What am I doing wrong? Or is this to be expected. (video)

[SpaceGrey]

Here's how I spent my weekend:
Decided to build up my 11 inch TopHat into cyclone, but end up completely redoing it along the way.
As a result build Pentz style cyclone out of cheapest possible materials (hardboard for top part, galvanized flashing for a cone, plus plywood and 2x4 and some 3d printed pipes)
I was planning to use it just as a test (hence the cheapest material) to see how much better it is compared to TopHap design.
Along the way I spent so much time and got mentally exhausted, so I think I will stop here and will use what I build till it falls apart.
This cyclone on top of the bucket is so tall, it's almost touching my garage ceiling also I was barely able to squeeze it into the corner by garage door rails, but it worked at the end.

[SpaceGrey]

Here are my air flow measurements for all the modifications done to this point (both TopHat and Cyclone are 19" diameter, Harbor Freight impeller with original 5 inch intake plate):
- Impeller only nothing added: ~7000 ft/min which equals to 950 cfm (5 inch inlet)
- 6 inch TH with no outlet pipe: ~2800 ft/min - 380 cfm (60% loss)
- 6 inch TH with 3D printed straight outlet: ~ 3000 ft/min - 410 cfm (57% loss)
- 6 inch TH with 3D printed Parabolic outlet: ~ 3500 ft/min - 470 cfm (50% loss)
- 11 inch TH with long parabolic outlet: ~3600 ft/min - 490 cfm (48% loss)
- 11 inch TH with short straight outlet: ~ 3500 ft/min - 470 cfm  (50% loss) - this is interesting, taller TH doesn't care if outlet straight or bell-mouthed

Finally cyclone numbers from the pic in previous reply:
- Cyclone with impeller assembly as is (5 inch intake into blower assembly): ~ 4600 ft/min - 620 cfm (34% loss)
- Cyclone with impeller intake replaced by 8 inch intake adapter (see pic):  5200 ft/min - 700 cfm (25% loss)

At this point I'm thinking that if I would do the same thing with TopHat separator (remove 5 inch intake and replace it with 8 inch intake) it should improve performance (at least for Harbor Freight impeller).

[retired2]

SpaceGrey,

I want to compliment you.  You are one of the few people who have built different configurations and then captured data to show what each one is doing.  That is one helluva lot of work, but the data is so valuable.  There has been a lot said about the performance of 2X separators, but unless my memory is failing me, it has all been subjective.  Yours is the first I have seen with actual data.

I agree, it is interesting that the 2X doesn?t seem to care about the outlet pipe ending.

Thanks again for a great piece of work.  Unfortunately, you still don?t have a conclusive answer to why your performance losses are so much higher than mine.

[SpaceGrey]

SpaceGrey,

I want to compliment you.  You are one of the few people who have built different configurations and then captured data to show what each one is going.  That is one helluva lot of work, but the data is so valuable.  The has been a lot said about the performance of 2X separators, but unless my memory is failing me, it has all been subjective.  Yours is the first I have seen with actual data.

I agree, it is interesting that the 2X doesn?t seem to care about the outlet pipe ending.

Thanks again for a great piece of work.  Unfortunately, you still don?t have a conclusive answer to why your performance losses are so much higher than mine.

Appreciate your compliment.
I think I'll try to get to the bottom of it once I get my mental capacity recharged :)

Few question about your setup:
- What size inlet pipe goes into your impeller? Mine is 5".
- Have you measured your impeller flow in bare configuration? Nothing attached to it at all, no filter, no nothing?

BTW, adding that additional lip to create additional flow curve before it hits air from intake (per your link above) didn't provide any improvement in my case.

[retired2]

SpaceGrey,

I want to compliment you.  You are one of the few people who have built different configurations and then captured data to show what each one is going.  That is one helluva lot of work, but the data is so valuable.  The has been a lot said about the performance of 2X separators, but unless my memory is failing me, it has all been subjective.  Yours is the first I have seen with actual data.

I agree, it is interesting that the 2X doesn?t seem to care about the outlet pipe ending.

Thanks again for a great piece of work.  Unfortunately, you still don?t have a conclusive answer to why your performance losses are so much higher than mine.

Appreciate your compliment.
I think I'll try to get to the bottom of it once I get my mental capacity recharged :)

Few question about your setup:
- What size inlet pipe goes into your impeller? Mine is 5".
- Have you measured your impeller flow in bare configuration? Nothing attached to it at all, no filter, no nothing?

BTW, adding that additional lip to create additional flow curve before it hits air from intake (per your link above) didn't provide any improvement in my case.

My inlet is 5".  I have not measured the air flow with nothing attached.  However, one of my tables lists airflow with and without a separator.  All my measurements were taken at the end of a drop about 20 feet of 5" pipe away from the blower.  I believe one of my data points is without the exit filter.  That was done to determine a best case scenario in replacing the bag filter with a pleated Wynn filter.  I can't remember, but I may have posted flow data after I installed the Wynn.

[alan m]

SpaceGrey,

I want to compliment you.  You are one of the few people who have built different configurations and then captured data to show what each one is doing.  That is one helluva lot of work, but the data is so valuable.  There has been a lot said about the performance of 2X separators, but unless my memory is failing me, it has all been subjective.  Yours is the first I have seen with actual data.

I agree, it is interesting that the 2X doesn?t seem to care about the outlet pipe ending.

Thanks again for a great piece of work.  Unfortunately, you still don?t have a conclusive answer to why your performance losses are so much higher than mine.

i have to agree with retired 2 . this information is very helpfull. it really needs to be highlighted someway (along with a few other posts that are very important)

there is a real lack of hard data on all aspects of these seperaters. there are too many variables  and build styles to be able to acuratly compare everything.

that is some interesting data. its interesting to note that the single and double  had the same cfm
given they are the same cfm it would be interesting to see if they had diferent seperation efficiency

as somone who has been very vocal about double height seperaters (without hard data to back that up) i would like to see the results

i have meant to rebuild mine in a way that can easily be adjusted and test it better.
i have recently bought a cnc router adn have been toying witht he idea of making a modular version like a jigsaw 

[DustySanders]

Wouldn't the bellmouth effectively increase the diameter of the opening, and thereby reduce the measured CFM across it's mouth? If you are measuring air speed, and converting it to CFM by multiplying that by the area of the pipe, what diameter do you use for the bellmouth? The outgoing pipe size, or the outer lip of the bellmouth? 

Also, to me it looks like you have more air straightening material in the pipe than you need, and that may be reducing flow. The article linked to below points to a three vaned deice with a streamlined body at the center of the pipe as being the most efficient design.

http://www.jpthien.com/smf/index.php?topic=1048.0

[retired2]

DustySanders brings up the question of how you are measuring performance, which is a good point.  You could be seeing differences greater than mine due to your methodology.  I followed Phil?s procedure, and that is to take readings at the end of a straight pipe.  Take several measurements along the edge of the pipe and several in the middle, then average the results.  Because of the way air flows in a round pipe, the readings will be different along the edge than in the middle.

If you are measuring air speeds that are near or exceed the stated range of your anemometer, you should increase the pipe diameter with a reducer,(increaser), and then add a short section of the larger diameter pipe and then take new readings.  The FPM will be lower, but the volume is greater so the calculated CFM should be similar, and most likely more accurate if you are pushing the limits of your cheap anemometer.

[SpaceGrey]

DustySanders, I took that into account, didn't include any measurements over the bellmouth opening, they were quite off and I couldn't calculate properly what those CFMs would be.

retired2, I think the difference you and I see is because I did all the measurements without any hose attached into the input (which I admit is not a real life scenario). Getting blower measurement with say 10 ft of hose attached directly to the blower will lower flow significantly. Then if I'd do the same with 10 ft hose attached to separator, it'll be lower as well (vs. no hose), but the different wouldn't be as significant since we already got some flow loss over separator. As a result the flow difference (Blower+hose) vs. (blower+separator+hose) would be less comparing to scenario (blower) vs. (blower+separator).

[retired2]

DustySanders, I took that into account, didn't include any measurements over the bellmouth opening, they were quite off and I couldn't calculate properly what those CFMs would be.

retired2, I think the difference you and I see is because I did all the measurements without any hose attached into the input (which I admit is not a real life scenario). Getting blower measurement with say 10 ft of hose attached directly to the blower will lower flow significantly. Then if I'd do the same with 10 ft hose attached to separator, it'll be lower as well (vs. no hose), but the different wouldn't be as significant since we already got some flow loss over separator. As a result the flow difference (Blower+hose) vs. (blower+separator+hose) would be less comparing to scenario (blower) vs. (blower+separator).

I don?t disagree with your comments.  I just thought it it?s worth mentioning that in your final configuration use as little hose as possible since hose has about 3x the line losses of a smooth pipe of the same size.

[SpaceGrey]

retired2, makes sense.
I'm right at the point now where I'm trying to convert 5in to 2x4in slitter setup directly off the system into 6in galvanized pipe with short 4in hose sections nested on that pipe.

[earnsdorff]

I'm new to the Thien baffle and I'm building my 2nd version to see if I can improve on my first one from the lesson's learned.
I saw that this thread showed using a single measurement of air velocity from a vane anemometer to somehow calculate CFM of the flow.  I have a great deal of experience with measurements of air flows and such (so far none of the nuclear plants that have relied on my analysis have had an issue and many other systems as well).  I don't see how this measurement method will produce anything close to accurate measurements.  If you maintained the measurement in the same location while changing only other aspects far removed from this flow area then it may be effective in seeing a relative effect.  But making changes at the inlet section where the measurement is made and using the vane anemometer to see the maximum air speed (I guess this is what is being done) doesn't equate to volumetric air flow. 
I think the differences are due to the measurement method and this is why the results don't follow logic.

[retired2]

I'm new to the Thien baffle and I'm building my 2nd version to see if I can improve on my first one from the lesson's learned.
I saw that this thread showed using a single measurement of air velocity from a vane anemometer to somehow calculate CFM of the flow.  I have a great deal of experience with measurements of air flows and such (so far none of the nuclear plants that have relied on my analysis have had an issue and many other systems as well).  I don't see how this measurement method will produce anything close to accurate measurements.  If you maintained the measurement in the same location while changing only other aspects far removed from this flow area then it may be effective in seeing a relative effect.  But making changes at the inlet section where the measurement is made and using the vane anemometer to see the maximum air speed (I guess this is what is being done) doesn't equate to volumetric air flow. 
I think the differences are due to the measurement method and this is why the results don't follow logic.

About the only thing you?ve said that I agree with is that using a cheap anemometer at the end of a straight pipe is not the most precise method to measure air flow for a particular setup.  That said, we are dealing with home shop dust conveying systems here, not rocket design.  If our measurements are accurate to within 25-50 CFM, it is plenty good.

I hope you agree it is a straight forward calculation to convert FPM to CFM for a given size pipe.  Secondly, the precision of the measurement is less important than the repeatability.  For a given setup and a given anemometer, you should be able to take a hundred readings and they should all be reasonably close - for sure close enough for dust conveying systems in a home shop.  So the methodology described here is plenty accurate to determine relative performance, i.e., how changes to the setup effect the flow rate.

In the thread on my top hat build there are several tables of data presented for various configurations I was testing.  Most of the tables also include power measurements to help validate the flow measurements.  Even noise levels are included.  Admittedly power and noise are not very sensitive measurements for measuring flow, but they should support the flow measurements.

So, I rest my case by stating I have taken many, many measurements testing my various configurations and plumbing.  They were all done with the same cheap anemometer.  The results have always been consistent, and made perfect sense with the CFM and SP rating of my blower.  So I would continue to recommend this methodology to anyone wanting to know something about the relative performance of their system and various configurations.  When someone gets a number that is out of wack, it is far more likely to be a design or construction problem than it is a measurement issue, assuming the methodology described in previous posts is followed. 

Oh, and for what it is worth, and it is not much, for several years I worked in the organization that designed, built and operated the Savannah River Plant for the U.S. Government. 

[WorksInTheory]

Great work and thank you for sharing! Helps a lot folks out in their decisions. I wish you did a 5" to 6" before 8" on the intake so we can see if there is a big diff and on the flipside if the 6" is a better balance of stress, velocity etc or should just go straight to 8" Any thoughts?