19 replies to this topic

[Makuna]

I know it doesn't really matter as what you look for is getting them all similiar; but I was looking a pile of pressure sensors and the electronic gadget part of brain thought; I could build a sync tool if these pressure (vacume) sensors were in range to what the FJR willl draw.

Anyone have an idea?

[03HiYoSilver]

http://www.carbtune.co.uk/c.com/

[garyahouse]

I know it doesn't really matter as what you look for is getting them all similiar; but I was looking a pile of pressure sensors and the electronic gadget part of brain thought; I could build a sync tool if these pressure (vacume) sensors were in range to what the FJR willl draw.

Anyone have an idea?

Indeed you could. What would be the purpose, to simply use it? Why not just borrow one? Or maybe you wanna market it? Keep in mind that a TBS isn't real high on most priority lists.

Gary
darksider #44

[Pepperell]

Don't pay any attention to those guys. It sounds like a real interesting project. If you could find small 6 or 8 led bar graph displays it could be permanently mounted on the bike. Keep us posted on your project!

[Fred W]

Well... The spec* for throttle vacuum at idle is 33.3 kPa (250mm Hg, 9,8 in Hg) which translates to negative .333 bar or 4.8 psi. But that is just at idle.
If you want to use the vacuum gauges "real time" as Pepperell mentioned, you'd need to know the maximum vacuum that will be seen0. The highest vacuum the engine will produce is when you get it spun up to redline and then chop the throttle closed. That is likely to be considerably higher vacuum than the idle pressures.

As a point of reference, the Morgan Carbtune scale goes from 10 to 40 cm Hg (1.9 to 7.7 psi) and I've never seen mine max out no matter what I did to the throttle. That would probably be a good estimated working range for a real time vacuum gauge.




* that most people rightfully ignore

[garyahouse]

Is that what you were thinking, to mount it on the bike and use it daily? It never occurred to me that you might have that in mind. Once upon a time, I had a car with a vacuum gage that was supposed to help me get better gas mileage. I'm thinking it was in the late 70's. Now there's a thought.

Gary
darksider #44

[Powerman]

Mercury don't need no calibration!

[Makuna]

I know it doesn't really matter as what you look for is getting them all similiar; but I was looking a pile of pressure sensors and the electronic gadget part of brain thought; I could build a sync tool if these pressure (vacume) sensors were in range to what the FJR willl draw.

Anyone have an idea?

Indeed you could. What would be the purpose, to simply use it? Why not just borrow one? Or maybe you wanna market it? Keep in mind that a TBS isn't real high on most priority lists.

Gary
darksider #44

Purpose? Thats like asking a FJR owner why they ride motorcycles; because well, I like to do to it. Plain and Simple.

..., the Morgan Carbtune scale goes from 10 to 40 cm Hg (1.9 to 7.7 psi) and I've never seen mine max out no matter what I did to the throttle. ...

Exactly what I needed, thanks!

Edited by Makuna, 10 April 2012 - 06:11 PM.

[Makuna]

Mercury don't need no calibration!

I thought I would update this on the state of the project. I have prototype up and running on an Arduino and connected to my bike for final UI tweaking.

Why did I respond to Powerman, becuase he speaks alot of truth in those few words.

Although the sensors don't need much calibration (just calibrate to current altitude), the problems come due to them being digital sensors. Mecury is very analog, with some mass that dampens the movement to get a flat reading. Digital doesn't have mass, it also only reads for a small time slice every so often (like looking at a single frame of every second of the movie, it doesn't represent the movie).

I had to increase my sample rate to 1000 hz and add alot of filtering to see usefull readings. My visualization needs some tweaks and basically I am done.

The actually pressure ranges from ~ -40 kPa to ~ -25 kPa every second (those cylinders aren't standing still and the pressure of course is different at each point on the stroke).

Videos and photos will be coming soon (or maybe just a link to a Maker/Instructional writeup).

[RossKean]

Put a small orifice restrictor between the vacuum caps and your sensor. Consider adding a reservoir after the restrictor - maybe 25cc or so. These will greatly dampen the pressure fluctuations.

[Fred W]

But, but, but....

Since that is actually what is happening to the vacuum in real time, those pressure fluctuations well may have some added significance. This is information that is completely unavailable using a mechanically damped system like the mercury sticks or sensors with vacuum restrictions inline. If you have the actual raw data you can always process it to find the means, but can also filter it for finding the peaks (high and low) and also the peak to peak swing. I'm sure that this kind of information would be indicative of how the individual cylinders are behaving under varying conditions.

One can never have too much data.

[gixxerjasen]

So, you going to feed these digital sensors into a computer? I mean, ya gotta store and graph the data. I'm not sure what else you do with data but graphing it is a MUST!

[NBB]

You should be able to do this as a differential pressure measurement. Make sure you use a sensor with a true analog output or you will get a ADC digital step chop in your output. Look at the output on a O-scope and you will see what I am talking about especially if you are using the digital output sensors. I did a project like this using vacuum pumps with reed valves to measure flow a while ago. When set up properly we could even see a flutter that was traced to the reed valves causing some leakage.

RossKean has the right idea with a restrictor, the reservoir acts like a air capacitor.

Good luck with you project.

[Makuna]

So, you going to feed these digital sensors into a computer? I mean, ya gotta store and graph the data. I'm not sure what else you do with data but graphing it is a MUST!

The Arduino sends the data to my laptop; but there is a limit to the bandwidth so it doesn't include every sample; just the filtered and smoothed data, I have software that takes it and graphs in four bars with usefull text data to the mean reading and delta between channels.

[Makuna]

Instructable

http://www.youtube.c...bed/HkCHG0MW7HA

[Denver_FJR]

Mercury don't need no calibration!

Sure, but you'll never suck mercury or some other fluid into your engine using electronic sensors.

Instructable

http://www.youtube.c...bed/HkCHG0MW7HA

Very impressive. The youtube link is horked up but found a working link on the Instructable website.

Edited by Denver_FJR, 20 May 2012 - 03:34 PM.

[Fred W]

Kule!!!

[mcatrophy]

Instructable

http://www.youtube.c...bed/HkCHG0MW7HA

I like it. I'm an electronics engineer (officially retired but still doing two days a week), so I fully appreciate what you've done.

When the engine is idling (or held with a low speed), the four pressures rise and fall in a slow manner. Is this caused by beating between the actual instantaneous pressure and the rate at which the data is transferred?

If so, how about a bit more Arduino intelligence to smooth this out?

You could even get it really clever, let it determine the engine speed from the cyclical variation and average over that period?

Or even compare the shapes of each cylinder's pressure cycle which could give clues as to compression loss, bad plugs, who knows what? Even if you have to make it "modey" so it can transfer a history of instantaneous pressure readings so that the PC can do any hard sums or comparisons?

It's like farkling a bike. You can always do more.

[Powerman]

I'm an electronic kinda guy so I can appreciate this project.


Constantly remind myself to keep it simple. I like the natural damping action of the mercury.
A long piece of hose looped high over the garage door rails keeps the mercury out of the engine.

[Makuna]

Instructable

http://www.youtube.c...bed/HkCHG0MW7HA

I like it. I'm an electronics engineer (officially retired but still doing two days a week), so I fully appreciate what you've done.

When the engine is idling (or held with a low speed), the four pressures rise and fall in a slow manner. Is this caused by beating between the actual instantaneous pressure and the rate at which the data is transferred?

If so, how about a bit more Arduino intelligence to smooth this out?

You could even get it really clever, let it determine the engine speed from the cyclical variation and average over that period?

Or even compare the shapes of each cylinder's pressure cycle which could give clues as to compression loss, bad plugs, who knows what? Even if you have to make it "modey" so it can transfer a history of instantaneous pressure readings so that the PC can do any hard sums or comparisons?

It's like farkling a bike. You can always do more.

Yeah, the current problem is that the samples are averaged over a constant and consistent time period, but the period it self does not conform to the engine cycle curve, so the values will vary over time.

I have an update in the works in the Arduino that solves this by finding the peek of the curves (cyclinder cycle) and average the samples over that period. The period will adjust with RPM natually. This provides not only the real RPM but will also provide a consistent value.