Fan testing

[Tim Mackey]

I too have experienced duff motor data from BRC - they are aware of some motors being supplied that have a duffer amongst them, and this then negates the supplied test data sheet. Rare, but does happen.

I imagine adapting the stands we have done to accept the whole model should not be difficult, and as for construction,  well I probbaly over engineered mine a little with ball bearings and the like, but the main thing is rigidity, and accuracy of moment arms etc. Simple hardwood strip seems the easiest option for materials, and of course, a set of digi scales which can be zeroed AFTER the load is set up. I also shose to use a system where all ancillary gear such as wires and meters etc are supported on the rig, to avoid any possibility of drag or binding affecting the readings - if all you want is thrust measurement then obviously you will not be hooking in meters and the like.

[Erfolg]

It is useful to know the watts in to kg force out, so I will use my watt meter. I really should get a rev counter, to see the effects of dia on revs/power consumption.

I was a little surprised to think that there might be poor performance examples of a brushless outrunner, given their simplicity. Magnets and a few windings.

I had put the poor apparent performance down to me using a folder (with aeronaut blades) against a rigid prop APC/GWS. My immediate solution is a bigger dia folder. Blades on order.

I have noticed another model (a Zaggi pusher type) apparently performing poorly (in my opinion) with the same motor, this was using a very small dia prop, although the revs were high.

Erfolg

[Tim Mackey]

Not so much "poor performance" more a DUFF motor, IE: windings shorted etc.

[Erfolg]

I have been thinking while starting my flashing (that is as in lead work on roof).

Timbos design is 1:1 leavers, and Richard talked about force multiplication. I thought 1:10 seems a good magnifier factor. I then thought Richards set up seemed much more compact than the Timbo design.

Compact means good in my situation, not enough storage space. so how compact can it be I thought?

Would a 100 and 1000mm leavers do it? Mmm seems a little big. Anyway looking at my scales (wife's really) they only go up to 5kg. That really means that 1:2 is the max I can practically have.

It then dawned on me, if I just placed the motor on top of the scales, would things become a lot simpler? I know that it is now 1:1, but I guess that is adequate for my needs.

What do you guys think? Can it work? Is the arrangement as simple as I think?

Your thoughts will be much appreciated

Erfolg

[Myron Beaumont]

When you think about leverage etc   Isnt it better to mount the fan / whatever onto the base plate (receptor of weight/thrust)  directly thus avoiding any hysterisis in the set up .Think about how much scales actually move even with the weight of a sticky bun eater like Timbo even ,Its only a few mm(1/4 inch)

[Erfolg]

Myron

I think we are advocating the same set up.

That is the motor thrust acting along the axis of the scale. There will be no hysteris effect as there is no moving parts other than the load cell of the scales, which is essentailly a solid state strain gauge device (yes we lashed out £5 at Leidels for the latest technology).

I am thinking of a light weight platform to hold my wattmeter, the speed controller, RX and lipo. The motor is to be conceptually mounted on the top of a section of fibreglass tail boom, to keep the propeller arc out of harms way.

I need to be able to build the lot for the minimum weight as this needs to come of the maximum weight capability of the scale, if 1kg leaves 4kg head room.

No one has pointed out the disadvantages of this set up yet, there must be some?

I am some what suprised that I have not been taken to task over units and descriptions, ie using kilograms not Newtons talking about weight when i should say mass. Are you all slipping , not noticed, or pragmatic, regarding termiology.

I need the practical feed back, to avoid, avoidable cock ups in my set up.

Erfolg 

[Tim Mackey]

You could always simply suspend the whole model on a "cradle" arrangement nose down on the scale platform and light the fire let the nose push down on the scale and read off the thrust. ( EDF or pusher only of course )

[Myron Beaumont]

Here is me thinking I had an old fashioned spring in 'er indoors' weighing M/C.--I was right-And she told me just now that it doesn't measure mass ! I do know the difference even though I don't live on Mars where its a bit different but isnt!.There is one problem with your set up (you were asking about)that is   ---Er indoors's scales you have taken over !    .I suggest you change the "chip" in it to make 'er think that she is losing weight ( I just stretch a spring to make it "fit"). Dont forget the obvious by the way in that you can sit everything vertically for testing to make life easier -Air flows in all directions equally & up & down & sideways as well    

PS Whats termiology    __  Is it the study of termites?

[Richard Sharman]

If I may join in?  It seems to me that we need a PRACTICAL device which can be used for COMPARATIVE tests, rather than a state of the art scientific measuring system with output to umpteen decimal places in units like Newtons which most people don't understand anyway?

So, Timbo's set-up of a horizontal airflow is just rght (since that's the way the prop/fan will be orientated, at least for take-off and landing) and it can be set up so that inflow and outflow are well away from any obstacles which could have an effect on the measurement.  It so happens that convenient digital scales(kitchen types) are available for WEIGHING (which is vertical, gravity being what it is), and therefore a lever to convert the direction of force is just fine.  The level can have mechanical advantage <1, =1 or >1 to taste.  Ok, there could be losses in the lever, pivot, etc, but they are minimal relative to the size of force being measured, and anyway it's comparative tests we want (= is fan/motor A better or worse than fan/motor B ?) 

So, I would support the Timbo-style device, possibly adapted to enable complete airframes to be mounted.  I have just done a test with an "open" fan(as in the photo above) and the same fan in a complete duct within a model.  The difference between the two tests must represent the "duct losses" (I doubt you could ever get a "duct advantage" at the scale we are working at?).  Surprise, surprise, I have measurable duct losses -- so, now I know this, I can concentrate on THAT problem rather than chasing some other illusion.

[Tim Mackey]

Exactly Richard - and was the point I made right at the start - I merely wanted comparative tests in a like for like controllable test environment to evaluate one component against another, and in fact accuracy of measurement was not particularly important, but did in fact turn out to be very accurate.

[Erfolg]

Thanks for the discussion.

I am trying to make a simpler unit. Not a different unit for the sake of it. I am trying to build on the work of others, for advantage.

My note on units etc. was a little tongue in cheek. To some extent, I was highlighting the same points you are making, Newtons have not reached the real world, nor does the difference with mass and weight. As we have not gone into outer space I do not care either. It was meant to rib you, I think I did it!

There is a problem with the concept I have outlined. If the motor is mounted in the most obvious orientation, then the apparent weight will reduce when the motor is running. Which is no problem for the scales as it has negative readings as well as the positive. If the thrust reaches the same or greater value than the weight, the rig will become unstable and potentially take of. Mounting with thrust in the other axial direction is less convienant.

A major consideration of any device I build is that it must take up little space. The no moving parts is a benefit which was not immediatly sought. Of course I want the minimum build, in time and possibly materials, I do want it save to operate

I to want a practal unit, and am not seeking one decimal point of accuracy, I thought i made that clear.  Must write clearer messages

Good to see you are commenting.

Erfolg 

[Former Member]

[This posting has been removed]

[Erfolg]

Whilst taking care of the hips, cutting to shape. No not physical exercise but am roofing.  I spotted an old abandoned plastic plant plot, which is just over 12" dia. I am now thinking of utilising this.

I will be next chasing out the brickwork for the flashing, I have £256 of lead which is 15m, just wonder if there will be any left over, to hold the rig down. That is a joke (as the wife says not very funny).

Anyway back to the roof, whilst still dry.

Erfolg

[Paul Williams]

I pinched timbo's design ( sorry mate you will get your reward in another life as for most things) and left out the lever bit, I mounted a cradle on top where his fan was and simply sat what ever plane I have on it. The measurement is by a simple digital pull mounted horizontally. Bingo. A versatile machine that cost 7 quid plus a few screws to make. I can mount fans motors or the whole model, which is great for a quick prop change then test.

Mount the digital pull so the pivot arm is slightly past 12 o clock set the tear and fire up. I have not stopped using it since.

[Erfolg]

Paul

A picture saves a thousand words.

I am sure that with a picture all will be so obvious.

Erfolg

[Richard Sharman]

While the builders, plagiarisers, lead flashers and others are busy at work, may I ask a fan question?

  How much thrust is lost in a typical duct ?

Is it neglible? a bit of problem? or a serious issue?  Clearly, the longer the duct the more resistance to air flowing through it, and there is the compressibility of air to consider as well.  At some length of duct there will be so much resistance that nothing will come out at all, but we probably don't have to consider that case. 

 For your normal duct of a few inches (say, up to 12 inches) it would be nice to know:

• how the resistance increases with length 
• whether the resistance increases with a narrowing duct (e.g. tail pipe)
• whether the resistance decreases with a widening duct  (e.g. an inlet throat)

any ideas? gratuitous opinions gratefully accepted!!

Richard

[Erfolg]

I have recommended this bit of light reading before and it answer's in principal a lot of your questions.

"Woods of Colchester" - "Ventilation Design Book"

I know we are not designing ventilation systems, but we are moving air, the issues are either the same or similar.

1. Resistance is directly proportional, that is the longer the duct the greater is the resistance, in proportion to the length.
2. Resistance increases as you narrow the duct, following our old friend Bernoulli's theorem
3. Again Bernoulli, is lower as duct increases- the concept of "Conservation of energy"

The book deals with inlet losses, tends to be a bit vague on outlets, but in HV they often seem to expand the discharge duct quite dramatically to regain some of the energy.

What the book does not deal with is the momentum effect and the impulse effects from a discharge at a highish speed. This is generally covered in fluid flow text books, where momentum and effects of fluids leaving pipes are examined. I just cannot bring the names of the principles to mind (senior moments).

The trouble is that these books and branches in science can tell you what is important and what to expect. But do not produce a model.

I think studying models that work well and those that do not, will give you more information quickly, than a theoretical study, which will take days or weeks. Probably years if you did not take physics at school etc.

DF are not an efficient way of flying a model, if you are going to build one it is a question of limiting losses.

Erfolg

[Richard Sharman]

Dear Erfolg,

  thank you for your prompt reply! and the detailed response - much appreciated! I was dimly aware that you had mentioned these points in a previous post, so apologies for getting you to repeat it. Of course, your principles are right at a conceptual level (I will try and find a copy of this book).  I should also declare my hand in that I have a Ph.d in fluid dynamics (but I did not specialise in flows in/out of pipes - so do not claim to know the answers here).  However, as you say, understanding the Navier-Stokes equations, Laplacians, Bernoulli and other principles, does not actually help you design a simple duct for a model aircraft.  Or more particularly, modifying an existing duct to improve its performance.

So, I was therefore thrown on the empirical scrap heap of fortune, and reduced to asking if anyone has a "rule of thumb" based on experience?  We have the rule-of-thumb for power loading (e.g. 150 watts/lb minimum).  We have the rule-of-thumb for thrust/weight ratios (e.g. 1:1 for jet scale performance).  It's not unreasonable to ask if anyone has a design rule-of-thumb (at the risk of running out of thumbs) for duct shapes(e.g. make your ducts increase in the direction of flow to reduce resistance). But is this what people have actually found?

Alternatively, we could do some experiments with our various fan testing devices?

regards, Richard

[Erfolg]

Dear Doctor

Surely you must have considered the effects of entries and exits to systems.

You almost certainly will have access to a FE program for fluids. In the past I had a stripped down package, which ran on my PC. Unfortunatly I had not returned to modelling and now seem to have lost it.

My own opinion is that most of our DF systems are very simple, very few elements.

• Boundary needs to be defined outside of system
• Inlet
• straight duct
• Convergant duct
• airpump/fan whatever
• divergent duct
• changes in cross section
• bends
• outlet, which i suspect can be considered the boundary

This is the sort of problem which I loved to investigate in my lunch hour, to find out what in principal mattered and the maximum performance that would be possible if everything was right and optimised.

The real problem was, while I did studies of this type, I had no time for modelling or even a hobby.

I once did study on DF using conventional relationships and came to the conclusion that you might get about  level of thrust that Byron modells subsequently achieved  but they certainly used more power by the use of Nitro, tuned pipe, faster revs, than i had calculated.

You could well be the guy who can help by developing a software package that helps in the prediction of  performance. Particularly if the airframe drag etc can be taken into account, a two part interrelated issue. What seems to put people off is the lack of money to be made from this type of software. Most of these predictive packages are Foreign free ware, aerofoil section plotting, motor performance etc.

Regards

Erfolg

[Richard Sharman]

Dear Erfolg, 

Thank you for your thoughts and cogitations - I will consider your insights and suggestions at length (over a beer) so I will not reply instantly. However, continuing the thought process....

In the past I would have built a computational model of the entire duct geometry, bends, edges, and all (probably in 2 dimensions first, but later in 3-d) and then applied collosal computing power to solving the finite element equations iteratively under a large number of different assumptions.  The project would have kept a parrallel computing engine busy for at least several months, and I would write a paper on the result which would be published in a technical journal.  I would get a generous grant from a funding agency, most of which I would spend on writing justifications for why this was a good thing to do, and why I could do it better that anyone else.  The money I spent on actually working would be a small fraction of this.  Now I am retired I don't have to do this kind of thing, and anyway no longer have access to serious computing resources.  All I  want is to make a small change to my model so that I can fly at the weekend. 

So, I have resorted to empical methods.  Here is a simple problem:

If a duct is 200mm long, and the fan diameter is 65mm. What should the tailpipe exit diameter be?

I have done some testing with my rig, using a rolled plastic sheet as a tailpipe, adjusting it to different exit diameters. Here is the result:

(exit, fraction of entryThrust )  (1, 14.85) (0.9, 15.15) (0.8, 15.75) (0.7, 16.2) (0.6,15.9) (0.5, 15.6) (0.4, 14.7)

so the answer is 30%  QED

[Erfolg]

Richard

I do not have a clue.

I have know experience at all, only seen DFs fly.

Looked at them, the features that sucessful models seem to have. are

• Large intake
• parallel internal duct
• large outlet
• the shorter the duct the better
• no nasty edges inside, significant changes in direction or configuration.

In the case of computer modelling, I really do not see the need for anything fancy. Given that all the base data for air as a fluid is well known, given that our operating window is not excetional in velocity, temperature or any effect from passing through a fan etc. When in the past I have commissioned FE, the stumbling block was either having the data to hand or using a company with experience of the the medis etc. I had thought and wondered if one of the commercial ventillation software packages would surfice, alhough the volumes are low, and the velocities are tending to as fast as we can get it. Air curtains probably have velocities far higher than we can achieve (we just do not have a power station at the other end of the lead).

Like you, I retired and no longer have access to these types of packages or any rael inclination to go in the loft to find the  relevant text books.  When I do, I end up thinking far to much, re-remembering about so much that I had forgotton, delving into other things. Nothing gets done , that I want to really do.

I generally preffer to deal with a real physical problem, for a technical solution I post an enquiry. Generally the USA RCGroups modelling furum and the German RC Line furum, seem to be full of scientists and engineers who come back very quickly with a numerative result. I think many are using data bases linked to conventional relationships to solve.

Erfolg

[Richard Sharman]

Erfolg wrote (see)

I do not have a clue.

I have know experience at all, only seen DFs fly.

Dear Erfolg,

Please do not despair -- this is only the fan testing forum, not the Witgenstein philosophy forum, or even the life, unverse and everything forum. Your insights and comments have been very interesting, thought provoking, and helpful, even wildly irrelevant sometimes - long may you continue!

Your thought about Heating & Ventilation engineers knowing more about moving air around than we do is inspired. Sadly, I think it is going to be hard to apply that knowledge to ducted fans since I suspect that they(H&V) are ultimately more interested in the volume of air to move around, the velocity of the air being secondary, whereas we (ducted fan nerds) are principally interested in the velocity of the air we move, and are not much interested in the actual volume moved. There are other considerations of course.

My point about computing models was not entirely flippant, but the same point you make stands: defining the boundary conditions in all respects is a rather tricky, not to say time-consuming, part of the problem definition, and one we would rather not do. Which then makes the subsequent modelling rather redundant. For now, theoretical models, and computed models, are not really suitable.

I will return to address the ideal duct shapes question later, for now I just repeat the experimental observations I made with the data above plotted as a graph. This is a single example of a 200mm duct leading from a 65mm fan, to an exit of variable size. Note that as the exit area decreases so the thrust increases (counter-intuitively) to a maximum, after which it decreases again. This shows that at first the Bernoulli effect is more powerful than the increasing drag due to resistance in the (constricted pipe) Eventually the drag overcomes everything, as usual.

A tailpipe exit with an area about 70% of the fan area would seem to be indicated in this case. (Interestingly, my uninformed guess a priori was 78% - not too far out??!)

[Tim Mackey]

Erfolg wrote (see)

Richard

I do not have a clue.

I have know experience at all, only seen DFs fly.

Looked at them, the features that sucessful models seem to have. are

• Large intake
• parallel internal duct
• large outlet
• the shorter the duct the better
• no nasty edges inside, significant changes in direction or configuration.

In the case of computer modelling, I really do not see the need for anything fancy. Given that all the base data for air as a fluid is well known, given that our operating window is not excetional in velocity, temperature or any effect from passing through a fan etc. When in the past I have commissioned FE, the stumbling block was either having the data to hand or using a company with experience of the the medis etc. I had thought and wondered if one of the commercial ventillation software packages would surfice, alhough the volumes are low, and the velocities are tending to as fast as we can get it. Air curtains probably have velocities far higher than we can achieve (we just do not have a power station at the other end of the lead).

Like you, I retired and no longer have access to these types of packages or any rael inclination to go in the loft to find the  relevant text books.  When I do, I end up thinking far to much, re-remembering about so much that I had forgotton, delving into other things. Nothing gets done , that I want to really do.

I generally preffer to deal with a real physical problem, for a technical solution I post an enquiry. Generally the USA RCGroups modelling furum and the German RC Line furum, seem to be full of scientists and engineers who come back very quickly with a numerative result. I think many are using data bases linked to conventional relationships to solve.

Erfolg

• Large intake
• parallel internal duct
• large outlet
• the shorter the duct the better
• no nasty edges inside, significant changes in direction or configuration.
• AND REALISTIC AND LARGE POWER TO WEIGHT RATIOS WITH LIGHT AIRFRAMES

[Myron Beaumont]

Doctor Richard   When you arrive at about 70% exit area do you mean as a proportion of the ACTUAL swept fan area covered or the simple area of the front of the entire unit .I'm thinking that the motor/casing in the centre obviously obstructs the flow -probably to the extent of about 30%. So am I to deduce that all we are doing by "restricting " the exit is actually keeping the effective cross-sectional area constant ?

[Paul Williams]

Enough I have run out of headache pills, the fan dia for my hawk is 89mm what lenth and exit dia should I fabricate the tube? Just the 2 dimentions would be great. Thank you thank you thank you