Walt - in terms of size, cost, ease of calibration, the best answer is almost certainly a flume tank. The late Dr. N.G. Calvert of Liverpool University produced a brilliantly simple ‘back-yard’ design. My brother was actually one of his students at the time and I think assisted in the project - in which case he certaly remembers the details better than I do. I shall try to get hold of him and see what he has to say,
:zbeer::zbeer:
Angus,
Thank you. Any information would be helpful. We probably need a flow channel at least 2 or 3 feet wide, and 1 or 2 feet deep. I had initially been thinking about a test tank about 20 feet long, with a tow motor mounted on a pendulum to measure thrust. Flavio has sent me a paper describing a 12 foot tank, with weights and pulleys to provide thrust, and very accurate electronic position and speed measurement using very simple electronics. But it is still quite large and heavy (at least 1,400 lb of water). A tank with moving water could probably be significantly shorter and lighter, but would need a system of vanes and baffles to straighten the flow, and a flow rate of several thousand gallons per minute. So any ideas to simplify this project are very welcome.
A garden towing tank seems to be a very intersting device ( even if wife point of view are usualy a bit different from ours )
a ) fixed tank - moving model :
b ) fixed model - circulating water :- a quite powerful engine will be needed to move the water
( by the way an italian modeller already tested it in its garden, but - unfortunately - without satisfactory results )
Alternate way of testing :the “test tug”
A remote controlled motorboat model fitted with :
speed measurement :
an hand held GPS such as a Garmin etrex ( price around 100 euros )
drag measurement :
a digital kitchen scale - precision o.1 gr , price less than 20 euros
data recording :
a digital camera pointed to gps and scale at same time
Such deadweight could be loaded on a model less than 1 meter long
During a fair weather day only one hour will be needed to collect drag data at various speed, with several runs in order to collect enough data points to make a fair average, or to delete “dirty” informations.
At the end of the day the tug model will be showed as a tabletop piece for the happines of wife 
Flavio
Folgore ITA 5
Flavio
I like your lateral thinking:D
I think you have invented the “Falocci infinite towing tank”
And THAT (the lateral part) led me to a thought which is not quite baked, but here it is:
One of the factors I debate with myself and anyone else who will play is the efficiency of the fin/rudder setup - and best section for the foils.
It seems to me that a boat requires some rudder angle to turn the fin to a positive AOA to the waterflow. So the boat will have an angle to the course of some small positive angle (maybe 1.5 degrees - this is my guesstimate based on aircraft and the relative thickness of water)
As always lowest drag requires a slim section but lift, especially if the AOA is changing, is better with relatively thick section, and to avoid a vicious stall the LE must to have a large radius
Within your tugboat/GPS/Kitchen scales/camera setup it should also be possible to investigate the behaviour of drag for small angles of boat to the direction of travel (it would need a tow-beside linkage with precise adjustments)
It might also be possible to measure the “lift” force under these conditions.
andrew
Flavio,
I like your idea, but have reservations about the accuracy of GPS for low speed measurements. When sailing full size boats in drifting conditions, I have found GPS speed and heading data relatively unreliable, despite excellent position and “bearing to next mark” data. This is because GPS can only measure position; everthing else is derived mathematically, so the speed calculation is limited by the resolution of the position data, which changes very little at low speeds. Nevertheless, you have presented the germ of a good idea. Perhaps towing over an accurately surveyed course, with a stop watch, would give the required accuracy. But I also have reservations about keeping the scale level on a moving tug boat.
But your thoughts have also re-directed my thinking, to a shore-based station that would tow the hull across the pond with the required accuracy. This might use some of the techniques described in the paper you had sent me. This might be a good winter project.
Flavio
I think you can rule out GPS for far more fundamental reasons than Walt suggests. The time taken to send one ‘frame’ of GPS data is 30 seconds. This represents the minimum granularity of a position in respect of time. Worse, I think that the maximum is fractionally under 60 seconds (start of frame from one satellite only barely overlaps end of frame of another: since each satellite transmits continuously, additional satellites necessarily fall within this ‘super-frame’). Worse still, I’ve never seen a receiver (certainly not a cheap one) that would tell you very much about what it was receiving and when (apart - oviousl - from the IDs of the satellites) - so we don’t know the granularity in any particular instance.
I would like to think there was a way round this - it’s such a nice idea. Can anyone tell me I’m wrong?:graduate::zbeer:
GPS
angus
I don’t know if speed measurement is accurate enough ( most probably it should be mentioned on manifacturer specifications ). I hope so.
About the “frame time” it seems to me that it can be adjusted in order to reduce battery consumption - if needed -
My garmin seems to show very quickly changes of speeds ( within 5 sec )
In any case :
drag and speed should be measured on stationary condition, on a long and straight run, so this is not a big problem.
IAnd, if needed speed can be measured as space/time between two buoys at fixed distance
Fin & Rudder
andrew,
On modern yacht ( not long keeled ) it is very easy , and useful, to split hull and appendages properties.
extensive investigations have been carried out on :
different hull shapes - using exactly same keel and rudder -
different keels , on same hull + rudder
different rudders , on same hull + keel
In order to tow a footy with heel angle and leeway, unfortunately a “test tug” should be much much bigger than the model.
Almost a real size dinghy ( may be this another “germ idea” )
The lift force on footy keel should be considered as a “floating rudder” moving tug stern to on side, and such force should be counteacted by rudder opposite force in order to have a straigh line course.
But there is another possibility : The electric test footy
removing the rig, and installing a motor driven air propeller not far from sail center of effort and rotating such “thing” in order to generate driving force as well side force ( an electric sail force ! ) model beaviour will be very close to a sailing one.
testing several keels , with same engine setting, and measuring time on a fixed distance it should be possible to investigate keel performances under “almost” correct working condition , and with a “scientific procedure”
A fore and aft propeller force shall be considered ad an “electric downwind”
As far I know this idea has already been tested
A lot of nice ideas for this winter 
Flavio
Folgore ITA 5
ps :
a shot of the unsuccesful italian garden test tank
and a sketch of the “electric footy”
Flavio!
What I mean by the time frame is that the block of data sent by one satellite as its input to the position determination algorithm (implemented on the receiver) takes 30 sec to transmit. This seems a long time, but the transmisson rate is very low (I think about 50 bps) and there is a lot of data: the complete ephemeris is transmitted in every block.
Obviously we can iron out these errors by making the test area/run length bigger, but that brings its own practical problems of instrumentation /
observation.
:zbeer:
Flavio,
Would this be a possible design of tank to avoid the problems you mentioned?
Hi Walt - In rowing and flatwater kayak training a devise for measuring speed through the water is temporarily attached to the hull. It takes the form of a small propeller in the water and a readout for the athlete to monitor mounted on the deck. In the case of your towing motorboat, the training devise could be attached to the hull and propeller monitor and the drag scale could be positioned for the camera.
Speed measurement
it seems that GPS speed reliability is around 0.1 kts : not enough for footy testing.
( my idea was based on a similar test years ago I carried out towing a 5 meter boat -full size - at 10 knots in order to measure effective drag. In that case, one tenth of knot was a good accuracy )
Niel proposal about a commercial knotmeter as used for kayaks seems good, even if our range of investigation ( around one knot ) is on their lower working field
An hand made paddle weel fitted with a 10 euros bycicle speed indicator as revolutions counter could be a cheap alternative.
( It should be calibrated before use, most obviously )
Tank testing
The circulating flow channel , most probably, in order to move a lot of water should be fitted with a huge engine, and it should be fitted with a lot of screens in order to have a “clean” flow
It could be a succesful method, but also a very expensive and complicated one.
Flavio
Folgore ITA 5
this is a shot of a professional circulating flow channel