Computer Flow BenchControl and Data Logging

Anything that does not fit into pitot or conventional orifice flowbench design
Tony
Posts: 1438
Joined: Wed Mar 03, 2010 3:40 pm
Location: Melbourne, Australia

Re: Computer Flow BenchControl and Data Logging

Post by Tony »

Live motor air volume sampling would be a very good application, and a low cost data acquisition system such as you suggest quite doable, provided the user had suitable programming skills to get it all together.

The "funny" calibration curve problem is one problem, another is that these flow meters are very sensitive to up stream air condition, as they measure air velocity in only one small place.
That link you provided earlier:
http://www.microsquirt.info/mafmap.htm
Goes into the upstream air quality problem, and they suggest any small changes to the up stream air box, filter element, or pipework can have a significant effect on MAF calibration.

An up stream air straightener would probably be quite sufficient, but fitting something to the top of an exposed carb during very high speed road testing may be problematic.
Should work on a dyno though.

It should work in a flow bench too, but may be a problem if located immediately below the test hole without a settling chamber or a flow straightener.
All these slight complications can be readily overcome if they are known about and understood.

I had hoped to do this myself at one stage, but at the time I just did not have the resources here to make it work.
Also known as the infamous "Warpspeed" on some other Forums.
1960FL
Posts: 1338
Joined: Fri Jan 08, 2010 10:36 pm
Location: Maryland

Re: Computer Flow BenchControl and Data Logging

Post by 1960FL »

S2Z,

Where are you on this?

Rick
S2Z
Posts: 37
Joined: Tue Mar 08, 2011 9:49 pm

Re: Computer Flow BenchControl and Data Logging

Post by S2Z »

I agree that the MAF Volt curves earlier discussed are not linear, but at what flow rate were the curves created at?

With the MAF calibrated, I can develop my own flow curves.

However, some questions arose about the calibration process:
What flow rate to you calibrate at?
What resolution or increment of flow? 10 cfm?

My answers:
28" WC, because that is the industry precident, and the rate we normally test at.
I also believe that the Voltage to CFM is more linear at 28" wc.
10 cfm resolution should be more than acurate enough to develop the flow curves

From the testing that I have accomplished thus far, I have found that the voltage readings were stable, very predictable,
and more importantly:
Repeatable.
I can only say that until the MAF is calibrated with known flows, one can only speculate what the actual values mean.
I can also state that the flow numbers I achieved on a set of H-D Twin Cam Heads, were within +/- 1% of dyno results for the subject engine.
Hard to argue with that...
Tony
Posts: 1438
Joined: Wed Mar 03, 2010 3:40 pm
Location: Melbourne, Australia

Re: Computer Flow BenchControl and Data Logging

Post by Tony »

Agree, the readings are probably very stable and very repeatable for a fixed test setup.

But what are these readings exactly ?

Suppose I set up a cylinder head, and measure a MAF voltage of 2.754 Volts at 0.2" valve lift.

Is that a good air flow number for that type of cylinder head ?
How much airflow s 2.754 volts ?

That is the problem.
You have built a totally unique measurement system that has no reference, no units of calibration, and cannot be compared to anything else.

My car flat out on a home made chassis dyno develops 1.207 Volts.
Is that REALLY powerful or what ?
Also known as the infamous "Warpspeed" on some other Forums.
S2Z
Posts: 37
Joined: Tue Mar 08, 2011 9:49 pm

Re: Computer Flow BenchControl and Data Logging

Post by S2Z »

Tony wrote:
"Suppose I set up a cylinder head, and measure a MAF voltage of 2.754 Volts at 0.2" valve lift."

S2Z response:
The Z32 MAF is worth approx 450 hp, so ~150 CFM=100 hp making the Z32 MAF good for 675 cfm.
The MAF signal voltage range is between 0.5 VDC and 5.12 VDC
Assume less than or equal to 0.5 vdc = 0 cfm and 5.12 vdc = 675 cfm

Resulting in the equation:
(Signal Voltage-0.5)*(675/(5.12-0.5))=cfm
or
(2.754-0.5)*(675/(5.12-0.5))=329 cfm


Tony Wrote:
"Is that a good air flow number for that type of cylinder head?"
"How much airflow s 2.754 volts?"

S2Z response:
Obviously one would be required to do baseline testing, record the voltages / cfm in steps, make the changes and retest.
As an hypothetical example, lets say the baseline test of the same port, at the same 0.200" valve opening (@28"WC) was 2.54 VDC. (298 cfm) The modified port during re-test readings were 2.754 vdc or 329 cfm or 10.4% increase in air-flow.

Tony wrote:
"That is the problem. You have built a totally unique measurement system that has no reference, no units of calibration, and cannot be compared to anything else."

S2Z response:
There is a direct correlation between HP & CFM. The Signal voltage range is known as well as Minimum and Maximum HP (and therefore CFM).
Plus the fact that a head that I ported was tuned on a dynomometer, which yielded less than 1% difference in what my flowbench predicted, I respectfully disagree...
Additionally as a disiplined process is utilized and BASELINE testing is conducted, important information about the percent change in flow is obtained, again, I respectfully disagree...

In my previous post, I did state that IF the MAF is calibrated then I would be able to know EXACTLY what voltage for a specific CFM flow would be...
At this point though, it's a pretty darn good guess-timate considering the dyno results.
Tony
Posts: 1438
Joined: Wed Mar 03, 2010 3:40 pm
Location: Melbourne, Australia

Re: Computer Flow BenchControl and Data Logging

Post by Tony »

S2Z wrote:
S2Z response:
There is a direct correlation between HP & CFM.
There is certainly a trend between air consumption and power.
But not a direct correlation.

Over scavenging, incomplete combustion, wrong mixture, wrong valve or ignition timing can all cause a dramatic power loss, but you claim there is a DIRECT correlation of power to air consumption.

If that were true, engine tuners have been completely wasting their time.
Also known as the infamous "Warpspeed" on some other Forums.
1960FL
Posts: 1338
Joined: Fri Jan 08, 2010 10:36 pm
Location: Maryland

Re: Computer Flow BenchControl and Data Logging

Post by 1960FL »

S2Z,

Sorry if you are feeling attacked I do not think that is anyone’s intention, but more like helping you refine what you have. I take your responses as you are good with what you have and are not looking for input??? Help me out here. I think that what you have is a clean easy low power consumption approach to a DIY flow bench and with good electronics and some simple code could be a rock solid bench. If the bench does function the way you say it does you should easily be able to use a few different sharp edge orifice plates and flow them across a range of depressions say 1” to 30” and build a nice flow mapping/calibration table I/E comparing the orifice theoretical flow at given depression against the voltage reading/Flow. Also you should look into the DataQ DI145 is a 10 bit ADC that you could connect to trough Excel with VBA/Script to automate your output no the CFM side. 4.62/1024 = .0045v resolution! Could be nice? Would not be dificult to make this bench full digital.

Rick
1960FL
Posts: 1338
Joined: Fri Jan 08, 2010 10:36 pm
Location: Maryland

Re: Computer Flow BenchControl and Data Logging

Post by 1960FL »

SZ2,

I do think that the only Fly in the ointment here is MAF’s and any hot wire device are very sensitive to Air density and just as important Relative Humidity. The moisture content in the air is directly related to its ability to conduct heat and this can trick a MAF to believe more air is flowing than truly is or far less even worse. So your final math will need to adjust for these factors as large swings in RH will have major affects on output and you should be able to get some data by tracking weather and output voltage flowing the same orifice same depression on different days.

Thoughts,

Rick
S2Z
Posts: 37
Joined: Tue Mar 08, 2011 9:49 pm

Re: Computer Flow BenchControl and Data Logging

Post by S2Z »

Rick wrote:
"are very sensitive to Air density and just as important Relative Humidity. The moisture content in the air is directly related to its ability to conduct heat and this can trick a MAF to believe more air is flowing than truly is or far less even worse."

S2Z:
I couldn't agree with you more! At this point, I just record the humidity as I'm looking to integrate the humidity factor in my MAF equation... The first step is getting the MAF calibrated...

Rick wrote:
"Sorry if you are feeling attacked I do not think that is anyone’s intention, but more like helping you refine what you have. I take your responses as you are good with what you have and are not looking for input??? Help me out here."

S2Z:
I am looking for input, as I'm looking for ways of improving my bench and making it easier to operate.


Rick wrote:
"I think that what you have is a clean easy low power consumption approach to a DIY flow bench and with good electronics and some simple code could be a rock solid bench."

S2Z:
Thanks Rick, that's my goal, just working on these last few details to take it to the next level...

Photos of the motor installation:
http://s1134.photobucket.com/albums/m61 ... G_0204.jpg

Note that all of the air from the first stage exits into the suction side of the 2nd motor stages.
4 motors in Parallel and 4 motors in series.

http://s1134.photobucket.com/albums/m61 ... G_0203.jpg
1960FL
Posts: 1338
Joined: Fri Jan 08, 2010 10:36 pm
Location: Maryland

Re: Computer Flow BenchControl and Data Logging

Post by 1960FL »

S2Z,
Note that all of the air from the first stage exits into the suction side of the 2nd motor stages.
4 motors in Parallel and 4 motors in series.
Did you do this because you are looking to flow at higher than say 36" depression?

Rick
Post Reply