Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Iacopo]

A little test, I spun this top with and without a fairing covering the gap between the flywheel and the stem.

Best timing has been 70.9 seconds, from 1900 to 1600 RPM, for both of them.

Because of tip friction variability, I spun the top 26 times, and avaraged the results. 
Average 1900-1600 RPM spin time without fairing: 64.6 seconds.
Average 1900-1600 RPM spin time with fairing: 65.8 seconds.

The difference is very little. 

[Jeremy McCreary]

Case C1. Lower fairing only, ≤ 3,800 RPM, no shroud ..... 188 s
Case C2 = Case B1 under shroud .................................  54 s
Case D1. Both fairings, ≤ 3,600 RPM, no shroud ............ 206 s
Case D2 = Case D1 under shroud .................................  82 s

So much less spin time under the shroud .. !?

Did the top under the shroud topple down anticipately, because of instability induced by the shroud ?
Or the top under the shroud really slowed down more rapidly ?

Maybe there is a Venturi effect between the side of the shroud and the top coming closer to it, making it less stable.
I too will try a test like this.

Yes, or maybe too much lateral viscous coupling to the relatively tight-fitting shroud. Don't forget that tread pattern. Unlike the blue top, these tops never settled into quiet sleep inside their shroud. But they did outside.

I think part of that had to do with the concave lens. On concave surfaces, I find that tops like these having length/radius ratios in the 1-2 range have to be released just so to go right to sleep. Even though they may sleep readily on flat surfaces. Something inside the shroud seemed to exaggerate this tendency on the concave lens. But I needed the concavity to keep the tops from traveling on their fine but non-spike tips.

Tops C and D never scraped the shroud before falling on their own. But I think all that hula-hooping and precession under the shroud diverted a lot of energy from spin.

[Jeremy McCreary]

Starter and base: Very fast unidirectional (CCW) wind-up starter,

That means you can't do a measurement with  only the lower fairing and opposite spin?

Correct, but if I have time, I'll try this one fairing configuration in both directions with the electric starter. This experiment was time-consuming enough in just one direction!

[ta0]

Case B1. Upper fairing only, ≤ 3,800 RPM, no shroud .....  64 s
Case B2 = Case C1 under shroud ................................ 122 s
Case C1. Lower fairing only, ≤ 3,800 RPM, no shroud ..... 188 s
Case C2 = Case B1 under shroud .................................  54 s

You either mixed up the B's and C's, or you are trying to confuse me 

Your top seems to have three balance screws on the inside but look too small to compensate for the asymmetric arm. Anyway, I have seen similar tops with no screws.
How is it balanced?

- I ripped both covers off and did another measurement on the bare top
This curve ended up somewhere between the others! Which means the whole measurement as I performed it so far can not tell us much.There must be some different effects involved here so that we can not see the aerodynamic effect by the fairings clearly.Might be some imbalance introduced by the fairing or maybe balance by chance. Through that different wobbling that results in different tip related spin decrease......

Good that you took a second measurement.

Because of tip friction variability, I spun the top 26 times, and avaraged the results. 

Iacopo is the gold standard in experimental measurements!

Average 1900-1600 RPM spin time without fairing: 64.6 seconds.
Average 1900-1600 RPM spin time with fairing: 65.8 seconds.
The difference is very little. 

With the cover the area in contact with the open air decreases while the von Karman pumping probably increases. Perhaps they compensate each other.
I'm printing a 3D top to try something related.

[Jeremy McCreary]

@ortwin: Interesting experiment. Not sure why the no-fairing case behaved so inconsistently.

But I do see one common finding in all the experiments in this thread: For the spoked flywheels we tested, covering the spokes with lightweight fairings above and below improved spin time.

[Jeremy McCreary]

@ta0: Nope, those findings for Tops B and C were consistent over at least 5 repetitions each.

This top's trying to confuse us all, and I think it's doing a pretty good job.

A lesson in humility, perhaps.

[ta0]

@Jeremy You call Case B2 = Case C1 under shroud and Case C2 = Case B1 under shroud. 

[ortwin]

Your top seems to have three balance screws on the inside but look too small to compensate for the asymmetric arm. Anyway, I have seen similar tops with no screws.
How is it balanced?

Yes I slightly modified this one: replaced the steel ball tip for a ceramic ball of the same size from a bearing. I glued three little magnets to the inside so I could change the balance by adding magnetic stuff to those magnets wherever needed. Across from the arm there is a second magnet. The weight is enough since the arm seems to be only plastic.

[Jeremy McCreary]

@Jeremy You call Case B2 = Case C1 under shroud and Case C2 = Case B1 under shroud. 

Rats! Double-checked my notes, and the times for Cases B2 and C2 are correct. Only the "= Case X under shroud" parts were wrong.

Best spin times (case numbers follow top names above)
Case A1. No fairings, ≤ 4,100 RPM, no shroud ..............  95 s
Case A2 = Case A1 under shroud .................................  *
Case B1. Upper fairing only, ≤ 3,800 RPM, no shroud .....  64 104 s
Case B2 = Case C1 B1 under shroud ................................ 122 s
Case C1. Lower fairing only, ≤ 3,800 RPM, no shroud ..... 188 s
Case C2 = Case B1 C1 under shroud .................................  54 s
Case D1. Both fairings, ≤ 3,600 RPM, no shroud ............ 206 s
Case D2 = Case D1 under shroud .................................  82 s

But the time for Case B1 (upper fairing, no shroud) was off as shown above. The shroud still helped Top B -- and Top B alone -- but by a much smaller margin than originally shown. More importantly, the corrected B1 time shows that adding only an upper fairing brings a 9% spin-time gain outside the shroud.

So any fairing helps this spoked flywheel spin longer outside the shroud -- especially both fairings. Ditto for the blue top's spoked flywheel.

Conclusion: Spoked flywheels are great for critical speed, but not for air resistance. Best of both worlds? Shrouds aside, probably a thinly spoked flywheel with very low-mass fairings above and below.

[Iacopo]

The results of Jeremy are so confusing that I wanted to repeat a similar test, with a very tall flywheel.

I simulated a tall flywheel (44 mm) with a cylinder of paper wrapped around my top Nr. 15.
Top diameter is 52 mm.

Does the tall flywheel air drag increase or decrease under a shroud ?

The shroud is a simple paper cylinder with a hole for the laser tachometer.
Clearance between the sides of the top and the shroud, 10 mm.



Here I have very good quality contact points and with tops not too heavy, like this one, the tip friction variablity is low, so in this case I don't need many measurements:

1400-1200 RPM
- No shroud:     63.3 seconds
- With shroud:  72.5 seconds
- No shroud:     62.4 seconds
- With shroud:  72.8 seconds

The top spins 15% longer with the shroud

So nothing strange is happening here.
 
Jeremy, I think that it is just all that wobble you have that makes your top to slow down so rapidly under the shroud.
My top does not wobble at all with the shroud, but I have a spiked tip, and a very little base.

[Jeremy McCreary]

A little test, I spun this top with and without a fairing covering the gap between the flywheel and the stem.... The difference is very little.

I think I now see a trend in our experiments...

Removing central mass generally lowers critical speed by increasing AMI per unit mass. If you like, you can then add the mass removed to the outer part of the rotor -- thus further increasing AMI per unit mass without increasing weight on the tip. If you remove central mass by digging a deep well next to the stem from above with no fenestrations (as in your test top here and in ortwin's Spartan), you also lower the CM with an additional reduction in critical speed -- a double win!

The good news: The aerodynamic penalty for removing central mass this last way now appears to be quite small. But if you remove it with rotor fenestrations (like the holes between spokes) exposed to the air, the aerodynamic penalty is quite high. And that's where low-mass fairings might help -- as you might be able to remove more central mass with fenestrations, or a fenestrated well, than with a well alone.

Again, it all comes down to playing the many trade-offs between your critical speed and air resistance reductions just so. The classic Simonelli design plays them beautifully, though in a somewhat different way.

[ortwin]

Starter and base: Very fast unidirectional (CCW) wind-up starter,

That means you can't do a measurement with  only the lower fairing and opposite spin?

Correct, but if I have time, I'll try this one fairing configuration in both directions with the electric starter. This experiment was time-consuming enough in just one direction!

Only the lower fairing, no shroud! That would be enough I think. Quite curious If a directional effect is visible.

[Jeremy McCreary]

Only the lower fairing, no shroud! That would be enough I think. Quite curious If a directional effect is visible.

No clear difference in this limited test of Top C (lower fairing only), no shroud, spokes blowing...

Upward: 166, 172
Downward: 165, 163

[Jeremy McCreary]

It's the little things...

Test top: 168 mm, 133-137 g flywheel top with long spokes, a very low critical speed of ~95 s, and a strong predilection for quiet sleep.

Driver and base: Fast electric driver with drop-out chuck, concave lens of mild to moderate curvature

Measurements: Best spin time of 3 runs from the same release speed with 2 different sets of spokes.



Measured spin times ran from release to first scrape. Started the clock in each run at approximately ω₀ ~ 94.2 rad/s (900 RPM). Remarkably, once I took steps to keep release speed constant, spin times for each spoke design varied by only 0.6%! Also did 1 run with the thick spokes under the same cylindrical shroud used last time with this test top. As before, this shroud cleared the rotor by 10 mm of laterally and 46 mm above.

Top A. Same spoked flywheel top as last time: Thin splined black spokes 4.8 mm in diameter.




Top B. Thicker yellow spokes 7.2 to 7.5 mm in diameter adding 4.3 gm mass, with a cross-section somewhat similar to that in Case A. No other changes.


How the spoke cross-section changed...



Varied parameters
1. Total spoke mass (5.9 vs. 10.2 g)
2. Frontal spoke thickness (4.8 vs. 7.5 mm)

Ignoring the 3.8% difference in total spoke mass relative to flywheel mass...

Controlled parameters (reasonably constant from case to case)
1. Spoke length and attachments
2. Flywheel, stem, hub, and tip assemblies
3. CM-contact distance ~ 24 mm
4. Flywheel AMI I_3F ~ 6.9e-4 kg m²
5. Release speed ω₀ ~ 94.2 rad/s (900 RPM)
6. Critical speed ω_C ~ 9.9 rad/s (95 RPM)
7. Tip-related braking torque vs. speed curve (guessing a very gradual decay of some kind)

Best spin times
Top A (thin spokes), no shroud ................ 184 s (3:04)
Top B (thick spokes), no shroud ............... 170 s (2:50)
Top B (thick spokes) under shroud (1 run) ... 314 s (5:14)

Conclusions
When you're looking to maximize spin time, small things clearly count. The 14-second no-shroud spin time difference between Tops A and B was small but consistent. And a lesson for endurance top designers. You know those intricate carvings on many pricey EDC tops claiming record spin times? Well, they're shooting themselves in the foot endurance-wise.

Note that the flywheel's large AMI of 6.9e-4 kg m² failed to protect spin time from the spoke change. This AMI falls roughly halfway between the AMIs Iacopo measured for his Nrs. 23 and 9. As tops go, the latter's AMI is huge. Fancy EDC tops generally have much smaller AMIs.

Shroud: As with the blue test top above, the shroud increased spin time dramatically. Whatever air flows the shroud might have forced here, the spokes were clearly getting push-back, as spin time increased by 75% with the thin spokes and by 85% with the dirtier thick spokes.

Personally, I'm going with the "von Braun space station" look with the thicker yellow spokes. It's gonna cost me 7.6% in spin time, but spin time isn't everything.

[ortwin]

Jeremy, so you say by this I should really go for the 0.04 mm fishing line instead of the 0.3 mm line I am currently using?