Loos Gauge calibration testing

[Tomfoolery]

Kinda disappointed in the results, though.

I took a stack of Ivanko Olympic weights, put them on a duck walk bar my youngest son made, hung that assembly from my shop crane with 5/32" 1x19 SS rope (old shrouds I had replaced and kept for stuff like this), and tested with my new Loos PT-1 gauge.

This is a shot of the heaviest configuration, at 377 lb.



Close-up of Loose gauge -



- reading 33.5 on the scale, which translates to 435 lb (interpolating between 33 and 34 in the little conversion table in the manual). That's a 15.5% error, reading high (obviously).

The configuration included six 45 lb plates, four 25 lb plates, and a 7 lb tee-handle (6.8 lb, but I threw in the extra for the little bit of rigging under the gauge). Those plates are reported to have a +/-2% error.

Next post, I drop 100 lb and test again.

Second photo fixed. Tinypic is playing games - no matter how many times I upload it, it still returns some weird image. Same one every time. Had to use photobucket to get around that.

[Tomfoolery]

I took off the four 25 lb plates -



and rechecked the gauge. It read '29' on the scale, which translates to 320 lb with 5/32" rope. Actual tension force was 277 lb, for an error of around 15.5% again.



I have an old backstay, so I'll try the 1/8" rope another time and report back.

[Tomfoolery]

I'm a bit disappointed, of course, especially since Loos claims a +/- 3% accuracy at mid-range, though I'm not clear on what they call 'mid-range'.

On the plus side, the readings are very repeatable. I could stack/unstack the 25 lb plates and after vibrating the wire by banging on it like Stanley Clarke playing the bass, the needle always went to the same place.

But the spring is obviously a little weak, indicating a higher index number than it should for the tension it's attempting to measure. I could pad it out a little, and may do just that, but it would have made a bit more sense if Loose had made the fixed spring attachment an eccentric so it could be rotated to calibrate the instrument.

Also, the actual preload tension isn't all that critical as long as I'm happy with what I end up with, and the gauge will still be useful for checking every so often, and at the beginning of the season, since repeat-ability is high. And if I know what index number corresponds to what actual tension (calibration), then I can still work with it. We have to calibrate line-running tensionometers (similar concept, but with wire rope running through three sheaves at high speed) with calibrated test weights under controlled conditions. Calibration of any instrument is hardly a new concept.

But it's just a bit aggravating. And who knows how close to accurate other similar tension gauges are.

[Phil M]

I had no idea that the loos gauges were not that entirely accurate. I suppose the fact that I have the port stays equal to the starboard stays, and, the lower stays slightly less then the upper stays is more critical to Accurate tension for my

[C Buchs]

Does the fact that the stay isn't fixed at the bottom and the weights move have anything to do with the reading being off?

Jeff

[Sumner]

I'm wondering if the problem is the shortness of you test stay.

Wonder if "Loos claims a +/- 3% accuracy at mid-range" refers to the area around 22 1/2 on the gauge and the further you get above or below it decreases accuracy??

I have one and used it on my rigging and as mentioned like the fact that at least I know the value of one stay vs. another. I found out that what I thought was quite a bit of tension after pinning and then turning the turnbuckle wasn't near as high as I thought it was. I also realized that If I could use the mast raising tackle to pull hard enough where I could pin it and then "not" tension more with the turnbuckle that it was really low. I know a number of people pin and sail like that with no problems, so I'll bet there is a big 'safety range' of tension for our boats.

I have my tension set quite high at about 330 or so on the side stays and forestay (5/32) if I remember right.

Sumner

============================
1300 miles to the Bahamas and back -- 2015

The MacGregor 26-S

The Endeavour 37

Trips to Utah, Wyoming, Idaho, Canada, Florida

Mac-Venture Links

[Tomfoolery]

Does the fact that the stay isn't fixed at the bottom and the weights move have anything to do with the reading being off?

That's an interesting question, which I've been thinking about all night.

The stays are big extension springs. The harder you pull (tension), the more they elongate elastically (stretch), and as long as they're loaded in the elastic range, they go back to their original length when the load is removed.

So if you anchor both ends and preload it, then put the Loos gauge on and pull the spring back and hook it, the rope is forced over the two radiused blocks at the bottom into an S-bend, with just a little bend at the spring end, to get back to the straight line above and below.



The bends in the rope will actually increase the tension in the rope, but being elastic, the amount of increase would depend on how long the rope is. If it's very short, then shortening it by .001" will have a greater effect on the tension at either end than if the rope is very long. The longer the rope, the lower its stiffness, just like coil springs or any other tension spring (all other variables held constant, of course).

A weight hanging off the end of the rope and not in contact with anything else would be the same as an infinitely long rope, since no matter how much you shorten it by bending the rope over the pegs, the tension does not change.

If the gauge was calibrated to take into account the increase in tension when you measure the tension (from the induced bend and therefore increased stretch), then it would be reading lower than the actual tension. And that would require knowing how long the stay is, since that spring rate differs with length. So the built-in error correction would have to be in the other direction, reading lower than actual, and would be dependent on stay length, which would be impossible to know in advance. Uppers are about twice as long as lowers, for instance.

To Sumner's question, the short stay length may be a factor in that the mechanical bending of the rope over the pins causes the strands (which are just individual wires in a 1x19 construction rope) to slide over each other, and not giving enough length at either end may effect that. I had the bottom pins more than 200 rope diameters away from the termination, so I don't think that was causing it, but I'll recheck it later today, as I'm going to run a series adding/subtracting one plate at a time, to plot a curve. Nerdy, I know, but it's actually the world I live in (for work), and I just gotta know.

I'll also try the 1/8" rope. Different scales, which obviously take into account the mechanical resistance to bending (independent of tension) the different sizes offer. And in actual fact, the 1x19 class is stiffer in bending than say 7x19 class rope, which is why Loos states that repeatability will be there for any construction rope, but direct tension measurement is only valid for 1x19 rope.

[Sumner]

Maybe run what you are doing past the manufacture and get their input ,

Sumner

============================
1300 miles to the Bahamas and back -- 2015

The MacGregor 26-S

The Endeavour 37

Trips to Utah, Wyoming, Idaho, Canada, Florida

Mac-Venture Links

[Tomfoolery]

I probably will. I just want to get more data points, including with the smaller 1/8" rope. But it is possible the spring is off. Or the plates don't actually weigh 45 lb each (I'll check them, or borrow a dynamometer from work to check the whole stack).

The unit relies on an extension spring, and there's an index hole in the aluminium backing plate that the scale (a sticker) aligns with (the scale decal also has a hole), so it's not likely the unit is calibrated on a unit to unit basis. So any variation in the spring will cause variation in the reading.

Of course, it's not terrible that it may (or may not) give accurate information, as long as it's repeatable, which it is. And I can calibrate it myself, by putting marks on it, or just making my own conversion table. Because in the end, if the lee shrouds remain taught in exciting sailing conditions, and the inners are tensioned to prevent mast pumping and still have the right mast pre-bend, then I'll know what I need to read on the gauge to duplicate that in the future.

But I'm just a little bummed that a $100 plus gauge is 15% off, at least as I see it now. But keeping an open mind that I'm doing something wrong, or not seeing something (like short test rope).

[grady]

In dealing with cable tensions and calibrated cable tensioometers for a while now here is my idea on your numbers and setup.

The hanging of the weight vs attached does not have any affect on the readings. This is exactly the method most calibration companies use to calibrate tensiometers. They have fixed calibrated weights and different size cables. Our tensiometers are build different, we have 2 fixed points and put pressure in between those points to measure deflection. This method we are able to check tension in a short span of cable. Even with this method when you get close to the end of the cable the tension will start to read higher and give you false readings. Sailboat loss gauge is measuring the deflection in a lot larger span since both points are close together then the measurement of the deflection is farther away. With this setup I beleive stay length could be a contributing factor. Cable size is a big factor in readings. Have you double checked your stay size?

As far as the manufacturer claiming a +- 3 percent acuaracy may be optimistic. Our $1000 electronic gauge is only guaranteed to be within 2 percent.

[Tomfoolery]

Our tensiometers are build different, we have 2 fixed points and put pressure in between those points to measure deflection.

That's how the 'line rider' style running tensiometers work on some cranes. Three relatively small sheaves with fixed geometry, and the middle one has a load pin with strain gauges built into it. The boom tips with a load pin in them use three much larger sheaves, including the point sheave (the one the rope falls over toward the ground, with a fixed geometry and load pin), and handle ropes to 1-3/4" dia. They tend to be more accurate, but both system have to be calibrated with certified test weights (tens of thousands of pounds usually) to calibrate them for the single-part line, with electronics to multiply readings for multiple part operation.

This method we are able to check tension in a short span of cable. Even with this method when you get close to the end of the cable the tension will start to read higher and give you false readings. Sailboat loss gauge is measuring the deflection in a lot larger span since both points are close together then the measurement of the deflection is farther away. With this setup I beleive stay length could be a contributing factor.

I just now checked near the bottom, and got very little if any difference. I thought that might be a factor, and I haven't convinced myself that it's not (and your comments only reinforce that suspicion), but I couldn't read any different just now. It's hanging on 1/8" rope right now, and I just came down here to get my notebook to record readings as I peel weights off and got distracted with work emails.

Cable size is a big factor in readings. Have you double checked your stay size?

Did that last night. 5/32" measured a few thou under, if that much. I'll recheck. Same with the 1/8" former backstay that's on there now.

As far as the manufacturer claiming a +- 3 percent acuaracy may be optimistic. Our $1000 electronic gauge is only guaranteed to be within 2 percent.

Yeah, I had my suspicions about that tolerance, especially with a production extension spring at the heart of it. But if nothing else, readings are very repeatable, which is half the battle I think.

I'll report back what I find with the 1/8", but at a glance, it too was reading high with the full 377 lb hanging off the crane. I'll take them off one by one, down to a single 45 lb plate, and see what happens.

[C Buchs]

In my previous job, Loos supplied the stainless wire rope that we used for operating dam gates. I used to work closely with a Loos distributor here in the Northwest and I need to give him a call to catch up anyway. I'll run this by him to see what he thinks about your test.

Jeff

[Tomfoolery]

That would be great.

Here's the results from retesting both 5/32" and 1/8" rope.



I did get some quirky results when I moved the gauge, but then I realized that one portion of the heavier rope had some minor kinks. If the spools are where the bends are, you get a different result depending on how the gauge is oriented, and when it's even on the curved section vs a straight piece. Which makes sense of course, but I didn't see the minor bend in the rope when I was messing about trying it near the bottom termination to see if it did make a difference. Can't see the curve when it's under load, so I didn't realize it was even there until I slackened it.

Moral of that story being, obviously, only measure on a section of shroud that's straight when slack. Preferably, far from any kinks or curves, and better still, perfectly straight and unmolested.

[C Buchs]

I finally reached my Loos distributor. Unfortunately he has no experience with these little gauges. 3/4" diameter is small to him. He did say that he was going to try to get one as a sample that he would let me "borrow" I'm hoping we can set it up in his calibrated test bed, pull to a specified value and test the gauge. I'll follow up.

Jeff

[Tomfoolery]

Here are some plots of the data. I fudged a data point on each of the two graphs. I believe it was Plate 7 on each of the two red 'Measured' plots. It was hard to read the small scale on the unit, and I remember they weren't all exactly (whatever 'exactly' means) on a mark, and had to be visually interpolated. A little error makes for a noticeable dip in what should be a straight line. But I don't feel dirty having done that, as the 'Index' values in the tables from which they were translated (referring to the higher-resolution table in the documentation that came with it) were eyeballed to the nearest whole or half-number in the first place. And the curves should be straight anyway, as an extension spring normally is for reasonable range of values, and the scale on the unit is also linear.



But it's just interesting to see how the lines cross, for both sizes of rope, and makes me wonder why the tables or the scale on the gauge didn't take that into account.

Or my rope was messed, and/or too short. I may try again with the boat in the driveway by hoisting an old upper stay on the main halyard, held over the side by the MRS pole, and tested again.

Just out of curiosity, of course.