IMHO that is not a very good test. Some points;
1) torque settings below 3Nm are useful on some bikes; something that is rarely found in 3/8" drive clicky wrenches.
2) main benefit of a torque wrench is that inexperienced mechanics don't overtighten stuff; they made little discrimination between torque wrenches that (safely) undertorqued vs ones that overtorqued.
3) They described the Topeak torque wrench as 'under reading' but this is not the case; it
over-reads the actual torque, which is safer than not.
4) their determination was made over a relatively short length of time with a sample size of one. This is not representative of long-term use/storage and it is not statistically significant to draw conclusions on that sample size.
5) most of the wrenches on test were 3/8" drive ones that need two hands to operate (in most cases on bike parts) without generating errors. The errors will vary with exactly how the wrench is held, and may be vastly greater than any likely basic calibration errors. (The torque tester couples differently to the wrench and so won't be subject to the same constraints )
6) The benefits of consistent one-handed operation are not given enough emphasis
7) the need for long reach bits for many parts is not mentioned
8 ) the potential for checking and adjusting the wrench calibration is not mentioned.
9) IME with torque wrenches ( which is both extensive and lengthy) any clicky type wrench should be checked pretty much every time it is used. Common faults include;
a) that the wrench takes a knock and goes out of calibration
b) that the wrench is left preloaded and goes out of calibration
c) that the wrench is left adjusted to less than zero and vital parts of the overcentre mechanism go out of position (wrench broken, or worse yet way out of calibration)
d) that the wrench dries out in storage and the first few clicks out of storage are at anomalously high torque values
e) that the wrench dries out and gives false readings all the time.
f) that a cheap wrench will wear internally and fail to work at all well after a time.
g) that dirt gets into the mechanism and renders it useless.
Thus in aerospace use (for example) there is a gadget to test torque wrenches on the shop floor and fitters (who are trained to use the wrenches and won't generate errors through holding them in an odd fashion) at required to use the tester before and after every job. Clicky wrenches fail quite often.
By contrast a home mechanic with a seldom used clicky wrench is on a pretty sticky wicket. The wrench will probably overtorque for the first few times it is used after any storage, and that is if it is still in calibration anyway (which it probably won't be).
IMHO that a clicky wrench (that overtorqued at low values... and only goes down to 3Nm anyway....
) 'won' their test without such caveats is pretty appalling.
A beam type wrench is a much better prospect in the long run; once set it basically can't (easily, some can still be held badly) go wrong. With a very simple arrangement of a lever and a dead weight you can check the settings of any torque wrench, and (unlike the others on test) the topeak one's calibration is easily adjusted in the case of an error from new.
BTW one is likely to get +/- 30% variations in fastener tension (even with the best torque wrench) depending on whether the bolt is old or new, what it and the thread are made of, and/or how it is lubricated. In general if you slightly undertighten low torque parts this is 'safe' in that the worst case is usually that they move slightly when you test the security of (say) a brake lever, saddle or handlebar. Errors in the other direction are rather less forgiving; parts are easily damaged and may then fail catastrophically, without warning, when the bike is being ridden.
cheers
) to give the torsion beam effect. The calibration can be changed by sliding the yellow scale up and down the shaft. If the shaft takes a 'set' at high torque then this gives a zero point error, which is obvious.
; I was about too mention that. In the past couple of months I have had a bar-end shifter and an STI move because I didn't apply enough enough torque using an allen key the 'wrong' way round!
