SPICA fuel delivery graph

[vangos]

I was wondering if anybody ever attempted to create a graph of the fuel delivery characteristics of SPICA pumps. It seems that for a given engine temperature, atmospheric pressure, and position of the fuel cutoff solenoid, the only variables that determine the amount of fuel that goes through the pump are the engine rpms and throttle position. I would love to see a 3D plot of this "surface".

Why am I asking? I am trying to determine what's going on with my SPICA pump and if the behavior I am observing is to be expected. I installed an oxygen sensor and an air/fuel mixture guage in order to set up the mixture strength. The pump gap is adjusted properly, the thermostatic actuator works, the pump timing is correct, and the rubber washer at the idle air port is there and working. The behavior I get is the following:

- If I set up the mixture to be close to stoich (using the fuel cutoff solenoid) at 2500 rpm, I cannot get the car to idle (too lean).

- Enriching the mixture turning the fuel cutoff solenoid improves things although at 2500rpm the mixture is rich. The car runs fine, the idle is good when the engine is cold but gets a bit lean and uneven as the temperature goes up. Even worse though, at 2500 and under highway cruising conditions, the mixture is rich. If I try to lean it out again, every time I step on the gas the guage shows full lean until the engine rpms catch up... I have to really push the gas pedal to the floor to get a non-lean mixture while accelerating.

So I am curious as to what the logic behind the fuel delivery of the pump is. Should the pump deliver more fuel while trying to accelerate (open throttle, low engine rpms) and then taper off as the engine speed catches up?

Thanks,

[tubut]

You may find the information in Wes Ingram's SPICA manual described at http://www.wesingram.com/hp.htm. Give them a call to verify.

[Roadtrip]

You didn't say what you are using as stoich, but Ingram recommends:

Engine at running temp. Pump gap .019" 2500 rpm:
Lowest combination of HC and CO
CO .7 to 1.2% at lowest HC

If idle is too slow, open the barrel until the CO drops below 2.0% HC should be below 350. Of course the ignition system must be in good condition. If emissions go down by removing the breather hose from the cam cover, the oil is diluted with fuel.

When the engine is at idle, can you kill the motor by putting your finger over the air distr barrel? If not, maybe the throttle butterfiles are open a bit when they should be closed.

As far as the fuel scheduling goes, of course there are 10 different 3D cams. Your 73 has one of the higher fuel delivery rates (about 15% more) at >5000 rpm than later emissions restricted cars. Here's the only graph I've seen of a camoid schedule.

[vangos]

Hi John,
Thanks for information. The graph is quite interesting.. I've never seen one of these before.

I'm using Wes' Dynograph air/fuel mixture gauge and a standard Bosch O2 sensor to check my mixture strength. The Dynograph has 3 red lights (lean), 4 green lights (close to stoich), and 3 amber lights (rich). The only way I can get the engine to avoid the red lights regardless of the load is to set it up such that at 2500rpm it's in the middle amber. This setting gives me strong acceleration but I was expecting that while cruising at highway speeds I should see something closer to the green region. The exhaust definitely smells rich (I have a leak in the trunk seal and when I open a window the exhaust gases get sucked into the car).

As for the idle, if I close the idle air port with my finger the engine starts running really rough, I don't think enough to die though. Doesn't the hose coming from the oil separation canister go into the idle air as well? Interstingly enough, my idle is smoother when the Dynograph reads red or leaner than when I close the idle air more so that it reads green. That has me thinking that I don't have enough gas supplied at idle speeds.

Cheers,

[Roadtrip]

I think you should check your throttle angles again as well as the throttle rod ends for slippage or maladjustment. I never used one of those LED meters that you have. Perhaps Wes can fill you in on the correct way to use it to tune a SPICA system. If you're setting the mixture correctly at 2500 rpm, the idle fuel delivery should be fine. That makes me suspect of that meter. It wasn't until 75 and later that the idle delivery was leaned out somewhat.

For interest here's a picture of a '71 T237/2 1750 3D cam. First pix is of the idle profile, then mid-throttle, then full throttle. Note the profile at the very top edge of the camoid.

[Roadtrip]

Half Throttle

[Roadtrip]

Full Throttle

[DriverJon]

Hey Vangelis...

Check out the report/instructions on tuning a Spica using the Dynagraph, at this link from the Oregon Alfa club:
http://www.alfaclub.org/tech_library.htm

It's the last entry on the page.. for some reason it's a Word file rather than a web page.

He documents his experience gettting the tuning right, with numerous calls to Wes for tips on how things should look.

There are a couple of Spica tuneup articles there as well

Get Wes's Spica book, if you haven't already.. there are several mechanical things to get right, and in the right order, so that the pump will properly calibrated when you do your final mixture setting.

If not, you can end up like you're describing things.. ok at one condition but not at another.

Jon

[vangos]

Thanks so much for the pictures of the 3D cam, John. I was great to see an actual picture of the various regions of the cam.

And yes, I have to second Jon's recommendation for the AROO website's tech library. There's a wealth of very interesting tips and other information in there.

As far as my SPICA pump goes, I have a feeling it just might be ready for a rebuilt. I've read Wes' manual a number of times, followed the AROO instructions, I have set and re-set everything I could: pump gap, throttles, the rods, the fuel-cutoff solenoid and everything seems up to spec, except from the resulting mixture!! The car had about 170K miles before I rebuilt the engine and although I have receipts dating back to 1974(!) there's no evidence that anything was done on the SPICA pump. So maybe it's past it's useful lifespan. According to what I read, poor gas mileage and an overly rich mixture at cruise speeds that cannot be corrected is one of the symptoms of a failing pump.

Now, I am toying the with idea of possibly switching over to webers. I know that some people (especially some of our most knowledgeable members of this forum) would probablyl not recommend the switch. I have no doubt that a freshly done SPICA pump by Wes will provide years of dependable performance, I just think that it might be fun to experiment with something different. I will try to source the necessary parts during my upcoming trip to Europe, add up the costs and go from there.


Cheers,
Vangelis

[dwc]

I've got a dynograph clone (home-built for ~$15, using the same chip) on my '74 Spica Spider, and I would have to comment that it isn't ever going to be useful as a tuning device. For one, you've got no baseline to reference. With regards to any standard O2 sensor, by design, it's too slow and its bandwidth is too narrow. The display is fun to watch, and clearly shows that the TA and cutoff solenoid are working properly, but I can change the mixture over a huge range and see identical readings from the LED's. Decent exhaust gas analyzers can be had for as little as $150 new, and really nice used Sun units show up on eBay from time to time. Regardless, I don't feel that either is needed to tune a Spica system well.

Regards,

Dean
Lutz, FL
'74 & '87 Spider Veloce's

[DriverJon]

Well, you sorta have a baseline, in that the fast change/center of the range part of the standard type O2 sensor is pretty much accurately on stoich...

Unfortunately, how fast it changes off that can vary quite a bit due to operating temperature and probably other influences... so it's hard to be accurate how much richer or leaner you are.

It might be possible to accurately calibrate a narrow band O2 sensor, but it would be a project, and probably each one would have to be specifically done. Not to mention added complexity in montioring the temp and changing the cal.. it's gone from being a simple voltage readout to needing a microprocessor and a calibration map..

That said, the procedure mentioned in that article seems to be a relative kind of thing, in that you orient the mixture range to be mostly centered on the stoich ratio and not too far off on each end. It probably gives you a decent idea where you are.

And I agree you probably can get pretty close by doing the manual settings and checking the engine response, (like in Wes's book) as well as looking at plug deposit color..

Would be interesting to get a car equipped with one on a "good" NDIR exhaust analyzer to see just how coarse or fine of a reading the LED's show..

[vangos]

In my experience, the O2 sensor/Dynograph system is pretty sensitive to changes in the mixture strength. For example I can clearly see the effects of even 1/4 of a turn rotation of the cutoff solenoid, of artificially extending the accelerator rod etc. And during driving you can see how suddenly stepping on the gas gives you a lean condition, and how modulating the accelerator changes the resulting exhaust gas composition. Since I use a single wire sensor it takes a few minutes till the sensor gets up to operating temperature. From that point on the readings become really consistent. A heated sensor might be more useful for tuning the mixture during warmup.

Unfortunately as Dean mentioned, the system has a pretty narrow range of operation but ideally the car should be running within that range, and definitely towards the middle of it. I can smell unburnt gas in my exhaust when the rich LEDs light up and feel the car hesitating when the red (lean) LEDs light up. It might not be the perfect setup for designing a custom fuel delivery curve, but it should be adequate for fine-tuning a system that's most of the way there.

Vangelis

[DriverJon]

Hey Dean...

Was wondering which type (2 wire, 3 wire) of sensor you were using, and where did you have it mounted?

From my researching of O2 sensors it sounds like they should have a fairly quick response time (>1 sec) and that makes sense for them to be able to be part of a closed loop control system that has to respond to fairly quick throttle changes.

I'm wondering if your placement somehow might have resulted in the sensor not getting up to full operating temp, and thereby being sluggish and perhaps less sensitive..

On the other hand, if it reaches it's standard voltage does that say it's up to temp? (not sure on this)

I did some reading a while ago, have to check back on the characteristics.

From what I've read in tech documents, and personal reports, it sounds like it SHOULD (famous word) be pretty sensitive and responsive.

[Roadtrip]

There really isn't any reason that you can't get a correct running mixture with the FCS adjustment even on a well worn pump. I'm assuming that the plungers and bores aren't worn to the point of leaking gas either. Remember the running mixture is set at 2500 rpm. Start with 10 turns in on the FCS and adjust from there. The FCS just pushes down on a lever that changes the fulcrum of the rack and compensator link. You should be able to lean it down to "fluffy" and no power. Maybe your pump logic section is really corroded up, but to be that bad, I doubt the pump would function well at all.

If it's running ok now, but sooting the plugs a little, try leaning an eighth to a quarter turn or so and drive it to see if the plugs to sooty again. Repeat as necessary.

[Roadtrip]

Before you give up on the pump, remove the rear inspection plate and using an inspection mirror, see if you can determin which "notch" the rack spring is in. It should be in about the 7th notch from the top. If it's higher than that, and your not in Death Valley, then that could be a cause of rich running. I would suspect a maladjusted barometric compensator. To check, I'd pull the compensator and use an accurate caliper to be sure it's in spec for the present pressure altitude.