Way to measure band saw frame flex?

[Curt Harms]

This is not my original idea but I'm curious about what others think. A guy said to start with no tension on the blade. Tape a dowel to the upper wheel housing so it's ridgid and sticks down to within an inch or so of the table. Mark the table where the dowel would be if it came all the way down. Now apply tension and see if and how much the dowel moves. I was thinking a laser pointer fastened to the upper wheel housing and pointing at the table might be better than a dowel.

Stupid idea? Too coarse to be of any use? I'd think any bandsaw would have some upper wheel movement when tensioned. How much is too much?

[David Kumm]

Curt, would a dial indicator touching the guide post accomplish the same thing? I'm thinking a oneway on the table with the indicator touching the bottom of either the post or the guides. I'm not close to the shop now to visualize but seems like that should telegraph frame flex. Dave

[Paul Symchych]

Just curious.... what would you do with that information?

[David Kumm]

Just curious.... what would you do with that information?

Paul, most saws advertise more capacity than they really have. When you get to 3/4" and wider resaw blades, particularly the carbide ones that need higher tension, many saws will flex before attaining optimum tension for the blade. You don't want to tension enough to flex the saw even a minute amount. My older Laguna ACM LT 18 was advertised to handle a 1" blade but a carbide Trimaster 1/2" was all it would take before flexing just enough to screw up the tracking of the blade. You could get an acceptable cut with a wider blade at lower tension but nothing like the almost planer quality cut the blade was capable of. If there was one piece of valuable info a manufacturer could give it would be the amount of deflection at a standardized point but I doubt it will ever happen. Dave

[Mike Wilkins]

A past issue in FineWoodworking outlined the procedure in one of their test features on big bandsaws. Don't remember which one or the model they liked. Google is your friend.

[Stephen Cherry]

You don't want to tension enough to flex the saw even a minute amount.

You may not want to do this, but it's going to flex anyway. the question is how much. Structures flex with applied force, somewhat linearly at first, so

flex=k times tension.

For a given force, the flex is approximately the tension times a constant, k.

It's two sides of the same thing:
1- tensioning the blade flexes the saw
2- flexing the saw tensions the blade.

Kind of like a bow for archery-- flexing the bow tensions the string, and you can't tension the string without flexing the bow.

[David Kumm]

Stephen, you are correct. I was overstating. there will always be some flex but is is kind of built in like the bow example. Exceeding that is probably more of an issue when upgrading the spring. That is sometimes given as an answer to providing higher blade tension which brings the frame stiffness into question. The bad news is there are many saws underengineered for the blades put on them. The good news is the manufacturers have become more aware of this and beefed their frames up over the past few years. Dave

[Van Huskey]

First it makes a difference whether you are attempting to get data to compare two or more saws OR trying to gather it to determine if a particular saw has "too much" flex. The first is easy, the latter is hard since there is not standard for testing thus no standard go/nogo number. I see the whole process as what I call a "Dick Chaney" problem, there are known knows, known unknowns and unknown unknowns. I think the first issue is you have to decide what you want to learn, if it is just to compare two or more saws to decide which one has lower deflection the process is somewhat easy because you can dispense with a lot of the variables.

Roland Jones wrote the article in FWW that Mike mentioned. It was written about 8 or 9 years ago, based on the saws I remember in the test, they were 18" saws. It is the article that shed some light on the weakness of the Jet square spined saws, what is useful as a whole from his data was that the subjective resawing test roughly followed the results of the frame deflection test. In general my conclusions from that article, which are really common sense, is that HP and frame stiffness are the two key factors to quality resawing (when the blade is the same).

Mr Jones used a dial indicator with a push pull gauge to determine twist in the frame when pushing back on the guides like wood does in the cut. I assume he standardized heights above the table etc. This is actually only half the story, you need to know how much the frame deflects in another key plane, the vertical plane the wheels reside in. I would measure this at the bottom center of the guide post with a dial indicator with a known amount of tension between the wheels. If I was really interested I would take an old band I could cut short and attach some sort of strain gauge between the two cut ends of the band (could be something as simply as a BIG fish scale clipped into holes drilled near the ends of the band) then I could repeat the tension.

In the end you have at least two planes ones needs to measure and each can be done reasonably easily but I see it mainly for comparision between two saws. To determine if a single saw is acceptable then one needs to determine how much deflection in each plan is "too much". Mr Jones determined the number for the Jet and a couple of other saws were too much and unless you replicate his test very closely his numbers are useless.

You may however be able to determine where to stop in terms of gauge and width for a particular saw. Use any sort of method to get a objective number for flex with each blade you use and you may see a trend where the flex number gets to a certain point and you see resulting poor cuts.

I have a lot of thoughts on this and I apologize for the jumbled post, the phone keeps ringing! My main point is you have to look at something like this from the bottom up, what are you trying to prove? Testing the deflection can be done in a number of ways but what is your goal with the data you collect?

[ian maybury]

My thoughts are similar to Van and Stephen. Once you apply any load you're going to get a deflection which should be proportional to the load. (at least until there's enough deflection that the mode in which the saw is flexing starts to change a bit (determined by the layout of the chassis), or it starts to take a permanent set. Which would require a lot of force.

Stiffness is typically measured as force/unit deflection. e.g. springs are often measured in lbs/inch, kg/mm or similar - although the deflection in this case would be small and need measuring using a dial gauge. Probably as the guys say between the upper frame and the table.

A clamped steel bar would probably be a better mounting than a bit of dowel. You wouldn't need to load the saw all that much - just enough to take up any slack and then produce maybe three or four decently separated points over the normal range of the stock tension adjustment. Graph them, and draw a best fit straight line though the points and you can read off a value for the vertical stiffness of the saw in whatever units you are using.

Repeating for another saw and extracting the same number (the graphed points don't have to be for the same loads) would produce a directly comparable number for k or measure of stiffness.

The other issue is how the load is measured - the real deal would be some sort of strain gauge based load cell bridging a joint in the blade. Not sure if you could get a stiff enough one, but some sort of high load spring balance or other blade tensioning force measuring device mounted the same way would also work.

As Van says all saws will deflect under load, no matter how small it is. Chances are the issue that matters is whether or not the wheels remain in the same plane as this happens. It seems likely that the combination of the top wheel, it's shaft assembly and the frame will sooner or later start to deflect out of plane in one direction or the other.

It might be possible to mount a high accuracy electronic angle measurement tool off the face of the top wheel or its stub axle (with the free leg pointing in the direction the blade cuts), and use a precision level to keep returning the free leg to horizontal to get a measure for the degrees tilt taking place as you move through the available range of blade tension adjustment as before. It depends a bit how much the wheel will tilt, the sort of grade of electronic angle gauges we use usually read down to about 0.1 deg. You could also use a long pointer mounted off the wheel, and measure the movement of the tip relative to a fixed mark on a floor mounted reference - but this would need a little maths to convert to degrees.

You could then graph the numbers as before and extract a measure for the saw, only this time the measure would be torsional or rotational stiffness of the top wheel of the saw about a horizontal axis at right angles to the side of the blade - in lbs or kg blade tension/degree tilt out of the plane of the wheels. Again to extract a number capable of being compared to the same reading from another saw.

The wheel could also tilt about a vertical axis (to move move the blade closer to or further from the thrust guide), and would probably need its stiffness measuring in that direction too. Then there's the question of what's going on around the lower wheel too - although my guess is that the first two measurements would tell a lot.

The one problem with the above is that the measurements while relatively easy and probably useful to compare saws in a practical sense would not be all that useful in mathematical terms.

Pardon the length again.

ian

[Curt Harms]

What would be useful would be to have a method to determine how wide/thick a blade a bandsaw can really use, not what the manual says. Like Van says, there are no standards that say "You should get good rip & resaw performance as long as frame/wheel deflection doesn't exceed X in this axis and Y in this". One could then tension a few different blades and determine how much tension could be applied without exceeding those values.

[David Kumm]

I agree Curt. You can experiment now and find the sweet spot with different blades but only after you have purchased the saw. I saw some pictures the other day on another forum of veneer cut with a 3/4" trimaster on a new Felder N4400 and although uniform, the cut was not nearly as good as the blade will produce. Looked to me like what is typical when the Lenox is undertensioned. The OP was happy with the result but it showed me that when you get to the .035 thick blades you need a lot of saw to get full benefit from the blade. Would be nice to have a standard to compare to before spending $$. Dave

[Myk Rian]

The laser pointer will work as well as the dowel.
Try it. No need to get complicated about it.

[Phil Thien]

The only problem w/ a laser pointer is you need one that stays on when you take your hands off the thing.

[Van Huskey]

The only problem w/ a laser pointer is you need one that stays on when you take your hands off the thing.

Tape?.......

[ian maybury]

A laser sounds good, it's got to be a lot easier to mount and take readings from than mechanical stuff.

You do need to combine force and deflection to measure stiffness Curt.

It'd be hard anyway to establish by calculation what the max permissible deflection on a given saw would be. Comparing the deflection to that of a fully tensioned saw known to just be able to adequately tension the blade would get a reasonable number though. Any more deflection that that on your test saw would suggest it probably wasn't as stiff.

Depends too I guess on how good max tension/chassis stiffness is as a predictor of saw capability.

I suppose another route would be to use the saw's own tension spring as a measure. Friction etc would play a part, but well equipped motorcycle race suspension shop (doing shock absorber rebuilds and set up) would probably be able to test the spring in a few minutes and give you numbers for deflection in inches vs. force. (they often use a nifty little hand operated press for it) Then see how much the spring needs to be compressed to get the saw working, having first compared the number (taking account of any leverage effects in the mechanism) to the recommendation for the blade.

Some sort of lead cell or spring balance in a dummy blade would allow a more direct calibration of the spring/saw tension indicator though....

Just tossing stuff about

ian

[Andrew Joiner]

I think you could minimize frame flex. A post or fence that fits between the table and upper frame should do it. Something like this http://www.falbergsawz.com/Other/accessories.html. For consistent thickness resawing this could help.

[John Coloccia]

What does it matter how much it flexes? It should matter how it cuts. Put a blade on it and try it. If you like the cut, then it handles that blade well. If it doesn't cut well, then it needs adjusting or something is too flimsy.

I think at the end of your experiment you will have a lot of numbers that are more or less meaningless and you will have run the actual blades anyway to anchor the results. All you really want to know is if it will run a certain blade well, and you can test that easily by just running the blade (which you need to do anyway) so why bother with any of the measurements? Unless you intend to do this for a lot of different bandsaws and blades and then publish the results somehow, or you're really just curious about determining a correlation for no practical reason whatsoever, I honestly think this will end up to be a complete waste of your time.

I have to ask. Is there some project that you're supposed to be working on? Sometimes I go down these rabbit holes myself when I'm STALLING because I don't feel like doing something. All of a sudden, it becomes critically important to me that all of my wood is labelled by origin, date and cost. It never mattered before, but I REALLY hate sanding so now it's far more important I organize my wood. Top priority, in fact...and it will take me days to do it, fueled by a healthy dose of beer. I even wear my shop apron while I'm thinking about the organizational system, because this is serious business and I'm a professional "upping" my game. What does that have to do with sanding? I hate sanding, that's what. LOL.

So stop stalling, if that's what you're doing :)

[Paul Symchych]

thank you John.
I come back to my original question. What will you do with this information?
While I understand the posts above, suppose I test for flex and come up with some number in whatever units you want. Either something close or out to 3 decimal places. Now what? Does it work?

A lot of this reminds me of the pixel-peeping on photography sites.

The plural of anecdote is not data.

[Van Huskey]

Data points would be relatively easy to gather and normalize in this instance but interpretation in how it plays on the practical world world would be far more difficult, but that is not unusual. For me intellectual exercises like this are as much fun as woodworking and just part of the fun of a hobby centered around machines. In the end the numbers are meaningless without some reference as to what is and isn't acceptable for a given situation. I have seen this included in one comparison test, by Roland Jones 8-9 years ago in a test of 18" saws, he tested twist but not crush (I don't know a good word but the frame allowing the wheels to come closer together) and made a distinction between 3 saws who's numbers were unacceptable. He did not give a cut off number nor did he explain how he derived this cutoff. One of the saws tested did gain a reputation for being overly flexible and was redesigned by its manufacturer.

Many of the things like this that I like to think about are akin to testing a F1 car for 0-100mph times (or 0-120kph) it is one metric regarding how the car will perform but a minor one and not one that can be used to predict the overall performance of the car nor its how well it will perform its sole purpose which is to win races.

PS The problem is that if you look hard enough and you don't really know what you're looking for, you'll probably find something.

[ian maybury]

The big issue Van is as you say that you could end up needing to do a lot of measuring of variables to figure out which one matter, and how they affect the process.

It's often a case to on something like a bandsaw that it's not a case of more (stiffness) is necessarily better - more a matter of tuning the structure to avoid vibration harmonics.

That said it seem likely that a given blade will have a minimum tension requirement if it's to work properly, so that a fairly basic force/deflection in the line of the blade tension and a check for twisting leading to wheel misalignment should tell quite a lot...

ian

[Erik Loza]

...You can experiment now and find the sweet spot with different blades but only after you have purchased the saw....

IMHO, this is actually the more germane question than "how to determine frame flexion". As others have asked, "What value would that be in this application?". For example, I honestly cannot say that I have ever tried to measure frame flexion on one of our saws but on the other hand, I don't know why I would have reason to. That's not meant to sound smug but rather, to say (again, just my personal opinion...) that if that is a spec which really concerns someone, it probably means that you need/ought to be looking at a bigger saw.

For example, our S45N bandsaw, on paper, will handle a 1.0" blade and has a 12" resaw capacity. I have used the S45N quite a bit, with many different blades and under many types of cutting circumstances and will tell you that I counsel customers against using any blade wider than 0.75" on theirs and also tell customers that while this machine will indeed resaw up top 12" on paper, if we are talking about a 12" tall piece of maple that you want to make a veneer out of, this is probably not enough machine for you.

Case in point, I took a call from a gentleman a couple of weeks ago who was interested in the S45N and also looking at that other machine that Dave mentioned, above. I asked him what he planned to do with the saw and his response was to cut 12" tall veneers out of hardwoods. Now, let's set up a hypothetical scenario: Suppose a prospective owner of the above machine asked his rep how much frame deflection their saw had with a blade tensioned to xxx-many psi and the rep was able to answer that. I'm sure that this saw Dave has mentioned probably has less frame deflection than any of the Asian ones that are similar in spec. But, what does that guarantee in terms of actual cutting performance? I basically told the gentleman I was speaking with that regardless of what the specs looked like on paper, if this was the type of work he envisioned, a saw of this class was not the ideal choice and that he ought to save his money and get an MM16, instead, because that machine would do what he wanted to do in stellar fashion.

And the bummer is, as David pointed out, that once you buy it, you are stuck with it. The reason I mention all this is as a company rep, I have heard all sorts of questions from potential customers in regards to spec on our saws that cause me to scratch my head (a gentleman asked me the other day, "What diameter are the arbor shafts for the wheels on your 24-inch saws?", the first time anyone has asked me that in almost 10 years with Minimax...) when, really, the discussion they need to be asking is, "Here is what I plan to do with the machine and here are my expectations. What saw do I need to do that best with?". I guess what I am trying to say is there are specs, and then there are specs. While any saw can be made to look good on paper or give some numeric value for a certain thing, I do believe there would be a lot more satisfied owners out there if we approached these decisions from the point of the end-goal, rather than, for example, "Such-and-such will save me a few hundred dollars". Because I can pretty much guarantee you that the owner of the saw Dave is referring to would be dissappointed if he compared the results he gets now to what he could could with a 1.0" Tri- or Woodmaster under proper tension. And what is a few hundred bucks for a machine you expect to own for 10+ more years? Just my 2-cents, as always.

Erik Loza
Minimax USA

[Kevin W Johnson]

I think you could minimize frame flex. A post or fence that fits between the table and upper frame should do it. Something like this http://www.falbergsawz.com/Other/accessories.html. For consistent thickness resawing this could help.

And depending on the material of the table trunnion, you may not want to do that either......

[ian maybury]

Hi Erik. Minimax have I think been to the fore is making available saws with whatever it is it takes to do a good job of tensioning larger blades for re-sawing.

That said it's not hard to think of a motive for the original post. (which i've no reason to think had anything to do with Minimax) Many manufacturers make OTT claims for their products that don't stand up when put to the test. Or at least are weasel words that are true for one highly specific circumstance of little use to the user, and which are irrelevant if not highly misleading when it comes to their trying to figure out if the item will meet their real world needs.

I've personally been taken for €000s that way as a result of buying a band saw based on claimed blade width and maximum depth of cut capability - it turned out to be able to handle perhaps half the re-saw depth the numbers suggested. By which point the blade was up to max tension, the saw was squealing and vibrating all over the place and it most certainly was not tracking straight.

The eventual switch to a 24in Agazzani fixed the problem instantly, but the point is that there was absolutely nothing in the available literature for either saw which accurately defined its capability. The band saw manufacturers have meanwhile been quick to hop on the bandwagon by offering specifically heavy duty (i.e. more rigid) saws, thus reinforcing people's perception that re-sawing requires more chassis rigidity.

There can be genuine reasons for this (e.g. when testing is very difficult, or so complex as to not be a possibility), but I guess what I'm driving at is that woodworking (especially DIY) is plagued by manufacturers of all sorts of products and systems making misleading claims about their products, while at the same time not providing remotely adequate data on capability.

People know this, and are running round in circles trying to find ways to figure out/test what the real capability of products is....

ian

[Van Huskey]

First, as Erik says this ultimately is a results driven pursuit but as a consumer we don't have the ability to take a bunch of saws on a test drive. I have been lucky (or unlucky depending on how you look at it) to own a bunch of bandsaws over the years and get a feel for which ones are good and which ones are poor, I also have crashed the shops of friends and Creekers to use saws I haven't used. Much of my experience was driven by learning about the saws but it still takes a lot more use than I have with all the saws I didn't actually own to get a feel for what the saw is capable of in terms of flex of the frame. I would love to have the numbers if nothing else just for fun and the numbers Roland Jones came up with despite being limited to one plane and ambiguous as to where the magical "this saw sucks" line is really does bear out in real life. The saws he found to have the lowest deflection numbers do the best with wide/thick/high tension blades, those on the other end won't nearly handle the blades the advertising cuts sheets say they will.

As we are now we are often stuck with the manufacturers to trust which can be a slippery slope, it is often hard to determine when info is advertising fluff (our 14" saw will tension a 1"x.035" carbide tipped band) and what is real. On the other hand when we see or hear a statement like Eriks regarding the S45N are we to think wow thats honestly or hmm is this an upsale? (BTW his statement about the S45N vs the MM16 is absolutely accurate IMHO) So while I agree without lots of data from different machines gathered in a standardized way the number mean nothing (short of what I would see as a fun intellectual pursuit). If we DID have that data it would be one more bit of information to shed light on what claims are over reaching and what claims are accurate regarding useful blade size on a particular saw.

This whole thing reminds me of the tourque vs hp debates surrounding cars, while some are arguing one is more important than the other and some are comparing peak values for different motors I am screaming that it is neither it is the total area under the curves of both the torque and hp traces between the specific cars shift points. (btw I am correct :D) In the end it truely only matters which car makes it 'round the track in the lowest time, like it only matters in the finished results here.

Maybe we need a tester guy... Now it is time to turn it over to our tame sawyer, some say he has carbide inserts for teeth, some say he shaves veener thin enough and uses it for toilet paper, all we know is he is called the bandsaw Stig.

^ poor reference to the BBC show Top Gear.