Author Topic: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine  (Read 129374 times)

Offline DRT

  • Full Member
  • ****
  • Posts: 64
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #345 on: March 03, 2019, 10:15:48 AM »
Jim
Have you ever run any of you saw nitro props on your gas engines?  And have you ever considered a rotary attachment similar to your helical pitch gauge to mill your own props? I'm very sure with your skills you could do something similar to this.You have access to cnc and if you dont you will soon.  ;) You'd mill both side out of a block. I def don't have to tell you how to reference position.



"These are made in the Ukraine by multiple champions Andreij and Volodymyr Smolnikov. The props are custom CNC machined from both sides in the blocks as seen to whatever diameters and pitches are specified. The rectangular block is a group of A2 in varying sizes and the circular block A3 props showing the machining from either side. A1 sizes are also on offer. Props are available either as a block at a sizeable discount or finished and ready for use. The blades in the block are finish machined to very high tolerance, a matter of microns, with just a small spur at the tips and more substantial fillets on the boss keeping them in position."

~on the wire uk~

« Last Edit: March 03, 2019, 11:16:54 AM by DRT »

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #346 on: March 03, 2019, 03:52:07 PM »
@Lohring,
All you have to do is take the scatter plots and connect lines of best fit and then you can see better your data variations and comparisons. Excel should have this ability. Also are you checking seeing which regression most accurately fits you data points? You should also be able to toggle between the data points and lines of best fit. Power, Logarithmic, Cubic, etc all have a  linear regression, so there will be one that likely fits your data best.

I did that as well as average the data from at least three runs.  Some of my plots have these average lines on them.  Our early tests had more scatter than our later tests, but the dyno's precision was about +- .1 hp.  I measured the standard deviation between several identical runs at between 1 and 3%.  Scatter plots keep people from making conclusions from the data that aren't justified.

Lohring Miller

Offline lohring

  • Full Member
  • ****
  • Posts: 163
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #347 on: March 03, 2019, 03:58:48 PM »
I built a helical fixture to mill props.  I haven't used it, though.  It's described in the article at http://namba.com/content/library/propwash/2014/April/#/4/ 

Lohring Miller

Offline strictlybusiness1

  • Full Member
  • ****
  • Posts: 266
  • ME., Tool Maker
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #348 on: March 03, 2019, 05:48:53 PM »

Jim
"Have you ever run any of you saw nitro props on your gas engines?  And have you ever considered a rotary attachment similar to your helical pitch gauge to mill your own props? I'm very sure with your skills you could do something similar to this.You have access to cnc and if you dont you will soon.  ;) You'd mill both side out of a block. I def don't have to tell you how to reference position."

The answer to all of these questions is absolutely no! There isn't enough time available to address in depth propeller making. I have posted some photos of what is presently being worked on. It is true that the carbon fiber prop test stand saves time, but I'm only looking at what happens at WOT settings.

Jim  Allen


Offline DRT

  • Full Member
  • ****
  • Posts: 64
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #349 on: March 04, 2019, 05:42:35 AM »
@Lohring,
All you have to do is take the scatter plots and connect lines of best fit and then you can see better your data variations and comparisons. Excel should have this ability. Also are you checking seeing which regression most accurately fits you data points? You should also be able to toggle between the data points and lines of best fit. Power, Logarithmic, Cubic, etc all have a  linear regression, so there will be one that likely fits your data best.
I did that as well as average the data from at least three runs.  Some of my plots have these average lines on them.  Our early tests had more scatter than our later tests, but the dyno's precision was about +- .1 hp.  I measured the standard deviation between several identical runs at between 1 and 3%.  Scatter plots keep people from making conclusions from the data that aren't justified.

Lohring Miller

Hi lohring,

Have you determined which regression has the best coefficient of correlation and determination to your data? This not the average of the data point you've taken from several runs. If your building a graph piecewise from data  which you have to be if its's taken on an inertial dyno there's no point to be worried about the variance in accuracy of line smoothing. Accuracy is already at the will of how well you build it from piece wise from discontinuous data which will certainly have variance. The linear regression doesn't have to have line smoothing per se and if you have selected the one with the highest coefficient of correlation and determination it will be as accurate any any scatter plot you have. It is built from you scatter data and gives you the line that actually fits it most accurately. When you want clear visual representation to compare data this is the required way.People should never make conclusions from just the data points or lines on a graph. It should be teamed with real world results as well as other confirming mathematical models. They should all be fair agreement to feel confident about making any assertions thereof.

@Jim,
A reaction torque sensor mounted on a linear slide on axis with the motors shaft with a propeller forced into a load cell with a tach should tell you what you want to know at wot. The setup can be hobby grade or IEEE grade simply using  "TEDS" sensors that meet the IEEE 1451.4 protocol. Your other option is an eddy current brake . You can use an induction motor with magnets and a copper disk in a touchless design and VFD with braking resistors. The torque cradle previously described is really all you need to see improvement though . You can certainly mount an inertial disk to the motor versus a prop. You may need to build a carbon fiber retention sleeve. You wrap it on an undersized arbor. Finish it to final od an length on the lathe and press fit it on the flywheel which you can make from 304 stainless, titanium, or Inconel 718 SPF @(1393 MPa). A undersized fit of .134mm with the carbon fiber sleeve will give you a zero speed pre stress pressure of 476.6MPa.It has to be balanced of course. I have a lot of information on these kinds of things but your a difficult man to get in touch with just a state a way. So I give you a little here without getting too off topic.

« Last Edit: March 04, 2019, 06:16:44 AM by DRT »

Offline Admiral_dk

  • Full Member
  • *****
  • Posts: 3741
  • Søften - Denmark
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #350 on: March 04, 2019, 11:24:49 AM »
Jim has a very high accuracy water brake - see earlier in this thread.

But some of the rest of us might find a cheaper / simpler measuring device interesting - more details please - though maybe in a separate thread.

Offline DRT

  • Full Member
  • ****
  • Posts: 64
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #351 on: March 04, 2019, 01:56:01 PM »
Hi Admiral,
Have no fears. Jim is a close friend Im very aware of all he has. Using Teds sensors means it's accurate enough to be approved by the IEEE. Look up the TEDS system and understand all it entails. It's alot more than just an accurate sensor. It has to do with stringent calibration schedules etc. He has a water brake but the sensor system I suggest isn't something that was available when he adopted and upgraded the brake from Ed Kalfus. It also speaks to the transformation from raw data the brake measures to what is visually represented through graphs charts etc. The brake is not fun to use without this side of the coin fully integrated. No one want to write down data points and derive regressions time after time. He needs a self working data acquisition software package to do this work automatically. There is no question I can help him complete that side of the coin.  Cell or registry math in MCU's brains are more my thing. U need these things for instance like a dedicated optical tach to get what the water brake reads easily up on the screen so he can simply examine a graph versus having to build it after every pull of his various dyne meters.

Jim and Lohring are both well respected highly valued technical minds in our sport. I know them both quite well away from here. They have both taught me many things about the subject matter at hand. I only consider my offering as paying it forward. I'll be more than happy in time to show you in a separate thread  that side as I develop my electric motors.

Thanks,
Hubert
« Last Edit: March 04, 2019, 02:48:10 PM by DRT »

Offline Niels Abildgaard

  • Full Member
  • ****
  • Posts: 167
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #352 on: March 04, 2019, 04:37:24 PM »
Jim has a very high accuracy water brake - see earlier in this thread.

But some of the rest of us might find a cheaper / simpler measuring device interesting - more details please - though maybe in a separate thread.

Aircraft engines has been tested for ages with  rotating wood blocks

http://4.bp.blogspot.com/-2DEoFo4Va9g/VfctF7G_DtI/AAAAAAAAP2c/FVKGF4-vSP0/s640/MOULINET_COLOMBAN-RENARD.gif

Offline strictlybusiness1

  • Full Member
  • ****
  • Posts: 266
  • ME., Tool Maker
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #353 on: March 05, 2019, 04:39:51 AM »
I wouldn't use any thing made of wood or glass filled nylons to test my engines. The photos show the last test stand I built which can test both gas & nitro engines on the same test stand. The loading devices are carbon fiber propellers of various pitches & diameters. They are cut from existing Falcon brand propellers & re-balanced in a magnetic balance. The same shafts with ball & pin universals are used for both type engines. This test stand weighs approximately 35 lbs & must be secured to the bench during WOT runs at typical RPM's of 20,000 to 30,000+. No torque is measured with this test stand. I load the engine to give the same RPM as measured on the water. I typically run an engine for approximately 4 to 5 minutes at WOT. If a modification is made which enables the standard load to turn faster, the engine must be making more HP. The last photo shows the electronic tach permanently mounted behind the propeller.

Jim Allen
« Last Edit: March 05, 2019, 12:39:47 PM by strictlybusiness1 »

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #354 on: March 05, 2019, 10:39:22 AM »
Hi DRT, thanks for that explanation, it’s great to know a little of where you fit in.

It is clear that some amazing work is being documented in this thread, and I normally just look on in awe (and silence).  But I am interested in why, when you were talking about identifying those standing wave frequencies in excell, you went to regression analysis instead of the Fast Fourier Transform function.  If it is a big enough data set, and that may of course be the problem, the FFT quickly identifies any natural frequencies in the data set.

Oh, and also, please add my vote to Admiral DK for your build thread on a simpler dynamometer for us mere mortals.

MJM460
The more I learn, the more I find that I still have to learn!

Offline strictlybusiness1

  • Full Member
  • ****
  • Posts: 266
  • ME., Tool Maker
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #355 on: March 05, 2019, 01:15:38 PM »
Hi DRT, thanks for that explanation, it’s great to know a little of where you fit in.

It is clear that some amazing work is being documented in this thread, and I normally just look on in awe (and silence).  But I am interested in why, when you were talking about identifying those standing wave frequencies in excell, you went to regression analysis instead of the Fast Fourier Transform function.  If it is a big enough data set, and that may of course be the problem, the FFT quickly identifies any natural frequencies in the data set.

Oh, and also, please add my vote to Admiral DK for your build thread on a simpler dynamometer for us mere mortals.

MJM460

I see you're from Australia. Do you know someone named Rod Smith?

Jim Allen

Offline MJM460

  • Full Member
  • ****
  • Posts: 1647
  • Melbourne, Australia
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #356 on: March 05, 2019, 09:07:05 PM »
Hi Jim,

Not that I can recal at the moment.  If he is a model engine maker like yourself, almost certainly not as I don’t think I know anyone in that field.

It is always worth asking though, as coincidences do happen.  Was once flying between Toronto and Sarnia, and was asked a similar question.   Turned out I had worked with the guy in Melbourne, and knew him quite well. 

I even worked with a Jim Allen, but that was a very long time ago.

MJM460



« Last Edit: March 06, 2019, 11:05:20 AM by MJM460 »
The more I learn, the more I find that I still have to learn!

Offline Zephyrin

  • Full Member
  • ****
  • Posts: 766
  • near Paris, France
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #357 on: March 06, 2019, 07:39:54 AM »
Quote
the FFT quickly identifies any natural frequencies in the data set.

You can use Audacity software, that give the frequencies spectrum from the sound track (or noise track...) 
may be helpful when tuning exhaust.

Offline strictlybusiness1

  • Full Member
  • ****
  • Posts: 266
  • ME., Tool Maker
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #358 on: March 06, 2019, 06:15:57 PM »
The simple fixture shown allows the stock piston bosses to be re-machined to a particular dimension & square to the stock wrist pin axis. What could be the reason that this is being done?

Jim Allen

Offline Admiral_dk

  • Full Member
  • *****
  • Posts: 3741
  • Søften - Denmark
Re: .90 cu in, 30,000 RPM, 7.2 HP custom built nitro engine
« Reply #359 on: March 06, 2019, 09:45:09 PM »
Hmm - since the skirts are further down, it isn't clearance to the crank web ....  I guess reducing weight.

 

SimplePortal 2.3.5 © 2008-2012, SimplePortal