<p align="center">Rev hang and ways of eliminating it
    Česky English
Rev hang and ways of eliminating it

(QREVS device)

This website has been updated several times in recent months and there are 4 versions of it. The text is mostly the same but diagrams and videos refer to different car makes and models. 2007 Mazda 3 SP23 (L3 engine) is covered in this page, 2011 Honda Civic Si (K20 engine) is covered here, 2012 Honda Accord Euro (K24 engine) is covered here and 2012 Toyota Yaris (1NZ engine ) is covered here.

 

What is rev hang?

Rev hang is colloquial expression describing an undesirable characteristic of modern cars. When throttle is closed quickly, engine speed does not drop instantly but decays slowly due to artificial delay programmed in the engine control module (ECM) by car manufacturers.

Why do they do it?

Car makers are forced by legislation to build cars meeting ever tighter exhaust emission limits, typically defined in grams/km of test for HC (hydrocarbons), CO (carbon monoxide) and NOx (oxides of nitrogen). It is a known fact that NOx tend to rise with sudden leaning of fuel mixture as it happens when throttle is closed abruptly.

Car manufactures tried since late '70s of the last century to prevent such a condition by installing vacuum operated dashpots on carburetted cars just for the purpose of slowing down the closing of carburettor butterfly valve. It was before cars with fuel injected engines became common. Now it is much easier for car designers to control many aspects of EFI (electronically fuel injected) engine operation, including prevention of sudden drop in engine speed by embedded software control. Unlike PC software, the code in ECM is written by its designers just for that ECM version. Typically, a car ECM is suitable for just one particular car make, type, model and variant. These days car ECMs are even coded to individual vehicles by their VINs (vehicle identification numbers) and this makes them non-transferable between vehicles. Specialist electronics companies usually build ECMs for world car makers (European Bosch and Siemens and Japanese Denso and Mitsubishi Electric Co. spring to mind, but there are many others) and ECM program code is complex, proprietary and generally kept secret.

It is worth noting that ECM control of the combustion process has brought motorists many benefits (cheaper unit cost, lower emissions, reduced fuel consumption, improved driveability, no need for regular tune ups and improved reliability, to name just a few) but it also gave rise to features like the rev hang.

This issue is caused by what one major car maker describes as "non-linear throttle control". It is illustrated in the picture below.

image of non-linear throttle         

What it means for the car driver is that the ECM controls how the car engine runs, or more accurately how much the throttle valve opens, without full regard for the driver's commands issued through the throttle pedal. For example, when the driver floors the throttle pedal, the ECM directs the engine to increase its speed at a more leisurely pace and the same thing happens in reverse when the driver lifts his foot off the throttle pedal and the ECM directs the engine to keep running under power for some time longer.

Apart from the exhaust emission compliance issue accounting for the presence of the rev hang, a small degree of it may in fact improve drivability of cars with manual gearboxes by reducing the amount of drive-line snatch. This is the accumulation of clearances and small free-plays in the gearbox and differential that could otherwise be felt as an unpleasant harshness, jolt or thud/shock when engine power was suddenly applied or reduced.

Misleading information in other Web sites.

Flawed explanations appear on many bulletin boards and Forums, asserting either that the purpose of the rev hang is to "burn unburnt fuel" or that the rev hang is caused by a slow reaction time of the analogue throttle valve. Sometimes even "a heavy flywheel" is blamed :-). It is beyond comprehension how could burning of unburnt fuel be possibly assisted by adding even more fuel after the throttle was closed. As for the slow throttle valve, in EFI engines it only controls supply of air to the engine. Throttle hang would not function without injection of fuel and the fuel supply to injectors is capable of being controlled instantly, if it was intended by car designers. So much for posts by Internet experts.....

Which vehicles are affected?

Many modern motor cars and possibly also motor cycles fitted with 4-stoke spark ignition combustion engines with electronic fuel injection and drive-by-wire throttle. Diesel engine cars are not known to be affected.

Perhaps not surprisingly, it is often lower-end models suffering most, as against more expensive models. With upmarket cars more time and effort is being spent on designing solutions to avoid such a problem and car buyers are charged premium price for the extra effort. Also, buyers of expensive cars understandably have higher expectations and would not accept explanation that "rev-hang is a design feature of new vehicles".

Cars fitted with automatic transmissions suffer the same rev hang problem as cars with manual gearboxes but drivers are much less aware of it because of the design and function of the auto 'boxes, especially those fitted with torque converters.

What is the effect of rev hang?

Basically the problem for drivers used to driving cars built before the rev hang existed (circa 2005-07) is one of annoyance and mental adjustment when driving new cars with rev hang. In the old days, we used to accelerate in low gear, back off the throttle, hit the clutch, select the next higher gear, release the clutch again and apply power, all within about 0.6 to 1 second, or faster or slower, as desired. Cars with the rev hang require a forced pause between backing off throttle and depressing the clutch to change gears. If the pause is not allowed, depressing the clutch immediately after backing off the throttle results in clearly perceivable increase in the engine speed at the time when it is least needed. This is caused by the ECM induced slow drop of revs and is not helped by the fact that disengaging the clutch too early lifts the load off the engine, until then working hard to drag the mass of the car forward. Clearly, the engine revs would shoot up when load is lifted off the engine that is still being operated under power by instructions from the ECM and not from the right foot of the driver.

What can be done about rev hang?

From the above description it should be clear by now that the rev hang is not a problem or fault that could be solved by changing parts (under warranty or otherwise) because it is truly a "design feature". It is coded in the ECM program by way of specifying particular amounts of fuel to be injected in the engine under various conditions, including the change of gear scenario. Car ECMs are now very clever and receive inputs from many sensors, including the throttle pedal position sensor, engine speed and road speed sensors, clutch and brake pedal switches and many other sensors and inputs.

It is also a wasted effort trying to modify other parts of the engine induction tract to reduce the rev hang by blocking/opening various ports/hoses in the idle control circuit. Also, fitting lighter flywheels does not help because the underlying issue still remains: the ECM feeding engine fuel when it should have stopped.

Elimination/reduction of the rev hang is possible by modifying software in the ECM. There are various after-market companies offering "tunes", predominantly for the purpose of improving performance by increasing power. Reports are available on line of successful "tunes" for some makes and models that can fix the dreaded rev hang.

Is there a solution?

In theory a solution was easy. A signal post-processing box was needed that would take inputs from the ECM about the duration and timing of fuel injection pulses, as well as information about other parameters involved in the gear change scenario. A microprocessor computes the required changes and outputs them to the fuelling system, effectively altering the composition of fuel mixture supplied to the engine as required.

Practical difficulties were manifold but able to be overcome. This was demonstrated by our successful design of just such a device called QREVS. Major concern was the ability of original ECM to cope with signal changes introduced by QREVS to avoid generation of internal ECM error codes, leading to possible illumination of warning lights in instrument panels (CEL) and ending with transition to limp mode.

Does the QREVS device work?

Yes. Here is a video of successful application of Qrevs device to a 2007 Mazda 3 SP23 with L3 engine and manual gearbox :


Movie file (8.6 Mb) - please allow sufficient time for video buffering.

Graphical representation of 1st to 2nd gear change is below. (Clicking the image opens full-size chart)

image of 1-2 gear change chart

It shows that when upshifting from 1st to 2nd gear at 3800 RPM in unmodifed car after the throttle was closed the engine speed dropped slowly. At the time of clutch engagement after the 2nd gear was selected the engine revs were higher than the revs required to match the road speed at that moment and sudden drop in revs ensued. This mismatch is clearly visible in the graph between 0.85 sec and 1.0 sec. time mark when revs were forced by clutch slippage to drop quickly from 3000 to 2500 rpm. In the same car fitted with QREVS the engine speed dropped faster and at a fairly steady rate from the moment the throttle was closed for the gear change.

Graphical representation of 2nd to 3rd gear change is below. (Clicking the image opens full-size chart)

image of 2-3 gear change chart

It shows that when upshifting from 2st to 3rd gear at 3800 RPM in unmodifed car after the throttle was closed the engine speed dropped slowly. At the time of clutch engagement after the 3rd gear was selected the engine revs were higher than the revs required to match the road speed at that moment and sudden drop in revs ensued. This mismatch is clearly visible in the graph between 0.75 sec and 0.9 sec. time mark when revs were forced by clutch slippage to drop quickly from 3100 to 2800 rpm. In the same car fitted with QREVS the engine speed dropped faster and at a fairly steady rate from the moment the throttle was closed for the gear change.

The above results show definite improvements in engine response times when changing gears and near elimination of rev hang in the car fitted with the QREVS device. (The exact delays were derived from videos made at 30 frames per second, i.e. from video frames spaced 0.033 second apart.)

An interesting aspect of the use of QREVS is the reduction of time to accelerate between gearchanges. A typical chart (this one is for Honda Civic Si with K20 engine) only shows the upshift from 1st to 2nd gear but the benefit is fairly obvious and would also extend to upshifts to higher gears. (Clicking the image opens full-size chart)

 
image of upshift & acceleration chart

 

Does it work by "fooling" the computer (ECM)?

Not really. Other ECM signal processing solutions rely on changing values of one or more input parameters, be they RPM signals, ignition or valve timings, coolant or intake air temperatures or any other engine sensor value to force the ECM into producing a desirable output, different from the original design. In contrast, the QREVS device does not interfere with any input sensor values and this means that the ECM output remains at all times the same as originally designed.

What about the continued exhaust emission compliance?

Knowledge of exhaust emission testing protocols enabled us to change engine fuelling in an unobtrusive way to maintain delicate emission compliance when QREVS was fitted. Official dynamometer test in a Government approved emission test cycle conducted in the vehicle testing facility operated by the Roads and Maritime Services (RMS) of New South Wales confirmed there was no change in any of the monitored pollutants, namely the hydrocarbons (HC), carbon monoxide (CO) and oxides of nitrogen (NOx) when QREVS device was connected to the tested vehicle fitted with 4-cylinder petrol engine.

Is the QREVS device suitable for all makes and models?

Many new cars have rev hang and the QREVS device was designed to eliminate it. The development was done on Honda, Mazda and Toyota cars and the device can be fitted to other makes and models provided their basic design parameters are similar. Such cars would need to have:

  1. petrol EFI engine (number of cylinders is unimportant)
  2. indirect or port fuel injection (no DI)
  3. normally aspirated air intake (non-turbo)
  4. manual transmission (no CVT, automatics, dual clutch or robotized manuals like DSG, PDK and similar)

The production ready QREVS device looks like this

image of QREVS module1       image of QREVS module2

and its typical connection diagram looks like this:

image of QREVS wiring diagram

Picture of a possible install in Mazda 3 SP23:

image of QREVS install

Final caution:

You may find the following words unreal or expressing exaggerated warnings but you may wish to consider them anyway, as I based them on long term observations and experience.

Many Internet sites and the so called Forums that appear legitimate places for general discussion of automotive affairs by enthusiasts are in fact operated by business entities for the main purpose of promotion and sales of products pushed by their sponsors. These sites do not allow any room for discussion of alternative products and their benefits (real or imagined) that may adversely affect sales of their own products. The QREVS module discussed above has been seen as such by many Forum administrators and this explains why rarely a mention is found there. If a question is posted, related links may get altered so as to be broken, positive replies get deleted and even PM messages in response to questions do not get delivered to the inquirer. I have experienced all of these actions personally. Of course, the ultimate punishment for trying to discuss alternative products is the cancellation of Forum login to avoid any disturbance to the preferred sale pitch.

One remaining example is here.

e-mail: info@startrade.org