Torque wrench tests, El Cheapo, PI, and Cman beam

pipsters · Jan 2, 2013

Used an Eastwood digital Torque Adapter to test out the various torque wrenches I have.
.

1/2" El Cheapo 30-150 ft-lbs clicker
3/8" Precision Instruments split beam 100-600 in-lbs
3/8" Craftsman beam 15-75 ft-lbs beam

Surprisingly, the most consistent torque wrench was the Craftsman beam. The most accurate seems to be the clicker (above a certain range). The Precision Instruments is easy to use, but doesn't seem to be that accurate or repeatable unfortunately.

There are two % ratings. The top one is percent difference between the lowest and highest value it clicked on relative to the torque set. The second % is the greatest variance from the set torque value desired.

marlinspike · Jan 2, 2013

Well, that's disappointing. I guess that's why the PI's are cheaper than other quality torque wrenches. I guess I'll stick with Beta for my next one.

pipsters · Jan 2, 2013

On the PI I set it to 20 ft-lbs. Did some slower pulls. Got 17.5 and 18. That is 2.5 and 2 ft-lbs off. This is a new wrench, I sent my old one in for warranty work due to rusting on the inside.

The PI staff were extremely helpful had great communication, but I am really disappointed in it.

Turbo_Prop · Jan 2, 2013

Very interesting. Would not have guessed the PI to come in at 9% off. Hmmm.

pipsters · Jan 2, 2013

Turbo_Prop said:
Very interesting. Would not have guessed the PI to come in at 9% off. Hmmm.

It wasn't, it was 9% off from the lowest torque to highest torque setting. Example the set torque was 40 ft-lbs, it clicked at 38.5 and 42.1, a difference of 3.6/40 = 9%.

It's basically a rating for the consistency of the clicks. If a torque wrench clicks at 40, 40, 40, and 50, and you want it to be 40, it might be accurate most of the time but not repeatable consistently which would be bad.

beerdog · Jan 2, 2013

From an engineering standpoint a beam or dial style will always be the most accurate. Clickers and most other styles give you speed at the expense of accuracy and repeatbility. One of the benefits of more expensive wrenches is durability and ability to hold it calibration over time. Your analysis is well organized and a great start. I actiually have extensive experience in industry analyzing measurion systems and gauges including torque wrenches. If you really wanted to analyze them you would want to do a full blown Measuring System Analysis (MSA). Similiar to what you did but it would factor in randomness, multiple, people, multiple parts, and more samples. My experience is beam styles will get you within 2% of the reading. Other styles and brands will get you within 3%-8% of the wrenches reading. The cheapo also appears to have linearity issues. Meaning it is more accurate on the higher end of its range.

andywander · Jan 2, 2013

You're also assuming the Eastwood Digital Adapter is accurate and consistent.

It may not be.....

beerdog · Jan 2, 2013

Very true. Another point is calculate the standardard deviation for each data set. This is the tyoical measure of varioation. If you used excel just type "stddev=(data range)". Also calculate the average and subract it from the target. We will obviously pick at this for hours.. Ha HA..

I commend you for the effort and initiating a good conversation.

Turbo_Prop · Jan 2, 2013

pipsters said:
It wasn't, it was 9% off from the lowest torque to highest torque setting. Example the set torque was 40 ft-lbs, it clicked at 38.5 and 42.1, a difference of 3.6/40 = 9%.

It's basically a rating for the consistency of the clicks. If a torque wrench clicks at 40, 40, 40, and 50, and you want it to be 40, it might be accurate most of the time but not repeatable consistently which would be bad.

I got that. 9% variation in consistency is pretty high, not what I would have expected from PI.

pipsters · Jan 2, 2013

andywander said:
You're also assuming the Eastwood Digital Adapter is accurate and consistent.

It may not be.....

I agree. However I am fairly satisfied with the results on the other wrenches. I think it shows consistency. This was a test I did for myself but figured I'd post the results.

pipsters · Jan 2, 2013

Turbo_Prop said:
I got that. 9% variation in consistency is pretty high, not what I would have expected from PI.

OK was just making sure, no worries

Hiball · Jan 2, 2013

beerdog said:
I commend you for the effort and initiating a good conversation.

+1...

Me personally.. When I was in the market for New torque wrenches, I went with PI. I wanted something that was serviceable and with any calibration device, its only as good as its last calibration and quality components.

SMKS · Jan 2, 2013

There are a couple things to note-

Most torque wrenches are only rated as accurate from 20% - 100% of full scale. So, in the case of the el-cheapo wrench, you should really only consider tests above the 20% mark. You'll see that the el-cheapo was actually very accurate above the 20% mark. Surprisingly accurate, actually.

What exact brand was the cheap wrench?

Steinmetz · Jan 2, 2013

beerdog said:
From an engineering standpoint a beam or dial style will always be the most accurate. Clickers and most other styles give you speed at the expense of accuracy and repeatbility. One of the benefits of more expensive wrenches is durability and ability to hold it calibration over time. Your analysis is well organized and a great start. I actiually have extensive experience in industry analyzing measurion systems and gauges including torque wrenches. If you really wanted to analyze them you would want to do a full blown Measuring System Analysis (MSA). Similiar to what you did but it would factor in randomness, multiple, people, multiple parts, and more samples. My experience is beam styles will get you within 2% of the reading. Other styles and brands will get you within 3%-8% of the wrenches reading. The cheapo also appears to have linearity issues. Meaning it is more accurate on the higher end of its range.

Agree. Interestingly, the sensory type beam torque wrenches made by Sturtevant were a hybrid between the beam type and the "clicker". The user thus benefitted from a tactile input while having the 2% FS accuracy of a beam type torque wrench. In addition, it never needed recalibration.

pipsters · Jan 2, 2013

stephen9666 said:
There are a couple things to note-

Most torque wrenches are only rated as accurate from 20% - 100% of full scale. So, in the case of the el-cheapo wrench, you should really only consider tests above the 20% mark. You'll see that the el-cheapo was actually very accurate above the 20% mark. Surprisingly accurate, actually.

What exact brand was the cheap wrench?

Yes I agree. I posted the 20 and 30 ft-lbs just for S&G's, and it's a great way of showing what you stated. It's an Ampro but I think they are all basically the same. It was bought maybe 6+ years ago.

pipsters · Jan 2, 2013

Hiball said:
+1...

Me personally.. When I was in the market for New torque wrenches, I went with PI. I wanted something that was serviceable and with any calibration device, its only as good as its last calibration and quality components.

How do you calibrate it?

pipsters · Jan 2, 2013

Also, if I'm working on engine components I'm going to be using a beam from now on! I really like the consistency of it. I'm really surprised how tight the readings were. Also, you note that they were all slightly high, I think this was due to the parallax error as I was looking slightly down on it. With a good sight line I bet it would be basically nuts on.

Hiball · Jan 2, 2013

pipsters said:
How do you calibrate it?

I have never "personally" calibrated a torque wrench, my go to torque wrench is my my jcpenney (NB head). It's been sent out for service twice since I've owned it. My PI wrenches are only a couple years old.. And based off everything I've reads out split beam wrenches is that they are very reliable and don't need freuquent service like your typical clicker..

pipsters · Jan 2, 2013

Hiball said:
I have never "personally" calibrated a torque wrench, my go to torque wrench is my my jcpenney (NB head). It's been sent out for service twice since I've owned it. My PI wrenches are only a couple years old.. And based off everything I've reads out split beam wrenches is that they are very reliable and don't need freuquent service like your typical clicker..

That's what I thought. Ironically I use that el cheapo for lug nuts all the time...and it's held it's consistency. PI was great about communicating with me, I'm going to email them my results and ask them what they think. I'm hoping I can get this worked out, I really like suing the PI but NEED it to be consistent.

vgs8606 · Jan 3, 2013

This is so funny, I am literally laughing my *** off!!!

BUT WE ALL KNOW PI IS THE BEST BECAUSE WE PAID THE MOST FOR IT!!!

GeneralDisorder · Jan 3, 2013

I would like to see how the CDI's stack up against the PI. CDI being owned by Snap-On and with Snap-On no longer doing business with PI I would be interested to see if there's much difference there. CDI is typically a little cheaper than PI and I have several of their click-type units. I haven't had a problem with one yet.

As for calibration - chuck the anvil in a vise and hang weights from the handle till it clicks on a given value. Do a little math and you can determine how accurate it is without using some other digital device that may introduce further calibration errors into your measurements.

It's a commendable effort but I think the use of the Eastwood digital torque adapter which itself has a calibration nullifies the results for the most part. You can't introduce additional variables into the system if you want accuracy. Best to go back to basics - hanging weights on the handle eliminates a questionable variable from the equation.

GD

wafrederick · Jan 3, 2013

I don't like the beam type torque wrenchs,hard to read.This is true with tight spots that are dark.There are goofy torque specs out there,I know Ford and Subaru are two with head bolt torque specs.Torque them down and loosen up a little bit,retorque again finally turn with an angle gauge for the final torque spec.

Steinmetz · Jan 3, 2013

GeneralDisorder said:
I would like to see how the CDI's stack up against the PI. CDI being owned by Snap-On and with Snap-On no longer doing business with PI I would be interested to see if there's much difference there. CDI is typically a little cheaper than PI and I have several of their click-type units. I haven't had a problem with one yet.

As for calibration - chuck the anvil in a vise and hang weights from the handle till it clicks on a given value. Do a little math and you can determine how accurate it is without using some other digital device that may introduce further calibration errors into your measurements.

It's a commendable effort but I think the use of the Eastwood digital torque adapter which itself has a calibration nullifies the results for the most part. You can't introduce additional variables into the system if you want accuracy. Best to go back to basics - hanging weights on the handle eliminates a questionable variable from the equation.

GD

And I suppose that you'll be using the weights from that barbell set you received at Christmas? That's a "calibrated" set for sure!

03protege · Jan 3, 2013

I feel like someone on here has to A. have a decent selection of torque wrenches and B. have a good relation with a tool truck dealer that has a real torque tester on board and can use that for a test.

Steinmetz · Jan 3, 2013

The dial-type CDI's are cheaply made. I've disassembled a few of them. The old Snap-On Torqometers were bulletproof.

pipsters · Jan 3, 2013

For those interested in the Eastwood Torque Angle adapter, I have another video here showing the angle function. It's pretty neat.

One thing about the unit - it's very bulky. I don't know the size of similar units for sale at places like Amazon or Harbor Freight, but you couldn't use this for everyday torquing. I got it mainly to see how my torque wrenches stood up and also for the angle function, as I'm doing more angle torquing these days.

For those who don't want to believe the accuracy of the unit, I can see where you are coming from. The research I have done however shows these strain type digital gauges to be extremely accurate. Also of note is how consistent some of the torque wrenches I had were, ie the beam style. The digital adapter basically showed how consistent it was, if something is 5% off not a big deal I would rather have something be 2% consistent vs 9% consistent when torquing multiple bolts as long as the accuracy is somewhat close.

The PI in comparison was neither accurate nor consistent.

Thunderbisciut · Jan 3, 2013

GeneralDisorder said:
As for calibration - chuck the anvil in a vise and hang weights from the handle till it clicks on a given value. Do a little math and you can determine how accurate it is without using some other digital device that may introduce further calibration errors into your measurements.

It's a commendable effort but I think the use of the Eastwood digital torque adapter which itself has a calibration nullifies the results for the most part. You can't introduce additional variables into the system if you want accuracy. Best to go back to basics - hanging weights on the handle eliminates a questionable variable from the equation.

GD

I agree here, always eliminate as many variables as possible to test the accuracy of anything, never add more.

You would have to math out the leverage from where you place weights on the handle.

And I suppose that you'll be using the weights from that barbell set you received at Christmas? That's a "calibrated" set for sure!

Come on, seriously? You could measure with a truckload of severed thumbs if you wanted. Even the crapiest shipping scales are very accurate, and if all else fails you could annoy the post office to weigh some stuff for you.

pipsters · Jan 3, 2013

Thunderbisciut said:
I agree here, always eliminate as many variables as possible to test the accuracy of anything, never add more.

You would have to math out the leverage from where you place weights on the handle.

Come on, seriously? You could measure with a truckload of severed thumbs if you wanted. Even the crapiest shipping scales are very accurate, and if all else fails you could annoy the post office to weigh some stuff for you.

If the crappiest shipping scales are accurate why would this not be, using the same tech?

How do you explain the consistency of some of the wrenches? Blind luck?

SMKS · Jan 3, 2013

GeneralDisorder said:
I would like to see how the CDI's stack up against the PI. CDI being owned by Snap-On and with Snap-On no longer doing business with PI

There have been a couple threads on this recently. It sure looks like the Snap-on split beams may still be made by PI.

No one has really ever produced any evidence either way. This is an article that talks about the split between Snap-on and PI, but it's not clear that Snap-on stopped buying every single torque wrench from PI. They could have just started sourcing the micrometer wrenches from CDI (which they own) and continued to source split beam wrenches from PI.
http://www.garagejournal.com/forum/showthread.php?t=136911

In the past, PI may have made more than just the split beams for Snap-on, but the Snap-on, and even the CDI split beams, sure look like they're still made by PI.

More pics/info in this thread:
http://www.garagejournal.com/forum/showthread.php?t=136911

Steinmetz · Jan 3, 2013

Thunderbisciut said:
I agree here, always eliminate as many variables as possible to test the accuracy of anything, never add more.

You would have to math out the leverage from where you place weights on the handle.

Come on, seriously? You could measure with a truckload of severed thumbs if you wanted. Even the crapiest shipping scales are very accurate, and if all else fails you could annoy the post office to weigh some stuff for you.

You must be kidding.

Farmall450 · Jan 3, 2013

andywander said:
You're also assuming the Eastwood Digital Adapter is accurate and consistent.

It may not be.....

Very true....but it should at least be consistent.

woody 73 · Jan 3, 2013

The thing that still confuses me the most is that if I add just a dab of Anti-Seize the reading is off (at least that is what all the engineers have said). Does that mean if I add a few drops of oil to the wheel lug threads the reading will be off?

I only know from past experience that the super cheap beam type scales were always very good; and if they got upset with me using the anti-seize, it sure made my job easier the next time I had to remove those rusted lug nuts.

sdguy55 · Jan 3, 2013

vgs8606 said:
This is so funny, I am literally laughing my *** off!!!

BUT WE ALL KNOW PI IS THE BEST BECAUSE WE PAID THE MOST FOR IT!!!

I don't think that was point being made at all. Good try though

Sent from my HTC One X using Tapatalk 2

GSteg · Jan 3, 2013

Most torque specs are for dry threads. Anti-seize will reduce the friction coefficient, ultimately changing the necessary torque value to achieve a certain bolt tension.

mikebramel · Jan 3, 2013

WOW you got a good torque wrench from Harbor Freight. I buy alot there but the first one I had was junk and so was the second one. No thanks

hchtractor · Jan 3, 2013

mikebramel said:
WOW you got a good torque wrench from Harbor Freight. I buy alot there but the first one I had was junk and so was the second one. No thanks

You make a good point. It would be a mistake to draw any inferences about accuracy or consistency beyond the three specific tools Pipster used... ie., test3 identical but different tools and the results may be totally inverted. Would be hard to draw any inference about the brands/types in general without 30+ of each brand/type. Would be interesting, though, to see if an experiment could be designed so that GJ members could independently test and report on their wrenches (with consistent methods) and enough data gathered to draw any conclusions. Would be tough to manage consistency of testing, though.

Snapped-off · Jan 3, 2013

woody 73 said:
The thing that still confuses me the most is that if I add just a dab of Anti-Seize the reading is off (at least that is what all the engineers have said). Does that mean if I add a few drops of oil to the wheel lug threads the reading will be off?

I only know from past experience that the super cheap beam type scales were always very good; and if they got upset with me using the anti-seize, it sure made my job easier the next time I had to remove those rusted lug nuts.

"Technically" the oil from your fingers is a no-no on threads too. When we work on military orders at work, we have to wear latex gloves.

One reason being dirty fingerprints on their CARC paint, and the other for handling the bolts before torquing.

I think it's a little absurd, but whatever the government wants I guess...

vgs8606 · Jan 3, 2013

sdguy55 said:
I don't think that was point being made at all. Good try though

Sent from my HTC One X using Tapatalk 2

If the results were contrary to the posted one, do you think anybody would have given benefit of doubt to the el cheapo one??? Fat chance!!

Treeman · Jan 3, 2013

pipsters said:
It wasn't, it was 9% off from the lowest torque to highest torque setting. Example the set torque was 40 ft-lbs, it clicked at 38.5 and 42.1, a difference of 3.6/40 = 9%.

Isn't the % calculation incorrect here?

At 38.5: 1.5/40 = 3.75%
At 42.1: 2.1/40 = 5.25%

The second reading is out of the +/- 4% calibration level claimed by P.I. User error could easily account for this: wrong hand position, not pulling slow and consistent, not pulling parrallel/in line.

Here's a good place to send tools for calibration. But, based on her hand position, I believe her expertise in "Makin' the Hard Job Easy" lies elsewhere.

pipsters · Jan 3, 2013

Treeman said:
Isn't the % calculation incorrect here?

At 38.5: 1.5/40 = 3.75%
At 42.1: 2.1/40 = 5.25%

The second reading is out of the +/- 4% calibration level claimed by P.I. User error could easily account for this: wrong hand position, not pulling slow and consistent, not pulling parrallel/in line.

I calculated the result two ways as you can see. What I did for accuracy was take the furthest out of spec one and divide by the desired torque.

I also calculated how inconsistent the results were. In the example above, there was a 9% difference (roughly) in the two torque values (3.6 ft-lb spread over 40 ft-lb).

Hand position was held steady throughout and I tried various methods of pulling. In fact, the PI was absolutely horrible when pulled slow and steady, I got a result that was over 10%-12% off doing it that way (17.5 and 18.0 for desired torque of 20 ft-lbs).

It's funny because the clicker and beam have no special attention that needs to be given to them where as the PI has to almost be babied, everything had to be "perfect" in order to get a result close to desired spec. IMO not a good feature.

I just can't come up with any excuses here. I wanted the PI to perform but I just can't get a consistent reading out of it.

Torque wrench tests, El Cheapo, PI, and Cman beam

Well-known member

Attachments

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Well-known member

Member

Well-known member

Well-known member

Well-known member

Well-known member