to get it low quickly) and that is also why enclosures are important for high temperature printing. The moment the heater turns off it drops quickly.

The question is can the Noctua fan cool quickly enough at various temperatures to ensure overall close to consistent temperatures and that is something I think I need to test further. Based on my limited printing experience with the fan so far the answer is 'Maybe?' as it only has to do a few degrees and I didn't see temperatures run too high but it did I notice tend to run hotter. But then I also never bothered to watch the temperature range as closely as I am right now so I don't have a dataset to back this feeling up.

But based on the fact Prusa has switched to this fan they must feel it up to the task. Granted they have the fan blowing directly on the heatsink where we need to use a tube but it is still doing the job with all the same specs. Needless to say come tax time I think I will be getting an upgrade kit lol. I really want to know how the 2.5/3 systems do with a similar test just to see how much the ducts make a difference.

If you can answer #4, I'll try to match your ambient with AC or room heater and run same test on my NF-A40x20 FLX 12V unless you already have one on order. Unfortunately I lost my stock heatsink fan so can't do any apples-to-apples comparisons but I can see if pull-config will damage Noctua. Noctua support stated max operating temperature was 160° but they didn't specify if that was C or F.

Update: Noctua got back to me, max op is +70°C with nominal max lifespan at +40°C so that will be our range.

Great thanks! I have one on order but it will be a while before I get it. Amazon is currently backordered. In the meantime I guess I will keep using the 10. I don't feel like really splicing in the Prusa fan again.

Re: Extruder fan specification ?

hi guys been all through this fan spec stuff .. best I could find in a 40x40 for the hotend cooling is the san ace 109P0412H901 which has slightly higher cfm and better static pressure in mm/hg than the original 30mm fan but at only slightly higher noise than the noctua.. for my parts cooling fan I originally used a multicomp MC36285 (you'll have to see my design for cooling to see how i made the 40x40x20 fit) this fan would only run at 100% or off.. in fact anything under 225 out of 254 on the fan scale even with the 'blip to max" option checked and it wouldn't start at all.. you need a fan with a low startup voltage, preferably in the 4-13.6v range to be able to use the full 0-100% choices of fan speeds.. I bought EB40201S2-000U-999 from RS components and that now gives me 8cfm parts cooling and will spin reliably from 10% to 100%.

here's the link to my mods

https://www.thingiverse.com/thing:2297721

Re: Extruder fan specification ?

@Gary
Great hunting Gary! I was going to create or source a V to PWM convertor but looks like you've found a fan to work with mini-RAMBo out of the box! At ~$4 USD a pop, I'll grab a couple since Sunon don't last as long as Noctua. Good mods too.

Re: Extruder fan specification ?

I've manually tested that fan on my dual parts cooling fan setup from my thingiverse page using pronterface to send individual fan speed commands and it will spin from as low as 25 all the way to 254 or whatever the max was.. not sure why you'd want it that low but it works.. at 100% speed because its a 20mm fan not 10mm it shifts some serious air, (although i have seen some pages quoting it at 13.6cfm which isn't correct, its 7.7cfm)..not entirely quiet at that speed but again with the larger blade height its a the low whoosh of the air moving rather than the high pitched whine of the fan motor.. I run my printer all night on a regular basis in my house and its all but inaudible.

Re: Extruder fan specification ?

@Paul
OK, so I've tested my fans. It would have been great to test stock fan since my numbers differ from yours slightly. I've lined the inner walls near heater block with some gold leaf insulation tape for automobiles as well have E3D silicone sock on. I've had stock nozzle melt and crash a 20+ hour print so, once burned twice shy and all. Otherwise, I've matched your ambient best I can to 23°C ~ 74°F with my room heater.

My operating theater.

I think we have thoroughly debunked the reverse fan performance gain with empirical data save for one fringe case. Not only that, actual testing shows heatsink is alarmingly hotter when in "pull" (aka suck) orientation.

@All
Here is a comparison of push and pull of all my fans.

If you noticed that the MK2.5's Noctua NF-A40x10 5V is the best performer, you might want to consider this.

Thank you Paul for the stock fan and passive cooling data, I've added my findings in a CSV format below. The primary data is in centigrade seconds recorded in gap form (lap timing).
°C No Fan Stock Push NF-A40x10 Push NF-A40x10 Pull NF-A40x10 Stator Vanes NF-A40x10 (5V) Push NF-A40x10 (5V) Pull NF-A40x20 Push NF-A40x20 Pull NF-A40x10 Push NF-A40x10 Pull NF-A40x10 (5V LNA12) Push NF-A40x10 (5V LNA12) Pull
50 336.7 158.2 202.6 199.1 199.7 172.6 182.2 205.8 212.9 220.3 219.5 223.5 217.7
75 167.3 96.8 124.7 117.9 122.7 100.1 109.2 120.4 124.1 127.6 129.3 129.4 126
100 106.6 66.5 85.2 84.6 86.2 71.9 76.5 87.2 88.9 89.7 91.5 89.5 89.4
125 76.1 51.5 66 66.9 65 54.9 60.9 63.8 66.7 68.7 69.8 68.7 68.6
150 57.9 41.3 49.8 52.9 50.9 44.5 47.9 51.2 53.7 53.6 54.8 53.7 55.9
175 49.2 35.8 44.1 43.9 45.5 36.5 41.3 43.7 45.6 45.7 46.9 45.6 45.8
200 38.3 28.7 34.4 33.3 35.1 29.7 32.1 34.1 35.3 35.4 35.5 36.6 35.4
225 32.1 24.4 32.8 32.9 35.9 31.4 33.5 34.3 35.1 35 37.5 35.9 36.2
250 0 0 0 0 0 0 0 0 0 0 0 0 0

Sum Time 14:24.2 08:23.2 10:39.6 10:31.5 10:41.0 09:01.6 09:43.6 10:40.5 11:02.3 11:16.0 11:24.8 11:22.9 11:15.0
Recorder PS PS PS PS PS SP SP SP SP SP SP SP SP


The conclusion for myself is that NF-A40x20 12V was not entirely necessary. Although I still feel the 5°C delta in exhaust of 20-thick vs 10-thick Noctua means more heat is being pumped out (mitigating thermal runnaway) that isn't captured by this testing method, I must concede all the other corrections such as designing my own parametric adapter duct vs. random Thingiverse part, expensive gold insulation in the block well, my own nozzle designed with focused flow well away from hotend vs. stock nozzle and always running my part fan even for PETG or ABS may have been far more effective than the fan alone. As for the recommendation to reverse the heatsink fan, I wouldn't.

....although (pregnant pause later) the more direct approach would be to measure if brake's cool zone temperatures are different between fans when at 250C? Nah, probably the same...or are they...to be continued.

Re: Extruder fan specification ?

@Gary
Great hunting Gary! I was going to create or source a V to PWM convertor but looks like you've found a fan to work with mini-RAMBo out of the box! At ~$4 USD a pop, I'll grab a couple since Sunon don't last as long as Noctua. Good mods too.

Thanks... i did some testing on cooldown times from 200°C to 50°C (might be posted earlier in this thread) and even the duct can make a difference. I tested 3 different ducts and the one in my mods was a full minute shorter than the next closest. I'm very happy with the setup i've ended up with.. the dual nozzle fan gives great cooling with the 20mm fan plus I can see the nozzle clearly which i personally like as compared to trying to look under the original parts cooling duct. oh and with the 4 bearing carriage I've never again had a layer shift like i was getting with the original 3 bearing cable tie setup.

Re: Extruder fan specification ?

@Paul
OK, so I've tested my fans. It would have been great to test stock fan since my numbers differ from yours slightly. I've lined the inner walls near heater block with some gold leaf insulation tape for automobiles as well have E3D silicone sock on. I've had stock nozzle melt and crash a 20+ hour print so, once burned twice shy and all. Otherwise, I've matched your ambient best I can to 23°C ~ 74°F with my room heater.

My operating theater.

I think we have thoroughly debunked the reverse fan performance gain with empirical data save for one fringe case. Not only that, actual testing shows heatsink is alarmingly hotter when in "pull" (aka suck) orientation.

@All
Here is a comparison of push and pull of all my fans.

If you noticed that the MK2.5's Noctua NF-A40x10 5V is the best performer, you might want to consider this.

Thank you Paul for the stock fan and passive cooling data, I've added my findings in a CSV format below. The primary data is in centigrade seconds recorded in gap form (lap timing).
°C No Fan Stock Push NF-A40x10 Push NF-A40x10 Pull NF-A40x10 Stator Vanes NF-A40x10 (5V) Push NF-A40x10 (5V) Pull NF-A40x20 Push NF-A40x20 Pull NF-A40x10 Push NF-A40x10 Pull NF-A40x10 (5V LNA12) Push NF-A40x10 (5V LNA12) Pull
50 336.7 158.2 202.6 199.1 199.7 172.6 182.2 205.8 212.9 220.3 219.5 223.5 217.7
75 167.3 96.8 124.7 117.9 122.7 100.1 109.2 120.4 124.1 127.6 129.3 129.4 126
100 106.6 66.5 85.2 84.6 86.2 71.9 76.5 87.2 88.9 89.7 91.5 89.5 89.4
125 76.1 51.5 66 66.9 65 54.9 60.9 63.8 66.7 68.7 69.8 68.7 68.6
150 57.9 41.3 49.8 52.9 50.9 44.5 47.9 51.2 53.7 53.6 54.8 53.7 55.9
175 49.2 35.8 44.1 43.9 45.5 36.5 41.3 43.7 45.6 45.7 46.9 45.6 45.8
200 38.3 28.7 34.4 33.3 35.1 29.7 32.1 34.1 35.3 35.4 35.5 36.6 35.4
225 32.1 24.4 32.8 32.9 35.9 31.4 33.5 34.3 35.1 35 37.5 35.9 36.2
250 0 0 0 0 0 0 0 0 0 0 0 0 0

Sum Time 14:24.2 08:23.2 10:39.6 10:31.5 10:41.0 09:01.6 09:43.6 10:40.5 11:02.3 11:16.0 11:24.8 11:22.9 11:15.0
Recorder PS PS PS PS PS SP SP SP SP SP SP SP SP


The conclusion for myself is that NF-A40x20 12V was not entirely necessary. Although I still feel the 5°C delta in exhaust of 20-thick vs 10-thick Noctua means more heat is being pumped out (mitigating thermal runnaway) that isn't captured by this testing method, I must concede all the other corrections such as designing my own parametric adapter duct vs. random Thingiverse part, expensive gold insulation in the block well, my own nozzle designed with focused flow well away from hotend vs. stock nozzle and always running my part fan even for PETG or ABS may have been far more effective than the fan alone. As for the recommendation to reverse the heatsink fan, I wouldn't.

....although (pregnant pause later) the more direct approach would be to measure if brake's cool zone temperatures are different between fans when at 250C? Nah, probably the same...or are they...to be continued.

Nice job! Thank you. I am surprised the 20mm didn't fair any better then the 10. I will liekly just do one of the official upgrades in the near future still on the fence if I want to do the full Mk3 or not. But this certainly shows the 5V fan works very similar to the old fan so I can now see why Prusa was ok going in that direction. Of course now I got a 10mm 12v fan with no idea what to use it on lol. I think I will return the 20mm when I get it.

Re: Extruder fan specification ?

Just got my Sunon 40x40x20 fan I will give that a shot.

Re: Extruder fan specification ?

@Gary and Paul
I got Sunon EB40201S2-000U-999 today and plugged it into my printer right away...after attaching standard 3 pin connector of course 😉

The 2 pin Sunon is oddly much more of a flow fan than expected. It exhibits a lot of cast-out (what I call air that gets ejected due to ineffective suction into thrust volume) when confronted with any resistance such as a heatsync fin array. I'll test this guy same as others and post a new chart comparing Noctua 40x20 against Sunon but I think Noctua will win because it has better fin gap, more lower AOA props and a better tuned thrust volume (aka duct thickness). Sunon is only slightly louder than Noctua but has noticible vibration increase though not enough to shake or make audible noise on attached parts. Unfortunately I'll have to wait because I snapped my thermal probe trying to test brake temperature cool down behavior, still mad at myself for being lazy and not removing PTFE liner prior to testing.

And though this is the wrong thread for this, and I'll post there as well, the Sunon may be OK as a non-PWM mod way to get axial part cooling fan. Again, Sunon's cast-out is concerning but as the two pin Sunon can kick over as low as Octet 26 (sorry I thought 255 was a hex thing) or 11% fan speed to full power, this could be a great part cooler.

The specs and where I bought it is here.

Re: Extruder fan specification ?

OK, this might be the last test I do but as I found my old stock fan and with new 2 wire Sunon 40x20 in hand, I ran the tests with my set-up to be apple-to-apples. As attempts to do in-heatbrake tests went horribly wrong, I'm going to be content with our previous yet clear method.
Again, I have an E3D sock and gold lined insulation on the heatblock well's plastic walls. Also, it may not be a fair sample of what Sunon can offer in terms of cooling as the two wire 40x20 was actually purchased as a axial fan alternative for part cooling purposes due to its ability to sustain constant RPM as low as 11% relieving the need for PWM conversion.

And the charts. Bigger this time so the older folk can read without spectacles. 😉

No fan cooldown is off the chart "literally" reaching 50°C from 250°C in 19:18.6

Note that the no-fan heatbrake temperature continued to rise as thermal runaway was taking hold. I lost my nerve and stopped as soon as the thermometer flicked down a degree. I'm not willing to melt my printer to know how bad it can get.

And for old times sake, more evidence reversed fan not a good idea.

CSV below.
°C,No Fan,Stock 30x10 12V,Noctua NF-A4x10 5V FLX,Noctua NF-A4x20 12V FLX,Noctua NF-A4x10 12V FLX,Noctua NF-A4x10 5V FLX with LNA12,Sunon EB40201S2-000U-999
50,484.7,148,172.6,205.8,220.3,223.5,195.8
75,231.8,86,100.1,120.4,127.6,129.4,112.2
100,134.9,62.8,71.9,87.2,89.7,89.5,78
125,94.9,48.1,54.9,63.8,68.7,68.7,59.4
150,71,39.9,44.5,51.2,53.6,53.7,48.1
175,55.9,33.4,36.5,43.7,45.7,45.6,39.1
200,43.5,26.2,29.7,34.1,35.4,36.6,32.3
225,41.9,28.1,31.4,34.3,35,35.9,32.4
250,0,0,0,0,0,0,0


Re: Extruder fan specification ?

I've got the Noctua 40 5V version on the way. Can some explain how I hook this up? I know the stock fan is a 12V version. I read I need a capacitor and a resistor?

Re: Extruder fan specification ?

Sander, if you're asking that question, you may be in for a bit of a hassle. But before any of that, the 5V isn't designed for a 12V line and since none of us getting the MK2.5 have gotten our packages, I believe the 5V will be heavily overvolted potentially reducing lifespan to months than years. What I have found is that the stock E3D fan which is obviously made in China by a sub-vendor has about a year life span, you heard me right, a year. So, maybe Prusa feel a high end fan rated for 3-5 years of life can be overvolted and last a year at least. We won't know until 2.5 lands.

Before I get to connectors, you still have an option. If you've purchased other Noctua fans you may have a free LNA (low noise adapter). This is basically a resistor, there is a 10 version and a 12 version. Either one will make the fan much safer to install.

As for connecting you have two choices.
1. splice and tape/shrink wrap.
2. get hard to source inline Molex KK style connector in two pin or three pin configuration as well as uncrimped pins and live with a big hump in your carriage cable bundle. Oh, right, you'll want a crimp tool or be amazingly skilled with needle nose pliers. Sorry.

I hope that helps.

Re: Extruder fan specification ?

Then I completely misunderstood. In the charts it was mentioned the 5V was the best performer. So, I went for that. I didn't realise that was only for those who are going with the MK2.5 upgrade. I'll reorder the 12V version.

Re: Extruder fan specification ?

This is right Sander, the 5V is here because it is part of MK2.5 kit, and 2.5 doesn't change the mainboard. As our mainboard is a mini-RAMBo with 12V fan lines (for the most part), MK2.5 owners will be covered under warranty. You will not. And if you've read Noctua's warning on such use "Being a 5V fan, the NF-A4x10 5V can not be run at 12V and will be damaged when used with typical 12V power sources such as PC mainboard fan headers." I'd say be safe and get a 12V. You can read the warning yourself here.

I've emailed support to forward my question to dev team to get answers and will update here on why there is an odd conflict of opinion on 5V use between the two companies.

The charts only indicate performance for the 30-45 minute test run per fan. This means the 5V fan isn't going to instantly blow up, but that fan isn't supposed to be that fast, and isn't supposed to be that loud. Then the warning from Noctua that wasn't there 3 months ago and well, WTF, YOLO, YMMV.

I think you've made the right call until we get clarity on what is going on.