POWER-TRIPP Performance

POWER-TRIPP Performance 205.663.4304 103 8th Ave. NW Alabaster Alabama 35007

AFR Meters and Multi-Gas Analyzers

There is a lot going on in the exhaust system. As the engine is running, there are pulses of exhaust gases each time one of the cylinders opens an exhaust valve that converge and interfere in all of the primary tubes and in the collector. There are also acoustic tuning pulses (positive and negative) that are bouncing around, converging, diverging, and interfering, as well. The temperatures go from reasonably mild to red hot and back again... all in less than the blink of an eye. This is a lot to ask of any sensor.

Despite many sources stating that gasoline burns at a stoichiometric air/fuel ratio of 14.7:1, what is not often understood is that this if for "perfect gasoline" that you cannot buy at the pump, and that this gasoline must be perfectly vaporized, and the engine must be 100% efficient, not the 95% efficiency where a really good engine normally operates. For best performance, engines actually operate at 10% (or more) rich conditions, and oxygenated fuels need 1% more fuel. This throws even more confusion into the search for best power.

Wide band oxygen (wbo2) sensors have drastically improved the ability to tune modern engines. Used in conjunction with a controller that takes the sensor voltage and converts it to a reading, the fuel-air ratio (AFR) or Lambda meter is a tuning tool that belongs in every tuning toolbox. However, as with any tool, to get the most out of it, one needs to understand how it works and its limitations. Without understanding the operation and limitations of the tool, one cannot expect the best results.

TechEdge has a great explanation of the operation of the 5-wire (6-7 pin) Lambda sensor on their site at www.wbo2.com. If you have not read through this, take the time to do so, it is an excellent article.

The basic narrow band sensor is designed to operate within a very limited range (band) between 14.5:1 - 15.0:1 AFR. Outside of this range it is inaccurate, and the fact that it is not heated and temperature compensated, only makes it less accurate. According to Bosch, their wide band sensors are calibrated to stoichiometric at 14.57:1 AFR, not the 14.7:1 that is often claimed. These sensors are reasonably accurate in a range from ~10.2:1 - 22.0:1 AFR, and the Bosch 4.9 sensors have an even wider range from ~9.0:1 - 36.0:1. But they are still designed for accuracy at stoich ratios, and tend to become less accurate the further the tuning varies from stoich.

We must then consider that there are different sensors that operate in different electro-chemical ways, and each controller takes the sensor voltage and corrects it to an AFR or Lambda meter reading in different ways. The result is that many of these meters can display as far as 1 to 1.5 numbers off of the actual ratio. Despite many claims to the contrary, these meters cannot be calibrated accurately in free air, but need to be done in rich gas mixtures of known value(s). Add in the fact that the meter is not actually displaying AFR but a reading of residual exhaust gas that can be altered by many factors including a simple change in ignition advance, when the actual amount of air and fuel entering the engine have not changed. This should lead one to realize that the AFR or Lambda meter is a very good tool with good response, but the readings should be taken on a relative basis, not as an accurate representation of the AFR. The really nice thing about these units is that the better ones give a relatively quick reading of exhaust gases. But on a vehicle that can accelerate quickly, the reading of the AFR meter can still lag behind the engine speed by 1000-1500 rpm on a sweep test.

Once one realizes that the optimum AFR or Lambda reading will be different at idle, maximum torque rpm, maximum horsepower rpm, etc., it should come as no surprise that tuning to best power, torque, or fuel consumption with these devices can take a reasonably long time. As a result of this, some companies have added software that lets one target different AFR or Lambda values in a table and then uses the sensor readings to alter (or suggest alterations) to the fuel curves. This is good if you know approximately what values a given engine will want at the various engine speeds, throttle positions, MAP readings, etc. But it will only get you in the ballpark, and there are big differences in watching from the nose-bleed seats, playing left field, and knocking home runs from the plate. One must still adjust fuel values to find best power, and this does not even mention finding best ignition timing throughout the entire map.

The down side of an AFR or Lambda meter is that the sensor needs to be mounted where it will see flow from the cylinders, but stay within its operating heat and pressure ranges. The meter must average the readings taken in the maelstrom mentioned in the very first paragraph of this article. And the reality is that any sensor (even a thin EGT thermocouple) disrupts the pulses in the exhaust simply by being in the flow or path. The sensors live a short life using leaded fuels, don't like silicone sealer residue at all, oxygenated fuels can cause inaccurate readings, fail with liquid water, clog with two-stroke oils, and fail without notice. While this may not sound like much, it is a lot to ask of the sensor, and it is amazing that the AFR meter can work as well as it does.

At this point, it should be obvious that for accuracy, speed, and best results (not to mention less wear and tear on the engine and dyno), we need a way of knowing exactly what the exhaust gases are, and what they are telling us in order to reduce time and wear. Luckily, just such a tool exists... a multi-gas analyzer.

A multi-gas analyzer uses a pump to pull in a sample of the exhaust gases and scans the sample with an infrared laser in the spectrometer. The way the light is diffracted lets the spectrometer know what gases, and how much of them, are in the sample. It is simple and accurate, without any guessing or correcting. The problem is that many analyzers can take 10 to 12 seconds to give a reading of the gases in the sample. However some units for performance tuning require 6-10 seconds, and others use a chiller to increase the density of the filtered gases and condense water out of the sample, while incorporating the shortest sampling tube possible to give results in even less time. Interestingly, it can take several seconds for the flow and pulses through an engine to stabilize when loaded at a steady rpm and load for step-testing using a gas analyzer or AFR meter. So with step testing, the multi-gas analyzer is a tuner's dream.

But what are the different gases, and what do they mean?

As general overview,A multi-gas analyzer generally gives 4 or 5 gas readings:

CO2 - Carbon dioxide: This is the relative efficiency of the burn from complete combustion of the fuel. At all engine speeds, the best power will generally be found within less than a 0.3% change.
O2 - Oxygen: If the percentage is high, it indicates that more ignition advance can likely be used, or that the different cylinder offsets/staggers need correction due to a lean condition in one cylinder.
CO - Carbon monoxide: Mixture strength from partially burned fuel. If a given throttle position makes best power at a given percentage. All other throttle positions will be very close to this reading at best power.
HC - Total hydrocarbons in parts per million: Unburned fuel - Shows general state of engine health with lower readings for good large bore engines, and higher readings for good small bore engines.
NOx - Nitrogen Oxides in parts per million: High readings indicate high combustion temperatures and can be a precursor to detonation. Among many things, this hydrocarbon count indicates compression and squish/quench conditions, as well as spark strength and combustion dynamics.

Generally you can look at the relative values of the gases and see even more, i.e.: high HC with low CO2 readings can indicate that too much ignition advance may be occurring and that less advance may improve power. Another issue is that it is easy to detect misfires in the exhaust traces with a bit of experience. This can be of great value when tuning cruise regions for best mileage, or diagnosing ignition issues. Once you know what a given engine family wants in each gas reading for best power, it is quick and easy to tune the fueling and ignition tables, allowing more time for the details.

For detailed individual cylinder tuning, a hole can be drilled in each header primary and either braze/weld or Rivnut a small tube in place for the exhaust sampling, that does not interfere with the pulses in the header, for best results.

Pump gas, leaded or oxygenated race fuels, two-stroke oil, alcohol fuels, etc. can all be tested and tuned using a multi-gas analyzer. Liquid water does not hurt the unit, and the units lasts for decades with regular cleanings after hundreds of test/tuning sessions.

The biggest down side is cost. With good multi-gas units costing 10 to 15 times the cost of a good AFR meter, it is not for the faint of heart. But the results speak for themselves, and it will not take long for the tuner or the customer to ask how he ever got along without it.

At this point, it could be said that the AFR/Lambda meter is quick, but relative, and the gas analyzer is slow, but accurate. Together both a multi-gas analyzer and an AFR meter are excellent tools that can be used in step and sweep tests to get results that continue to impress. Add in a form of knock detection and a good bit of experience and the results are safe, quick, accurate, and FAST.

The above chart shows the gases in the exhaust that an infrared exhaust gas analyzer reads and how the gases change as the A/F mixture changes. The 14.7:1 Stoichiometric calculation value is for an engine with perfect fuel vaporization and it assumes you are using “pure” gasoline. Reformulated and/or oxygenated gasoline will have a slightly richer Stoichiometric A/F mixture ratio value. Bore diameter and combustion efficiency can alter the values as well. Since actual operating conditions in motorcycles are different, this graph is for reference use only.

Bridge Analyzers has written an informative paper on using EGA to tune a motorcycle that is worth reading - Bridge White Paper 17.

It is not all about the equipment used, and not all tuning sessions are the same. It takes time, effort, and knowledge wrapped in experience to correctly map each and every cell in fuel, MAP, throttle position, and ignition tables. And working out issues with starting, idle, temperature and pressure compensation, and all the other details required for best performance, smoothness, and throttle response takes time. Time is a valuable commodity that we all must pay for with hard earned money. Make sure you are paying for the best tuning you can get, and not just a few alterations at wide open throttle and a pretty graph on paper.

- POWER-TRIPP Performance