Low cost digital readout of velocity and flow

[Moriniman]

That's the new Corsaro from the reformed Moto Morini company

Corsaros new and old.



The yellow bike in the forground is the very bike that's going to be racerised. Strictly speaking it's not a Corsaro, but the American model called the Hurricane.

This is my 1977 3 1/2 Sport at Cadwell park (sorry about the size) at my very first track day a couple of years ago.



The new 9 1/2 is in fact a detuned 1200, just the name harking back to the quirky 3 1/2 designation for the original 350.

is the offical factory site and lists the historic models.

My website is (needs updating).

One of my other rides; very rare they only built 250.

[Tony]

[Moriniman]

Yes, a uProcessor is really the way to go and is probably cheaper these days, but you rapidly get to the stage where it's easier to output data to a PC rather than have a stand alone display. Then you're on the slippery slope to a full DAQ/ Control system.

[86rocco1]

Someone that clever's and good at writing code, could probably do some very interesting thing with an . For $20 you get a fairly powerful 8-bit microprocessor with a couple available 10 bit a/d channels, an LCD display, joystick input, rs232 i/o, all in one tidy little package.

[77-1201624802]

[86rocco1]

I look forward to seeing what you can come up with, I wish that I was capable of writing some code myself but given my current skill level, the learning curve would be just too steep considering the amount of time I'd be able to devote to the project.

In terms of sensors, I'm thinking more of the Freescale MPX5000 series of sensors rather than the MPX2000 series. The 5000 series sensors have an output signal that's already conditioned so you don't need any external amplifiers. That makes them a little less versatile but much easier to use. For most flowbench applications,a 10 kPa sensor would probably be adequate if not 50 and 100 kPa sensors are available as well. The current street price on the MPX5010dp is about $16 Cdn, roughly 10 Euros.

[77-1201624802]

That sounds even better.

I will need some weeks/month (due to work load at the real life job), but I am sure I can build something here.

Cheers,

Daniel

[Tony]

As explained earlier, there are two very different ways to generate an inverse square law function in software. The first is a full lookup table, but that requires a lot more memory storage capacity than most simple microcontrollers have.

The second method is to number crunch. Adding and subtracting are usually the only two direct commands available on a basic microcontroller. Multiplying and dividing can be done with multiple additions or subtractions, (plus shifting). Finding the square root of a number is certainly not an easy straightforward process.


There is a third method that uses both, a much smaller lookup table, plus interpolation, but that is probably going to be even worse!

I did go through all these various options in my mind, once the initial idea I had turned out not to work. I eliminated all of them as being impractical as a very low cost simple home construction project. Ready access to a personal computer has made a lot of other ways of doing this, much too complicated, impractical, or expensive.

[86rocco1]

I'm not sure how good the machine code that's generated is but there are high level programming languages with math libraries available for that processor, they include square root.

[Tony]

[bruce]

Here is a sensor that has a square root function:

[Tony]

[Tony]

Another thought about that sensor Bruce.

It appears to measures pressure indirectly, by sensing the (cooling) airflow through a small internal orifice, instead of measuring the actual pressure against a strain diaphragm.

That has one huge advantage, in that it will not be damaged by a gross overpressure accident. That overcomes one of the biggest problems of very low pressure measurement rather neatly.

The only potential problem I can see is it might be susceptible to contamination over long periods from very fine dust, condensed oil vapor or anything similar. The thermal sensor might eventually get coated with crap, but a very simple in line air filter on the high pressure side should get around that problem.

[bruce]

But, they are pricey! I happened to stumble on them last evening price was around $90 US if I recall.

Might be able to use a Honeywell sensor for static pickup to keep the cost of a unit down?

[Tony]

Actually that price is pretty good Bruce for what it is, I was expecting more like double that.

But if you wanted to build a digital readout of CFM for an orifice flow bench, without needing to use a PC, one of those sensors plus a low cost digital voltmeter module would do it.

The best, and most economical solution still keeps coming back to a low cost pressure sensor, a commercial analog to serial Chinese "black box". and a spread sheet program loaded into an old PC.