Dyno operation
#1
Member
Thread Starter
Dyno operation
Hey guys. I am new here but not new to working on and building cars. I wrote this post for the LX guys when I was one:
http://www.lxforums.com/board/showth...ting-Explained
I have included it here in its original form:
I have seen a lot of posts from members unsatisfied with dyno numbers and dynos in general so I thought I would write up my experiences with them and what I have seen.
Some background first. I was involved very early with chassis dynos for motorcycles and automobiles. I have well over 10,000 runs over 10 years of running. I have dynoed up to 60 cars in one day at events.
I also have experience with engine dynos. I have been involved with several successful racing endeavors in NASCAR, BAJA, Road Racing, Bonneville and NHRA.
I was directly involved in dyno testing for all three NASCAR divisions.
I no longer am involved in any way, shape or form in vehicle testing, racing or tuning. I do not represent any manufacturer or sanctioning body of any kind.
This is not a “Why dyno’s are better than the Track” thread. If you are not going to look at this with open eyes then you might want to go to the track section of the forums because nothing here I say will convince you, nor will I try. All I will do is explain what I know and have observed firsthand from my experience. I don’t want this to become a “Track vs. Dyno” thread and will not entertain that direction.
That out of the way, I will explain what I know and what my experiences have been.
This post will be broken up into 4 parts:
Theory, Dyno construction, Dyno practices and Common Dyno questions.
Theory:
There are two main ways to measure horsepower or torque. Depending on which one you measure, the other is calculated from the first.
Load has been used since the very first for measuring torque. If you apply a load to an engine and stop it from accelerating you can measure thetorque available required to hold speed (or RPM) at that point. This is how engine dynos do it. They hold the engine at a specific point and read a load cell. Using the formula: HP=TQ*RPM/5252 you arrive at a horsepower number; horsepower is not actually measured… its calculated.
Inertia testing and the need for Chassis Dyno testing came later. Some more background as I know it:
The first commercial inertia chassis dyno that I know of was created for the purpose of tuning motorcycles, specifically vacuum slide carburetors. Since they are dependant on rate of acceleration and work off of vacuum they proved to be very hard to tune under steady state load.
Street testing was very difficult due to the speeds involved and ever increasing police presence. The inertia dyno idea was born. If you take a bike out and time a run in the 1 to 1 gear you can put it on a rotating drum and add mass until you get the same run time. If you can accurately measure time, mass of the drum and its circumference you have all the components needed to measure horsepower.
Force = Mass x Acceleration
If you know the mass, time of the run and change in acceleration you can measure horsepower. But F=MxA is not the end of the equation. Force has to be converted into HP by the equation HP= Force x Distance/ Time then divide by 550 (hp is the force required to move 550 lbs 1 foot in 1 second or 330,000 feet per minute).If you plot this against time you have a power curve.
A requirement to do this type of testing is accurate measurement of time. This is where the computer comes in. Inertia was only a theory until we had a way to measure time accurately.
So if you know the distance the drum travels in 1 revolution (circumference), the mass equivalent being rotated (since the drum rotates the mass rotated is not the mass static) and time you have all the parts to measure horsepower.
Notice that RPM and torque are not mentioned. You do not need either to measure horsepower. But to calculate torque you need RPM. An inductive pick up is used to measure RPM. From there you enter RPM into the equation and you get torque. Torque is part of the process as the force on the dyno drum is torque but it is not engine torque but wheel torque. Wheel torque in first gear can be in the thousands of pounds at the wheels due to gear ratio. This torque is measured indirectly as work in the equation work=force x distance.
Because we are measuring acceleration we don't need a brake to hold a steady RPM or speed. We just "Let 'er eat" or go full throttle. As long as there are no physical restrictions in the linkage this is a very repeatable state.
That’s it. No other measurements are needed for inertia. But how do we compare results in Las Vegas to results in Indy or North Carolina? Enter the correction factor.
The correction factor uses a standardized set of conditions to compare a run at different conditions.
If you look at the standards we are concerned with they are:
SAE= 77 degrees 29.234 inhg 0% humidity
STD or STP = 60 degrees 29.92 inhg 0% humidity
If you make a run under those conditions you will see a correction factor of 1.00. This is a multiplier. As conditions improve or degrade you will see the multiplier add or subtract power from all your numbers.
This is what makes comparing vehicles in varying conditions possible. This is the same as Density Altitude correction that drag racers use for the most part.
So to summarize, if you can time the run, know the mass equivalent and circumference of the drum then you can measure horsepower. Simplicity usually equals repeatability.
A quick aside about repeatability vs. accuracy:
The experiment theory is basically measuring your variable against your control. The more you control, the more accurate your results. But as you get to smaller and smaller measurements it gets harder to repeat your results.
If I measure a bore and get 4.065 and then measure it again I may get 4.067. If I measure only to inches the results repeat at 4. But is that good enough? No engine builder will agree. But if my measuring equipment is a pair of calipers and not a bore mic what will my results be?
In our case the vehicle we are testing is the variable. When tuning a carbureted vehicle you can get 1 to 3 hp repeatability because there is little adjustment potential. Timing curve is fixed as well as fueling.
Enter the ECU or PCM in the LX world. Now timing is variable as well as fueling. Depending on a myriad of variables you will see power change to best support what the ECU wants to accomplish. This is not always compatible with the highest horsepower! Now you have a whole new set of changes to your test subject. This means that 1 to 3 horsepower is the range you may see from run to run if not more as the computer adjusts and your load value changes due to temperature and other variables.
The higher degree of accuracy the less likely we can control all the variables and the less likely we are to repeat to a consistent number.
Just some things that can change a reading are: What gear you complete the run in, engine temp, trans temp, rear diff temp, atmospheric conditions, heat soak, state of tune, voltage, tire size, tire pressure, wheel weight…
The more things we control the more repeatable our numbers. Trying to repeat to 1 to 3 horsepower is unlikely due to the sheer number of variables. But if we can repeat within 3 to 5 horsepower we can measure a change above that number. More on this subject in Dyno Practices.
Load vs. Inertia
Why inertia? Because there are fewer variables and they are simpler to operate. They also reduce set up time and that means greater efficiency in testing. More cars in a day = More money for the shop. And not only that, it’s far less stressful on the car being run; this becomes especially important as the dyno pull count for a particular car increases.
Inertia needs fewer instruments to measure and they are not subject to calibration. You only need Mass, Circumference, a way of measuring time and a way of counting drum rotation.
There are no control loops to try and hold a vehicle at a steady state. This is a difficult thing to do as you must stabilize the RPM or speed before you can measure it.
Since both styles measure different values under different test conditions the results are not comparable. I have spent many years trying to make an engine dyno correlate to what the chassis dyno showed. I even wrote acceleration test profiles for the engine dyno that simulate what I saw on a chassis dyno run. I never had a match or even a ball park. Comparing Dynamic motion to Static motion just doesn’t compare easily.
Since inertia has fewer variables you will see higher repeatability. No one I know will say an engine dyno repeats within a couple horsepower without an extremely skilled operator. Since all you are doing on a chassis dyno is driving the car there is less need for the skills required of an engine dyno operator. You still need skills and I will cover that later.
Where load excels is in tuning fuel injection. Most EFI systems use tables for the ecu to look at for what fuel and timing values are needed based on load and vehicle conditions. You hold the vehicle at a preset rpm and adjust a/f and timing until you show the greatest torque output. Lather, rinse repeat and you have tuned a fuel and timing map. (This is greatly simplified but should give you the idea).
Inertia only hits so many places on the map so you will not be able to get to those other values. Since the drums only model a set amount of rotational mass you may end up having a lighter load on the vehicle. If you set timing to this load you will find quite a bit of detonation on the street when real world conditions are added in.
The best dyno is one that has both worlds. Inertia for overall testing of the finished product and Load to get the maps right. You still only need inertia for test comparisons but you need load for tuning.
Second Part:
Dyno Construction:
There are 2 types of chassis dynos:
Twin roller and single roller chassis dynos.
The twin roller chassis dyno was the original. Clayton is the first brand that comes to mind. In use the vehicle is driven over the rollers and the tire sits between the front and rear roller. It is essentially “pinched” between the rollers. The vehicle is run up to a certain RPM and a brake is applied to the rollers to stop the vehicle from accelerating. Two types of brakes are used. Water brake uses an impeller and water flow to provide a braking force against the drum. This is the choice for higher horsepower vehicles as it is hard to control the water flow needed for lower power braking. Eddy Current retarders are used in lower horsepower applications and work by current inducing a magnetic field to slow rotation (eddy current retarders are not my specialty so if I have that wrong understand that the principle is to provide a braking force).
Measurement is taken by a load cell. The load cell measures force from the rollers trying to slow down the tires. It measures this torque and calculates horsepower. Load cells need to be calibrated periodically but newer styles have a much greater ability to maintain calibration. My rule of thumb was to calibrate based on what info I could afford to lose. I calibrated weekly. Nothing is worse than testing and finding an anomaly and finding out that the load cell had drifted.
The twin roller has a problem in that by placing the tire between the two rollers it wants to climb the front roller. Standard practice is to tie down the vehicle in a vertical direction to place as much force as possible to combat this tendency. This places a greater load on the vehicle and can vary the power measurements considerably. It can also lead to premature tire failure as you have created two stress points at two small contact patches on the tire. This forces the internal belts to conform to the smaller radius and puts stresses in the tire carcass that it was never meant to comply with. Advances have been made in this arena but I have no practical experience in this area so I will leave that to those more knowledgeable than me.
The single roller dyno uses 2 drums from 24 to 52 inches in diameter. These sizes were designed to give the tire a greater contact patch and more realistically recreate road conditions. The larger size also meant that more mass could be placed in the drum to better simulate the inertial load on the vehicle.
The only brake on an inertia single roller dyno is to slow the drums down. It is not used for any calculation in the inertia mode. The instrumentation consists of a hall effect sensor used to read number of rotations (usually 1 or 2 tab wheels) and accurately measured circumference and mass as well as bearing drag numbers.
Dynojet uses a single shaft and Superflow uses a differential. Superflow accounts for the drag on the differential and Dynojet accounts for the drag of the bearings. This is the main difference in the physical makeup of the two dynos.( I have no working knowledge of the Mustang dynos or any other chassis inertia dyno.)
Both Dynojet and Superflow have capabilities for load testing using eddy current retarders. Since the drums and inertia are higher they make a water brake difficult to use.
In order to calculate a correction factor a weather station is needed for calculation. This is only for corrected numbers and not needed to get raw numbers.
Third part:
Dyno operation:
To get the most out of a dyno you need to control your variables. For inertia testing this means:
Make sure your battery is charged. Make sure you have the right air pressure. Start the runs at the same coolant and oil temp if possible. Use the same gear in the transmission.
When I test I follow this guideline:
I remove the floor mat. I get the car to a coolant temperature I can return to consistently. Coolant will cool down quicker than oil in most cases. I start the test at an RPM that won’t bog the car down but will give me the greatest range. I hit start, count to 1 and floor the accelerator. When I hit my top RPM I back off the throttle and end the test. I repeat this process two more times and let the car cool down while I look at the data.
I compare the three runs. Which is higher and why? Are there any anomalies in the curve? Is there anything that would point to a problem with the run?
From this point I have established a baseline of the vehicle. If I am testing something I will make the change and repeat the process. I will compare the same run in each series to get the real gain. If I have 2 groups of 3 runs I will compare run 2 against run 5. This helps average out some of the differences caused by temperature during the runs.
Using this model I have tested the same vehicle in Las Vegas and South Carolina and my results have repeated to within 3 horsepower and the graphs have overlaid. I made sure that all the variables I could control were controlled. No new tires, no gear ratio change, same fueling and timing, ect…
If you test your car and then add 22’s, a Predator and a different exhaust, and then try to compare your new CAI against your old runs you will fail as you have changed your variables.
You cannot measure the effect of your changes if you make more than one change. Vehicles are a combination of all the parts they are made up of and you have no idea what effect the changes have in concert with other changes.
Fourth part:
Common Dyno Questions
Why don’t I have a Torque reading? Why does my graph have funny squiggles on it?
You need an RPM signal to get a torque reading. If you do not have an RPM signal the dyno can not calculate torque. Problems in this area could be a bad signal, wrong wire or an operator who doesnt want to set up the indicator.
Why is there a gap in the torque reading?
You will see a gap if the pick up looses signal or there is slippage between the drums and the tires. The first reason could be a high energy ignition or a dirty signal. The second scenario leads into the squiggles.
Why does my graph have funny squiggles on it?
If you have slippage the curve will change because your rate of acceleration changed. If you have tire or converter slip you will see a spike or a backwards squiggle. This is the actual acceleration curve, positive and negative. This is very prevalent in automatics where the converter slips between gears. You also see it when shifting a stick car on the dyno. The anomaly is caused by rpm going up while acceleration stays the same and this causes an error in the torque calculation. Torque = Hp x RPM/ 5252. With no slippage your RPM goes up in a ratio to acceleration. This is the dyno calculated gear ratio. If you slip on the dyno either at the tires or in the transmission you will have an ever changing ratio. Since the ratio changes so does the calculation and this results in a funky graph. A spike is usually caused by a hard shift changing your rate of acceleration for a second. This is not a true number. I have had spikes over 600 hp on a 200 hp car by shifting hard. The only way to combat this type of graph issue is to lock the vehicle in one gear.
Will different gears show different horsepower?
Yes, for a couple of reasons. If you dyno in a gear below your 1 to 1 gear in the tranny you will see a lower number. Some of your power is absorbed by the inertia of the components. You are also changing your rate of acceleration. Your highest efficiency is in the 1 to 1 gear ratio. Overdrive also has this issue.
What about rear gears?
Same thing. You will see a change. The reason is more complicated than it seems. When you change rear gears you have new fluid, new gears, new bearings, shims, etc… All of theses can affect your power as they need to bed in. Using broken in rear gears on a ford 9” with new fluid and monitoring temps I have seen a change of about 25 hp going from ratio’s spanning 3.00 to 6.00. So why did you lose 25 hp from going to a 3.23 from a 2.82? Break in. Once you put some miles on that gear set you will get most of your power back. But since you changed your rate of acceleration you will have some losses as you have changed your time component in the equation.
What is smoothing?
This has caused some problems for people and is easily explained. When you measure a data point you do so over time. This is called the sampling rate. The greater the rate the more data you get. Since there is a space between data points you have no value in between. When you graph this you are looking at a line connecting two points. Since there are areas where the data has a high difference the graph is spikey. Smoothing interpolates the distance between two points to smooth out the line between two points. I usually set the smoothing at 5 so I don’t have a spike give me a false hp or torque number. I am interested in the whole graph, not one point.
I dynoed on one dyno and then on another and the numbers were different! Why?
Well, here is the crux of the problem with dynos. You need to make sure that nothing has changed. Did you dyno hot at one and cold at another? Was the air pressure in the tires the same? What other changes did you make? If you are sure you have made no other changes then what could happen with the dyno? Here are some things an unscrupulous dyno operator could do to show higher power:
• Temp probe placed in an area of higher temp than ambient.
Correction factor will be higher. Look at your cluster temp before dynoing and compare to your reading.
• Run took place in a lower gear.
Specify what gear you want the run to take place in. If you have older runs that seem low, check the gear ratio to the new runs. They should match within a couple of tenths.
• Not going full throttle.
This should be obvious but the curve will dip down as the acceleration slows. A big culprit here may not be the operator but your floor mats! I dynoed a Viper with heads and it made about 450hp. When I took out the mats and tried it again it made 505hp! The mat wouldn't let full throttle happen and since the pedal was floored the curve was good, just lower.
• Place the dyno weather station in a hyperbaric chamber and raise the pressure.
Unlikely. If the weather station is in a separate room, ask why. If you have a programmer, read the pressure and compare it. Or find another operator.
• Dragging the dyno brakes.
Since this will result in lower power this is also unlikely. On the dynojet they use train brakes to stop the dyno. If the operator has proportional air control to use for loading the brakes will slow the drum based on speed or rpm. Unless they are looking to screw you it is possible to have this happen but I have never seen it and I have dealt with many issues related to dyno operation.
The best way I know of to validate a change is to look at run time. If you have a run on two different dynos and they are different look at the run time. If the times are identical you should look at the correction factor for your answer. If they are not then the change was with the vehicle.
One other thing to look at is fans. Some dyno owners use them and some don’t. Some use carpet dryer fans and some use large blade fans. In NASCAR most shops duct air directly to the intake of the carb.
Fans can change your readings from dyno to dyno depending on how much air they flow. If you need to compare dyno to dyno, try asking the operator to turn off the fan. Your second run in the series will probably show a lower number but you will be more repeatable and a valid comparison could be made if the temp swing was less that 20 degrees. Again, we want to control our variables as much as possible for repeatable results.
What is correction factor? Which one should I use?
Correction factor is used to even the playing field when testing at different locations and conditions. This is similar to Density altitude. We all know that higher atmospheric pressure means more oxygen. So does lower heat and humidity. If we agree on a standard temp, pressure and humidity value then we can compare runs at different conditions. You multiply the data by the correction factor to get corrected results.
The two I use are STD (or STP) and SAE. The most common one in use on an inertia chassis dyno is SAE. Both correct to different values so you can not compare one run at one factor to another with a different factor. STD (or STP) will always show more correction because it assumes standard to be 29.92 inhg at 60 degrees. Most conditions are worse than these values so you will see a greater multiplication factor and therefore higher corrected numbers. SAE corrects to 29.234 inhg and 77 degrees so the values are lower and the difference is less. STD (or STP) was and is still used by most engine dynos.
What inputs are needed to test my car? If the operator puts in a bad value will my numbers change?
Well, no. Most software inputs the conditions from a measured value. The dyno weather station reads temp, absolute pressure and humidity and automatically puts this data in the software. A note here; absolute pressure is what pressure you have uncorrected for altitude. If you listen to the news you could see 29.92 inhg in Denver. How? It is corrected for altitude. Your dyno absolute pressure would be around 29.10. The gear ratio is calculated by RPM vs. Drum Speed and there is no way to change the mass of the drum so nothing is needed to be input which greatly reduces the issue of wrong information. Older dyno systems had to have humidity manually input but even if you were at 100% humidity and the input was 0% you would only change the overall correction factor by 7%.
If you get into load dynos then the answer could be yes depending on what information they are using to simulate load. If you try to estimate coefficient of drag, frontal area, and or weight and this is an input in the software you can change the values. Since the dyno is not the street I don't concern myself with an accurate number for "real world" conditions for testing. I can't control the "real world" so why try to measure it? I need repeatability first. I don't even need it for tuning as I will hit all the tables on the map anyway. But if it makes you feel good, go for it.
Can I run on a Dynojet and then run on a Superflow and hope to compare the results in a meaningful way?
Yes and no. Both manufacturers use different calculations proprietary to their respective software. There are corrections for this but you are adding variables. If you have no choice you may have to adjust your expectations.
What is the percentage of loss from the flywheel?
I have no idea. I have seen anywhere from 10% to 15% for manuals and 15% to 25% for automatics. You can not compare numbers generated on a load engine dyno to an inertia chassis dyno. They measure different things. There is no way to generalize the losses between trannys or the efficiencies of the engines. They are different. When a manufacturer gives a hp rating it is an average generated on an engine dyno. Since it is an average you have no way of knowing where you fall on that scale. Are you a factory freak at 360 hp or a Friday car at 325 hp? Does the factory use an efficiency of 100%? Why compare? You will never get a real answer and that conjecture can lead to wrong assumptions. Many people have run their vehicle and come up with a percentage that they are comfortable with. It is only accurate to their car. Add a modified tranny or rear gears and you are now in unfamiliar territory again.
My car only dynos X but at the track I turn Y’s. Why?
Don’t know! There are many variables at the track. Different drivers can make wildly different results. If I run a 13.9 in my car and someone else runs a 12.9 in the same car did my power change? Or do I just suck? I know it is the latter. Dynos simply do not have all the variables the track has. If the track was the “be all end all” then we would all run the same times with the same equipment. That is not the case. I have nothing against track testing as long as it is not compared to dyno testing. The two are completely different and there is no real correlation due to the variables involved. Compare track data to track data and dyno data to dyno data.
The idea behind dyno testing is to quantify your changes in a repeatable manor. That means you need to minimize as many variables as possible. That is very hard to do at the racetrack. If I gain 10hp on the dyno I will have 10hp more at the track. But the reverse is not true. You may cut a better light but that didn’t give you more horsepower. Neither did that change you made to your shift point which let you carry more RPM over the line. But your times will change.
In conclusion (finally!):
This is my experience with dyno testing. Please feel free to post any questions you may have and if I can’t answer them I will let you know. I haven’t seen it all by a long shot and learn every day. Math is not my strong suit so if you see an equation that is wrong, post up and I will change it.
If you want to post up your graphs, feel free. But the only way to truly compare results is if you have the dyno files. Superflow and Dynojet both have free dyno run viewers on their websites. When you dyno your car, ask for the run files. Then you can compare your runs. I can’t look at a graph and tell you why you have two different numbers from two dyno’s. But if you have both run files I can look at your run time and see if that changed. If you change your rate of acceleration you change the work (power) measured. It’s as simple as that.
Happy testing,
Sean
Hope this is of some help to you guys. I didn't want to be one of those people that took from the knowledge here without giving back.
I have no problems looking at graphs but my understanding of the programming of these cars is very limited right now. If you have run files from the dyno we can look at what may or may not be happening. This again is only good for Superflow or Dynojet. I have no working knowledge of Mustang or any of the twin roller or wheel attached dynos.
http://www.lxforums.com/board/showth...ting-Explained
I have included it here in its original form:
I have seen a lot of posts from members unsatisfied with dyno numbers and dynos in general so I thought I would write up my experiences with them and what I have seen.
Some background first. I was involved very early with chassis dynos for motorcycles and automobiles. I have well over 10,000 runs over 10 years of running. I have dynoed up to 60 cars in one day at events.
I also have experience with engine dynos. I have been involved with several successful racing endeavors in NASCAR, BAJA, Road Racing, Bonneville and NHRA.
I was directly involved in dyno testing for all three NASCAR divisions.
I no longer am involved in any way, shape or form in vehicle testing, racing or tuning. I do not represent any manufacturer or sanctioning body of any kind.
This is not a “Why dyno’s are better than the Track” thread. If you are not going to look at this with open eyes then you might want to go to the track section of the forums because nothing here I say will convince you, nor will I try. All I will do is explain what I know and have observed firsthand from my experience. I don’t want this to become a “Track vs. Dyno” thread and will not entertain that direction.
That out of the way, I will explain what I know and what my experiences have been.
This post will be broken up into 4 parts:
Theory, Dyno construction, Dyno practices and Common Dyno questions.
Theory:
There are two main ways to measure horsepower or torque. Depending on which one you measure, the other is calculated from the first.
Load has been used since the very first for measuring torque. If you apply a load to an engine and stop it from accelerating you can measure thetorque available required to hold speed (or RPM) at that point. This is how engine dynos do it. They hold the engine at a specific point and read a load cell. Using the formula: HP=TQ*RPM/5252 you arrive at a horsepower number; horsepower is not actually measured… its calculated.
Inertia testing and the need for Chassis Dyno testing came later. Some more background as I know it:
The first commercial inertia chassis dyno that I know of was created for the purpose of tuning motorcycles, specifically vacuum slide carburetors. Since they are dependant on rate of acceleration and work off of vacuum they proved to be very hard to tune under steady state load.
Street testing was very difficult due to the speeds involved and ever increasing police presence. The inertia dyno idea was born. If you take a bike out and time a run in the 1 to 1 gear you can put it on a rotating drum and add mass until you get the same run time. If you can accurately measure time, mass of the drum and its circumference you have all the components needed to measure horsepower.
Force = Mass x Acceleration
If you know the mass, time of the run and change in acceleration you can measure horsepower. But F=MxA is not the end of the equation. Force has to be converted into HP by the equation HP= Force x Distance/ Time then divide by 550 (hp is the force required to move 550 lbs 1 foot in 1 second or 330,000 feet per minute).If you plot this against time you have a power curve.
A requirement to do this type of testing is accurate measurement of time. This is where the computer comes in. Inertia was only a theory until we had a way to measure time accurately.
So if you know the distance the drum travels in 1 revolution (circumference), the mass equivalent being rotated (since the drum rotates the mass rotated is not the mass static) and time you have all the parts to measure horsepower.
Notice that RPM and torque are not mentioned. You do not need either to measure horsepower. But to calculate torque you need RPM. An inductive pick up is used to measure RPM. From there you enter RPM into the equation and you get torque. Torque is part of the process as the force on the dyno drum is torque but it is not engine torque but wheel torque. Wheel torque in first gear can be in the thousands of pounds at the wheels due to gear ratio. This torque is measured indirectly as work in the equation work=force x distance.
Because we are measuring acceleration we don't need a brake to hold a steady RPM or speed. We just "Let 'er eat" or go full throttle. As long as there are no physical restrictions in the linkage this is a very repeatable state.
That’s it. No other measurements are needed for inertia. But how do we compare results in Las Vegas to results in Indy or North Carolina? Enter the correction factor.
The correction factor uses a standardized set of conditions to compare a run at different conditions.
If you look at the standards we are concerned with they are:
SAE= 77 degrees 29.234 inhg 0% humidity
STD or STP = 60 degrees 29.92 inhg 0% humidity
If you make a run under those conditions you will see a correction factor of 1.00. This is a multiplier. As conditions improve or degrade you will see the multiplier add or subtract power from all your numbers.
This is what makes comparing vehicles in varying conditions possible. This is the same as Density Altitude correction that drag racers use for the most part.
So to summarize, if you can time the run, know the mass equivalent and circumference of the drum then you can measure horsepower. Simplicity usually equals repeatability.
A quick aside about repeatability vs. accuracy:
The experiment theory is basically measuring your variable against your control. The more you control, the more accurate your results. But as you get to smaller and smaller measurements it gets harder to repeat your results.
If I measure a bore and get 4.065 and then measure it again I may get 4.067. If I measure only to inches the results repeat at 4. But is that good enough? No engine builder will agree. But if my measuring equipment is a pair of calipers and not a bore mic what will my results be?
In our case the vehicle we are testing is the variable. When tuning a carbureted vehicle you can get 1 to 3 hp repeatability because there is little adjustment potential. Timing curve is fixed as well as fueling.
Enter the ECU or PCM in the LX world. Now timing is variable as well as fueling. Depending on a myriad of variables you will see power change to best support what the ECU wants to accomplish. This is not always compatible with the highest horsepower! Now you have a whole new set of changes to your test subject. This means that 1 to 3 horsepower is the range you may see from run to run if not more as the computer adjusts and your load value changes due to temperature and other variables.
The higher degree of accuracy the less likely we can control all the variables and the less likely we are to repeat to a consistent number.
Just some things that can change a reading are: What gear you complete the run in, engine temp, trans temp, rear diff temp, atmospheric conditions, heat soak, state of tune, voltage, tire size, tire pressure, wheel weight…
The more things we control the more repeatable our numbers. Trying to repeat to 1 to 3 horsepower is unlikely due to the sheer number of variables. But if we can repeat within 3 to 5 horsepower we can measure a change above that number. More on this subject in Dyno Practices.
Load vs. Inertia
Why inertia? Because there are fewer variables and they are simpler to operate. They also reduce set up time and that means greater efficiency in testing. More cars in a day = More money for the shop. And not only that, it’s far less stressful on the car being run; this becomes especially important as the dyno pull count for a particular car increases.
Inertia needs fewer instruments to measure and they are not subject to calibration. You only need Mass, Circumference, a way of measuring time and a way of counting drum rotation.
There are no control loops to try and hold a vehicle at a steady state. This is a difficult thing to do as you must stabilize the RPM or speed before you can measure it.
Since both styles measure different values under different test conditions the results are not comparable. I have spent many years trying to make an engine dyno correlate to what the chassis dyno showed. I even wrote acceleration test profiles for the engine dyno that simulate what I saw on a chassis dyno run. I never had a match or even a ball park. Comparing Dynamic motion to Static motion just doesn’t compare easily.
Since inertia has fewer variables you will see higher repeatability. No one I know will say an engine dyno repeats within a couple horsepower without an extremely skilled operator. Since all you are doing on a chassis dyno is driving the car there is less need for the skills required of an engine dyno operator. You still need skills and I will cover that later.
Where load excels is in tuning fuel injection. Most EFI systems use tables for the ecu to look at for what fuel and timing values are needed based on load and vehicle conditions. You hold the vehicle at a preset rpm and adjust a/f and timing until you show the greatest torque output. Lather, rinse repeat and you have tuned a fuel and timing map. (This is greatly simplified but should give you the idea).
Inertia only hits so many places on the map so you will not be able to get to those other values. Since the drums only model a set amount of rotational mass you may end up having a lighter load on the vehicle. If you set timing to this load you will find quite a bit of detonation on the street when real world conditions are added in.
The best dyno is one that has both worlds. Inertia for overall testing of the finished product and Load to get the maps right. You still only need inertia for test comparisons but you need load for tuning.
Second Part:
Dyno Construction:
There are 2 types of chassis dynos:
Twin roller and single roller chassis dynos.
The twin roller chassis dyno was the original. Clayton is the first brand that comes to mind. In use the vehicle is driven over the rollers and the tire sits between the front and rear roller. It is essentially “pinched” between the rollers. The vehicle is run up to a certain RPM and a brake is applied to the rollers to stop the vehicle from accelerating. Two types of brakes are used. Water brake uses an impeller and water flow to provide a braking force against the drum. This is the choice for higher horsepower vehicles as it is hard to control the water flow needed for lower power braking. Eddy Current retarders are used in lower horsepower applications and work by current inducing a magnetic field to slow rotation (eddy current retarders are not my specialty so if I have that wrong understand that the principle is to provide a braking force).
Measurement is taken by a load cell. The load cell measures force from the rollers trying to slow down the tires. It measures this torque and calculates horsepower. Load cells need to be calibrated periodically but newer styles have a much greater ability to maintain calibration. My rule of thumb was to calibrate based on what info I could afford to lose. I calibrated weekly. Nothing is worse than testing and finding an anomaly and finding out that the load cell had drifted.
The twin roller has a problem in that by placing the tire between the two rollers it wants to climb the front roller. Standard practice is to tie down the vehicle in a vertical direction to place as much force as possible to combat this tendency. This places a greater load on the vehicle and can vary the power measurements considerably. It can also lead to premature tire failure as you have created two stress points at two small contact patches on the tire. This forces the internal belts to conform to the smaller radius and puts stresses in the tire carcass that it was never meant to comply with. Advances have been made in this arena but I have no practical experience in this area so I will leave that to those more knowledgeable than me.
The single roller dyno uses 2 drums from 24 to 52 inches in diameter. These sizes were designed to give the tire a greater contact patch and more realistically recreate road conditions. The larger size also meant that more mass could be placed in the drum to better simulate the inertial load on the vehicle.
The only brake on an inertia single roller dyno is to slow the drums down. It is not used for any calculation in the inertia mode. The instrumentation consists of a hall effect sensor used to read number of rotations (usually 1 or 2 tab wheels) and accurately measured circumference and mass as well as bearing drag numbers.
Dynojet uses a single shaft and Superflow uses a differential. Superflow accounts for the drag on the differential and Dynojet accounts for the drag of the bearings. This is the main difference in the physical makeup of the two dynos.( I have no working knowledge of the Mustang dynos or any other chassis inertia dyno.)
Both Dynojet and Superflow have capabilities for load testing using eddy current retarders. Since the drums and inertia are higher they make a water brake difficult to use.
In order to calculate a correction factor a weather station is needed for calculation. This is only for corrected numbers and not needed to get raw numbers.
Third part:
Dyno operation:
To get the most out of a dyno you need to control your variables. For inertia testing this means:
Make sure your battery is charged. Make sure you have the right air pressure. Start the runs at the same coolant and oil temp if possible. Use the same gear in the transmission.
When I test I follow this guideline:
I remove the floor mat. I get the car to a coolant temperature I can return to consistently. Coolant will cool down quicker than oil in most cases. I start the test at an RPM that won’t bog the car down but will give me the greatest range. I hit start, count to 1 and floor the accelerator. When I hit my top RPM I back off the throttle and end the test. I repeat this process two more times and let the car cool down while I look at the data.
I compare the three runs. Which is higher and why? Are there any anomalies in the curve? Is there anything that would point to a problem with the run?
From this point I have established a baseline of the vehicle. If I am testing something I will make the change and repeat the process. I will compare the same run in each series to get the real gain. If I have 2 groups of 3 runs I will compare run 2 against run 5. This helps average out some of the differences caused by temperature during the runs.
Using this model I have tested the same vehicle in Las Vegas and South Carolina and my results have repeated to within 3 horsepower and the graphs have overlaid. I made sure that all the variables I could control were controlled. No new tires, no gear ratio change, same fueling and timing, ect…
If you test your car and then add 22’s, a Predator and a different exhaust, and then try to compare your new CAI against your old runs you will fail as you have changed your variables.
You cannot measure the effect of your changes if you make more than one change. Vehicles are a combination of all the parts they are made up of and you have no idea what effect the changes have in concert with other changes.
Fourth part:
Common Dyno Questions
Why don’t I have a Torque reading? Why does my graph have funny squiggles on it?
You need an RPM signal to get a torque reading. If you do not have an RPM signal the dyno can not calculate torque. Problems in this area could be a bad signal, wrong wire or an operator who doesnt want to set up the indicator.
Why is there a gap in the torque reading?
You will see a gap if the pick up looses signal or there is slippage between the drums and the tires. The first reason could be a high energy ignition or a dirty signal. The second scenario leads into the squiggles.
Why does my graph have funny squiggles on it?
If you have slippage the curve will change because your rate of acceleration changed. If you have tire or converter slip you will see a spike or a backwards squiggle. This is the actual acceleration curve, positive and negative. This is very prevalent in automatics where the converter slips between gears. You also see it when shifting a stick car on the dyno. The anomaly is caused by rpm going up while acceleration stays the same and this causes an error in the torque calculation. Torque = Hp x RPM/ 5252. With no slippage your RPM goes up in a ratio to acceleration. This is the dyno calculated gear ratio. If you slip on the dyno either at the tires or in the transmission you will have an ever changing ratio. Since the ratio changes so does the calculation and this results in a funky graph. A spike is usually caused by a hard shift changing your rate of acceleration for a second. This is not a true number. I have had spikes over 600 hp on a 200 hp car by shifting hard. The only way to combat this type of graph issue is to lock the vehicle in one gear.
Will different gears show different horsepower?
Yes, for a couple of reasons. If you dyno in a gear below your 1 to 1 gear in the tranny you will see a lower number. Some of your power is absorbed by the inertia of the components. You are also changing your rate of acceleration. Your highest efficiency is in the 1 to 1 gear ratio. Overdrive also has this issue.
What about rear gears?
Same thing. You will see a change. The reason is more complicated than it seems. When you change rear gears you have new fluid, new gears, new bearings, shims, etc… All of theses can affect your power as they need to bed in. Using broken in rear gears on a ford 9” with new fluid and monitoring temps I have seen a change of about 25 hp going from ratio’s spanning 3.00 to 6.00. So why did you lose 25 hp from going to a 3.23 from a 2.82? Break in. Once you put some miles on that gear set you will get most of your power back. But since you changed your rate of acceleration you will have some losses as you have changed your time component in the equation.
What is smoothing?
This has caused some problems for people and is easily explained. When you measure a data point you do so over time. This is called the sampling rate. The greater the rate the more data you get. Since there is a space between data points you have no value in between. When you graph this you are looking at a line connecting two points. Since there are areas where the data has a high difference the graph is spikey. Smoothing interpolates the distance between two points to smooth out the line between two points. I usually set the smoothing at 5 so I don’t have a spike give me a false hp or torque number. I am interested in the whole graph, not one point.
I dynoed on one dyno and then on another and the numbers were different! Why?
Well, here is the crux of the problem with dynos. You need to make sure that nothing has changed. Did you dyno hot at one and cold at another? Was the air pressure in the tires the same? What other changes did you make? If you are sure you have made no other changes then what could happen with the dyno? Here are some things an unscrupulous dyno operator could do to show higher power:
• Temp probe placed in an area of higher temp than ambient.
Correction factor will be higher. Look at your cluster temp before dynoing and compare to your reading.
• Run took place in a lower gear.
Specify what gear you want the run to take place in. If you have older runs that seem low, check the gear ratio to the new runs. They should match within a couple of tenths.
• Not going full throttle.
This should be obvious but the curve will dip down as the acceleration slows. A big culprit here may not be the operator but your floor mats! I dynoed a Viper with heads and it made about 450hp. When I took out the mats and tried it again it made 505hp! The mat wouldn't let full throttle happen and since the pedal was floored the curve was good, just lower.
• Place the dyno weather station in a hyperbaric chamber and raise the pressure.
Unlikely. If the weather station is in a separate room, ask why. If you have a programmer, read the pressure and compare it. Or find another operator.
• Dragging the dyno brakes.
Since this will result in lower power this is also unlikely. On the dynojet they use train brakes to stop the dyno. If the operator has proportional air control to use for loading the brakes will slow the drum based on speed or rpm. Unless they are looking to screw you it is possible to have this happen but I have never seen it and I have dealt with many issues related to dyno operation.
The best way I know of to validate a change is to look at run time. If you have a run on two different dynos and they are different look at the run time. If the times are identical you should look at the correction factor for your answer. If they are not then the change was with the vehicle.
One other thing to look at is fans. Some dyno owners use them and some don’t. Some use carpet dryer fans and some use large blade fans. In NASCAR most shops duct air directly to the intake of the carb.
Fans can change your readings from dyno to dyno depending on how much air they flow. If you need to compare dyno to dyno, try asking the operator to turn off the fan. Your second run in the series will probably show a lower number but you will be more repeatable and a valid comparison could be made if the temp swing was less that 20 degrees. Again, we want to control our variables as much as possible for repeatable results.
What is correction factor? Which one should I use?
Correction factor is used to even the playing field when testing at different locations and conditions. This is similar to Density altitude. We all know that higher atmospheric pressure means more oxygen. So does lower heat and humidity. If we agree on a standard temp, pressure and humidity value then we can compare runs at different conditions. You multiply the data by the correction factor to get corrected results.
The two I use are STD (or STP) and SAE. The most common one in use on an inertia chassis dyno is SAE. Both correct to different values so you can not compare one run at one factor to another with a different factor. STD (or STP) will always show more correction because it assumes standard to be 29.92 inhg at 60 degrees. Most conditions are worse than these values so you will see a greater multiplication factor and therefore higher corrected numbers. SAE corrects to 29.234 inhg and 77 degrees so the values are lower and the difference is less. STD (or STP) was and is still used by most engine dynos.
What inputs are needed to test my car? If the operator puts in a bad value will my numbers change?
Well, no. Most software inputs the conditions from a measured value. The dyno weather station reads temp, absolute pressure and humidity and automatically puts this data in the software. A note here; absolute pressure is what pressure you have uncorrected for altitude. If you listen to the news you could see 29.92 inhg in Denver. How? It is corrected for altitude. Your dyno absolute pressure would be around 29.10. The gear ratio is calculated by RPM vs. Drum Speed and there is no way to change the mass of the drum so nothing is needed to be input which greatly reduces the issue of wrong information. Older dyno systems had to have humidity manually input but even if you were at 100% humidity and the input was 0% you would only change the overall correction factor by 7%.
If you get into load dynos then the answer could be yes depending on what information they are using to simulate load. If you try to estimate coefficient of drag, frontal area, and or weight and this is an input in the software you can change the values. Since the dyno is not the street I don't concern myself with an accurate number for "real world" conditions for testing. I can't control the "real world" so why try to measure it? I need repeatability first. I don't even need it for tuning as I will hit all the tables on the map anyway. But if it makes you feel good, go for it.
Can I run on a Dynojet and then run on a Superflow and hope to compare the results in a meaningful way?
Yes and no. Both manufacturers use different calculations proprietary to their respective software. There are corrections for this but you are adding variables. If you have no choice you may have to adjust your expectations.
What is the percentage of loss from the flywheel?
I have no idea. I have seen anywhere from 10% to 15% for manuals and 15% to 25% for automatics. You can not compare numbers generated on a load engine dyno to an inertia chassis dyno. They measure different things. There is no way to generalize the losses between trannys or the efficiencies of the engines. They are different. When a manufacturer gives a hp rating it is an average generated on an engine dyno. Since it is an average you have no way of knowing where you fall on that scale. Are you a factory freak at 360 hp or a Friday car at 325 hp? Does the factory use an efficiency of 100%? Why compare? You will never get a real answer and that conjecture can lead to wrong assumptions. Many people have run their vehicle and come up with a percentage that they are comfortable with. It is only accurate to their car. Add a modified tranny or rear gears and you are now in unfamiliar territory again.
My car only dynos X but at the track I turn Y’s. Why?
Don’t know! There are many variables at the track. Different drivers can make wildly different results. If I run a 13.9 in my car and someone else runs a 12.9 in the same car did my power change? Or do I just suck? I know it is the latter. Dynos simply do not have all the variables the track has. If the track was the “be all end all” then we would all run the same times with the same equipment. That is not the case. I have nothing against track testing as long as it is not compared to dyno testing. The two are completely different and there is no real correlation due to the variables involved. Compare track data to track data and dyno data to dyno data.
The idea behind dyno testing is to quantify your changes in a repeatable manor. That means you need to minimize as many variables as possible. That is very hard to do at the racetrack. If I gain 10hp on the dyno I will have 10hp more at the track. But the reverse is not true. You may cut a better light but that didn’t give you more horsepower. Neither did that change you made to your shift point which let you carry more RPM over the line. But your times will change.
In conclusion (finally!):
This is my experience with dyno testing. Please feel free to post any questions you may have and if I can’t answer them I will let you know. I haven’t seen it all by a long shot and learn every day. Math is not my strong suit so if you see an equation that is wrong, post up and I will change it.
If you want to post up your graphs, feel free. But the only way to truly compare results is if you have the dyno files. Superflow and Dynojet both have free dyno run viewers on their websites. When you dyno your car, ask for the run files. Then you can compare your runs. I can’t look at a graph and tell you why you have two different numbers from two dyno’s. But if you have both run files I can look at your run time and see if that changed. If you change your rate of acceleration you change the work (power) measured. It’s as simple as that.
Happy testing,
Sean
Hope this is of some help to you guys. I didn't want to be one of those people that took from the knowledge here without giving back.
I have no problems looking at graphs but my understanding of the programming of these cars is very limited right now. If you have run files from the dyno we can look at what may or may not be happening. This again is only good for Superflow or Dynojet. I have no working knowledge of Mustang or any of the twin roller or wheel attached dynos.
The following 2 users liked this post by seanol:
BoostedAero (07-25-2016),
sexyxe (07-26-2016)
#3
Member
Thread Starter
Testing is all about consistency. I went down to testing specific gravity of the fuel every change.
Sean