Engine Capacity

Krop/ David, has there been any other paddock clues as to what the duke is, if your analysis is ambiguous.



Perhaps it doesn,t matter, because even if it is currently 930cc, the new 75 degree V engine may well be 999cc anyway...?? Thoughts........?

Kropotkin said:

They are not. It is a massively complicated operation, sorting the signal from the noise. I am rather dismayed by just how difficult it is.... I'll post a shot of the sound file later.

Click to expand...

Post access to the sound file and I will analyse it if you like ...... seriously.

( PS. don't make it mega huge it doesn't have to be ultra high res. I'm assuming the big bang config they have will make it pretty easy to decipher )

Is it trackside? or just pulled off a TV quali session? cos I can pull it off anyway I suppose, if thats the best quality I can get.

BarryMachine said:

( PS. don't make it mega huge it doesn't have to be ultra high res. I'm assuming the big bang config they have will make it pretty easy to decipher )

Click to expand...

That's what I thought. I was wrong. Recordings from trackside extremely difficult to decipher as it is not clear what 1 cylinder firing looks like.

You are just looking for a repeated wave sequence, it really doesn't matter what each cylinder is doing. So once the regular sequence is identifed you can measure the period of several hundred waves. Or I believe have a bit of spectrum analyser software that will find the frequency of the sequence, even easier.

Might see if I can find a sequence from Quali of Rossi going past at full tilt.



Edit. I think I see what you describe as the problem ....... the big bang fires in such a way that 3 cyl fires will probably blend as one beat ( hence the long drawn out sound of say the M1 and the Duc. But realistically you can take a punt at finding the sequence because it can not be far away from a guesstimate. It could seem half the rev.s but that would be glaringly obvious. ........ kinda like using an old strobe tach. to chase rpm's. I'll see what I can do

Two views of the same piece of audio, Rossi at full tilt:











Spot the mixture being ignited.

Can you put the actual sound file up? Are they just Audacity recordings? Been a while since I've used Audacity but I believe it even has some fourier stuff in the menus ...... might be worth trying a few smoothing filters.



Its a pretty ragged sound bite. Is their fences and posts in the way? Maybe an onboard sound bite maybe better.



Have you just tried playing it into a spectrum analyser, may not even need to fiddle around with the wave form.

130dB and a cheap audio recorder. Nothing between me and the bike, down Jerez' back straight.

Been messing around visually and I'm justifying 15,000 and 18,000. Depending on what I see as the firing order.



I think I'd better get an actual sound bite It really is a lot messier than I expected!

Kropotkin said:

130dB and a cheap audio recorder. Nothing between me and the bike, down Jerez' back straight.

Click to expand...

Must be a lot of echo and bounce of walls maybe ?



Anyway I'll see if I can get some off an onboard vid ...... if there is one ? they seem to be very scarce with the Rossi footage at present



Do you have a video or sound file storage site? maybe you can put what you have there and I'll download it and play with it. You can probably even load it on youtube as a blank video with sound.

Ok visually from what you posted here:







Event Sec.s A 3.3631 B 3.37065 C 3.3783



Are you seeing that?? maybe?





A-B is 2 revs gives 15894 rpm



B-C is 2 revs gives 15686 rpm



A-C is 4 revs gives 15789 rpm



But I wouldn't declare that in a court of law ......... sample too small ........ assumed timeline etc. etc. etc.







Your other curve was recorded too loud and it has suffered a bit of over-saturation ......... bit hard to tell with that one. Edit: on second thoughts the second curve is just a normalized version of curve 1? Would yield the same result.

Another way you could do it .........



Cut out a good sample section and copy and paste it several times over then play it to a frequency meter.



Again ...... irrespective of the frequency read only one sensible answer will be obvious. You could be measuring 1/4,1/2,3/4, 4/4s of the actual revs. It can't be out by miniscule amounts, only out by steps of 25% an then even knowing which 25% is superfluous.





The Ducati would sound something like a mix of the musical notes C3 and C4. Mostly its the C3 we can hear cos of the big bang note.



If it was near



19,000 it would sound like Eb 3 or 4



18,000 .............................. nearer D 3 or 4



Upload a free guitar tuner with frequency meter on your Android phone ( thats what i use these days ) and measure it live with that next time.



Or a function generator with say square wave and saw wave options then play it as Rossi goes past ...... if you are a really good guesser you could tell by the harmonics Ok that ones a bit "out there" but I have done similar in the distant past.

BarryMachine said:

Another way you could do it .........



Cut out a good sample section and copy and paste it several times over then play it to a frequency meter.



Again ...... irrespective of the frequency read only one sensible answer will be obvious. You could be measuring 1/4,1/2,3/4, 4/4s of the actual revs. It can't be out by miniscule amounts, only out by steps of 25% an then even knowing which 25% is superfluous.





The Ducati would sound something like a mix of the musical notes C3 and C4. Mostly its the C3 we can hear cos of the big bang note.



If it was near



19,000 it would sound like Eb 3 or 4



18,000 .............................. nearer D 3 or 4



Upload a free guitar tuner with frequency meter on your Android phone ( thats what i use these days ) and measure it live with that next time.



Or a function generator with say square wave and saw wave options then play it as Rossi goes past ...... if you are a really good guesser you could tell by the harmonics

Ok that ones a bit "out there" but I have done similar in the distant past.

Click to expand...

Don't see that working out, dopplereffect and all, especially not with some cheap tuners software using a telephone. Nice thinking though.

Doppler effect won't throw it far off ( bearing in mind this is not a high sample low uncertainty measurement ) , but thats part of why the onboard sound bite is better. You can actually work out the doppler effect anyway, so long as you catch the changeover.

You can actually use the dopper effect as well, it points to the correct frequency. As with all tools you have to know how to use them.



I wouldn't underestimate free tuners/frequency software, its a pretty known piece of kit these days. I have used one to set the cam belt tensions on a Ducati.The manual says use the multithousand dollar Mathesis tune console to set the belt to 110hz ,,,,,,, 110hz is a guitar string tune ......... so cowboys can tune their Ducati's round the old campfire I suppose. I was sceptical at first but after doing it and checking it with several devices I have ( including actually picking the guitar string in harmony ) I'd have to say its a great way to set belt tensions.



Anyway they are far more accurate than the old strobe tachs i previously mentioned, all you had with those was an analogue tuning dial, and not much in the way of another standard to check them by.



Android phones are great! On mine I have loaded an oscilloscope, frequency meter, several spectrum analyzers, chromatic tuners, a function generator and an engine tuner which accesses bluetooth sensors where needed. Then there is the GPS programs some of which use the accelerometers to give a dyno reading from the known weight of you and the bike .... though thats pretty dicky

Dopple post!

http://motomatters.com/files/battaini-ducati-2.wav



http://motomatters.com/files/rossi-4.wav

...

Kropo, did Rossi change gear just as he got to you there? At about 2.35s

Need to get an onboard Krop, the following Honda? is influencing the waveform and I cant filter it out well enough at where I believe is the spot where he must have been adjacent to you.



Its pretty dicky but it almost seems that it got to 19,800 then he shifted gear just after your position and the revs dropped to around 15,800.



But again ...... its too dicky a wave. There may even be pulses from the honda in there.



If not an onboard I believe a soundbite taken at a larger distance away, running running across/perpendicular to the recorder position, with only the Duc. on track would be good.



It kinda shows that to get a decent sound bite reading, a lot of the readability would be in the setup.



I'll try for an onboard somehow.

BarryMachine said:

Need to get an onboard Krop, the following Honda? is influencing the waveform and I cant filter it out well enough at where I believe is the spot where he must have been adjacent to you.



Its pretty dicky but it almost seems that it got to 19,800 then he shifted gear just after your position and the revs dropped to around 15,800.



But again ...... its too dicky a wave. There may even be pulses from the honda in there.



If not an onboard I believe a soundbite taken at a larger distance away, running running across/perpendicular to the recorder position, with only the Duc. on track would be good.



It kinda shows that to get a decent sound bite reading, a lot of the readability would be in the setup.



I'll try for an onboard somehow.

Click to expand...

I got about 30 or 40 sound clips, and they are all the same. I also looked at sound from that onboard lap of Stoner's that they put online, and grabbed the audio from that. Same problem.

Gotta admit, the sound on that Stoner clip was a shocker ....... but mainly cos the slipper clutch makes a hell of a racket!



Do you have any where rossi is at one revs for a longer period?