Any fluid dynamics gurus on hand?
May years ago I studied fluids as part of my engineering degree and I have recollections about the properties of different types of fan and their applications, sadly years of neglect has left this knowledge too vague to be useful…
I ask as I’m considering options for a Kraken fan duct; There is pretty much no way you can manufacture a duct that surrounds each nozzle (as you see with single hot end applications), so I’m thinking of a single duct pointed just below the orifice for each nozzle. The 4 ducts would be connected via an appropriate flexible tube to a frame mounted centrifugal blower. The intent would be that the airflow would be quite fast/high pressure but would only cover a very small area below the nozzle so it cools the extrudiate quickly as soon as it exits the nozzle. I assume a blower is the best option for this application as they deal with back pressure better than axial fans.
Would this be viable?
Not a guru, but i think the better option would be to use one larger duct to just kraken mount, then a 4 way junction. the reasoning is that by having 4 small hoses, you will increase the surface area relative to the volume (IE more friction loss).
Wouldn’t high-speed air risk pushing the just-extruded plastic out of line?
A different approach would be to use a very directed flow and a nozzle for each one that actually pulls in a layer of the surrounding air. Look up air knifes for the gist of what I’m talking about.
The other route is if you are wanting to cool just the nozzle and not the plastic with this, perhaps use a blower and create a duct with just the nozzles protruding out.
To answer your first question, yes, you want some sort of fan or blower that supports pressurised flow. The ratings for any reasonable fan should be available from the manufacturer. A small centrifugal fan/blower is a good starting point. Alternatively, you might use an air pump, like for fish tanks, which is what is used for laser air assist.
I think it cools very well if you just have the fan / fans blowing onto the hotend/print. The ceramic heaters are so effective that it will keep the temp up even if the air also hits the hotend. On my std. E3D i have to 30mm fans attached on each side of the hotend cooler duct, blowing down onto the hot end. I did 8cm bridges ok, have not tried longer …
@Matt_Kraemer1 more pla focussed in this case.
@Camerin_hahn sorry, I explained badly but yes, that would be the idea.
Point is. Don’t overthink it. Just attached some fans that blows in the general direction of the nozzle 
@Tom_Oyvind_Hogstad I’m not worried about airflow across the nozzle particularly, allthough it makes sense to avoid this. This does require thought though, as Kraken poses new problems wrt extrudiate cooling, I really don’t want to mount 2 fans on the Kraken carriage if I can avoid it.
@Matt_Kraemer1 tried it the other day :). That’s what got me started thinking about this!
Nothing happening atm though as I’m slowly rewiring Ingentis. I hate doing wiring 
What if you had a duct like what they use on the Mendel max 2 but have it shaped like a four leaf clover instead of just a circle? You could even just make its big circle that blows air down.
See this page for a photo:
https://honeybee3d.com/about/blog/mendelmax-20-kisslicer/
I have been wondering about an airbrush compressor feeding compressed air to the extruder tip in a chamber. A PTFE tube from one of the Bowden extruders should handle the pressure and still be very light. The compressor may be overkill, though.
That is the same type that some laser air assist uses.
@Dale_Dunn In the short term i might try using a compressor as I have one on hand. Long term, I don’t need any more weight or noise added to Ingentis, (what with water cooling and dibond shell).
yes i vote for blowers, you should see if you can lay your hands on some obselete nidec gamma32 12/24volt. they are great imo. (edit : even the gamma30, 28 series and below are the smaller ones imo)
i use fan PDF alot to see their performance, pressure vs CFM. with this info, you can approx how it will be like assuming you have tube sizes all approximated out.
edit : just for fun, i power a gamma32 partially, to produce about 0.3 inwh = 0.56mmHG pressure, online there is a web calculator which gives you an approx airflow rate based on pressure and orifice size. from my approx pressure, squeezing into a 1inch diameter orifice, the calculator tells me i am moving 7.8 CFM of air. which is really good for my use. however, the longer my duct the more air resistance loss i have, approx 5-10% per foot (@1 inch diameter). however if you noticed, 7.8 CFM is about 3.7 litres per second, IT is alot of air shooting out of a 1inch orifice. (this is calculated based on abs mmHG problem is, on non-abs values, the results become 11CFM, unfortunately im not sure what is the diff between the 2)
. just to give an idea, at this pressure, i ramp the gamma32 @ about 20vDC, and the noise level is like a micro jet about 50dBA. and i have a choice to ramp this from 8vDC (fan starts to spin) all the way to 24vDC because i run this off a variable PWM supply. its a weird project i did to pump fresh air into a small room.
(sources used : engrtoolbox website, TLV.com website)
So here’s something that I’m actually educated in, finally. I’d lean toward @Dale_Dunn 's suggestion and go with the compressor. A blower is generally good for high for rate but still may not have the pressure you’re looking for. All a fan/blower/compressor is, is a pressure rise. Well just call them pressure sources, like a battery in a circuit. Every bit of ducting, turns, transitions, instructions, and the such, are resistors, which lower your current, or flow rate. Your voltage/pressure increase is constant as this is your pressure source. The relationship between your resistors (called head losses) is not really linear, each pressure source has a fan curve describing flow rate vs pressure.
You require very focused flow at a moderate speed, but flowing through very small flexible ducting. The small flexible thing is what hurts you. You could use really small nylon/Teflon/vinyl hose downstream of a valve to jet air at the respective nozzle exit. It doesn’t sound like you need a quick response of turning the air on and off, so your manifold does not need to be terribly close to exits. You can probably use a single valve even. Our you could cool each nozzle separately to save air/power/noise. You could either use a compressor out even the fish pump, but I’m not sure you’d get the power needed from the fish pump, guess it depends on the pump.
Good luck my friend. Always on the bleeding edge with your designs, great stuff!
EDIT—
so after thinking about it, I remembered the downfall of using an air compressor, if your compressor does not have a large enough accumulator and cannot support the flowrate, your pressure/flew will drop and you’ll have the compressor running constantly. So there’s a trade-off. The blower is always steady state, you just may be hard pressed to get one that has such a high flow at such a high pressure drop. I’d still go compressor, and minimize the flow the only the nozzle being used, and the flow should be minimal, especially since the pressure is related down pretty low before the valve
well if you can only use small diameter “hose”, it may be possible to explore aquarium type pumps. but the flow rate is very small. most pumps are jerky, i think even a tiny blower would be better off because it does not have the stuttering effect of a diaphragm pump. just a though i guess … 
Just spitballing some more:
@Camerin_hahn has a good point about using a large duct vs small tubing. I don’t know if you have it in NZ, but here in the 'States, plastic dryer vent hose is very cheap and not horribly heavy. At about Ø75mm, it could carry a lot of volume at ordinary fan pressure. It might still need a blower. The big hose could also keep the bowden tubes near room temperature if they are routed through it. I don’t know if that will ever be a problem. Will there ever be a low temperature material printed in combination with a material that requires a heated chamber?
That doesn’t help with figuring out ductwork at the nozzles though. I don’t see an alternative to creating a manifold that surrounds all four nozzles and floods the entire area. It would need to flow from one side across all four nozzles to avoid stagnant flow in the center, if a large low pressure duct is used, or from the center out if a small, high pressure tube is used. Maybe vacuum form the manifold over printed forms to save weight? Whatever the case, cooling the area around 4 nozzles is uncharted territory as far as I know. More than one iteration will probably be required.
Ok, thanks all for input. I think I’ll plan on a blower with a duct (the same id as the blower aperture) heading towards the Kraken, I’ll print a manifold that breaks the flow into two above the Kraken carriage and work out a duct design that goes on either side of the carriae to direct air under that side’s nozzles. I can see a lot of compound loft object work in my future :(. As much as possible I’ll keep the aggregate cross sectional area the same as the blower aperture right up to the point the air exits at the nozzle.