Don't buy SMA cable assemblies from CentricRF

[Someone]

Like I said, math or GTFO.

The math is not that complicated.
Start with a simple beam fixed at one end on the nut, and a load at the other.
The moment along the beam is equal to the torque, maximal at the fixed "nut" end and linearly tapering to 0 at the load.
Given the breaking "click" joint of the torque wrench is not coaxial with the fixed point, it will be breaking at a fraction of the torque imparted at the nut.
If you move the point of the load along the beam while keeping the click joint consistent in position and breaking moment then you get the result found in the video above.

Closer to the nut = higher than anticipated/intended torque.

Maths first year engineering!

[Smoky]

"We tried a few printed wrenches.  The biggest problem I see with these is they will wear fast, changing the torque.  You would be much better served going the route tautech shows.  Those wrenches can be adjusted.  You can always apply a little Kapton tape like that video I linked shows if the concern is with the surface finish."

I was just envisioning a tool that you could palm and imagine it's like a flare nut wrench where still the wrench goes around all corners of the hex but leaves a slot for the cable to slip through.

Some impact sockets have plastic liners so that chrome lugnuts avoid being marred or even softer metals having their corners getting rounded if the wrench slips.

Btw, I was just joking about scratching the nuts, but what the heck, if it can be avoided, why not

[KE5FX]

Calling out the thumbnail used for that video as evidence was pretty stupid on your part, but expected.   I suspect it was obvious he was removing the connector.   The math was fairly straight forward and so was the pivot point but again,  you appear more interested in pointing out your political views than discussing anything technical.   You're wasting your time.

What's the guy in the photo doing?  Removing, or installing?  I have no beef with the video, only the photo.

[KE5FX]

The math is not that complicated.
Start with a simple beam fixed at one end on the nut, and a load at the other.
The moment along the beam is equal to the torque, maximal at the fixed "nut" end and linearly tapering to 0 at the load.
Given the breaking "click" joint of the torque wrench is not coaxial with the fixed point, it will be breaking at a fraction of the torque imparted at the nut.

If it's "first year engineering," it should be easy to do the math.  What is the actual difference between the torque applied to the fastener at breakover when the wrench is held at the very tip of the handle, and the torque applied when the wrench is held at, say, the halfway point?

Unlabeled graphs might have sufficed at your engineering school, but not at mine.

[Someone]

The math is not that complicated.
Start with a simple beam fixed at one end on the nut, and a load at the other.
The moment along the beam is equal to the torque, maximal at the fixed "nut" end and linearly tapering to 0 at the load.
Given the breaking "click" joint of the torque wrench is not coaxial with the fixed point, it will be breaking at a fraction of the torque imparted at the nut.

If it's "first year engineering," it should be easy to do the math.  What is the actual difference between the torque applied to the fastener at breakover when the wrench is held at the very tip of the handle, and the torque applied when the wrench is held at, say, the halfway point?

Unlabeled graphs might have sufficed at your engineering school, but not at mine.

Unlabelled graph still presents the ratios, unlike your question with zero numerical input.

So you'd like someone to calculate the change in torque for an undefined length, with and undefined change in length, and an undefined relative breaking pivot ? I'm pretty sure the answer is undefined, but we can say the torque is larger when the force/load is at the halfway point compared to the endpoint.

Only if the braking joint distance to the rotating axis of the nut approaches 0 does the torque remain constant. Those ratios are trivial triangles!

[Smoky]

This comment I'm making here has nothing to do with Joeqsmith and KE5FX's conversation.

Torque wrenches should never be used to loosen.

Break-away torque is much higher when loosening.

[KE5FX]

Only if the braking joint distance to the rotating axis of the nut approaches 0 does the torque remain constant. Those ratios are trivial triangles!

The difference isn't zero, it just isn't relevant. 

For that matter, with these low-torque tools, the angle at which the wrench is held is likely more important.  My Fairview Microwave wrench will practically break over by itself if held horizontally by the jaws, as in the video Joe linked.  Not that the wrench should be held horizontally, but if you do, you'll quickly realize the absurdity of worrying about where you hold it.

For this conversation to go any further, I think someone's going to have to buy a scale and rig up a demo with the actual hardware in question.  And come up with a way to correct the results for the wrench's mass distribution and angle at breakover. 

[Someone]

Only if the braking joint distance to the rotating axis of the nut approaches 0 does the torque remain constant. Those ratios are trivial triangles!

The difference isn't zero, it just isn't relevant.

At which point the onus is on you to justify the magnitude of not relevant. With the breaking joint being 20-30% of the way up the shaft (Huber+Suhner or Rohde & Schwarz branded SMA torque wrenches) the change in torque from using the end vs the middle of the "handle" (bit beyond the joint) would be a 20-30% increase using the simple first order geometric approximation of the bending moments (zero thickness shafts etc).

If I've done the algebra correctly that specific case of comparing the load at the end or half way from the end to the breaking joint falls out to a neat linear: increase in torque = breaking joint distance from torque axis as a %

[joeqsmith]

Calling out the thumbnail used for that video as evidence was pretty stupid on your part, but expected.   I suspect it was obvious he was removing the connector.   The math was fairly straight forward and so was the pivot point but again,  you appear more interested in pointing out your political views than discussing anything technical.   You're wasting your time.

What's the guy in the photo doing?  Removing, or installing?  I have no beef with the video, only the photo.

Maybe cut back on the eggnog. 

[joeqsmith]

...
Torque wrenches should never be used to loosen.

Break-away torque is much higher when loosening.

While I agree in general, the style wrench mentioned is similar to mine and will hit a dead stop when tightening or loosening.   They offer that same wrench for different sized connectors (which require much higher torque).  The highest I found was factory calibrated at 14 in-lbs.  Looking at that video where I pushed to SMA connectors far beyond their limits, at 35 in-lbs the connectors along with the copper ground plane were tore from the test boards.

[joeqsmith]

For this conversation to go any further, I think someone's going to have to buy a scale and rig up a demo with the actual hardware in question.  And come up with a way to correct the results for the wrench's mass distribution and angle at breakover.

Well?

[nctnico]

Like I said, math or GTFO.

The math is not that complicated.
Start with a simple beam fixed at one end on the nut, and a load at the other.
The moment along the beam is equal to the torque, maximal at the fixed "nut" end and linearly tapering to 0 at the load.
Given the breaking "click" joint of the torque wrench is not coaxial with the fixed point, it will be breaking at a fraction of the torque imparted at the nut.
If you move the point of the load along the beam while keeping the click joint consistent in position and breaking moment then you get the result found in the video above.

Closer to the nut = higher than anticipated/intended torque.

The key is the pivot point for the click mechanism being away from the rotation point of the nut. The error in torque will go up exponentially when the force is applied closer to the pivot point. If the force is applied at the pivot point, the mechanism will never 'click'.

The video joeqsmith linked to in this reply, shows the math:
https://www.eevblog.com/forum/testgear/dont-buy-sma-cable-assemblies-from-centricrf/msg5746497/#msg5746497

IMHO it is worthwhile to teach inexperience engineers the correct way to work with a torque wrench so they will do it right. Instead of assuming people know how to operate a tool. Just look at how many people hold a hammer close to the head instead of using the entire handle.

edit: typo

[KE5FX]

But the difference is immaterial.  It's silly to pretend otherwise, and counterproductive to waste limited bandwidth between instructors and students by telling the students silly things.

If I weren't speaking from experience, I wouldn't speak at all.

[joeqsmith]

The discussion is similar to dealing with the believers in free energy, warp drives, all things Tesla, robots made with string and old socks with AI.  Hams seem to follow a rich history of deep rooted beliefs based on folklore.  I dedicated a thread to the Bird Wattmeter being the gold standard of accuracy for the hams. 

[nctnico]

I wouldn't go that far. But I do believe that people should be instructed properly. Sometimes to the point where you might think something is too commonly known to need explaination. But that is a trap for old players 

I'm more and more involved in bigger projects with bigger teams which includes coaching / teaching junior engineers. One of the things I realised quickly is how much knowledge I take for granted. I'm not somebody who is going around labelling people silly or stupid though. If somebody under my supervision is doing something wrong which I didn't explain (well enough) then that is my fault. A long time ago an intern at a former employer managed to put a PCI express card into an AGP slot (or the other way around -doesn't matter-) as part of putting a computer together. As a computer savy 'crowd' we all found that a silly mistake. But in hindsight the fault was on the person supervising the intern.

[Someone]

Like I said, math or GTFO.

The math is not that complicated.
Start with a simple beam fixed at one end on the nut, and a load at the other.
The moment along the beam is equal to the torque, maximal at the fixed "nut" end and linearly tapering to 0 at the load.
Given the breaking "click" joint of the torque wrench is not coaxial with the fixed point, it will be breaking at a fraction of the torque imparted at the nut.
If you move the point of the load along the beam while keeping the click joint consistent in position and breaking moment then you get the result found in the video above.

Closer to the nut = higher than anticipated/intended torque.

The key is the pivot point for the click mechanism being away from the rotation point of the nut. The error in torque will go up exponentially when the force is applied closer to the pivot point. If the force is applied at the pivot point, the mechanism will never 'click'.

The video joeqsmith linked to in this reply, shows the math:
https://www.eevblog.com/forum/testgear/dont-buy-sma-cable-assemblies-from-centricrf/msg5746497/#msg5746497

So why reply to that message? A big pile of furious agreement with what I said? It was KE5FX asking for a worked example.

[Someone]

But the difference is immaterial.

How can you justify the difference being irrelevant/immaterial? If a torque wrench has a 10% tolerance, and where you press on it adds another 30%, while the specification for the joint is +/- 33% then you've failed to meet the requirements.

[KE5FX]

But the difference is immaterial.

How can you justify the difference being irrelevant/immaterial? If a torque wrench has a 10% tolerance, and where you press on it adds another 30%, while the specification for the joint is +/- 33% then you've failed to meet the requirements.

If holding the torque wrench halfway up the handle adds 30% torque, the problem isn't where you're holding it.  The problem is that you bought a really crappy torque wrench.

[nctnico]

But the difference is immaterial.

How can you justify the difference being irrelevant/immaterial? If a torque wrench has a 10% tolerance, and where you press on it adds another 30%, while the specification for the joint is +/- 33% then you've failed to meet the requirements.

If holding the torque wrench halfway up the handle adds 30% torque, the problem isn't where you're holding it.  The problem is that you bought a really crappy torque wrench.

Looking at the R&S torque wrenches & sizes: https://scdn.rohde-schwarz.com/ur/pws/dl_downloads/pdm/cl_brochures_and_datasheets/specifications/3607_1994_22/ZN-ZTW_dat-sw_en_3607-1994-22_v0300.pdf

And going by the formula given in the video, applying force to an R&S torque wrench (the f R&S®ZN-ZTW model .10 on page 6) at 1/3 of the handle instead of the specified load point, gives an error of over 50%. Pushing it at half way the handle, the error is over 20% already. This while the accuracy of the torque wrench itself is specified at 2.25%. The bottom line is, if you want to use these torque wrenches within their specification -if that is important-, you have to apply the force exactly at the specified load point. Even deviating 1.5cm from the load point gives an error of 3%.

[KE5FX]

https://youtu.be/NVE_ztptYPY

TL,DW: Hold the wrench like a sane human being and it will be fine.  No, really.  It will be fine.

I was able to overtorque to ~12-13 in-lbs by tying a wire just below the halfway point and pulling on it (not shown, done on a different bench), so I can't support my original blanket assertion ("the breakover point is independent of (the handle's) leverage.")  Frankly, it feels more like plain old stiction than mechanical action because the result was so inconsistent -- in one trial it took more than 20 in-lbs to break over, which makes no sense mathematically -- but I will let Joe have that hill.  There are various other ways to do the same thing, all of which involve pulling or pushing the handle in odd directions. 

But unless you are just trying to break something, the handle exists for a reason and can be used as such.  There's no reason to hold it by the tip of the handle, or to instruct anyone else to do so.

[G0HZU]

To help try and resolve this issue, I waited until I was back at work today. There is a suite of Gedore torque testers there and I selected the 0.5 to 5 Nm model. It looks like a bigger and more expensive version of the one linked to below:

https://www.gedore.com/en-de/products/torque-tools/torque,-measuring-testing-appliances/torque-testing-appliances/capture-hub/ch-5---3124053

I did try and remember the model number but I've forgotten it. The datasheet said 0.5% accuracy across most of the range. It measures and holds the display at the peak torque achieved at the break point of the torque tool under test.

I tested three Agilent 8710-1765 torque tools taken from 85033E VNA cal kits. I held them the same way as recommended by my N4431B-60006  Ecal module manual as seen on page 4-19 here:

https://www.keysight.com/gb/en/assets/9018-03493/reference-guides/9018-03493.pdf?success=true

This means holding it between finger and thumb at the tip of the tool.

When I do this, I get really repeatable results each time. There is a slight difference between each torque tool but not much. When tested, they ranged from 0.897Nm to 0.917 Nm but probably none of them have been adjusted in many years.

In terms of repeatability, I could typically achieve better than +/- 0.01 Nm if I really tried to hold and rotate the torque tool in exactly the same way. This was quite impressive. I also saw about a 0.025 Nm difference if I reversed the tool as the hex cutout is at an angle.

If I then tried the same at halfway down the gold handle of the 8710-1765 then the reported torque was at about 1.07 Nm. If I tried at most of the way down the handle (where it says 0.9 Nm) then the torque went up to about 1.3 Nm.  However, it's unlikely anyone would use the tool like this.

If I held it the way most people would hold it fairly near the tip using several fingers and thumb (as shown as the 'wrong' way to hold it in that guide doc) then I couldn't get repeatable results. A lot depends on how the fingers or thumb apply the pressure. I could see anything from 1.0 Nm to 0.8 Nm. Usually it showed 0.85 Nm to 0.95 Nm which is quite close to the 0.9 Nm spec for this torque wrench.

I did also notice that if I held the torque tool like a golf club with my thumb along the shaft and several fingers wrapped around it, the breaking torque could sometimes go below 0.75 Nm.

So in summary, if you are a nerd and you really do want to achieve repeatable torque settings that are impressively close to 0.9 Nm then hold it at the tip between finger and thumb as recommended by Keysight.

However, if you are an engineer just doing everyday SOLT calibrations with a sub 8.5 GHz VNA then I don't see a problem holding it the 'wrong' way because the error in the torque isn't that significant in my opinion.

[Smoky]

G0HZU, if you have the time, please try this.

Tighten the nut of an SMA connector onto a fitting with a torque wrench to its proper torque.

Next, make a mark on the fitting where one of the corners of the nut is located (mark the corner of the nut too).

Now back the nut off 1/4 turn or so until it's slightly loose.

Then tighten the nut back with just your fingers and see how close you can get to the mark on the nut to line up with the mark on the fitting.

I have no idea what to expect but was just wondering what finger-tight is

[G0HZU]

G0HZU, if you have the time, please try this.

Tighten the nut of an SMA connector onto a fitting with a torque wrench to its proper torque.

Next, make a mark on the fitting where one of the corners of the nut is located (mark the corner of the nut too).

Now back the nut off 1/4 turn or so until it's slightly loose.

Then tighten the nut back with just your fingers and see how close you can get to the mark on the nut to line up with the mark on the fitting.

I have no idea what to expect but was just wondering what finger-tight is

I think I can try this at work with the torque tester. It's this model here:

https://gedoregroup.ir/wp-content/uploads/P30390-ET-Cal-Compact-Instructions-PDF-Aug-2011.pdf

It's the ET-cal-5  0.5 Nm to 5 Nm model. There's a load of adaptors that plug into it including one for SMA sized fixtures.

I also tried a classic Huber and Suhner SMA torque tool Z-0-0-21 on it yesterday. These are spec'd at 1Nm. 


https://www.farnell.com/datasheets/3157770.pdf


I was getting bored by the time I tested this one, but it read 0.9 Nm when held with just thumb and finger at the sticker at the end. This was quite repeatable but not quite as repeatable as the Keysight torque tool.

When held half way down the silver handle with thumb and finger the torque went up to about 1.3 Nm. I don't think many people would use it like this because it is not as easy to use it this way and it isn't a natural way to hold the tool.

I could reach 1.6Nm if I went closer to the pivot point. This really wasn't pleasant to do with just finger and thumb so I don't think this is a realistic test.

If used fairly casually near the tip using thumb and a couple of fingers (the way I usually use it) the torque varied from about 0.85Nm to about 1.05 Nm.

The golf club grip brought the break point down to below 0.75 Nm.

It was typically about 1 Nm though when used the way most people would use it. This is good enough for the majority of VNA based applications.

The Suhner tool definitely can't match the Keysight tool for repeatability and the torque break point is 1 Nm rather than 0.9 Nm. I'm not sure this matters much though...


Maybe if someone was doing a metrology grade comparison between calibration kits up at many GHz it might be worthwhile to swap across to the Keysight tool (and use the Keysight method) to achieve a repeatable torque throughout the SOLT process.

However, for everyday VNA stuff I'm happy to carry on using the little Suhner tool across a spread of about 0.85 Nm to 1.05Nm when held casually near the end using two or three fingers and a thumb.

[joeqsmith]

https://youtu.be/NVE_ztptYPY

TL,DW: Hold the wrench like a sane human being and it will be fine.  No, really.  It will be fine.   I was able to overtorque to ~12-13 in-lbs by tying a wire just below the halfway point and pulling on it (not shown, done on a different bench), so I can't support my original blanket assertion ("the breakover point is independent of (the handle's) leverage.")  Frankly, it feels more like plain old stiction than mechanical action because the result was so inconsistent -- in one trial it took more than 20 in-lbs to break over, which makes no sense mathematically -- but I will let Joe have that hill.  There are various other ways to do the same thing, all of which involve pulling or pushing the handle in odd directions. 

But unless you are just trying to break something, the handle exists for a reason and can be used as such.  There's no reason to hold it by the tip of the handle, or to instruct anyone else to do so.

A sane person follows the manufactures procedures. 

I was expecting  you to follow their instructions below, but you would need to ditch the cheap torque meter for a peak reading meter.   Of course, the absolute values are meaningless but at least you could get some relative data.    Then, don't cover up the handle with your hand as you show.  Maybe just use a finger and work your way up the handle in one inch increments. 

• To improve performance of a wrench that hasn’t been used recently, actu-
ate the wrench by clicking the wrench head back and forth several times
before use. This will spread the lubricant through the internal mechanism.
• Click type tools require smooth consistent application of force. Each torqu-
ing action should be completed within 2 seconds. Slower or faster torquing
action will detract from accuracy.


***
Maybe measure at six locations.   You did later state anywhere on the red handle, but with the 3 closer to the pivot point, I expect you will measure a big increase.  While you asked about where the pivot point was for your wrench, I think once you place your finger in the far left location and apply force, you will have answered that one for yourself.

[joeqsmith]

To help try and resolve this issue, I waited until I was back at work today.   ....

Does your company have a set of gages to repeat the OPs measurements?   Or, I wonder if you or your company has seen a similar trend? 

***
I'm curious if your company produces any products that use RF connectors during testing, and it so do they torque them?