PART 4 - Data Collection
Story and Photos by John Copeland
Over the past few months we've talked about different kinds of dynos and how they work. We've looked at the most commonly used fluid-brake systems, and at the highly sophisticated electronic dynos. Before we go on to the subject of data collection, several of you have called or written to point out an error in the story about hydraulic dynos. (I swear, sometimes I think I just slip these goofs in just to see if anybody's reading this stuff!) Anyway, when we were talking about transposing the pressure reading from the output of a hydraulic pump into foot/pounds of torque, I incorrectly gave the impression that a reading of 50 psi might relate, on a given pump, to 10 foot/pounds of torque. Sorry. I just dropped a zero or two. An input of 10 foot/pounds, at a pump-shaft speed of 2500 RPM, is more likely to generate pressure in the neighborhood of 2000 psi! Didn't mean to cause any confusion. And to any of you who might have hooked a 0 to 50 psi gauge onto the output of a hydraulic pump...... Sorry 'bout that.
Anyway, on to this month's topic, data collection. Regardless of what kind of dyno you're working with, you've got to have a good method to gathering the information it has to give you. For openers, let's look at what different kinds of information might be available. The most basic piece of information you're going to get from a dyno is torque. You'll remember that torque is the twisting force that your engine generates on the crankshaft. Whether you measure it with a spring scale, a pressure gauge, or an electronic strain gauge (a load cell) doesn't matter. Torque is the starting point for all your dyno work. Of course, torque doesn't mean anything without knowing the relative RPM at which that torque was measured. And it follows that, using the formula HP = Torque X RPM/5252.1, you can determine horsepower. A word of caution here; knowledgeable dyno users pay a lot more attention to Torque than they do to Horsepower. Torque is what drives your kart. It is what determines how fast you can accelerate, and for how long. Because horsepower is a function of RPM as well as torque, it's easy to fall into the trap of thinking that turning more RPMs will make you go faster. In practice, it just ain't so. A wise old engine builder once told me "You can get 20 HP out of a sewing machine motor, but you'll have to turn it a couple of hundred thousand RPM!" Torque is what matters most. Ever wonder why a 100cc Yamaha, with about 2 ½ times as much horsepower as a Briggs, doesn't go 2 ½ times faster? Take a look at the relative torque of the two engines.
So you'll need to measure torque and RPM. The design of your dyno will dictate how the torque is measured. RPM readings can be handled by any conventional tach commonly used in karting. A mechanical tach driven off the crankshaft will work too, but beware of trying to adapt an electronic tach unit from automotive use. It may not have the necessary electronic shielding to adequately deal with the RF interference from the ignition system. RF stands for "Radio Frequency" and it's the gremlin that is the biggest pain in the butt in dyno data collection. You see, every time your coil fires the sparkplug, a magnetic field builds around the plug wire, then collapses. There is also an electro-magnetic "shock wave" that emanates from the spark. Think of it as a tiny sonic boom. Even though the spark is inside the engine, the electro-magnetic wave goes right through the metal of your engine and into the air. This radiation is called "Radio Frequency" or RF interference and it's a real bugger of a problem around electronic devices. It's the same noise that you hear on the radio in an old car if the plug wires are going bad. Tach and Page 2 temp gauges intended for use in karting (and other engines without shielded ignition systems) have simple but effective systems to shield the electronics from this RF "noise". But a conventional automotive tach unit may not. Stick with what works. In addition to measuring torque and RPM, you'll find it very useful to measure engine temperature. You already know how critical engine temp is on the track. If you're going to simulate track conditions, you'll need to be able to measure and control engine temp on the dyno. Cylinder head temp is by far the most commonly used temp measurement in karting, and you'll certainly want that. But I'm a big believer in Exhaust Gas Temperature as a better measurement. EGT reads faster, gives a truer picture of what's going on in the combustion chamber, and is unaffected by outside air temperature. In an upcoming article we'll take a closer look at EGT and why every serious karter should be using it. On the dyno, it's more than just a plus, it's almost a necessity, particularly if you're dynoing other people's engines. The faster response of EGT gives the operator a better chance to run an engine to it's maximum without burning it up. And in a commercial application like ours, torching a customer's engine on the dyno is never popular. Cylinder head temp measurement devices are very susceptible to RF interference so again, your best bet is to stick with commercially made units build for karting. Exhaust gas temp measurement is largely immune from this problem, so most any thermocouple with the correct range (800-1500 degrees F) will do. But then, the karting stuff works just fine, so why reinvent it?
With these basic measurements covered, you can go to work. But, depending on your level of interest and sophistication, you may also be interested in intake airflow, fuel flow rates, or any of dozens of other measurements. You'll also find it very helpful to keep track of air temperature, humidity, and barometric pressure around the dyno. These factors have a major influence on the output of the engine and, if you want to be able to compare dyno runs made on different days and under different conditions, you'll need to record those conditions. Just remember, anything electrical or electronic that you use around the dyno will have to be properly shielded to prevent RF noise from messing up your readings. More about this later.
Now let's look at a basic dyno run and how you can get the data you need. In the most rudimentary dyno you'll need to have a means for measuring torque, RPM, and temperature. For the sake of simplicity, let's use your Digatron tach-temp. Before you even start the engine, write down the date, the engine number, the air temperature, humidity, and barometric pressure. If you're testing something in particular, like a different pipe or a new cam, write that down too. It's too easy to forget when you look at the data a day or two later. Once you've got the engine started and running you'll need to get some heat into it before you do any serious work. After all, it's going to get heated up on the track, right? Apply a moderate load and run the RPMs up and down until the head temp gets into the low side of the normal operating range, 325-330 for a Yamaha, 285 - 300 for a Briggs. Then apply full throttle and adjust the load to hold the engine at the desired RPM. Move quickly now. Holding the engine for an extended period at a fixed RPM is hard on the engine and may give you heat buildup problems too. Note the RPM, the Torque, and the Temperature. Write them down. Then move the load to adjust the RPM for the next reading and repeat. If the engine gets too hot, just back off the load and cycle the RPMs up and down until it cools back down. Try to take note of any other changes that you may notice. How fast is the temp going up? Does the engine surge, like it's straining to get enough fuel? Do these things occur at all RPMs, or just in certain ranges? You can take data at as many RPM points as you like but, from a practical point of view, every 100 RPM is plenty. Take readings from a couple of hundred RPM below where you would normally slip the clutch, to a couple of hundred above where you would expect to turn the engine on top end. That way you'll know, when you look at the data, if you might be missing some valuable torque on one end of the range or the other. When you've finished a run, shut down the engine and check to see if any of the atmospheric conditions have changed. It's not uncommon for the air temperature in a small shop to go up several degrees just from the heat of the engine and the exhaust, even with good ventilation. An inexpensive home weather station with temperature, humidity, and barometric pressure is a sound investment and should hang right next to your dyno.
So now you have your first dyno data. You can take the time to graph it right now, but you'd be smarter to make whatever change you want to look at and make another run. The closer together you can make your dyno runs, the more consistent and reliable you data will be. You can always graph it at your leisure. Manual data collection like this is tough, demanding work, and it's really best suited to a 2-man operation. That way, one can run the engine and control the load while the other writes down the data. One more thing, before you graph any of your hard-earned data, you'll need to correct it for the atmospheric conditions. There are standard correction factors to apply to the torque readings so that, even if you compare a run that was made in the middle of the winter to one that was made on the hottest day of the year, you can compare the two accurately. A few minutes with a calculator and you'll have good results. Just remember, write everything down. What may seem like an insignificant detail at the time, may turn out to be an important factor when you start looking for conclusions from the data. Nobody ever regretted having more information than they needed, but the opposite........
The availability of desktop computers, particularly since their cost has become so reasonable, makes the idea of going to automatic data collection irresistibly attractive. Older, slower PCs can often be had for only a couple of hundred dollars and will serve very nicely in most applications. The critical bit of hardware you'll need is a plug-in card for the computer to convert analog signals into digital signals, commonly called an A/D board. Unfortunately, an A/D card may cost you as much as an inexpensive used computer, but it's worth it. This device plugs into the main board of your computer and converts the signals from the load sensor, the cylinder head temp sensor, the exhaust gas temp sensor, and anything else you're measuring, into a digital signal that the computer can read. Then it's a matter of setting up the programming so the computer can correctly convert the digital signal into the format you want. For example, the EGT sensor will generate a certain current in millivolts at a given temperature. The same with the load sensor, etc. You just need to tell the computer how to display what it's reading. A little more programming and you can have the computer record the data it collects and display or print it in whatever format you require. I don't mean to make that part sound easy, it's not. The fact is, it's a job for a skilled programmer, but it's very do-able. On our dyno, the computer reads all the data 10 times each second, then averages all that data around each 100 RPM step. Taking lots of data and compressing it like this helps smooth out any "hiccups" in the readings and makes for a more readable result. There is also some excellent commercially available data collection and analysis software available. Most notably, Davenport Dynos in California offer an outstanding package at a reasonable cost. They can also provide all the necessary computer hardware to attach to your dyno.
The real beauty of automatic data collection is that it means you don't have to hold the engine at a set RPM to take the readings you want. Since the computer is collecting data continuously, you can simply increase or decrease the load gradually to move the engine RPM through the range you're interested in. As long as the computer is taking the data fast enough (like every 1/10 of a second or so) you'll have plenty of input to work with. Moving the RPM steadily up and down so that you cover about a 100 RPM range every 1 ½ to 2 seconds takes some practice, but it will give you a excellent picture of what the engine is doing. You'll have almost 1000 data points and it will take less than 2 minutes to collect. Try doing that manually!
The only real bugaboo with computer data collection is our old enemy RF interference. You'll remember we talked about all the havoc it can raise with gauges and instruments that are not properly shielded. Let me tell you, there aren't many environments more hostile to a computer than the area around a small engine dyno. The tidal waves of RF interference can play hell with the internal workings of the computer boards. And the smoky, oily exhaust blowing around isn't too good for the mechanical parts of the computer either. Even more troublesome is the fact that most of the input signals coming from the sensors on the dyno are very tiny currents, on the order of millivolts. By comparison, the electromagnetic interference, and especially the currents flowing around an electronic dyno, are relatively large. It's very tricky to keep these little input signals from getting "lost in the mud" of the larger signals. Electrical isolation, extensive shielding, and great care are the requirements here. There really isn't enough space here to get into a lot of the details of how this needs to be done, but if you have somebody sharp enough to set up the hardware and software, they'll probably know how to protect it. And, of course, I'm always happy to talk about it if you want to call me at 317-742-0935.
Next month we'll take a look at some real dyno data and how to start interpreting it. We'll talk about what the data tells you and how to put what you learn to the test on the racetrack. Till then, remember, knowledge is power. With a good dyno, it's horsepower.