|
|
This article describes a modification that, if misapplied, has the potential for causing early engine failure. The information presented in this article is correct to the best of the author's knowledge, but is not guaranteed in any way. Neither the author or Team FC3S can be held responsible for any damage to a car that occurs as a result of using this circuit, nor for failure of the circuit to perform as expected.
There has been a great deal of interest on the Team FC3S list in a cheaper (and if possible, superior) alternative to the expensive FCD's offered commercially. This is understandable given that no modification to the Turbo II that raises the possible boost pressures can be attempted without an FCD. This article presents a simple FCD circuit that can be built by anyone with moderate electronics assembly expertise. In addition, the article covers the reasons why an FCD is needed, why this FCD is superior to other FCD designs, and the theory of operation. By the time you are done reading this article, you will hopefully understand a little more about how your TII works, and a little electronics as well.
The FCD (acronym for Fuel Cut Defencer) is any device that prevents the stock ECU from cutting fuel to the rear rotor when the boost exceeds a preset boost level. This maximum safe boost level and behavior is coded into the firm ware in the engine management unit and can not be changed without reprogramming the ECU. The maximum allowed boost pressure is about 8.6 psi. The way an FCD defeats the fuel cut is by lying to the ECU about what the boost pressure is. The FCD is placed between the boost sensor and the ECU where it modifies the boost pressure signal by some amount so that as the boost pressure rises above the preset "safe limit", the ECU continues to see a signal that is below the limit.
There are a couple of immediate consequences to this fooling around. As the boost rises, the ECU must increase the amount of fuel being delivered to the engine in order to maintain safe and efficient operation. As the FCD starts lying to the ECU, the ECU will begin to under-compensate for the rise in pressure leading to a gradual leaning of the air-fuel mixture. The amount of error increases as the boost rises. For relatively small errors, the only penalty is efficiency. As the error gets larger, however, detonation becomes likely, exacerbated by the high boost pressures and accompanying high intake charge temperatures. Detonation under these conditions will quickly kill an engine. So, before we go any further, be forewarned that using an FCD and increasing boost pressure without also compensating for the ECU error with fuel enrichment (and preferably more efficient intercooling) can cause serious damage to your engine!
What you would like to have is an FCD that leaves the boost signal alone until it approaches the cutoff level, and then kicks in, holding the signal below the critical level. This can be accomplished with a circuit element called a clamp. The FCD circuit described in this article utilizes an active clamp which performs the necessary function very efficiently.
Following is a graph of the boost sensor signal without an FCD, with two commercially available FCDs, and with our cheap little DIY FCD.
As you can see from the graph, FCD #1 is a clamp circuit. The output follows the input until the clamp voltage of 3.33 V (about 6 psi) is reached. At this point the output stops rising. FCD #2 works by reducing the input by a percentage. This creates an error across the entire range, as opposed to only when the boost is over the limit. (At a safe 5.5 psi, the computer is only seeing 4 psi.) This is not the desired behavior. Our FCD works similarly to FCD #1, except we have raised the clamping voltage to 3.65V (about 7.9 psi).
A linear regression ran on the data gathered shows that the equation for the best fit line is:
Solving for Pressure, we get:
NOTE: If you look closely at the graph, you can see that our FCD has an output that is 0.05 V below the input up to the clamping voltage. I have since fixed this problem by changing R1 from 2.2k ohms to 680 ohms. The FCD output is now within 0.02V of the input, right up to the clamping voltage. Sorry, I didn't have time to rerun the numbers.
The measurements above were obtained with the FCD under test connected to the TII boost sensor and TII wiring harness in order to simulate actual operating conditions as closely as possible. Pressures were read on a diagnostic pressure/vacuum gauge. If you are interested in it, I would be happy to send you the Excel spreadsheet containing the raw data, the linear regression, and the graph above.
The circuit schematic is shown in the following figure:
There are a number of requirements on an FCD that is going to work well and survive the stresses in your Rx7. Here are the most important ones:
In our circuit, the clamping is performed by op-amp U1-2 and D3. These two devices form what is called an active clamp. While the signal at the (-) input of the op-amp is below the voltage setting at the (+) input, the output of the op-amp will be high, D3 will be reverse biased, and the output will follow the input. When the signal at the (-) input of the op-amp exceeds the voltage setting at the (+) input, D3 conducts closing the feedback loop and causing the output to follow the (+) input, which is set at the clamping voltage. This all happens so quickly that it might as well be instantaneous. Notice that the output impedance of the clamp is not zero. When clamping is not occurring, and the output is following the input, the signal is passing through R1, which makes the output impedance 680 ohms. Not to worry, this output impedance is substantially lower than the boost sensor's output impedance.
The clamping voltage is adjustable and is set by the trimmer at R3 which is wired as an adjustable voltage divider between the Vref lead to the pressure sensor and ground. Vref is supplied by a voltage regulator in the ECU and is maintained at 5V .
High input impedance is achieved by buffering the input to the FCD. U1-1 is wired as a voltage follower, meaning that the output just follows the input. The input impedance of the buffer is nearly infinite.
Noise suppression is achieved via the use of a low pass filter on the input, formed by R2 and C1. High frequency roll-off is at about 100 Hz, allowing the circuit to be responsive but effectively suppressing RF noise.
Input protection is provided by D1 and D2 which clamp the inputs to +12V and ground, effectively protecting the op-amp inputs. Stray voltage spikes above or below ground will be shorted to the appropriate supply rail.
Thermal stability is potentially a small problem for our circuit. The resistance of R3 can vary as a result of extreme temperature changes. For this reason, I would recommend installing the FCD inside the car, alongside the ECU. This article will provide instructions for installation in both places.
For the active component in our FCD, we have chosen the LM358 dual op-amp. This IC puts two single supply op-amps in a single package. It's not a high precision op-amp, but it serves the FCD circuit very well. Having two op-amps in the package makes it possible for us to include a buffer with the active clamp, which improves the input impedance as mentioned above.
The parts utilized in this circuit are readily available from any electronic parts supply house. Even Radio Shack. In order to help you in this process, however, I have included the table below which also includes the part numbers and prices from Jameco electronics. I chose Jameco because they provide excellent service, and because their prices are consistently very reasonable. You can reach Jameco at 1-800-831-4242, or you can order via the World Wide Web.
Please Note: These prices, part numbers and page numbers are from Catalog #971, Feb. - Apr. 1997. My experience has been that the part numbers remain good, even though the page numbers may change. Please contact Jameco for the latest catalog, availability, or pricing information.
|
||||
|
|
|
|
|
|
|
1N4001 Diodes |
35975 pg. 25 |
|
|
|
1N914 Diodes |
36311 pg. 25 |
|
|
|
680 ohm 5% Resistor |
31499 pg. 33 |
|
|
|
10k 5% Resistor |
29911 pg. 33 |
|
|
|
10k 10 Turn Cermet Trimmer |
43P10k pg. 34 |
|
|
|
0.01 uF Mylar Capacitor |
26884 pg. 31 |
|
|
|
LM358AN Dual Op-Amp |
120862 pg. 12 |
|
|
|
1.6X2.7 inch Epoxy Glass PCB |
105099 pg. 68 |
|
|
|
3.1x2.0x0.9 inch ABS Box |
18921 pg. 69 |
|
As you can see, building our FCD is not going to break the bank!
If for some reason you should have trouble finding the mylar capacitor, a ceramic capacitor may be substituted with very little impact on circuit performance. Similarly, any 10k trimmer may be substituted for R3, although the Cermet trimmer mentioned is a superior part (in terms of stability and durability) to most trimmers sold.
The perf board specified here matches the box well, taking one of the worries out of construction. However, any perf board or box may be used.
This circuit is not very sensitive to construction techniques. You can wire wrap it, use point to point wiring on a perf board, dead bug, PCB, or whatever works for you. If you use the perf board specified above, you will probably find that wire-wrap or point-to-point wiring makes the most sense. Connection between adjacent pads can easily be made using small pieces of wire (or even solder bridges), resulting in a pretty dense layout.
For those who aren't familiar with the dead bug approach, it involves epoxying the IC on its back on a piece of metal or plastic (metal is better because it can act as a ground plane) and soldering the components directly to it's leads. The leads are usually bent such that every other lead points in or out so that you have more room to work. The trimmer can also be epoxied to the backing since it is fairly large and needs to be mounted firmly. Once you have it wired up and tested, you can put it in a box. Usually, the cover of the box is what you use as the backing. Then you just put the cover on the box, parts down, and the parts end up inside the box, safe and sound.
Before beginning construction, have a look at the trouble-shooting section for some more things to watch out for.
After you have built the circuit, you will need to test and calibrate it, before installing it permanently in your car. Use the following procedure, being careful to avoid shorts between 12V and ground.
Don't get carried away fiddling with this adjustment. You are not going to achieve significantly better performance by the slight changes you make. You can, essentially, only make things worse. Adjust it to 3.7V, and leave it alone unless you experience fuel cut, then reduce it in about 0.25V increments until the fuel cut goes away.
Your FCD is now calibrated. Go on to Installation.
If you still have problems, contact me.
If you are going to install the FCD in the engine compartment, you'll need to decide how you want to hook up to the wiring harness. The obvious option is to tap right into the wires that connect to the boost pressure sensor plug. Since you only need to tap into the wires for Vref, V+ and ground, only the signal wire will need to be cut. The input to your FCD will need to be connected to the connector side of the cut signal wire, and the output will need to be connected to the wiring harness end of the wire.
You can tap into the other three wires without cutting them by removing the connectors from the plug, soldering your leads to them, then re-inserting them. To remove the connectors from the plug insert a small screwdriver between the plug housing and the connector (from the back) to depress the locking tab, then just push the connector through. Push your lead in from the back of the plug housing, solder it to the connector, and push the connector back in.
Solder all connections you make. Don't rely on just twisting wires together and taping them, and don't use crimp-on taps. Otherwise, as time goes on, your connections will get flakey. If you swear by crimp-on connectors, go ahead and use them, I don't want to hear about it!
If you would prefer not to cut any wires or otherwise modify your wiring harness, there is possibly another option. I have found the socket that matches the boost sensor plug in the Mazda stereo installation kit that AutoZone sells. It's the same as the the rear speaker connector. That gives you the connection for the plug side. The boost sensor side (the plug) might be had by scavenging for one at a wrecking yard, or you could use discrete spade connectors to make the connection to the sensor. (Just be really careful not to short neighboring pins!) Make sure that when you wire things up that the boost sensor gets the V+, Vref, and ground connections it needs, and that it's output is wired to your FCD's input.
If you're brave enough, I recommend installing the FCD alongside the ECU. The upside is that it is protected from heat and moisture. The downside is that you have to modify your ECU wiring harness, it's a little difficult to figure out what wires to tap into, and it is difficult to get at.
To get at your ECU, remove the passenger side door sill and kick panel, and pull back the carpet (gently) from the footwell. Underneath the carpet is a large metal cover, attached to the floor by 2 nuts on studs and 2 bolts. Remove the panel to reveal the ECU and the wiring harness underneath.
The connections are as follows:
Connector 2 is the center connector.
Sorry, this section is incomplete. If you feel comfortable working around the ECU, go ahead. If not, install it in the engine compartment for now. I'll be updating this section just as soon as I have an opportunity to get at my ECU. I don't trust the factory shop manual enough to go out on a limb about where the right connections are without actually checking them myself.
Special thanks to Ari Yarallon of Rotary Performance in Dallas who provided the FCD's for comparison (almost a year ago!), and to Brian McWhirter and Ryan Vaughn for trying out an early version of the design. And, of course, thanks to Ryan for all his hard work in keeping the TeamFC3S site up and running, and making it one of the best automotive sites on the WEB.
Paul Stoaks received a Bachelor's Degree in Physics from Eastern Oregon State College in 1986. After spending a couple of years as a microcomputer support technician, he and his wife spent a year teaching Computer Science at the Christian Academy in Japan to the children of missionaries stationed there. Since 1989, Paul has been involved in the development of software for Integrated Circuit design (at Mentor Graphics, Inc.) and currently in the development of software for electronic system design at Savantage, Inc.
Paul's hobbies are electronics and sports cars. He currently owns a Charcoal Gray, '87 TII which provides the perfect platform for the simultaneous enjoyment of both his hobbies.
Feel free to contact Paul via E-mail at pstoaks@savantage.com for help with TII electronics issues.
Return to the RX7 Article Page
Return to the Main Page