New Guy with a DC Pre-Separator

[KC7CN]

Hello Everybody.  I'm the new guy today.  I made this trashcan separator last October, and have since learned about Phil's design.  Needless to say, I'm impressed!

Click on thumbnail's to enlarge the picture.  When pictures are enlarged, comments are below the pictures.



Link to Pre-Separator album: http://www.ncwoodworker.net/pp/showgallery.php?cat=1022

I was not totally satisfied with how well the pre-separator performed, and decided to add Phil's baffle. 



Link to upgrade album: http://www.ncwoodworker.net/pp/showgallery.php?cat=1080

Initial tests with the upgrade are not very encouraging at this point.  I emptied the box on my table saw, and guesstimate about 30% of the dust went into the can. 

As you can see in this picture, I did cut the corner of the elbow per Phil's suggestion on this forum. In the following picture, note the white laminate (placed in the trashcan lid grove) that simulates the side of the can.  This orientation of the elbow was the first test.  The next two tests will be to move the elbow closer to the side of the can.

Enlarge the picture.  Click on the small arrow in the upper right-hand corner of the picture to view the next picture in the album.



My initial objective was to put the trashcan inside the frame of my Delta 50-760 dust collector.  This would require a special 5" to 4" adapter.  I made this one, but the was unable to make the can fit.  Win some, loose some!



Link to special adapter album: http://www.ncwoodworker.net/pp/showgallery.php?cat=1029

Comments, suggestion, observation, etcetera will be appreciated.

-Don

[D Romano]

Placing it under the blower on the 760 is a good idea. If there is enough room under there, it is really an ideal configuration. Why doesn't it work well enough? Does the DC have a 5" inlet to the blower? If so, why not use a 5" outlet from the pre-sep?

[phil (admin)]

I guess I would suggest double-checking all the obvious things, like that the baffle larger dimension is all the way against the edge of the can (no gaps), and that if you seal-off the input tube that your DC starts free-spinning (unable to get any air from a leak in the can or top).  But, your implementation look correct.

[phil (admin)]

Another thought occurred to me.  If there is insufficient CFM, or if the separator is "stuffed" (fed too quickly), you won't get optimum separation.  I have tested my units by pretty much burying the hose in a can of sawdust and find that separation efficiency is greatly reduced.  However, when running with a machine (table saw, jointer, planer, etc.) the makeup of air to particles makes the separators job much easier.

[KC7CN]

Placing it under the blower on the 760 is a good idea. If there is enough room under there, it is really an ideal configuration. Why doesn't it work well enough? Does the DC have a 5" inlet to the blower? If so, why not use a 5" outlet from the pre-sep?

I still haven't given up on this idea!  The DC has a 5'' input. There might be enough room.  With the trashcan placed on the DC stand (underneath the DC input port), the center of the trashcan lid is off-set about 2'' from the DC input port.  The shop-made 5'' to 4'' adapter almost works, but not good enough.  My research indicated the 5'' input would allow too much suction, and cause scrubbing (hope that's the right term).  You could still do this with a reducer inside the can!  I could not find a short section of 5'' hose, and would have to special order.  I think I can relocate the output port on the separator in order to line up the the DC, and make it work.  Need to get it work better first!

-Don 

[KC7CN]

I guess I would suggest double-checking all the obvious things, like that the baffle larger dimension is all the way against the edge of the can (no gaps), and that if you seal-off the input tube that your DC starts free-spinning (unable to get any air from a leak in the can or top).  But, your implementation look correct.

Thanks for the suggestions.  I don't think there is a lot of gap, but take a closer look today, and let you know.  The lid fits tight, and the seal is good.  I will also try the seal-off the input tube to check.

Another thought occurred to me.  If there is insufficient CFM, or if the separator is "stuffed" (fed too quickly), you won't get optimum separation.  I have tested my units by pretty much burying the hose in a can of sawdust and find that separation efficiency is greatly reduced.  However, when running with a machine (table saw, jointer, planer, etc.) the makeup of air to particles makes the separators job much easier.

Check out the photo below to see how I tested the upgrade.  The separator hose was connected to the bottom of the table saw collection box.  Disregard the white plug, it's not used with the hose connected!  The box was full of sawdust because I didn't have the DC connected!  For the test, I connected the hose, and removed access door (see picture), and used my hand to gently feed the sawdust into the hose opening.  There is a lot of suction, and it's possible the feed-rate was to fast!  But I tried!  ;D



Link to PowerMatic Dust Collection album: http://www.ncwoodworker.net/pp/showgallery.php?cat=845

My question for you:  Check out the three photo's below.  How close do you think the position should be for optimum performance (note the laminate that simulates the side of the can). I may be able to rotate the elbow while it's running to test it!

Test Elbow Orientation.


Elbow moved closer to trashcan sidewall.


Moved very close to trashcan sidewall.


Thanks for your input.

-Don

[phil (admin)]

I tend to think you may have been feeding too fast.  I would try again and this time slow down the feed rate of the debris.

I think the postition of your elbow is fine as-is.

[KC7CN]

I tend to think you may have been feeding too fast.  I would try again and this time slow down the feed rate of the debris.

I think the position of your elbow is fine as-is.

Yes, I think you are correct!  In addition, I feel the table saw sawdust used in the first test included a lot of fine dust, which would not drop in the can.

I conducted a series of tests today, beginning with a suggestion from Bas.

Here's the link to the photo album for reference: http://www.ncwoodworker.net/pp/showgallery.php?cat=1080

The first test was to shine a light through the viewing window, and observe the cyclone action; without the output hose connected, and then connected to the planner through a 10' hose.  You can definitely see the cyclone action; more action without the hose connected.

Separator Viewing Window


The next test was to run the dust collector for a minute to see what would happen to the sawdust in the can.  I spread out the sawdust and run the test DC for about 1 minute.

Result


I placed a copy of the baffle in the bottom of the can to show the pattern.


The next series of test were to determine the best position for the elbow: 1st (original), 2nd (middle), and 3rd (next to sidewall of can).  A 17-1/4' length of 2x4 was passed through the planer 3 times, at a cut depth of 1/64''.  I also made 3 passes at 1/32'' with same result.

Position #3 - Elbow next to sidewall


Position #1 - original position


Position #2 - middle position
This was the best orientation.  Although the picture does not show it, there was slightly more sawdust.


Detail of Elbow Position


The Baffle is 8'' below the top of the can.
Illustration

This should give you an idea of how well it fits.

In summary, I moved the elbow to the middle position.  I will be planing some Oak in the next few days, that should indicate how well it works.

-Don

[D Romano]

I looked a bit closer at your photos of the elbow locations, and I think that the reason why the performance was worse with it closest to the edge of the can is that the elbow is actually pointed toward the can. As phil has stated, turbulence is your enemy here. You need to keep the incoming dust stream tangent to the side of the can because you want it to slip through the gap in the baffle. As the dust is bounced ofr blown towards the center, it becomes more likely to get sucked up through the outlet pipe.

The ideal would be to mount inlet on the side of the can like a DC or cyclone would do, not the lid, but this is more difficult for sure. Not impossible though and it would make the lid easier to remove with only one hose stuck in it. What you are calling positions are really angles. To find the optimum position, you would need to offset the hole in the lid that the elbow fits through, while keeping the dust stream tangent to the edge of the can. What you will find in doing this is that closer the hose gets to the center of the can, the more turbulence will be created because the angle of attack on the side of the can will get steeper, i.e., in the center , the dust stream would be hitting the can head on no matter what angle it was placed at. (This function follows a sinusoid) The least turbulence will occur when the mouth of the elbow hugs the edge of the can.

David

[KC7CN]

I looked a bit closer at your photos of the elbow locations, and I think that the reason why the performance was worse with it closest to the edge of the can is that the elbow is actually pointed toward the can. As phil has stated, turbulence is your enemy here. You need to keep the incoming dust stream tangent to the side of the can because you want it to slip through the gap in the baffle. As the dust is bounced ofr blown towards the center, it becomes more likely to get sucked up through the outlet pipe.

The ideal would be to mount inlet on the side of the can like a DC or cyclone would do, not the lid, but this is more difficult for sure. Not impossible though and it would make the lid easier to remove with only one hose stuck in it. What you are calling positions are really angles. To find the optimum position, you would need to offset the hole in the lid that the elbow fits through, while keeping the dust stream tangent to the edge of the can. What you will find in doing this is that closer the hose gets to the center of the can, the more turbulence will be created because the angle of attack on the side of the can will get steeper, i.e., in the center , the dust stream would be hitting the can head on no matter what angle it was placed at. (This function follows a sinusoid) The least turbulence will occur when the mouth of the elbow hugs the edge of the can.

David

This makes sense!  The most interesting part of my series of test was to see the cyclone action when I looked through the view window.  I had to hold a light next to the window, shielded with my hand, to see, but you can definitely see the cyclone turbulence.

-Don

Edited for memilanuk

[memilanuk]

...but you can effeminately see the cyclone turbulence.

ROTFLMAO!!!

Might I suggest a different word... like maybe 'definitely' or 'certainly'?  ;D

[KC7CN]

ROTFLMAO!!!

Me too!

[phil (admin)]

ROTFLMAO!!!

Me too!

I can't believe I missed that the first time I read it.   :D

[Todd]

My initial objective was to put the trashcan inside the frame of my Delta 50-760 dust collector.  This would require a special 5" to 4" adapter.  I made this one, but the was unable to make the can fit.  Win some, loose some!

Don,

I have that same DC and I am very courious to see how this will work.

[Ken45140]

I have a question that all here might respond to...it is prompted by Todd's description of how he tested the unit. He essentially planed some wood to generate chips and debris.  But there is no way to determine the amount of chips (although it just occurred to me that I guess you could calculate the volume of wood planed off and use that as the theoretical input).

Here is how I conducted my test (which resulted in a 97.2% efficiency. I took the collected material from the bottom bag of my dust collector.  The majority of this stuff was fine particles (I have done no planing in the recent past).  I weighted out a certain amount of ounces of dust, then spread it out on the floor and sucked it up slowly, just like Phil did in his online video.  I then emptied the metal garbage can and weighted what had been collected there.  For completeness, I cleaned out the shopvac including the filter prior to the test, then weighed the filter. After the test I reweighed the filter. Results: Input amount = 60.7 oz, Separator = 59.0 oz, In filter = 1.7 oz, or 97.2% separation. Actually, there was probably some particles adhering to the sides of the ducts, the internal crevices of the vac, etc., but this seems like hair splitting. 

Do you think this was an "unfair" test in that the majority of the input debris was already pretty fine particles. Would the separation efficiency be higher with a more uniform mix of particle size in the input stream?

Ken