Hammering and Rolling Large Saw Blades

Saw blades can be pre-tensioned so that they take on a true, flat shape when spinning. The technique is an art that requires knowledge and experience. June 24, 2007

I have a problem with saw blades sometimes wobbling and creating a bad cut. I have heard mention of hammering a blade, but I am not sure what that means. Does anyone have any information on this?

Forum Responses
(Architectural Woodworking Forum)
From contributor H:
The opposite happens for me; after a few rum and cokes, I'm hammered and start to wobble.

Give the blade to your sharpener and let him fix it.

From contributor A:
If the blade is a $100+ Forrest quality blade, try sending it to Forrest to retrue it. If it's not a $100 blade, use it for rough cutting and buy another blade. But before you do the above, make sure the bearings haven't gone bad or the arbor is not true.

From contributor R:
Sawblade wobbling is a somewhat common issue. Hammering is a process of making sure the blade is true and flat. It is exactly how it sounds - the blade is put on a steel anvil and hammered with special hammers to accomplish this.

Unknown to a lot of people is that plate tension (the stiffness of the plate) is more important than the flatness of the plate. This is also done in the hammering process by pounding the plate in certain areas to increase or decrease the stiffness of the plate. Even a perfect flat plate can wobble if improper tension is not met. Hammering is almost an art. Years of training and teaching are required to be good at this. A lot of people can hammer a saw flat but have no clue how to tension. Just make sure whoever you are sending your blades to for service has the capability and experience.

From contributor J:
Let's say you are running a 12" blade. And let's say your arbor is out of true or bent by 0.0005". That is so miniscule an amount, you could probably not even tell it was out without an expensive "tenths" dial indicator. But by the time you got out to the rim of your saw blade, that miniscule amount now equals 0.006".

None of the blades you buy, even new, will be perfectly flat. Say the one you buy is not flat by 0.004". Your total runout after mounting on your saw could therefore be anywhere between 0.010 and 0.002 depending on where the saw blade wound up on the arbor.

This is why I use a cheap (under 20 bucks) dial indicator whenever I install a blade. I repeatedly tighten the blade and check the runout, move the blade 90 degrees with respect to the arbor, and do it again. I do this until I have found the spot where the runout is the least. I have found that I can improve the runout on every saw that I use by doing this procedure.

I would definitely heed contributor R's advice regarding hammering. A friend of mine has an ancient sawmill running a large blade. The blade, at rest, has been hammered and tensioned so it is shaped like a satellite dish when not spinning.

Once it comes up to RPM, it flattens out. He claimed there were three reasons for this: the blade was more stable at running RPM's, if the idiot (him) sawing logs was pushing the carriage too fast, the RPM would drop and the blade would come out of the log on the board side. This would prevent severe damage to the plate, the arbor, etc., and would also prevent ruining more than one board at a time. So, like contributor R says, there is more to it than just banging away with a hammer.

From contributor L:
The best solution is to buy high quality blades. We've had some blades in the past that developed wobble or screamed; don't buy that brand any more. Most modern blades have been tensioned by rolling (doesn't require a skilled artisan). I donít like to see heavy roll marks on a blade, as I consider it a sign of sloppy manufacture. When a blade is run, the outer perimeter heats more than the center. When the outer rim heats, it expands. That extra size has to go somewhere so the blade pushes out sideways. By tensioning the blade, it pre-stresses it to compensate for some of the stress that will develop in heating. Some of the expansion is taken up in the slots you see cut around the edge of the blade. Big sawmill blades have the stress of the high forces developed by their spin, which overcomes some of the pre-stressing built into the blades (which causes them to be cupped when stopped).

Like someone else said, your arbor may be partly to blame also. I put a readout on my Unisaw and was surprised by the amount of runout. I called their tech support and they told me that it was within specs! I did the same on our Schelling - barely detectable runout! (Of course it cost $130,000!) You never get more than what you pay for.

From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
When a circular blade begins to spin, the outside rim wants to get longer. (Remember the ball you had that was attached with a rubber band and you would spin it and it would increase the spinning diameter?) With the blade, the metal closer to the eye will not let the outside rim expand in diameter. So, how can this outside rim get longer?

An alternative to increasing diameter as the saw spins is to get wavy at the rim, which is what actually happens. (Note that this effect occurs progressively less as you move away from the rim.) In fact, there is a frequency at which the sawblade will actually create a "standing wave" along the rim. It is similar to what you might observe in a car antenna at certain speeds. To avoid this waviness, we can make the saw blade very stiff, but that means thick, which means lots of sawdust.

Another option is to hammer the saw blade so that when the blade is not spinning, the blade is dished slightly. (Hammering does not make a flat blade!) With a dished blade, as the blade beings to spin, the outside can get a bit longer and this increase in diameter actually causes the blade to come flat while spinning. Hence, when hammering a blade (sometimes hammering is called rolling or tensioning), one needs to know the amount of dish required, which means one needs to know the rpms that the blade will be working at. You cannot hammer without knowing the rpms.

Hammering is sometimes called tensioning the blade and you are putting in this dishing. If one were to check the ring that is heard when a blade is hit lightly, the flat blade (untensioned or unhammered) will be lower in frequency than a hammered (or tensioned or dished) blade. In fact, measuring the frequency is a quality check when rolling a blade.

Incidentally, this wobble is why we have difficulty using thin kerf blades and why as the rpms drop when sawing, the cut can actually get poorer.

Also note that heat will destroy the hammering or tension. So, a blade that is accidentally heated will possibly cut poorly it the future.

There is another option to hammering and that is adding heat to the eye and cooling the rim. This is often done with band saws in sawmills. They have the same centripetal force issues and resonance. Larger band saws are hammered so that they are slightly curved, edge to edge.

Hope this clears up some of the incorrect info posted previously here.

From contributor L:
Where does the expansion from heat go? I know blades heat as they are used; not all the energy is taken away in the chip. We've had rip saw blades that would cut fine but after running awhile, almost continually cutting, they would start to develop wobble. Put a good blade on and problem cured. This doesn't seem to be very affected by sharpness; even with the good blade being somewhat duller, it doesn't develop wobble.

From contributor J:
Dr. Gene, thank you for your excellent explanation. I always wondered if my buddy really knew what he was talking about. Thank you for clearing that up.

From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
Getting rid of the heat that might be generated is indeed a key. That is why the saw teeth are a bit wider than the blade... to avoid rubbing and heat generation. Excess heat is conducted toward the eye, but oftentimes too slowly. One option is to apply heat at the eye, but this is seldom done. Note that a little bit of heat can be conducted away through a frictionless guide. I did once see a guide that had cold air pushed through the back-side to apply cooling to the rim (contracting it) and also providing a cushion of air if the blade deviated toward the guide. We also have electronic guides that use magnets to pull the blade one way or the other with computer controls to keep it running straight. With smaller saws, guides are rarely used.

From contributor J:
Dr. Gene, I noticed a long time ago when running different blades on a 15" old time, regular, no frills, chop saw that I got the best cut with straight ATB blades. My supplier was out of straight ATB's once, so I bought an ATB with raker. This blade would tend to slide slightly when first engaging the top of the stock before straightening out and cutting the rest of the way fine. Unfortunately, this small initial gap showed when the joint was clamped, since the rest of the joint was mated properly and the "moon" could not close.

I took the blade back to the grinding shop where I bought it and the owner suggested grinding the rakers down to below the lower intersection of the two beveled teeth. Now when I cut, the bottom of the saw kerf had a center inverted "V" of stock sticking up which came to a sharp point just like a straight ATB without rakers. The top of the "V" was not lopped off like before. This not only eliminated the problem, but that blade cut better than the straight ATB's I had.

1) What exactly is the purpose of the raker? Is it to just clean the kerf of chips?
2) What height should the rakers be ground to relative to the intersection point of the ATB's on a cross cut saw?
3) Do you think the rakers were inadvertently left too high at the manufacturer?
4) Or do you think it was simply that my grinder did a better job grinding the teeth than the factory, and the raker height makes no difference?

Also, with regard to the rim expansion of say 15" blades, assuming there is not any heating problem, is that expansion generally consistent enough throughout the rim that edge grinding the teeth will result in a cleaner (smoother) cut given no change in the top to bottom or front to back angles?

From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
The answer to #4, part 1, is most likely "Yes."

Note that with small diameter blades, and with a large collar, the centripetal force problem is close to zero. The larger the blade, the higher the rpms and the smaller the collar, the more the problem..

From Professor Gene Wengert, Sawing and Drying Forum technical advisor:
A raker tooth is a specialized tooth that cleans up the cut after the other teeth have done their thing. For example, we often have 4 alternate top bevel teeth (ATB) followed by a raker that is flat across. If you are cutting through all the way, then the raker does little to help. But if you were cutting part way (like in a dado), with only the 4 ATB teeth, the groove left would have a slight rise in the center rather than being flat across the width. The flat raker removes this rise. The raker tooth is the same height as the other teeth, but making it slightly lower means that you will not have a perfectly flat bottom to a cut. Sometimes the raker is a tiny bit narrower, but it should never be higher or wider, as it is not really doing the main cutting job.

From contributor L:
"We also have electronic guides that use magnets to pull the blade one way or the other with computer controls to keep it running straight. With smaller saws, guides are rarely used."

That must be an interesting setup for the computer to control the blade wobble with magnets! What kind of saw uses that system? The blades we had the most problem with are only 14" in diameter, big national brand. Haven't bought any in quite awhile; maybe they've figured it out.

The comments below were added after this Forum discussion was archived as a Knowledge Base article (add your comment).

Comment from contributor C:
Just to clarify a point about the WZ/FZ group toothing design the finishing tooth is 0.3mm below the highest point of the angled teeth. The purpose of the finishing tooth is two-fold. It gives a flatter end cut and a cleaner cut too. The more aggressive a tooth design is, like the 15-20 degree bevels used in this instance (and usually with a shear angle too), the rougher the finished edge. The finishing (FZ) tooth helps keep the cut clean.