PDA

View Full Version : Dusty air flow - some measurements



BobL
22nd May 2006, 11:03 PM
This evening I made some air flow measurements on my 1HP Sherwood dusty and associated DC system. The dusty's impeller is nominally rated at 600 CFM (Ha!). The dusty is located in an enclosure outside my shed and uses a pair of standard 30 micron bags.

I used a TSI hot wire anemometer gas velocity flow meter. The problem with using this device is that the air speed is measured for a small 1cm long wire and the not consistent across an orifice. The easiest thing to do was to measure the highest speed. I converted the air speed to maximum CFM by using the cross sectional area of the orifi.

Experiment 1, Dirty versus Clean:
By accessing a 100mm diameter inspection port immediately in front of the impeller intake I can measure the airflow of the dusty by itself.
DIRTY: The first measurement of 260 CFM made was at the intake with the collection bag still containing about 150mm of sawdust in the bottom bag.
CLEAN: I then removed the sawdust and shook out the bags. The airflow at the impeller intake increased to 340 CFM.

Conclusion: Clean dusty regularly.

Experiment 2, collection run measurements
I have 3 machines hooked up to the dusty using a combination of 100mm straight PVC pipe and connectors and 100mm flexible PVC hose. I will describe each "run" in terms of length + degrees of turn. eg 90o 1m 90o 1F = a 90o PVC sewage pipe elbow, followed by 1m PVC pipe, a 90o PVC elbow and then 1m of Flexible hose. Because I only wanted to assess the effect of the different runs the connections at the machine ends were disconnected to allow measurements at the final 100mm diam opening

i) Joiner: 90o 0.5m 45o 45o 1m 90o 2.5m 90o 2F (total of 6m and 360o) gave 238 CFM or 70% efficiency

ii) TS: 90o 1.5m 90o 2.5m 90o 2.5m 90o 2F (total of 7m and 360o) gave 246 CFM or 72% efficiency

iii) Router:90o 1.5m 90o 0.5m 45o 45o 2m 90o 3F (total of 7m and 360o)gave 228 CFM of 67% efficiency.

Conclusion: 6-7m runs with a total of 360o of turn are about 70% efficient.

Experiment 3, Dual run measurements
Having both run ii and run iii described above open at the same time, gave reduced efficiency of 57 and 43% respectively for each run so together they give 100% efficiency overall!

Experiment 4, coiling up a flexi hose
I took the 3m of flexi hose and compared the air flow straight through and then coiled into 3 x 360 = 1080o of tight turns. This only dropped the air flow by about 3%!

Anyone got any other ideas for experiments?

Markw
23rd May 2006, 07:51 AM
Did you actually measure the performance of the turbine, ie no bags and no pipes. This probably would be a lot closer to the machine rated spec. which would be what they advertise with. The spec is probably a "free air" type capacity.

Once you stick a bag/bags on them or actually connect a pipe to it then efficiency must drop. Especially pipework because you need vacuum to make it work which can only happen once a venturi effect (the pipe) is created.

BobL
23rd May 2006, 10:09 AM
Did you actually measure the performance of the turbine, ie no bags and no pipes. This probably would be a lot closer to the machine rated spec. which would be what they advertise with. The spec is probably a "free air" type capacity.


The 340 CFM was effectively a "no pipes" measurement. The impeller intake has a permanently mounted guard on it so I attached a 100mm length of 100 mm pipe to the so I could get a clear reference diameter intake.

The no-bags idea is a good point, I'll do that as soon as it becomes light.

I was also thinking the supposed 600 CFM rating might be the RPM x swept volume of the impeller.

Cheers

Bob38S
23rd May 2006, 01:02 PM
Interesting - keep the info coming.
Bob

Skew ChiDAMN!!
23rd May 2006, 02:16 PM
I've noticed a few people comment about improved performance on a take-off if they crack open the end of the straight main to allow a bit of flow. Apparently it reduces the "dead spot" in the wye, improving overall flow.

Here's a suggestion to see if there's something to this:

Measure the flow at a take-off, then also open one further up the line and repeat the measurement at the first. Close the one you just measured, and take a third measurement at the one further up the line. Re-open the one you started with, taking a final measurement of the one further up the line.

To be honest, I wouldn't be surprised if the one further up the line loses more efficiency than t'other, even though "common sense" suggests they'd both drop by around 50% or so.

Markw
23rd May 2006, 04:23 PM
I was also thinking the supposed 600 CFM rating might be the RPM x swept volume of the impeller.

Cheers

I'd lay money on it. Of course the RPM is a no load under best conditions and the swept volume probably doesn't come with a co-efficient to provide a real life efficiency.

Certainly is pie in the sky

chrisp
23rd May 2006, 05:02 PM
I have wondered if we 50 Hz countries are being duped with the CFM figures. I've noticed that the specs on dust collector sold in the US and Australia,(take JET for an example) often quote the same CFM figures in both countries. But here our motors run about 20% slower - and therefore the CFM should be about 20% lower.

I suspect this may also have some impact on the HP figures that are quoted on some machines as well.

Chris

BobL
23rd May 2006, 09:27 PM
Measure the flow at a take-off, then also open one further up the line and repeat the measurement at the first. Close the one you just measured, and take a third measurement at the one further up the line. Re-open the one you started with, taking a final measurement of the one further up the line.

No problem, but let me just get this straight.
i) Measure the flow at take-off A, with another take-off further up the line at B closed.
ii) Open B, and meas A again
iii) With B still open, close A, and measure B
iv) Open A and B and measure B?

If so. I have done this but did not observe the effect your suggesting.
i) A was 230 CFM
ii) A was 195 CFM
iii) B was 246 CFM
iv) B was 145 CFM

The outcome of this is with A & B both open, the sum of A+B = 340 CFM > A (with B closed) OR B (with A closed). So you do move more total air but the air flow at A or B is still reduced. The effect you refer to may be more evident with one off-take more restricted than another.

BobL
23rd May 2006, 09:36 PM
After much phaffing about dodging rain and trying to get dinner started for SWMBO I have managed a no-bags/no input resistance restriction measurement. The air-flow out of the square x-section of the dusty was too fast for my meter so I had to systematically sample spots at the top and bottom of the bag holders. After 24 measurements and a bit of arithmetic I get . . . .

590 CFM! (Dusty is rated at 600 CFM)

So the outcome here is the 1HP dusty drops from 600 CFM to 340 CFM just by adding the 30 micron bags and drops to 260 CFM with about 150 mm of sawdust in it.

Skew ChiDAMN!!
24th May 2006, 01:21 AM
No problem, but let me just get this straight.
i) Measure the flow at take-off A, with another take-off further up the line at B closed.
ii) Open B, and meas A again
iii) With B still open, close A, and measure B
iv) Open A and B and measure B?

Yep. That's the basic idea.


If so. I have done this but did not observe the effect your suggesting.
i) A was 230 CFM
ii) A was 195 CFM
iii) B was 246 CFM
iv) B was 145 CFM

The outcome of this is with A & B both open, the sum of A+B = 340 CFM > A (with B closed) OR B (with A closed). So you do move more total air but the air flow at A or B is still reduced. The effect you refer to may be more evident with one off-take more restricted than another.

Actually, you've given me food for thought.

Look at it this way: with both A&B open, A dropped to 85% of just A open while B dropped to 59% of just B open. That's a significant difference. If the drop was more or less equal across both, I'd be inclined to say "Aha! They must be imagining it!" but as it stands... Hmmmm... From my reading of the other peoples posts, they don't open B all the way, they crack it open just enough to "improve flow at A" as it were. The theory seems to be that the airflow stops the vortex effect in the unused arm of the wye. I'm probably not wording it properly, but I know what I mean. ;)

Also, the increased cfm for both open indicates that the duct isn't the limiting factor... if the duct from the wye to the DC can handle the higher cfm, so can either of the ducts after the takeoffs, which points at the wye being the choke-point.

Unless you're using a 6" main and 4" takeoffs?

If I had a good flowmeter (I wonder if there's a quick'n'easy home-made job?) I'd be running some test myself, measuring at A while adjusting B from half-open to nearly closed. But I'm happy to sit in the sidelines and watch someone else do the work. ;):D

Markw
24th May 2006, 08:53 AM
Skew
By cracking the "B" side with "A" side fully open your reducing a "dead spot" at this junction.

Consider the fluid flowing happily through the "A" tube at a constant laminar flow. Suddenly it flows into a larger chamber losing the laminar flow and reducing in pressure. By entraining a small amount of air at this point, the chamber fills slightly so that the pressure reduction doesn't occur and the laminar flow isn't as severely effected. Its all to do with Bernoulli's Equation - p + 1/2pV^2 + pgh = Constant (very hard to type squared) but you probably go crazy trying to apply it.

BobL
24th May 2006, 10:15 AM
Also, the increased cfm for both open indicates that the duct isn't the limiting factor... if the duct from the wye to the DC can handle the higher cfm, so can either of the ducts after the takeoffs, which points at the wye being the choke-point.


Yeah interesting . . . . especially when there are still 2 x 90o bends between the dusty and the wye junction. After the wye junction both arms have 2 x 90o bends and a coupla meters of flexy hose. I starting to suspect the flexy as being the major contributor to line inefficiency. Will test this soon.


If I had a good flowmeter (I wonder if there's a quick'n'easy home-made job?) I'd be running some test myself, measuring at A while adjusting B from half-open to nearly closed. But I'm happy to sit in the sidelines and watch someone else do the work. ;):D

I'll do it, maybe on the weekend - work is busy busy busy.

Skew ChiDAMN!!
25th May 2006, 12:32 AM
Yeah interesting . . . . especially when there are still 2 x 90o bends between the dusty and the wye junction. After the wye junction both arms have 2 x 90o bends and a coupla meters of flexy hose. I starting to suspect the flexy as being the major contributor to line inefficiency. Will test this soon.

I'll do it, maybe on the weekend - work is busy busy busy.

I've only the one 90° in my main, a vertical from the overhead into my seperator; I'm undecided whether to replace it with a couple of 45° a few feet apart or a length of flexi bent to about a 3' radius... but I think it can wait.

There's a couple of other parts of my setup that can also do with improving, so I'm following your experiments with interest, trying to work out in what order to start tackling things... I see no point in fixing "minor" problems and leaving the major improvements 'til last.


Consider the fluid flowing happily through the "A" tube at a constant laminar flow. Suddenly it flows into a larger chamber losing the laminar flow and reducing in pressure. By entraining a small amount of air at this point, the chamber fills slightly so that the pressure reduction doesn't occur and the laminar flow isn't as severely effected. Its all to do with Bernoulli's Equation - p + 1/2pV^2 + pgh = Constant (very hard to type squared) but you probably go crazy trying to apply it.

Yup. I understand the concept, it's my vocabulary that failed me. ;) I've also found a big difference between what's theoretically possible (on paper) and what's actually feasible on my budget... I reckon you're spot on about going crazy trying to work it all out in advance. :D

BobL
25th May 2006, 01:03 AM
I've only the one 90 in my main, a vertical from the overhead into my seperator; I'm undecided whether to replace it with a couple of 45 a few feet apart or a length of flexi bent to about a 3' radius... but I think it can wait.


Do you want me to try and measure these differences for you. I suspect I won't be able to tell the difference outside measurement tolerance or uncertainty (estimate my measurements are only good to +/- 5%) but I'll give it a go.

I reckon I may have found a way to measure the efficiency of a single 90o bend. (NB I don't believe I can tell the difference, outside tolerance, between a single 90o bend and no bend). However, I found a big bag of 90o bends at work so I can connect up multiple 90o bends in series and plot a graph of efficiency versus number of bends. Maybe this weekend?

RE: I've also found a big difference between what's theoretically possible (on paper) and what's actually feasible on my budget.
Yeah, and there also appears to be a difference between what one might think works in practice and what actually does work in practice.

Skew ChiDAMN!!
25th May 2006, 02:02 AM
Do you want me to try and measure these differences for you. I suspect I won't be able to tell the difference outside measurement tolerance or uncertainty (estimate my measurements are only good to +/- 5%) but I'll give it a go.

It'd be interesting to see your results, but don't do it on my account. As I said, I think there are other things on my system that need improvements first... I don't expect that changing one 90° bend will make much difference. :)


I reckon I may have found a way to measure the efficiency of a single 90o bend. (NB I don't believe I can tell the difference, outside tolerance, between a single 90o bend and no bend). However, I found a big bag of 90o bends at work so I can connect up multiple 90o bends in series and plot a graph of efficiency versus number of bends. Maybe this weekend?

I think the difficulty there is the length of the straights between 'em. I believe an ordinary PVC 90° bend is too tight for smooth airflow (from memory a 3' radius bend is about right) and it creates turbulence. I'm not sure.
just how long a length of straight is needed after a 90° to smooth it out again. Anybody reading know? Just stacking 90's in a row will give you varying measurements, depending on how they orient to each other, etc., because you're just extending the areas of turbulence.


RE: I've also found a big difference between what's theoretically possible (on paper) and what's actually feasible on my budget.
Yeah, and there also appears to be a difference between what one might think works in practice and what actually does work in practice.

Danged right! :) I've a couple of lathes at the far end of my shop, so I ran the main duct direct to them, with a wye at the end to select between 'em. Of course, at the other end of the main there's other wyes running to machines near the dusty. It works, but there's a definitely noticable loss of vacuum at the lathe end.

Now, given your findings, it looks like I should shorten my existing main, removing the length going to the lathes and then running a seperate duct from the very first takeoff, back to the lathes. This'd mean that my longest run would only have the one wye in it's length (instead of 4 or 5 as it does now) and should work better. I hope! It looks counter-intuitive at first glance, but now that I think about it...

If you follow what I mean? A quick pic might be better... it's not to scale or actually what I have, but it should serve to show what I mean?