cps fix

[Ecomike]

The waveform shown in the VRS link provided earlier looks exactly like the waveform from the CPS and the Tranny output speed sensor. If you're looking at the picture in the Jeep Renix Fuel Injection Manual, the sawtooth pattern is labelled "magnetic field", so the inducted voltage is due to the rate-of-change of the field - you get the spikes at the tips of the sawtooth.

In any case, I agree that the cap can only reduce the voltage spikes as the cap will act as a 1st order low-pass filter. Adding the cap in parallel makes it a low-pass filter with a cutoff freq of about 80kHz if you assume the nominal sensor resistance of 200 ohms. The rep rate for the pulses is about 5000 pulses/second but the frequency content of the spike should be a lot higher. My guess is that it's simply cleaning up a noisy sensor signal - perhaps due to a cracked magnet in the sensor. The computer undoubtably has a cap on the signal lines anyway.

The voltage peek of the spikes is dependent on the engine speed too. They grow proportionally bigger with engine speed. When I was building a box to modify the tranny sensor signal to use an 89 TCU with a 99 trans, I measured the voltages from zero to nearly 30-volts at 60 mph.

I'm not a EE, but I'm a reasonably competent electronics tech and I happen to do a lot of work with acoustics, filters, amplifiers, and data acquisition systems.

If a 3-cent cap fixes your problem then by all means be happy. It might not work for everyone else and I highly recommend having a spare with you in case this bandaid stops working.

Re-reading this old thread, and your post got me to thinking that maybe the capacitor in his ECU failed, or its capacitance changed enough, that his Edisonian approach of trial and error brought the system capacitance back OEM spec?

 

[cruiser54]

Aparently Renix wasn't programmed to kill everything if a weak CPS signal tries to be counted. After my post last nite I came home today and re-aligned it closer to the flexplate.
ALL the horrendous syptoms have disapperared, as quick as they came.
The fuel additive was coincidence.
When I tried the original CPS again yesterday to nip the problem, I didn't push it down hard enough either. I have modified both as well. I think this requires a little extra push to compensate for the loss of the true holes. Must wanna climb up a few hundred microns or whatever. Read today that ideal gap is .82mm. Not like you're gonna gauge it. Maybe could attach a gauged piece of plastic with parrafin or whatever to be ripped off by the flex plate and push down Hard when installing....

When these rigs were new at the dealership we had many intermittent crank no starts. We would find that the CPS many times put out less than 250 MV while cranking. This was before the CPS bypass bulletin came out. At this threshold it would start intermittently and unreliably. We would drill one mounting hole of the CPS bigger and move it closer to the flywheel so it would generate a stronger signal. We always shot for 500 MV.

A cruder way we did the mod on the Eagle Premiers which used the same type of CPS but it was a bear to get to, was to have one guy crank the engine over while the guy under the hood had a voltmeter on the CPS connector. Holding a long bar against the top side of the CPS bracket, the guy under the hood would strike the bar with a hammer until the desired voltage output was generated.

 

[Ecomike]

Has any one tried this capacitor fix since the original thread was started? If so how well did it work? Is the original poster still using the capacitor?

 

[Ecomike]

Well I guess oldman, cygnus, and I need to eat our hats. Here's what the CPS signal looks like at idle on my 89 MJ 4.0 auto. The frequency and amplitude get proportionally higher as the engine is reved. It is spikes, but they are spaced close enough together that it looks like a sawtooth.

I add a 0.01uF cap across the sensor and the voltage increased roughly 5% from idle up to about 3k rpms. No distortion of the waveform, just increased amplitude. Stepping up to a 0.1uF cap increased the voltage roughly 20% at idle. I only reved it to probably 2k, but in hindsight I should have looked at higher rpms too. A 1.0uF cap cut the amplitude in half at idle and the engine started stumbling.

So I'm thinking it's a tuned LCR circuit and the extra capacitance is boosting the voltage at starting where the rpm and signal strength is lowest. This might also explain why beefing up your battery cables and getting better starting rpms helps too.

-Chris
(looking for the hot sauce to put on the ball cap)

Just dawned on me that if it works for starting, and interferes with higher RPM operation, one could add an on/off switch in series with the capacitor, turn it on for starting, and off for running, maybe wire it in with a start relay tied into the start ignition switch to remove it from the circuit after starting (run position).

 

[kastein]

I'm not an EE either, but I can run a Spice analysis.

The "right" cap value across the coil will create a parallel resonant circuit -- causing the voltage 'seen' across the indutor to increase.

Give the man a cookie....haters!

ding ding ding ding ding!

Just dawned on me that if it works for starting, and interferes with higher RPM operation, one could add an on/off switch in series with the capacitor, turn it on for starting, and off for running, maybe wire it in with a start relay tied into the start ignition switch to remove it from the circuit after starting (run position).

You could do this, but remember that those wires to and from the CPS sensor wiring are going to pick up noise, add capacitance, etc. If you want to do this position the capacitor and the relay that switches it into the circuit as close as you possibly can to the CPS wiring, keep everything short and wires well spaced if possible.

My bet on the large vs small notch issue you mentioned in this thread (but I forgot to multiquote) is:
1) the capacitor changes the resonant frequency of the CPS circuit such that the rising and falling edges of the pulse "tickle" it, rather than the whole pulse. I would blow a bunch of smoke out my arse about the heaviside unit step function and a parallel LCR circuit's response to it, but it's been years since I used that stuff and I don't remember enough to avoid spreading blatant misinformation, so you'll have to google it and brush up on electronic circuits 1, because I don't feel like doing it right now :spin1:
OR 2) the ECU only cares about 1 or the other size of pulse while trying to start, sets the ignition advance blindly, and hopes it fires, then starts caring about the ignition advance after the engine is warmed up. The capacitor happens to help enough on the index pulses that it makes it start, but it doesn't affect the other size pulses at other engine RPMs to cause a problem (well, unless you use the 1.0uF cap as listed previously, which caused stumbling problems because it put the resonance of the CPS system too far out and reduced the signal at high RPMs to the point that the ECU got confused.)

My bet is on 2), or maybe a combination of both.

 

[Ecomike]

Not sure the RF noise is a real issue based on the data I have seen on mine.

Well I had some experimental fun with the capacitors. I got over 30 volts at 2500 rpm with the .22 uF (versus 5 volts with out it!). I got about an 8% increase in voltage at cranking rpm with the .022 and .22 uF, and not much more the .22 uF. With out the capacitor the peak voltage at 3000 rpm was about 5 volts.

Th .22 uF has a peak voltage over 30 volts that falls off rapidly at 2000 rpm back to the 5 volt range and lower and at high rpms up near 3000 rpm the rpms become unstable on the 89. It did raise the idle cps voltage about 35%! The .22 and .1 uF made the engine stumble badly at around 3000 rpm and higher.

Even the .1 uF was running up over 30 volts at 3000 rpm and starting to act like a rev limiter acts in the newer cars.

The .022 worked OK, raised the idle voltage about 20-25% but did not move the cranking voltage of the cps up much, maybe 7%. Might work better on a bad CPS versus the new one I have?

The battery was nearly DOA at Autozone, tested as bad. I swapped and upgraded to their gold top, about 40% more cold cranking amps than the cheap one I had. It made a huge difference. So far no starting (got spark) problems with the new battery. It is starting, firing up much faster.

I suspect, in a pinch, having a large one, like .1 to .2 uF in the tool box, for an emergency start and an easy way to plug it in, would be a nice trick to get a hard a start one to fire up and run. Maybe even as a testing tool.

Also of note, the new CPS was reading closer to .30 volts cranking with the new battery, the old battery tested at .28 volts on the CPS peak voltage while cranking before the battery started really dying noticeably yesterday.

I think 5-90 is on to something, with his claim that the Renix ECU will not order spark and fuel unless the Renix reaches a minimum cranking rpm, good CPS voltage or not? But my 87 battery (it has been parked for weeks with out running daily (AW4 needs work on it), and the battery is getting old), battery is dying now, and it barely turned over the 87 Renix jeep tonight at 43 F, and yet it started with only .28 volts peak coming from the 4 year old CPS on it.

Maybe there is more difference between the 87 and 89 ECU's than we think? And the 87 still has its black goo filled C-101 on it, and its CPS is not soldered, but still in a weather pack, QC!:laugh3:

When all this no spark trouble started on the 89 jeep, 2 weeks ago, I got a good 15 minutes of cranking, over an hours time out of the 2 year old Azone Valucraft battery before it gave up and needed charging, that is the battery I replaced today. At that time the CPS was the old OEM from 89, and it was only giving .17 volts peak, so it got replaced first. Alternator is and has been working perfectly.

I hope this new battery is the last of my problems for a while with the 89 jeep. I decided not to use the capacitor as a permanent add on, for now. But the next time I have a no spark issue, I will pull out the .22 uF and give it a try.

 

[kastein]

Given that the .1 and .022 both don't do quite the right thing, I would try a .047 next. They should be available and fairly common.

The specific capacitance probably depends on the sensor somewhat.

 

[Ecomike]

Given that the .1 and .022 both don't do quite the right thing, I would try a .047 next. They should be available and fairly common.

The specific capacitance probably depends on the sensor somewhat.

I tried the .047 as well. Just did not mention it. It was not noticably different than the .022

But I am using a DVM on a 60 hertz AC scale, not an O-Scope, so I may have missed useful data. I was trying to use and post data the average O-Scope challenged person could use.

 

[kastein]

You probably are missing a lot, actually. Most DVMs don't do true RMS and won't sample fast enough to get you a real voltage measurement there - and they show RMS reading, not peak-peak reading.

 

[Ecomike]

You probably are missing a lot, actually. Most DVMs don't do true RMS and won't sample fast enough to get you a real voltage measurement there - and they show RMS reading, not peak-peak reading.

From everything I have found and read, the CPS target voltage for cranking voltage posts I found was based on using the high impedance DVMs most people use today. And the DVM AC signal reading is proportional to engine rpm. I would be curious as to the frequency of the CPS while cranking and at idle, but I am not sure what the exact tooth count is. 12 teeth per cylinder, 3 per 360 degree rotation I think. So 36 teeth per revolution times 750 rpm at idle, divided by 60 seconds per minute for teeth per second? 450 teeth, or cycles per second at idle, so not too far off from 60 cycles per second for AC power, at cranking rpms?

Maybe the DVM is counting the voltage of every 8th tooth? Aliasing?

IIRC correctly, I read a 20 some odd year old Chrysler TSB that referenced using a DVM for measuring the CPS cranking voltage on the Renix, and the target was .35 to .5 Volts AC on a high impedance DVM