J. Phil Thien's Projects

The slot width does matter as well.  The 3" (really? 3"? that seems pretty large) slot of CCrafts setup is pretty large and undoubtedly hurting his fines separation - although as he notes he doesn't care  ;D.

I ended up widening my slot to 2", which is still wider than normal, due to the behavior of the shavings off of my planer.  I noticed a difference in sucking up piles of sanding dust in my tests with that vs a smaller slow.  Basically I put a clean cloth  filter over the outlet and sucked up a big pile of sanding dust and saw how much made it to the filter.  At 1.25" it did better than at 2", but it was really the difference between practically invisible and fairly noticeable but still darn good.

I don't know if anyone has done any studies on slot width vs HP/CFM (vs all the other things that matter like: separator diameter - bigger separates more but hits your cfm more, inlet/outlet placement, etc... etc...) but more info is interesting.

Based on comparing reviews and observed performance at the low-mid range I'd say that Phils baffle works at least as well and in at least some cases better than a lot of the crappier cyclones and about as well as some of the better ones.  At the high end I haven't seen anyone who has done any sort of comparison so... for ease of build and straight up performance the baffle is still a huge win in my book and while a cyclone _may_ work better at fines separation it will also eat more CFM and is a lot harder/more expensive to build.

If I ever find a higher powered blower cheap I'm still planning on doing another one, hoping to find at least a 3HP or better a 5HP and see how it works  ;D

Well - did anything help?  The peanut gallery is curious  :D

On the smoothing comment, mostly I found it critical around the edges of the baffle into the drop slot.  I didn't so much have a turbulence problem as a small "catch" problem where one strand would be temporarily snagged and then it was all downhill from there.

Steve,

I had almost exactly the same problem.  http://www.jpthien.com/smf/index.php?topic=574.msg3213#msg3213

Things that I found helped (I'm currently separating pretty well at all sizes, but did take a wee bit of a hit on the small stuff):

• POLISH ALL THE EDGES EVERYWHERE.  The wood feathers really hang up on anything and once one does you're hosed as the rest build up behind it in a hurry.  I believe that this is really the #1 thing to try first.  I know this from having spent several hours throwing handfuls of shavings into my separator and then taking it all apart to see what got stuck where.  With your build this may involve adding some sort of surfacing to make it super smooth.
• bevel the edges of the slot to make it thinner.  A thinner baffle does work better, although I believe that there is a lower limit somewhere (I had started with a piece if 1/32 sheet metal over a piece of wood (various thicknesses from 3/4" to 1/4" depending on where I was in the troubleshooting process), and that seemed to be to thin - although see item #1 as its very possible that the two issues were aggravating each other, and I never tried thinner than ~1/4" with the wider slot so .. yeah mostly speculation.
• Open up the slot.  I ended up at 2" which is pretty wide.  But that was - in the end - what solved my problem.  I did it in several steps of < 1/8" or so until it started working  :o  Took time but was trying to save as much fine dust separation as I could.
  
• Drop the inlet size to 5".  The advantage of the 6" inlet was minimal for me and I moved to 5" - this didn't totally fix the problem (I had done this before opening the slot) and - I believe (without strong evidence) that the size of the inlet needs to be less that some percentage of the diameter of the separator (what those numbers are - dunno!! - but I have a relatively small separator and it did change the behavior for the good some for me...), I'd probably do this last as you appear to have a fair bit of piping invested

Just a quick update.  I've been pretty busy doing other things instead of playing with my dust collection :-\   Sadly most of them have been non-woodworking related, was trying to finish up some yard projects before winter hit :'(  With the recent change in weather here I've been back inside more.  Anyway I've pulled about 20g of mixed "sort of fine" (2-3mm) shavings to "really fine" (essentially powder) table/bandsaw dust through the separator and have only had about ~1/2C end up in the bag (in comparison the 20g weighs about 4x the 30g of fluffy shavings I pulled through while testing so its a lot smaller/denser).  This is certainly a lot more bypass in relative terms than I was getting before I opened up the slot to 2", but is still pretty darn good (I've seen videos of some cyclones that do a heck of a lot worse   ;D).

The 50g drum is still sitting on the side waiting for a new build, however given my current project list and the relative effectiveness of the current setup I reckon it could be waiting a while.  At this point I'm not sure how much benefit it would buy me (except possibly enabling an upgrade to 6" hose - which appeared to have a little more  flow).

Didn't actually get much work done on the new separator.  Instead I was running with the current one (builder gotta build, when swmbo starts asking when you're going to actually build something useful instead of building more tools  ::)). 

Even with the crazy wide slot on the small top hat it works remarkably well.  I ran about ~10 gallons with of lyptus chips/dust into it (mostly jointer/planer for volume but a more runtime on the TS which produces super fine bits with this wood).  These were narrow/smaller pieces and come out as small ~1/4" long crumblies from the planer which it handled well before.  The inside of the bag below the filter has a light dusting of the really fine red powder in the high impact spots but is surprisingly clean considering...

I won't have a chance to do the new build for another couple of weeks but will check back in once I get to it.

Rick: The chips are definitely less stringy on hardwood.  I don't have as big of a piece of that as I do soft at the moment (well, I do.. but I'm not willing to shave that piece down for testing  ::)) so I can't really "push the envelope" as much with that.  I do use a fair bit of softwood work simply because I can get it a lot more readily/cheaper for stuff that doesn't need the stability/appearance (and for some rustic pieces pine just looks good to me).  The dust port is actually 120mm, which is just a smidge less than 5", but I found a local plumbing supply place (or really junkyard but the most awesome junkyard ever - if you're ever in the columbia river gorge area check out Reds Trading Post - no affiliation, just spent a few afternoons wandering around there ;D) that has 5"<->120mm rubber fittings that make great quick connects.  You can also find the fittings online but Reds was actually cheaper.

I'm not sure how much another pre-filter would affect cfm, but it would take up even more space which is already at a severe premium so... yeah filed in the options box but exploring other things first.

I picked up a 55g drum earlier this week on the way home from work (free at the local recycling place, had foaming caustic cleaner so pretty safe once thoroughly flushed - know your chemicals and make friends at the local junkyards) and am working on a new build using that; both because of the greater diameter which I am hoping improves separation as well as the improved capacity.  Filling a 30g drum mostly full rough finishing ONE BOARD made me rethink things a bit there  :o  (ok so it was a pretty rough board, but still...). I'm hoping to have that pretty much complete this weekend and will report how it goes.  My current thought is that the initial build was way undersized for the chip volume.

In the interest of science^(tm) I went ahead and cut up my current top hat some more and opened up the slot to a crazy 2" wide.  The slot width as a function of the separator size is now somewhat ridiculous, but hey I'm not planning to keep this one at this point so why not try it :D.   As expected this threw the fine dust separation even further out the window (still amazingly good considering but down to maybe 90% from the almost perfect it was before), but it actually does seem to be able to most separate the planer shavings now.  I was able to basically stick the hose into a drum full of the shavings I've created so far and suck them all up with only two minor blow by events when a really big clump got caught.  Running the planer actually seems to work pretty well now as it makes little to no fine dust.   This supports my "chip volume vs slot opening area matters when chip volume reaches some critical level" hypothesis.  This also lends hope to the larger separator working better.  I believe that I can probably use a narrower slot on the larger top hat, because I think the required size is at least partially a function of opening area vs chip volume and that the actual slot width only matters as part of that function (after you've crossed the threshold where the chips fit through the slot which is another part of the function).  Its pretty clearly more complicated than that and I know I'm simplifying the heck out of it, but overall opening size does appear to matter for large chip volumes.  So my tentative plan is to figure the current opening area and cut a slot wide enough to have as much area as that into the new top hat to see how that works.  That will I believe either support or disprove my hypothesis (assuming that it ends up wide enough for the chips to fit.. multivariate functions are complicated  :P).  Even if I do have to open up the slot on the larger top hat, it should still separate better than the current setup because of the distance from the impeller inlet to the slot will be so much more (and I have some rough ideas on how the turbulence and indeed even the friction coefficients will change for the better).

I also think that the current top hat may be slightly too large for the barrel (yeah I know, but let me explain) - the current barrel has a bit of a lip (about 3/8" wide) around the inside about 2" below the top where the original top was (I cut the inside of the top off not around the sides - since it retained the more rigid piece on the very top that way).  This actually intrudes slightly into the slot and I think that that actually messes up the airflow based on where the sawdust lands and watching some dust go by it..  In effect this appears to make a narrower slot and perhaps worse may well cause some "kick back" pushing dust/chips back up into the slot.  When I cut the top off of the barrel I did what I could to round this off, but I believe that its still having an impact based on observation.  Its quite possible that had I chosen to use a barrel with smooth sides the current size may have even worked fine.  At this point I have no reasonable way to test how much that actually affects things.  Something else to put in the things that can go wrong file.

1) dry/wet? Its pretty dry - yeah I've seen shavings like this from green turnings, but not from dry wood.  I don't have a moisture meter so I can't say a lot more than - its very light for the size/thickness and drier than any wood you'd get from the local home depot  (not that thats saying a lot) but maybe a wee bit wetter than well cured furniture grade hardwood.  In any case its certainly not straight up green wood.
2/3) chipbreaker/adjusted?  Actually not sure.  I don't believe there are any adjustable parts for this in any case - or if they are its not obvious to me (fyi: its a hammer A3-31 - the knives are quick replace and pretty simple setup).

Thanks for the ideas all, much appreciated.  I've made a few mods and gotten a little more meat out of it.  Its still clogging up but not quite as bad - and now its clogging differently, so I'm calling this progress of sorts.  At the very least I'm figuring some stuff out :D

Quick summary for others, I believe my main causes of failure thus far are:

• Planer+wood combo is insane, "normal" tools wouldn't be (and indeed weren't) a problem here.
• top hat is to small for chip volume
• the metal skin baffle is a bit fussy.  It has to be uber smooth on the edges to work with this type of shaving.

I think its a little hard to convey with words the sheer volume of shavings I'm getting here so I've attached a few more pics.  The first pic is of a handful to show the general consistency and type, they also tend to occasionally clump up a little on the way out of the joiner/planer, which doesn't help any.  The second is of one shaving, they are micro thin and that 5" shaving is about average length - some are actually longer (how the heck a straight blade planer makes shavings like that is still confusing/amazing me).  The final picture is of a pile from 2 passes with my test board (about 11" wide x 7' long) - thats a 12" rule sitting next to it for scale - each pass takes roughly 1 minute so figure 50% of that pile in one minute  :o.  After only ~8 thin passes the 30g barrel was 75% full albeit super fluffy...  I do realize that I'm pushing the far edge of normal here, this is kind of a worst case scenario as far as the type/volume of shavings go (and practically speaking I normally work with harder woods that would take lighter cuts and thus have less volume and break up into smaller pieces) - so I figure if I can beat this one.. I'll have no problems with anything else :D  I'm actually not 100% sure what type of wood this is - it was part of the pallet the joiner came on - looks and cuts like pine but is a bit whiter than I'd normally associate with pine and has no observable pitch spots.

I'm starting to think that the 30g drum is to small for a whole bunch of reasons  :o  Retired2 - you are correct my separator is smaller than yours, its only about 15.5" ID.  I definitely agree that the size likely isn't helping any.  Observing the airflow with a temporary short (8" long) inlet it seems to "swirl" pretty good, but having done a wee bit of fluid dynamics modeling in a previous life I know enough to know that I don't know much in that regard and that what I think I know usually ends up being wrong  ;D  Switching to a 5" inlet did actually appear to help the smoothness of the airflow some so you are probably on to something there.

What I've done so since last post:

• Polished the edge by draw filing and then stoning (twice actually - once before and then again after opening the slot more).  dabullseye you are definitely correct here. This appears to be very important with the metal plate (and this type of shaving - matters not at all with sawdust/chips), as the slightest rough spot definitely causes these little strips to catch.  While testing with small amounts of shavings I also found a spot about 1/2 way around that was causing a small catch, and thus eventual backup.  This irregularity was barely detectable by feel, so even the smallest catch definitely mattered and demonstrating how touchy this config is
• Dropped the inlet to 5" (finally got the parts), this seemed to help some and definitely seems like the way to go here.  While bigger is usually better... its definitely not true in this case.
• Opened up the slot to a full 1.5" (was closer to 1.25" as of last post once I carefully measured).  I believe this was not really optional with this shaving volume after having spent some time spent dumping handfuls of shavings into the inlet at about the rate the planer runs them in (I'm not sure, but suspect that with a bigger top hat a smaller slot may well work ok because you have more slot opening overall).  While doing this (or more accurately while carefully measuring for it) I found out that the slot was actually about 3/8" narrower at the end of the slot.   I'm pretty sure this was also causing problems.  Opening the slot also made the edge a bit thicker because I removed all of the overhang and went into the wood a smidge.  Observation indicates that the slightly wider may well have actually helped some, at least for shavings.  I tried the wider slot with a 6" inlet (just for fun) and was getting quite a bit of bypass with the sawdust test - that seems to have stopped once I switched to a 5" inlet.
• Polished the edge again ::).

So.. after doing that.. what I'm finding is that the slot is still getting clogged up with shavings but its not necessarily (or even usually) clogging up at the end, but more randomly around the edge.  I've taken multiple passes, shut down and observed the clog spot and its usually at a spot about opposite the inlet but does move around some.  It could be that I have a few spots with a smidge of variation in width or smoothness, but I'm starting to believe that its just a simple volume vs opening size problem at this point.  I ended up cutting the guards out of the impeller inlet so I could  finish a full pass without being completely gummed up.  Before doing that I had one pass completely fill the inside of the separator plumb full when the clog hit early :o  Interestingly removing the guards also seems to have slowed down the apparent clogging, my theory is that the continued airflow actually pulls the clumps back out and eventually clears them unless they are really jammed in place.

Currently my plan is to move up to a bigger drum (50g) and thus a bigger top hat.  This is at least partially because I'm coming to the conclusion that 30g - which seemed huge - isn't actually all that big in this case!  I was slightly concerned that the little HF unit wouldn't be able to drive a larger separator (which was why I went with a 30g initially), but having observed the increased airflow from replacing the flex pipe between the filter/impeller with a short straight pipe and generally tightening things up I think it can drive it ok (and really I don't have much to loose at this point - other than some scrap plywood and about $20 for the barrel).

For the larger drum build I believe I'm going to steal galerdude's baffle sandwich idea because that will allow me to try different separator plate materials if I decide to change out for some reason (not that I'm having second thoughts on my choices so far or anything :D).  The thousand cuts sidewall skinned with sheet metal still seems like a win and is pretty easy to do.  I'm really not sure about the baffle though, likely will switch back to the tried and true tempered hardboard - why mess with success by being overly clever twice in a row.

Sorry to be so slow with the pictures, but the machines I was building this for showed up so I've been busy getting them setup and .. well.. playing  ;D  Attached is a picture of the inside of the baffle, the bottom of the baffle (well a close up of one part) - before I routed it down and two pictures of the cart with the baffle assembly removed for maintenance.  Its not nearly as pretty as some of the builds I see here by a long shot - nice work guys, them things are sexy!

First the good news  :), it works great with the table saw.  No measurable dust showing up from that and lots of suction.

Now the bad news :(  I'm having some problems with the new planer, it worked fine with my lunchbox and heck even picked up a decent size pile of hand plane shavings fine.. but... the shear volume of long stringies from the new floor planer is overwhelming it (from my test board the shavings come off in a heavy stream of paper thin pieces about 1/4" wide and 4-5" long - impressive in a way).  They build up starting at the "far" end  of the slot and eventually wrap all the way back around to the entrance.  I may well be pushing the far end of what the capabilities of this config are, but its interesting to see how/if this works anyway.

I've tried a few things so far:

• Routed out the underside of the baffle to slim the whole thing down to about 1/4" and shaved the edge of the bevel on the underside a lot thinner.  This helped a little but not nearly enough.
• Opened up the slot, I'm pushing 1.5" now.  Still clogs, but I can get most of the way through a wide board before it clogs up.

I'm thinking I'll try polishing the edge of the baffle some more tomorrow, small nicks in the metal really seem to catch the planer shavings.  If that doesn't work... well.. heh.. not sure yet.  Might try widening the slot more, but I suspect that I'm reaching some upper limit on functionality there.  I'm also wondering if I can do "something" around where the slot ends to reduce the amount of shavings that catch there.

Worst case I figure I can re-engineer the whole thing and move the baffle over the bag and I'm not really behind much and have learned some stuff.

I am certainly open to ideas on tweaks though.

  :D :D :D :D :D :D

First let me say thanks to all for posting the great ideas and excellent pictures and mostly to Phil for putting this idea and plan out there to begin with!

Nothing really new on mine and unfortunately I didn't take a lot^wany of pictures during the build as I was racing to get it done before some equipment showed up.  However I did try a few things along the way that I liked and a few I didn't like that I thought I'd pass on in case they prove useful for anyone else.

Started with the ubiquitous  harbor freight 2hp unit and added a Wynn 35a nano filter.  My barrel is a recycled bleach barrel I got from the local recycling place.  The price for taking one barrel was "take two", so I have an extra 30g barrel to do something with.  It pays to figure out where all the friendly local junkyards are.  Had to rasp the top "flat" (or close, some weather stripping makes up the difference), cut the top out and plug a few weep holes.

My build is a top hat made from 3/4" cabinet grade" (from HD, ok for shop cabinets I guess) plywood throughout.  The curved wall was made from the "death by a thousand cuts" and bend method.  The wall is set into a rabbet flush along the slot and glued in place with construction adhesive.  I skinned the inside of the wall with with sheet metal obtained from the ductwork section of HD (really easy, just cut to width with tin snips and construction adhesive in place, required two pieces so I overlapped them a smidge just past the curve on the inlet side).  Required a fair number of clamps to hold in place, but some strategically shaped/placed cauls could make up the difference.  The side sheet metal projects slightly (1/32 or so) below the bottom plate (its notched to fit flush with the bottom on the non slotted side) and was filed it "almost" flush on top (left a 1/100-2/100 sticking up to "seal" the top on).

For the floor/baffle (cut from one piece) I actually used 3/4 plywood but cut the slot 1/2" wider than it needed to be.  I routed a 45d chamfer on the bottom of the floating part and then skinned the top of the floor with more sheet metal with a 1/2" overhang.  The idea was that I'd get the strength of the plywood but the thin edge of the sheet metal to get decent separation.  This appears to be really solid and separates quite well (I've run both table saw sawdust and hand plane shavings through it no problem).  The impeller unit is simply "set" on top (they are actually both bolted to the stand..) so the intake run is super short/straight.  I used 5" stove pipe for the extension.  I tried a 5" to 6" reducer to try and increase the intake size a little, but that seemed to let more fine dust through (it was still really good but not as good as the straight 5").  I'll be watching the bell mouth builds with interest :D.  I think with a larger separator it would have been fine (my 30g plastic drum only allowed ~18" across inside). 

Currently I'm running a 6" inlet/duct, but am planning to drop that to 5" once I get some hose in.  Preliminary tests indicate minimal pressure drop at the distances I'm running (measured by the "from how far away can I suck up a pile of sawdust" method, no anemometer but still accurate enough in relative terms) and somewhat better separation, its also a more natural size for my machines (mostly 120mm).

I wanted the whole assembly to be movable as I'm still figuring out my machine layout and the barrel + everything wouldn't fit on the original HF base, so I built a new base/platform out of angle iron (mostly recycled bed frames I got for free - watch the local freecycle), some 1/2" plywood for the floor and some old casters I had laying around.  The top hat and impeller/motor assembly are both bolted to a stack of glue lam 2x3's that are bolted to 1.5x1.5" angle iron uprights.  I put a pair of angled cross brace to from the back to the front but just left those "floating" on the bottom (the top hat is also bolted to these to tie the front of it and the impeller together more).  The angle brace fronts sort of "clip" onto the front of the base with the way the angle iron goes together holding it in place but still allowing the the whole top hat + impeller to float a little.  This allows me to tilt the assembly back a 1/2" and squeeze the barrel underneath.   The inlet and outlet are on the same side, this puts the inlet right under the filter pushing the original plastic bag aside which is great for space efficiency and I can still get the bag on/off if need be (doesn't look like I'll have to worry about that).   I was planning on doing some sort of "lifter" mechanism to hold the barrel in place but found that just positioning it correctly, turning the unit on (to suck it up) and then wedging a couple of slats under it works quit well so may not bother with anything more complicated.

Overall performance is pretty spectacular.  My initial worst case test run was to sweep up all of the sawdust from the cuts for making the bendy piece and other pieces (roughly a 1' tall cone of pretty fine dust) and shove a 6" flex pipe into it, dump and repeat.  I had taped the original cloth bag over the outlet (it was still clean) to see how much fines actually came though.  With the 5"-6" expansion you could see "puffs" of dust on the outlet, with the straight 5" - nothing really visible and the dust on the bag didn't seem to increase measurably.  Yesterday I peeled a bunch of cedar shavings off with some hand planes (was sharping up/testing some old ones so was mostly just goofing around) and shoved them in the 6" outlet, nothing escaped.

Its not the prettiest beast, but performance is all I could ask for and more!  Thanks again all for the great details and information.

I'll try to take a couple of snapshots this week if I remember/have time around my other projects.