CIS Lambda Frequency Valve Signal

[Walt Fricke]

I hooked the red wire of my dwell meter to the green wire of the test plug for a 1982 US 3.0 motor, and the black wire to a ground. The only other wire is the inductive clamp for spark plug wires, and is used when checking RPM, and so has no role in checking duty cycles. The motor was at all times in open loop (no O2 Sensor hooked up), and at an idle of about 1,000 rpm unless otherwise noted. The meter read 90% duty cycle.



This didn't seem right. For one thing, I don't think the frequency valve can achieve that. So I started Googling around.

I found a discussion where the author explained that the frequency of the frequency valve is 70 Hz (70 cycles per second, or just a bit faster than a light bulb). And that this square wave signal is pulse width modulated. This means that the high portion of the wave varies in its length, and the low varies inversely. Which means that the duty cycle would range from 100% (always high), to 0 (always low). I believe that the frequency valve opens when it sees a high, and closes when that goes away. So the greater the duty cycle, the more fuel the FV directs back to the gas tank from the lower chambers of the system pressure connections to each injector, which lowers the pressure there, allowing more fuel to flowsthrough the upper chamber. Which results in a richer mixture. A leaner mixture would result from the duty cycle being smaller.

I then hooked up a small digital oscilloscope. It showed that the frequency was 70 Hz.



Then I used it in scope mode. It showed nice square waves, with an approximately 50% duty cycle.



Next I wanted to see what cracking the throttle just a little bit would do. When I had done this, the dwell meter needle never budged. But the scope reading did.



I believe that by this time I had inverted the signal by poking the buttons on the scope. It looks like a 61% duty cycle (11 segments /18 segments if I counted right). That doesn't seem quite right either for just off idle, but at least it indicated that the microswitch had communicated with the minibrain, and the system had changed what the FV was asked to do to a value within its capability.

I didn't want to advance the throttle to see what wide open throttle would do, and I hadn't reached around to put a jumper on the WOT switch to simulate this.

Why didn't the dwell meter react the way the scope did, and why did it give only a bad reading? In dwell mode the meter has no source of power other than the green test lead (its design does not include a battery, and in RPM mode it does get 12V from where you attach the red wire to such a source). Looking at some circuits which purport to let you measure dwell, I saw quite a few which got their power, so to speak, from the circuit being tested. But you will notice that the frequency signals that the green test wire provides are just over one volt. The dwell meter must be used to the usual 12V you'd see when looking at distributor points (which is what it is designed to do - on its backside is a chart converting duty cycle into dwell degrees for various configurations of engine.

I opened the dwell meter up. 3 transistors, a zener and a small signal diode, 5 pots from 500 ohms to 10K ohms, a 220 and a 1000 microfarad capacitor, 7 resistors, and 4 other small capacitors. No ICs. My not too well educated guess is that one or more capacitors help move the meter needle, and the relationship between their discharge rate and the charge rate from the incoming signal is what sets the duty cycle. If the pots are calibrated for 12V, a lesser voltage isn't going to work, and it is only the inherent capacitance (or something) of the meter which causes it to move off its rest position of 100.

Perhaps the dwell meters which include a battery, or which have a positive, ground, and signal lead (in addition, for some, to an inductive clamp) work as Mr. Probst and others say, as demonstrated by the numerous posts here by guys talking about what their meters read? And you can check your Lambda CIS just fine with the duty cycle numbers they give? Adjust the mixture, even?

I won't be able to do that because I don't have an O2 sensor signal to add in.

[tirwin]

Walt,

I had a thread about this a while back. Good work! You came across the same thing I did and scratched my head.

Here's the simple answer. All dwell meters are not created equal! A lot of them are junk frankly. Jim Williams posted a picture of the oscilloscope output. We had a good talk about this. Jim has noted quite a difference in accuracy when compared to an oscilloscope -- same as you've now noted.

What's worse is that some meters report the amount of time the FV is CLOSED and others report the amount of time that it is OPEN. If you don't know what your particular meter reports, you can get into trouble. For example, if you knew for an absolute fact that you were running way too rich then it would be easy to tell how the meter reports the duty cycle - a reading of 90% or 10% would be obvious. But if your meter says 55%, how would you know which way to adjust if you didn't know how your particular meter read?

I got a DVM with a duty cycle function built in that I trust. Well, I trust it 1000x more than any dwell meter, especially the cheap Chinese junk ones, but I trust it less than an oscilloscope where you can really see what's going on.

What brand/model is that scope? I've been looking at getting one myself.

[javadog]

JYE Tech 06203P 062 LCD oscilloscope: Oscilloscope Probes: Amazon.com: Industrial & Scientific

[schoward]

I have had similar experience using an old Sears three wire engine analyzer for looking at duty cycle/dwell. For me, my closed loop dithering readings seem to be good/usable for what I need to do (fine tune the mixture and set idle rpms) which I've checked with a CO meter also.

But reading Duty cycle on enrichment points seems to make the meter go a little bananas. Monitoring the fixed cold duty cycles with my meter seems to be directionally correct but they are always a tad off spec. I have used my meter with two different ECUs installed and get similar results which tells me its the meter, not a bad ECU.

[Walt Fricke]

Tim
Could you point me to the earlier discussion? I did a search, found your name on posts a lot, but wasn't finding one which specifically discussed details of measuring duty cycles.

I wish I could convince myself to call the various Lambda components an ECU.

That little scope (a kit you assemble, but the hard part - surface mounted devices - has been done for you, so not hard, though I bought mine assembled because the store owner had been intrigued, and put one together, then sold it for the cost of the kit).

It is closer to a go/no go device than a good oscope, but small and handy.

A hobbyist could find a schematic for a duty cycle meter by Googling, and make one pretty easily. They usually include a pot which you adjust to get the analog meter to read what it should.

[Bob Kontak]

Here's the link, Walt.

Low power in low RPMs after reconnecting O2 sensor

[tirwin]

Bob's got you covered. I was in the garage breaking the 22mm bolts holding the shocks on the banana arm. 4 ft cheater saves the day! Cripes!