Good quality (smooth) UHMW cut on PM66 table saw with power feeder

Short summary: Need to rip UHMW ranging from 2.0mm to 11.0mm thick (not flat) and about 154cm long. Both edges of the UHMW must be 90•+/-.5•, and most importantly the edges of the UHMW must be completely smooth and free of saw marks.

About the process: We work with UHMW material that is about 13mm thick, and with lengths ranging from 130cm to 160cm. We use either a CNC router or planer to machine a "thickness profile" into the material. This thickness will always be 2.0mm on either end of the material and in the center be in the range of 7.0-11.0mm thick.

After machining, we take the material to a table saw with power feeder to rip into 12mm wide pieces. Down the line, this UHMW will be bonded on 3 out of 4 sides - including one of the saw cut sides. Table saw cut quality is crucial and anything with blade marks showing will have to be sanded on a belt sander.

Since upgrading our table saw, we have not been able to achieve smooth quality.

Smooth cut quality sides are 90•+/-.5•, surface texture is free of saw marks, is smooth, and has a matte or slight melt appearance.

We have achieved smooth cut quality before: Our current saw, and the saw we are having trouble with is a Powermatic 66 from 1978. Our previous saw, a Grizzly 715 did achieve good cut quality but had some other inherent problems and had to go. Although we were cutting smooth with this saw, we were going through about a belt a month and would sometimes have bind ups that would trip the breaker on the motor starter. Even when running well you could hear the material bogging the motor down. The other factor that turned us off was that we could not move the trunions enough to achieve perfect miter and blade parallelism.

The main reason for the upgrade was to have a more powerful saw. The Grizzly has a 2HP motor and power requirements limit us to 3HP at max. We decided to buy a 3HP Powermatic 66. I have heard nothing but good reviews of the PM66 and we were able to trade our saw + some cash for the PM66.

What could go wrong? I mean the only practical difference in our process is the PM66 has a slighty faster arbor RPM, 4000RPM on the PM66 and 3850RPM on the Grizzly. I did mention the Grizzly was bogging down so the actual arbor RPM must have been slower.

What we have tried:

• Same set up that worked on Grizzly - ie same blade, same blade height, same feed rate, etc... Saw sounded smooth and quiet like cutting better. Cut quality is a little... rough. Our saw vendor and re-sharpener believes the Triple Chip Grind (TCG) (meaning that the carbide cutters have 3 different grinds) and that the UHMW is causing blade wobble as it hits each different tooth in succession. More on this below in "What is currently (kind of) working:"
• Slower feed rates, faster feed rates. We tried every feed rate possible with our power feeder and then bought more gears to try more feed rates.
• Different blades - Every "plastic" blade that worked in the past, carbide tipped wood blades, carbide tipped plastic blades for a circular saw, HSS high tooth count blade meant for cutting plywood. We tried all these blades forwards and backwards. The best cut quality was achieved using the plastic cutting blades backwards.
• Blade stiffeners - We now have a blade stiffener on both sides of the blade.
• Backwards feed - with the blade installed both backwards and forwards I set the power feeder to reverse and fed the UHMW through. This cut quality was not any better than feeding the correct way.

What is currently (kind of) working:

We are achieving a very good cut quality with our blade, Freud LU94M010 10", 80T, -3• hook, carbide tipped, triple chip grind, .110" kerf installed backwards. We are running this blade backwards with our power feeder set to the slowest speed, 13FPM. The cut quality is actually very good. The surface is very smooth and never requires touch up. The angle is always within .1• of 90•. We believe that the blades is melting its way through the material and the triple chip grind is actually centering the blade rather than let it wobble as it catches going forward.

Running with the blade backwards does have its problems. The most troubling is the stress this is putting on the motor. The motor is getting very warm after ripping and sometimes bogs down - still not as much as the Grizzly bogged down. We are getting about 25% of previous tool life. Before we were getting about 1200 rips per blade sharpening now we are down to about 300 rips.

What we are thinking about trying:

• Blade vendor wants to use HSS and make a blade with no hook angle, a "v blade" that has no forwards or backwards, similar to a cold cut saw blade. We could try this for the cost of materials and arbor bushing.
• Slowing down our table saw arbor RPM either through different motor and arbor pulleys or controlling the motor speed. Our motor is 3HP, 1phase. I am currently researching VFD with single phase output. I know that at full speed our Grizzly saw was spinning slower at the arbor and will be even slower when bogging down. With speed control we could slow the motor down to what the Grizzly bog down speed would have been. We could also start thinking about adjusting with chip load in mind.
• Use 2 of our Freud blades and grind one side of the teeth flush to the plate and grind the teeth flat to the plate on the other blade so that can "mate" into one thicker blade. Getting a quality cut is more important to us here than the saw kerf, my theory is that the two together will be less likely to wobble..
• "side" one of our Freud blades - grind the carbide teeth down to thickness of blade. Blade vendor is advising us against this, believing that the blade will quickly over heat.

Does anyone have any suggestions?

I'm open to any suggestions to try. Time is very valuable at the moment and money is tight. Any advice or insight is greatly appreciated.

How are the arbor bearings? You can put a sanding disk/plate on it and pull/push pretty hard to see what the play is. Try it cold and warm.

Fence is verifiably tight? Straight along its length? If it were bent away from blade due to pressure from feeder, your exit could be narrower than entrance.

How is the cut with hand feed?

Is the feeder in the same relation to blade and fence as it was on the Griz?

Have you tried feather boards on output? Splitter/knife after blade to keep cut open?

Have you tried the blade way high, so it cuts nearly straight down? (Less time in cut should mean less heat/warping?)

Or does it work better shallow, just barely peaking thru, for some added stability? (Prob. not since material is soft and heat sensitive.)

There are fiber-cement blades (6 teeth total on a 6 1/2" blade) and metal-cutting carbide tooth blades (almost flat rake, and definitely not thin-kerf) that you could audition without custom blade lead times.

Final weird thought: if the arbor is misadjusted so blade is slightly skew, and feed is relatively slow, can you get a smooth enough cut with a kerf that is actually wider than the blade, so binding doesn't occur.

From your post: "The other factor that turned us off was that we could not move the trunions enough to achieve perfect miter and blade parallelism."

Plexi guys will flame-polish edges to remove tooling marks and return clarity to cut edges. Is there a safe temp that could do the same for your material?
There is both fire/spoilage risk and environmental/operator risk in that, of course.

I've got a 66, and it's hard to believe it can't do what a Griz can...

Chip

I am not a table saw operation/setup guru, but I am having a hard time believing that running the blade backwards is giving you a smoother cut than the correct direction. Can you post up a couple pics of what you are getting and trying to get?

I think if the grizzly was giving you the results you need, then my first thought is the PM is not aligned properly, the arbor or flange may not be true, and as you said, the rpm is different.

Have you put a dial indicator on the arbor and the blade flange to make sure that they are running true? If the blade is wobbling it will leave marks.

I think you should be able to match the formula you had with the grizzly by changing diameter of blade and tooth count.

Something else to look into, is a hollow ground saw blade. I have ran some 3" dia ones on delrin alot and they give good results. I have not looked into whether you can get a hollow ground blade in 7-10" range affordably.

Use a dial indicator to verify the fence is parallel to the blade. Also, use an indicator to see if the fence deflects when using the feeder.

Illinoyance said:

Use a dial indicator to verify the fence is parallel to the blade. Also, use an indicator to see if the fence deflects when using the feeder.

Click to expand...

A number of alignments to check, put a 1/2" plate on the arbor and align it parallel with the table slots. On the PM 66 the trunnion is mounted to the cabinet and the table to the cabinet, it could have shifted in shipping. Align the splitter and then the finally the fence to the blade. Do not use a blade for the alignment process use a 1/2" ground plate that is verifiably flat. you should be able to get everything within 0.001" or less if you are careful. when you align the fence opt for being slightly wider on the out-feed side. if the material is pinched on the out-feed that will leave saw marks. you could have the fence tilting away from the back of the blade as much as 0.001 or 0.002 and still get parallel cuts, tilting in less then 0.001 will give you grief.

http://www.lexington.woodworkersguild.org/shop/manuals/Powermatic Model 66 Table Saw Manual.pdf


dee
;-D

have you looked into stress relieving the material? what are you using to get it down to the thickness on you cnc?

Wow, I didnt expect to have 6 replies over night. Thanks so much for the insight, I will try answer all questions and provide as much additional insight as possible.

One thing I'd like to clarify - and I will edit the original post for future readers. Our optimal cut quality is not totally void of saw marks. They are still visible (barely) and even harder to feel than they are to see. Although not perfectly smooth they are what most people would consider smooth.

• How are the arbor bearings? When I inspected before purchasing they spun freely. I have not yet wrenched on them. I will see what we have with a 5/8 bore that I can put on the arbor, maybe I'll sacrifice an old blade.
  
• Is fence verifiably tight and straight? Fence is very tight. We have a t track/bolt mounting system that we tighten down after the front side is tight. I have always assumed the fence was straight. I will check that today.
  
• Is the feeder in the same relation to blade and fence as it was on the Griz? Yes
  
• How is the cut with hand? Better when we hand feed very slow but incredibly inconsistent, every slight change in feed shows up in the product.
  
• Have you tried feather boards on output? Splitter/knife after blade to keep cut open? - Not exactly, we had a spring loaded device pushing down on the material near the blade but got rid of that and moved the power feeder so it sits on top of the blade and the blade cuts grooves in the wheels. We have used a riving knife in the past but found better luck with the power feeder on top of the blade in a fashion that did not allow clearance for the knife.
  
• Have you tried the blade way high, so it cuts nearly straight down? (Less time in cut should mean less heat/warping?) Only when hand feeding. Positioning the power feeder over the blade limits our ability to raise the blade very high.
  
• Or does it work better shallow, just barely peaking thru, for some added stability? (Prob. not since material is soft and heat sensitive.) The material is actually pretty hard and dense. (at least in my world, maybe not in relation to metals)
  
• There are fiber-cement blades (6 teeth total on a 6 1/2" blade) and metal-cutting carbide tooth blades (almost flat rake, and definitely not thin-kerf) that you could audition without custom blade lead times. Great, I will look into those today. Thanks!
  
• From your post: "The other factor that turned us off was that we could not move the trunions enough to achieve perfect miter and blade parallelism." - On the Grizzly we could not achieve true miter slot/blade parallelism.
  
• Plexi guys will flame-polish edges to remove tooling marks and return clarity to cut edges. Is there a safe temp that could do the same for your material?
  There is both fire/spoilage risk and environmental/operator risk in that, of course. - We already have a flame treating step that happens next. The UHMW is passed through 3 lit torches. We flame treat to increase the surface energy to help the plastic bond, the flame treating does smooth out the plastic slightly but not enough to rely on.
  
• I am not a table saw operation/setup guru, but I am having a hard time believing that running the blade backwards is giving you a smoother cut than the correct direction. Can you post up a couple pics of what you are getting and trying to get? When cutting backwards, the saw is actually melting through the material and leaving a smooth finish. I will post some pictures today. A bit crunched for time at the moment but will get some up today.
  
• I think if the grizzly was giving you the results you need, then my first thought is the PM is not aligned properly, the arbor or flange may not be true, and as you said, the rpm is different. - The PM is aligned properly. The first thing we did with the PM was deep clean, degrease, lubricate, and re-align everything. The Grizzly was never fully aligned as there was not enough play in the trunions to fully align the blade to miter slot. I guess this is another variable I overlooked until now. While the PM is aligned perfectly, the alignment is different than it was on the Grizzly.
  
• Have you put a dial indicator on the arbor and the blade flange to make sure that they are running true? If the blade is wobbling it will leave marks. - Other than to adjust the blade to miter slot parallelism, no. While checking this I did gauge off multiple spots on the blade and think I would have noticed if out of true. I will check again today only focusing on the trueness.
  
• I think you should be able to match the formula you had with the grizzly by changing diameter of blade and tooth count. - Interesting, good thought. I guess I'd look at SFM here. I would need to estimate the RPM of the Grizzly while bogged down as that was giving us our best texture.
  
• Something else to look into, is a hollow ground saw blade. I have ran some 3" dia ones on delrin alot and they give good results. I have not looked into whether you can get a hollow ground blade in 7-10" range affordably. - Interesting. I have never heard of a hollow ground blade. What is the advantage here? I will do research on my end also.
  
• Use a dial indicator to verify the fence is parallel to the blade. Also, use an indicator to see if the fence deflects when using the feeder. - We use a different method to align fence to blade, but I will try this and see how it checks out. Also interested to look at fence deflection with the dial, never done that before.
  
• A number of alignments to check, put a 1/2" plate on the arbor and align it parallel with the table slots. On the PM 66 the trunnion is mounted to the cabinet and the table to the cabinet, it could have shifted in shipping. Align the splitter and then the finally the fence to the blade. Do not use a blade for the alignment process use a 1/2" ground plate that is verifiably flat. you should be able to get everything within 0.001" or less if you are careful. when you align the fence opt for being slightly wider on the out-feed side. if the material is pinched on the out-feed that will leave saw marks. you could have the fence tilting away from the back of the blade as much as 0.001 or 0.002 and still get parallel cuts, tilting in less then 0.001 will give you grief.
  
  http://www.lexington.woodworkersguil...w Manual.pdf
  - We aligned everything as close to perfectly parallel as possible. I will look into shifting just slightly as you have mentioned. Also, we have always aligned with a blade. A true blade alignment plate would be a good upgrade for us.
  
• have you looked into stress relieving the material? what are you using to get it down to the thickness on you cnc? - we have not looked at stress relieving. Would this be another cutting tool on our line? We use a 2 "O" flute carbide cutter on the router.
  
  
  
  
  
  

Thanks so much for all the input so far. I have to run but will post pictures this afternoon.

Maybe try a straight flute endmill or straight carbide tipped and get a flat put on it for bottom cutting.(example onsrud 48-072) we have had great success with this in our shop. That and giving the UHMW enough time to rest, before moving onto the next operation.

JDMek said:

Maybe try a straight flute endmill or straight carbide tipped and get a flat put on it for bottom cutting.(example onsrud 48-072) we have had great success with this in our shop. That and giving the UHMW enough time to rest, before moving onto the next operation.

Click to expand...

sorry this will help keep the stress out of the sheet. The company we order from (piedmont Plastics) has a huge fly cutter that we had them use to get to our spec.
this takes alot of stress out of the sheet before we even get it. When we saw cut we aimed for crossed hatched saw marks, we found this worked the best for our saws ,motor and blade. Rubbin may be for racin but its the last thing you want to do with UHMW.

OP recently wrote this: "The Grizzly was never fully aligned as there was not enough play in the trunions to fully align the blade to miter slot. I guess this is another variable I overlooked until now."

Which I'm betting goes with this original statement: "The other factor that turned us off was that we could not move the trunions enough to achieve perfect miter and blade parallelism. "

And these pearls of whizz-dumb from yours truly: "Final weird thought: if the arbor is misadjusted so blade is slightly skew, and feed is relatively slow, can you get a smooth enough cut with a kerf that is actually wider than the blade, so binding doesn't occur."

I'm thinkin' a skew blade carves enough of a kerf and heats things up just a little so the cut bows open just a little, avoiding random re-cutting.

But maybe I just need another cup o' joe.

By the way... nicely informative original post. Not the norm around here.

Chip

A hollow ground blade is thinner in the center or middle than the teeth so that the body doesn't rub, again, not sure if you will find one in a large enough diameter to put on the saw, but I have used them on the mill for slitting operations.

Something else that I am wondering, what surface finish do you want? What led you to doing this on a table saw if you are already doing work on the cnc? I am assuming that you are cutting a large blank with whatever features on it then ripping it into the narrow strips. Are you trying to save material by the narrow kerf or are you not able to hold it to cut it on the cnc?

Would it make sense to set up a wood shaper with a power feeder to run the strips through to put a 'routed' edge on it? I know it would add another machine and operation, but it may be worth it as far as speed, just let the saw rough it out real fast and then put the nice finish on it with the shaper. Honestly you could use a hand router in a table so you get high rpms and higher feedrate since it would be taking off very little.

Looking forward to some pics of the parts etc

JDMek said:

sorry this will help keep the stress out of the sheet. The company we order from (piedmont Plastics) has a huge fly cutter that we had them use to get to our spec.
this takes alot of stress out of the sheet before we even get it. When we saw cut we aimed for crossed hatched saw marks, we found this worked the best for our saws ,motor and blade. Rubbin may be for racin but its the last thing you want to do with UHMW.

Click to expand...

Not sure if this is what you are talking about or not, but our material already comes flat, it is sanded flat on one side before it is sent to us.

Are your saw cut edges smooth? Would you mind sharing anything about your saw setup?

Chip Chester said:

OP recently wrote this: "The Grizzly was never fully aligned as there was not enough play in the trunions to fully align the blade to miter slot. I guess this is another variable I overlooked until now."

Which I'm betting goes with this original statement: "The other factor that turned us off was that we could not move the trunions enough to achieve perfect miter and blade parallelism. "

And these pearls of whizz-dumb from yours truly: "Final weird thought: if the arbor is misadjusted so blade is slightly skew, and feed is relatively slow, can you get a smooth enough cut with a kerf that is actually wider than the blade, so binding doesn't occur."

I'm thinkin' a skew blade carves enough of a kerf and heats things up just a little so the cut bows open just a little, avoiding random re-cutting.

But maybe I just need another cup o' joe.

By the way... nicely informative original post. Not the norm around here.

Chip

Click to expand...

I'm thinkin' a skew blade carves enough of a kerf and heats things up just a little so the cut bows open just a little, avoiding random re-cutting. - This is a very interesting thought and something that is starting to brew around in my brain.


I checked the arbor bearings today by tightening a sacrificial blade to the arbor and really wrenching on it. I was nearly able to move the entire saw without feeling any play in the arbor.

Other than checking the arbor and taking some pictures of the cut qualities I haven't put any time into the saw today. I will be out tomorrow so I wont get into things until Thursday.

Below is a picture of the sidewall cut with the blade installed backwards. The front facing side which my thumb is against in the picture is the saw cut edge, the other side is the sanded side from the manufacturer.


Below is a picture of the sidewall cut with the blade installed forwards. The front facing side is the saw cut side and the other side is the sanded side.



Thanks again for all the shared knowledge.

Hmm, learn something new everyday. I'd have never guessed you would have gotten anything but a welded together mess running a blade backwards.
So, now after seeing the pics, that edge looks to be almost sanded like the top. Is that what you want? I have never tried to sand uhmw, didn't expect it to work if I did, but if you want the sanded finish, can you run them through a widebelt sander?

Sanding UHMW is something new to me. At least not without melting. Since you are using a feeder I would try a zero rake or negative rake blade. "Cutting" with the blade backwards shortens blade life because at some point enough plastic gets stuck to the blade causing other problems. HDPE is not exactly like UHMW, but neither one have any abrasive content. Blade life should be in years if feeds and speeds are right. I have HSS end mills on my VMC for 15 plus years. Still making good parts. The deep drilling is different because some plastic gets stuck to the bit and periodicaly needs to be cleaned off. I would guess your blades just need a cleaning.
And Frued is not the blade company I would start with.
Burns Power Tools in Fall River, Ma could make one for you. Marshall Burns is no longer alive but he was a real master saw smith. His sons and grand sons run the business now. At one time many years age Teneryu (sp) paid for him to go to Japan and teach their saw smiths about hammer tensioning saw blades.

macgyver said:

Hmm, learn something new everyday. I'd have never guessed you would have gotten anything but a welded together mess running a blade backwards.
So, now after seeing the pics, that edge looks to be almost sanded like the top. Is that what you want? I have never tried to sand uhmw, didn't expect it to work if I did, but if you want the sanded finish, can you run them through a widebelt sander?

Click to expand...

Yes, that is exactly the look we want. I think the saw is melting right the UHMW right on the surface. You can sand UHMW. We have a belt sander that we use to touch up the material when we need to. Sanding does take some time though. Adding a widebelt, or I saw someone mention a shaper to smooth the sides is out of the question. We were getting a satisfactory result on a cheaper machine. I think the real issue here (re-enforced by my theory of the backwards blade melting through the material) is that the Grizzly was already a lower RPM at the arbor and was bogged down therefore could have been melting its way through the material, only going forwards. The PM66 is to powerful to bog down going forward so the material is not melting along the cut.

Scruffy887 said:

Sanding UHMW is something new to me. At least not without melting. Since you are using a feeder I would try a zero rake or negative rake blade. "Cutting" with the blade backwards shortens blade life because at some point enough plastic gets stuck to the blade causing other problems. HDPE is not exactly like UHMW, but neither one have any abrasive content. Blade life should be in years if feeds and speeds are right. I have HSS end mills on my VMC for 15 plus years. Still making good parts. The deep drilling is different because some plastic gets stuck to the bit and periodicaly needs to be cleaned off. I would guess your blades just need a cleaning.
And Frued is not the blade company I would start with.
Burns Power Tools in Fall River, Ma could make one for you. Marshall Burns is no longer alive but he was a real master saw smith. His sons and grand sons run the business now. At one time many years age Teneryu (sp) paid for him to go to Japan and teach their saw smiths about hammer tensioning saw blades.

Click to expand...

This PM66 has been in our production line for about a month now. We keep our blades clean and have them re-sharpened when needed. This saw has seen 4 unique blades in production in addition to the ones we have tested.

I do think you are on to something with the suggestion for zero and negative blade rake. I have a lot of different directions to go with blade trials and research. One avenue I plan to go is into a thicker kerf (steadier blade) and fewer teeth. To determine tooth count I will do an estimation of arbor speed at bogged down Grizzly speed to determine a "chip load" formula using the PM66 arbor RPM and solving for number of teeth.

You and some other members have mentioned different blade geometries such as the negative hooks and the neutral "v" hook, the hollow ground carbide, etc...


Thanks

I would be very interested in the material spec for that sheet. For sure not the natural color and it appears to be a filled plastic to boot. Could be the filler causing some of the results, good and bad.
Exactly what adhesives do you use?

I gotta ask - have you contacted any of the suppliers or manufacturers of UHMW for their input?

There's gotta be a way to get a clean cut without emulating a bogged down Grizzly. Or maybe not...

richard newman said:

I gotta ask - have you contacted any of the suppliers or manufacturers of UHMW for their input?

There's gotta be a way to get a clean cut without emulating a bogged down Grizzly. Or maybe not...

Click to expand...

try using a 7" blade made for plastic, just drill the center hole to fit the saw arbor or try to order it with the right arbor hole to start with. That will reduce the surface speed by about 30%.

No Melt Saw Blades for Plastic Sheets and Extrusions -

dee
;-D

Judging by the saw pattern on the part shown, it does not appear that the fence is parallel to the blade.

Another thought along the lines someone else mentioned: maybe the old grizzly arbor wobbled more than the PM66, possibly yielding a wider kerf that did not bind on the exit? Can you try a set of wobble collars? Not sure where to get them these days, but i have a couple old sets. Or, as an experiment, put a piece of electric tape across the bottom of one collar, and across the "top" of another before clamping the sawblade between, and see what happens.

I agree with dcsipo about the saw blade, shouldn't even need bored - aren't PM66's only a 5/8" arbor?

smt

HDPE is not exactly like UHMW, but neither one have any abrasive content. Blade life should be in years if feeds and speeds are right. I have HSS end mills on my VMC for 15 plus years. Still making good parts

Click to expand...

Scruffy always gives good advice, but I have to comment in case the OP is not familiar with cutting edge degradation.

In "soft" substrates including wood, the bulk of blade deterioration is due to chemical erosion of the edge, which leads to blunting, leading to heat, leading to faster chemical erosion. Due to the mechanics of cutting, abrasives in un-contaminated lumber (or plastics) are usually among the minor factors.

It may be that for many plastics there is a practical feed/speed at which the substrate has little chemical interaction with the tip material. But the lack of abrasive does not automatically rule out edge wear, in some cases even "rapid" edge wear. The famous example of this is that in the first couple decades after carbides started to become common in industry, they could be used on hardwoods, but not economically with pine. The chemical reaction at cutting tip temperatures with the resin in the pine caused quick dulling and failure - faster than HSS dulled in the same material.

smt