Engine dies around 4500 rpm

[Blue]

I'm certainly learning a lot thorough this process. One of the things I am learning about is oscilloscope operation, which is all new to me. I hooked the scope probe up to the +12V wire going to my single injector last night, and set the scope to AC coupled, as Peter suggested. My scope has an autoscale button, and I pressed it to see the signal. It showed a slightly fuzzy line about 14 V above ground. I didn't take detailed notes, but I think the time scale was around 5 ms/ div. An occasional vertical spike would appear either above or below the main signal. I think those spikes could be as much as 3 or 4 volts above or below the 14 volt signal. As I changed the time scale to the microsecond range, the fuzzy line became a regular pattern, but only a few millivolts high. When I changed the time scale to the nanosecond range, The scope pattern changed to a decaying sine wave. I should have recorded more details, but I was seeing so many new things. I didn't know what was important, and what I could ignore.

None of these aberrations got significantly worse when I raised the engine speed to the critical 4,000- 4,500 rpm range when the engine dies.

Can you give me a rough idea of what time scale I should be looking at - milliseconds. microseconds, or nanoseconds per division?

Am I looking for noise with an amplitude in the volt, millivolt, or microvolt range?

After that experiment, I pulled the MicroSquirt unit from the car and connected it to the JimStim for another experiment. I disconnected the Injector 1 wire from the JimStim and connected it to the injector control (ground) wire of a spare throttle body unit that I have. I ran two sets of wires from the battery that I used to power everything up. One set went through a 5 amp fuse to power up the JimStim. The other set of wires went to the positive wire of the injector and to the the three large gauge ground wires from the MicroSquirt harness. I was trying to prevent any current that the injector was drawing from going through the JimStim.

Everything had been working fine with the MicroSquirt on the Jim Stim the night before. When the MicroSquirt was controlling an actual fuel injector instead of the LED on the JimStim, it became unstable. It was getting late, and I didn't take detailed notes, but it appeared that the MicroSquirt was suffering from noise created by the injector.

A friend (who is a mechanical engineer) once told me that Mechanical Engineers add a spring to fix things. Electrical Engineers add a capacitor. Would it make any sense to add a capacitor someplace in the injector circuit to quiet things down? Even though the MicroSquirt technically shouldn't require a ballast resistor, is that something I should try?

As always, I am grateful for any insight or advice.

[Peter Florance]

Hi Brad
You might try a larger capacitor on the injector +12V line. 4700@25v would probably have enough energy to help

[Blue]

Would something like this be appropriate?

http://www.newark.com/30K6774/passives/ ... 478TTA025M

[Peter Florance]

Would something like this be appropriate?

http://www.newark.com/30K6774/passives/ ... 478TTA025M

Sure or Radio Shack 272-1022 in most stores

[Blue]

The mystery of the discrepancy between the value for pulse width displayed in MegaTune, and the signal displayed on my oscilloscope appears to have been solved. Bruce paid a visit to my home yesterday (Thanks Bruce), and the final determination was that the injector driver in the MicroSquirt is not capable of controlling my low impedance (approx 1 ohm) injector without some help. An approximately 1 ohm resistor in series with the injector reduced the current to an amount that the MicroSquirt injector driver was capable of handling.

Since my engine uses a single injector, we were also able to fix the problem by connecting both the injector 1 and injector 2 wires (green and blue in the MicroSquirt harness) from the MicroSquirt controller to the injector, dividing the load between the two drivers.

I will post more when I have more details and more time.

Once again, thanks Bruce. This one was tricky.

[Bruce Bowling]

Bruce paid a visit to my home yesterday (Thanks Bruce), and the final determination was that the injector driver in the MicroSquirt is not capable of controlling my low impedance (approx 1 ohm) injector without some help. An approximately 1 ohm resistor in series with the injector reduced the current to an amount that the MicroSquirt injector driver was capable of handling.

Yep, this one was tricky! But having the scope on the output available was the help needed....

The driver stage on the MicroSquirt is the VND5N07 from ST. A nice part in that it will current limit at 5 amps and will turn itself off during overtemp. Datasheet indicates that the temp shutdown is for a die temp at 150 DegC, restoring at 135 DegC.

When this part reaches the 5 amp point it goes into a linear mode to hold the current at 5 amps. The best operation is if the load is such that the driver almost reaches the 5 amp point or a bit over, because when the part is in linear mode it is generating more heat.

Here is an example - for a 2.4 ohm load low-Z injector and a 13 volt vehicle battery the ultimate current is 13/2.4 = 5.41. The part will hold the current at 5 amps, so the 0.41 amp remainder is basically dropped by the driver chip. To do this the driver increases its internal resistance such that the equivalent sum of resistance around the loop yields 5 amps at 13 volts. From ohms law it takes 2.6 ohms to give 5 amps at 13 volts (2.6 = 13 / 5). The injector has a 2.4 ohm resistance, so the driver chip introduces another 0.2 ohms to maintain the 5 amp limit. Now, since there is a 0.2 volt drop across the driver and 5 amps is flowing, the power dissipation is P = I * I * R = 5 watts (5Watts = 5 amps * 5 amps * 0.2 ohms). This power dissipation can be handled with uS due to the low duty cycle of injection, and the fact that there is inductance which slows the time it takes to reach current limit.

Now, take the 1 ohm injector case and 13 volts. The ultimate current thru this injector (unlimited) is 13 amps (13 volts/1 ohm). The driver will limit at 5 amps, and at 13 volts again this is a 2.6 equivalent series resistance to maintain this. The injector is 1 ohm, so the driver would have to introduce another 1.6 ohms of equivalent resistance to maintain the 5 amps holding current. The wattage is 5 amps * 5 amps * 1.6 ohms = 40 watts!

It is interesting to compare this to high impedance injectors. In my SBC, I have 4 high-Z injectors per bank in parallel. Each injector is 12 ohms, so four in parallel is 3 ohms load to the driver. At 13 volts the current is 4.3 amps at maximum - this is below the 5 amps current limit point so the driver will never go into current limit mode.

- Bruce

[Blue]

I can't believe it's been 2 months since Bruce visited me. After Bruce's visit, I purchased a one ohm and a two ohm power resistor to put in series with the injector to reduce the current draw. The 1 ohm resistor worked fine at low MAP values, but when I tried to run at full throttle and high RPM I started running into the old problem of not being able to get enough fuel into the engine, regardless of the entry I used in the VE table.

I connected both injector control wires from the MicroSquirt through the 1 ohm resistor to the injector, and that solved everything.

I haven't had time to experiment with the 2 ohm power resistor, or with using both injector control wires and no power resistor, or with using PWM.

Once again, thanks Bruce and everyone else for all your help. I'm a MicroSquirt devotee now, and I want to put one on everything I drive.