I've been using a couple of torque wrenchs, the kind that basically depends on the handle bending, while you read the torque by looking at a scale near the handle and seeing how much deflection there is, according to the pointer which is fixed to the socket end extends out to the scale.

I've never felt this to be 100% accurate, but hoped it was 90%.  And of course there are many times where you just have to go by feel.

Then when I went back to the tire shop because ONCE AGAIN they had used their impact gun to way over-tighten the lug nuts (to the point that my breaker bar wasn't enough), I had one of the guys say (as I was using my torque wrench at junior's insistence),  "You know, those things are inaccurate as hell".

I felt there might be an element of BS to this comment, after all, I had in effect questioned their competence by telling them they'd overtightened on all the wheels.

But it did get me wondering just how accurate this kind of torque wrench is.

What do you guys say?  And what other kind is affordable and fairly accurate, if this kind is marginal (Husky brand, from [gag] Homw Depot).

P.S.   It seems that lug nuts and tire shops are a lifetime issue.

1. Sears -  forgot to tighten them years ago.
2. Goodyear - so tight, the shop manager needed to jump up & down on a piece of pipe he'd added to his 18" breaker bar.
3. Local shop - decent guys, but had to get in the habit of making the owner write "hand tighten -  use 4-way" on the work order.  We just forgot this time.

Deflection bars are much more accurate than any standard impact wrenches.  Overtorqued bolts are a hazard.  I always redo my lug nuts after any service.

I heard that Costco was using torque wrenches but I've observed a local one falling back on bad habits.

Beware the Sears warranty for handtools extends to deflection type torque wrenches but not the "click" type. I've been burned for $90 there.  

Home Depot's Husky warranty includes their "click" type. That's the kind I am using since Sears wouldn't cover theirs under warranty.

That is total crap. The military almost exclusivley uses the deflector kind you describe. Now they have to be calibrated periodically to prove they are reading correctly. You are correct to assume they will be off by a percentage of some sort, but more like 5 -10 at the most. So the larger the range the more it can be off.
Discount Tire is the only one that claims to torque them correctly with dial type torque wrenches.
I always have to come home after they do tire stuff to my car and brake them loose with a big cheater bar to make sure I can get them off it I have a flat.

I don't think it really matters, since I think for 99% of applictaions, a torque wrench can be off by over 10% and you'll never have a problem.

At the places I worked at we used these...  

[img]http://www.kentool.com/img/TorqueIU.JPG[/img]

[url]http://www.kentool.com/Wrenches_IndividualSockets.htm[/url]

[b]prk[/b], just be glad you don't own a Chrysler product built in the '60s or '70's. Those left-hand threads on the lug nuts for the left side of the vehicle were a sure-fire recipe for snapped-off lugs, especially after every garage went to impact wrenches.

My little Mitsubishi pickup has a bad habit of galling the lug threads, making it easy to snap a lug off. So, I'm now using anti-seize lubricant on them — not recommeded by the "pros" (screws up the torque readings and may make them more prone to loosening), but still beats the heck out of replacing lugs...

Ex mechanic here. The beam type is the most consistently accurate. The click type requires periodic calibration, but can be used much more quickly, particularly at odd angles where it is difficult to view the scale, so most garages use the click type.

When dropping off the vehicle, ask the service writer politely to request that the lug nuts are tightened to the factory torque spec., and supply him with that info. Inform him that you intend to recheck them when you pick it up.

Torque wrenches are inaccurate when used at less than 20% of full scale readings.

Spent 22 years in Army Aviation and NEVER saw a deflector torque wrench used.

The pointer type torque wrench is just fine for checking lug nut torque and occasional use. I don't think they are 'off' by any consequential degree when new, but since they are cheap you could easily replace them after they break or get old. The next step up are the ones you described from Sears and HD, accurate enough for the pro mechanic just not as durable for frequent heavy load use as the top end tools. If you are a mechanic then I would recommend something a little better, I use the new digital model from Snap-On and it's very nice, it clicks, beeps and vibrates when you hit the right torque but I also have the standard click type in my box. These all must be calibrated at your expense so it's costly to maintain. It seems as though 90% of all mechanics do not torque your lug nuts, but the good ones know when to stop the impact gun. Try using anti-seize, it won't loosen the lugs while driving and it will help you remove them later on. I use it on the hub and the studs, especially with aluminum wheels.

Not to be picky, but I DEFINITELY saw this thread posted about 6 weeks ago...

Use your Search function...oh wait...

I CALIBRATE about 150 torque wrenches a week. The deflection beam is one of the most accurate but very hard to read while torque is being applied. The Dial/needle type are also very accurate but just as hard to read. The clicker types at times CAN be accurate but the design is flawed and not much can be done to help that.

Here is what you are [b]ROUGHLY[/b] looking at.

Beam Deflection  1.5 - 3 % accuracy

Dial Type 2 - 3 % accuracy

Clicker/Impulse Type 3 - 6 % accuracy.

Torque wrenches should be used between the 20-100 of the max range.

Always return the clicker type to the lowest setting after use.

The wrench should be exercised about 6 times to full scale BEFORE use.

NEVER apply torque ANYWHERE but the center of the grip/handle.

Do NOT use extensions to lengthen the wrench.

Do not use a jerking motion to tighten the bolt/nut

That about sums it up.

A wealth of information here.  Thanks for everything so far.

One trick I was shown for messed up lug nuts was to put them on backward and run them on & off to chase / clean up the threads, then put them on the regular way.   Of course, that doesn't help the problem of galling on the tapered portion.

Same here.  The only time I ever used a deflector type was determining tear torque on tail rotor driveshaft clamp bolts.  The rest of the time we use the click type.

My mistake. I should have been more specific. I was in the Navy and only had them there as well as the dial type, not the click type, but the rotating dial like a clock face type.

Not in Navy or Marine Corps Aviation.
Click type one way torque wrenches or dial gauge torque wrenches calibrated every quarter.

Here's the deal.  Someone fill in if I leave something out.  What you see pictured above will do just fine.  I use these on all cars with soft rotors.  The only reason you really need to apply the same amount of torque to all the lugs is to prevent the rotors (the disk part of disk brakes) from warping.  Most rotors won't even be affected by differences in torque between lugs unless it is very significant.  From my experience as a mechanic the only rotors that are truly affected by an impact gun are those found on Mercedes, BMW, Audi, Corvettes, Porshe, and any other vehicle with malleable rotors.  In other words, any vehicle with excellent braking.  Yes, I realize breaking has much to do with the size of the rotors and pads, but the issue here is the temper of the material the rotors are made of.  

Also, if you find that the lugs are too tight for you to remove with the POS tire iron the car manufacturer supplies you with, then it is in your best interest to buy yourself a 4-way lug wrench to keep in the car.

Cousin-Eddie beat me to it -
I calibrated several torque wrenches while on loan to the PMEL shop.  They must be exercised before use to have any hope of accuracy.

Other factors affecting torque reading accuracy are -
- the type of lube (if any) on the threads,
- the bolt material, and
- the nut material.

Knowledge of the torque on the bolt is an important factor in determining the amount of preload in a bolt that is used in primary tension in a joint, particular joints subject to repeated loads that induce fatigue.  The preload in the bolt causes the minimum stress in a load cycle to be larger than it would be in the absence of preload, causing less damage to the material during each load cycle.  [For more info, review Miner's Rule, Stress Ratio,  and Material Hysteresis Effects in an advanced mechanics of materials textbook or introductory fatigue analysis textbook.]

The last time I broke a lug wrench with about 350 ft lbf applied, I went to the tire shop and told them to retorque all the nuts on all four wheels.  When I asked the mechanic how much torque he had his wrench set for, he said "Only 90 psi" (!).  He didn't have a clue when I told him the wrench was at near maximum torque at "90 psi", not the 90 ft-lbf he was supposed to use.