Thread: Air flow calibration measurements

I have gone back over all the data comparing BMH to no-BMH measurements and realised there is on average ~25% smaller variability in the measure air speeds in the test duct when using a BMH.

This is highlighted by this graph where the error bars for the BMH results show up "on average" as smaller than for the naked duct.
This means the BMH must be making the air inside the duct at the measuring point less turbulent.
I knew this was the case at the actual air inlet itself but it's interesting to see that this transfers to the air down inside the duct and would explain the improved performance of the hood.

Hi Bob,

Interesting work. Are you going to do particle density (#/M³) and or particle removal efficiency studies as well?

Thanks,
Rob

Originally Posted by rob streeper

Hi Bob,

Interesting work. Are you going to do particle density (#/M³) and or particle removal efficiency studies as well?

I've done quite a lot of this already but I'm open to suggestions for more experiments. Just bear in mind I cannot place the particle detector inside a duct containing sawdust or it will die. I can only simulate %dust removal/efficiencies using the lower amounts of dust in regular air.

I've been looking for a way to independently calibrate the air speed measurements of these meters and I have and think I have come up with something that might work using a motor vehicle.

The speeds required are between zero and 30 m/s = 108 km/hr.

The plan is as follows
1) find a quite piece of straight road with a 110 lm/hr speed limit on a calm day

2) Mount the sensors outside the vehicle away from the effect of the car pushing air out of the way. The meter component can sit inside the car easily enough.

3) Have a driver operate the vehicle at a specific speed as shown by the GPS using cruise control.

4) As a passenger I will operate the mobile phone that communicates with the Testo meter and have it relay the data to the phone using blue tooth. The other meters will have to be read and recorded manual

The faster the air speed the greater the accuracy will be in theory although the effect of the car pushing air out of the way becomes greater and may affect the results.

Target calibration speeds of 18, 36, 54, 72, 90 and 108 km/hr should cover it.

If there is a slight breeze I could measure it and I will need to know the direction as well to allow for it.
Or if the breeze is stead enough do a reverse run and average the two measurements?

Anything else I haven't thought about or any other suggestions?

Many years ago Burt Rutan tested his aircraft designs with scale models mounted above a pickup truck. He stayed in the back to monitor the instrumentation. You're in good company.

The probe would have to be mounted a long way from the vehicle to not be affected by the air flow around the vehicle.

Maybe out on a long pole in front.

Originally Posted by Bohdan

The probe would have to be mounted a long way from the vehicle to not be affected by the air flow around the vehicle.

Maybe out on a long pole in front.

Yep I guessed that - I will have to experiment.

At slow speed this will be less of an effect so I will start at slow speed and vary the distance away from the vehicle until it plateaus out so I know far I will have to go.

I used a length of 12mm diameter galv tube to make a mast that clamps onto a roof rack.
The mast holds the Testo sensor 1.2 m above the car and convinced SWMBO to spare me ~30 minutes drive she drove and I operated the iPad that drive the meter.

First promising thing was it all stayed in place - nothing fell off or broke.

Not ideal measuring conditions.
Midday on Kwinana freeway, moderate traffic - enough to make running under cruise control away from other vehicles difficult.
Ambient breeze was variable around 9 kph coming from a ~45º direction
Blue tooth drops out unless I have the window open and hold the iPad next to the window

We did a run back and forth between the Narrows and Canning bridge and got some reasonably stead sections of 80 and 90 kph.
Then on the way home we did a 50 kph run in a side street .

Doing a run back and forth enables the ambient breeze effect to be minimised but it does not take into account gusts or changes in breeze speed or direction.

The 90 kph setting on the GPS gave 82.9 kph (22.4 m/s or 10% low), 80 kph on the GPS gave 75.5 kph (20.4 m/s, 5.6 % low), and 50 kph gave 46.5 kph (13.3 m/s or 7.1% Low).
These lower differences are consistent with the difference between the Testo and the TSI meter I have seen inside my test duct at lower speeds.

We will take it out for more testing on a calmer and quieter day.

More testing this arvo.

I've tested flexy before but its always been a simple - "this is better than this" sort of test without trying to quantify the losses involved.

Part of the problem is the losses from ducting and flexy are not linear e.g. if you double the length of the flexy it won't reduce the flow by two and you will always get some flow no matter how long the flexy is.

The losses are know to be logarithmic so a KPI in terms of a "%loss per metre" is not valid. Instead one has to think of the "length of pipe to produce a 50% loss in flow" or the "half loss of flow length"
e.g. if a 5m length of flexy reduces the flow to 50% then another 3 m (i.e. 6m in total) will reduce the flow to 25% etc.

I disconnected some of my bits of flexy and measure the flow for each length relative to a length of straight ducting.
Then i used the changes in flow data to compute the the "length of pipe to produce a 50% loss in flow"
Note these numbers are for my 3HP DC - the actual numbers will change for other DCs but the order should remain the same.

I had enough pieces to test 4 types of flex.
150 mm CV is the highly flexible stuff from ClearVue .
T150 mm HT is some stiff quite corrugated High Temperature flex I picked up through Gumtree.
It looks like the green and black flex in this photo I use it on my lathe, sander and combo jointer/thicky
GreenHTflextest.jpg


100 mm CT is the standard soft transparent stuff from Carbatech
100 mm TC is the grey stuff from Timbecon with the plastic as opposed to wire helical support. This is a length I had behind the shed and is pretty old and beat up with several feet having being crushed and squeeze open again.
I don't think it's quite the same as what they sell lately.

100mmflexytest.jpg

However the CT flex has corrugations that are much more pronounced that the TC stuff which is relatively smooth bothe
100mmflexy.jpg

On the graph you can see the 150 CV is about twice as "lossy" as the HT flex. I was quite surprised by this as I had purchased the CV stuff to replace the short lengths of the HT I am currently using on my machines but now I will keep the HT stuff on my machines and give the CV flex to the mens shed.

Remember, the higher the number on the Y axis the better the flex (longer length needed to reduce the flow).
The tolerance of the measurements are not shown but are about twice the size of the actual blue diamond points shown.

summary.jpg


I was surprised to see that the CT 100mm flex is only about 2/3 as effective as the TC flex.

It also highlights the fact that only about 4 m of CT flex will knock the stuffing out of the flow.
OK we knew this already but it highlights why we should keep the flexy as short as possible.

Looks like I'll be keeping my 100mm and 150mm TC type flex, only have short lengths of around 600mm of the 150mm, was going to get something more flexible but now thinking I should just put up with it.

Originally Posted by MandJ

Looks like I'll be keeping my 100mm and 150mm TC type flex, only have short lengths of around 600mm of the 150mm, was going to get something more flexible but now thinking I should just put up with it.

My understanding is there are 2 types of TC flex, their older stuff with the round white plastic supporting ribs while the new one uses get square cross section supporting ribs. I have not tested the newer stuff. If anyone in perth has a piece around 2m long that I could borrow for 1/2 hour that is enough for me to test.

Next week I will also go to the Mens shed and borrow a length of the current Grey 150 mm CT flex and bring it home and test that.

Both of mine have round white plastic ribs. Look forward to the 150mm testing, all of this is really good info and thanks for all the testing.

How was the change in pipe diam accounted for in the calcs?
Cheers

Originally Posted by Lappa

How was the change in pipe diam accounted for in the calcs?
Cheers

The measurements are relative to a similar length of PVC ducting of the same diameter - then all I had to take into account was the length.

So order of measurement for the 150 mm size was

~3m long piece of 150mm PVC ducting
Then various piece of 150 mm flexy
Then back to the 150mm PVC ducting

Same thing was then done with the 100 mm ducting and piece of flex.

I have organised a setup to directly compare the GPS with the Testo using my van.
Although the van doesn't have cruise control I will do it using the camera method.

I have transferred the mast that holds the test probe from SWMBO car to mt van roof rack - this will hold the Testo probe
The mobile phone running the Testo software sits in a cradle next to my GPS on the van dash.
In the back of the van poking through the open cargo barrier door I have set up a camera on a tripod clamped to the door frame.
This will enable me to photograph the GPS and the Mobile phone at the same time.
The camera focusses and is activated by a small remote I can easily operate while driving.
That way whatever the speeds are on the mobile and the GPS, the camera will record both speeds simultaneously.

Just waiting for a relatively calm day.