Stay with FI, or go to Carbs?

Road-race engines and ancillaries - general discussion
Post Reply
vandor
Posts: 108
Joined: November 30th, 2006, 3:24 pm
Location: Texas, USA
Contact:

Stay with FI, or go to Carbs?

Post by vandor »

I would like to know if there is anybody who has been able to get an engine with hot cams, individual throttle bodies, and programmable injection to run well. I just talked to someone who has the exact same setup as what I am working on (Fiat twincam engine, TWM 40mm throttle bodies, hot cams, SDS Fuel Injection), and he said the injection worked very well on milder engines, but once he switched to throttle bodies and hot cams it became very difficult to make everything work. Of course the car runs 'well enough', and has good top end power, but the reason we went with FI is because we wanted driveability like a modern car. Otherwise one could just use 44 IDFs carbs... Two other people at last year's big Fiat meet had programmable injection systems that had some driveability issues.

My question is, is it really worth it? We are thinking what to do next, spend a bunch of time tuning the injection, use milder cams, or go to carbs?

Any ideas, opinions, experiences would be appreciated.
Thanks,

Csaba
kpsig
Posts: 35
Joined: May 10th, 2008, 6:41 pm
Location: Greece

Re: Stay with FI, or go to Carbs?

Post by kpsig »

I do have some experience, although from a hugely different engine. In any case, I believe that the main points apply to many setups:
With mild cams you can use MAP sensor, which simplifies things a lot, both in terms of programming and driveability'; however, going to wild cams and throttle bodies means that MAP Sensor is out of the question, at least not until you can reach 3-4000 rpm. This is because vacuum readings are erroneous, full of noise and at the end, useless for the ECU. So you have to turn to Alpha N programming, meaning throttle position and load vs rpm.
IT is really tricky to make it work, a lot more than setting up carbs, but the end result is better than using carbs, both in torque delivery and fuel management.
On the other hand, you could use "Hybrid" Alpha N programming, although this is used for high revving motorcycles and turbo motors (eg Ferrari F40). With "Hybrid" you can actually enter MAP Readings as an additional factor to fueling from a certain point of rpm and upwards, depending on the engine.
My engine (Peugeot 1.6 8V TU from 106 Rallye Mk2) has a 12:7/1 compression, cam with 285 duration and 11,25mm lift (a lot wilder than OEM), throttle bodies 40mm coming from HOnda CBR 600 bike, highly modified head, coil pack from Punto GT and is using 8 injectors. 4 OEM at the original position and the 4 ones from the bike's throttle bodies, which start to operate in parallel to the OEM ones at a reduced duty cycle at 4100 rpm.
The result of this slightly complex setup? Very good torque from low down, steady idle at 950 rpm and very good powerband with peak at 7700rpm.It actually behaves a lot like a stock motor.
The key to everything is to use wideband sensor and spend something like 300 liters of fuel for tests on the road and in the circuit, in all traffic circumstances, temperatures and speeds.
vandor
Posts: 108
Joined: November 30th, 2006, 3:24 pm
Location: Texas, USA
Contact:

Re: Stay with FI, or go to Carbs?

Post by vandor »

Hi,

Thanks for your reply, that means there is hope :-) Yes, we tried the MAP sensor first, as the EFI maker said we should see if that would work. We were able to get a steady signal by putting a restrictor in the vacuum line, but the main problem was that the vacuum range was too small and we did not have adjustment points. The pressure range was less than half that of a normal engine.
We recently switched to TPS based sensing, and have not gotten it to run properly yet. I just installed a new intake manifold aswell, and I am having a horrible time trying to get the car to idle at all. What is weird is that these cams are not that hot, they are not race cams. They have a published 260 degree duration at 0.050" (1.25mm) clearance.
Possibly part of the problem is that this FI system does not use a table, rather it has a set of values for RPM, and another for TPS, and it merges the two to calculate the pulse width.
Thanks for your comments!

Csaba
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Stay with FI, or go to Carbs?

Post by Guy Croft »

What folk want to do is their affair, that I know.

But my advice on setups from the outset has been widely sought for many years and so I will say here and now that whilst I know prefectly well that FI has a 'certain magic' (bit like 16v heads and a few other things) the cost v benefit really only stacks up if you are competing at National level and must have the 10% or so extra power that the systems can give. I'm not suggesting it is the case here and I'm for open discussion but most often FI is a 'must have' among those who, frankly, have the least experience of building and driving tuned engines and who in almostevery case would be better off putting their meagre resources into much more important things that will actually ensure the long-term survivability of the package whatever its output. Few understand the loads imposed on even the most lowly engines on er, 'just trackdays' and argue to me that the standard clutch/cooling system/lubrication system will be 'good enough'. It never is.

The cost v benefit in ALL other FI applications does not 'stack up', not at all and sadly there is a large number of 'dyno operators' making a handsome living out of selling and 'setting-up' FI systems that, again, frankly, have no business being on the car in the first place. And as for the overall setup - since so much time-consuming calibration has to be done (because current systems cannot cope on their own for variations in demand with merely trim & Lambda compensation) you can easily spend a huge amount of time and money setting-up on an engine which is nowhere near optimised in terms of CR, cam timing and inlet-ex configuration. And how many folk know enough about engines to make the assertion that the core material is worth it? Notmany I can tell you. Thus for all that effort you only ever get a '3/8' solution. These are real issues that you MUST consider.

If FI had been invented before carburettors it would never have gained a foothold in the market place.

GC
kpsig
Posts: 35
Joined: May 10th, 2008, 6:41 pm
Location: Greece

Re: Stay with FI, or go to Carbs?

Post by kpsig »

I cannot disagree with what Mr. Croft says; and I can add to this that you should first consider how confident you feel with understanding all aspects of a modern injection setup, especially electronics, cabling and evaluation of results in relevance to your VE and ignition tables. It can be very surprising that even the smallest changes for example in dwell settings, can lead to better performance or burning out your coils and/or drivers of ECU.
In any case, if you or your tuner are really determined to make things work, test a lot on the track and read manuals and theory, the end results can be very satisfying.
vandor
Posts: 108
Joined: November 30th, 2006, 3:24 pm
Location: Texas, USA
Contact:

Re: Stay with FI, or go to Carbs?

Post by vandor »

Hi,

Yes, I agree that the cost vs. benefit ratios is rather small. But as you said, some people just have to have that last 10% :-)
The idea was that FI will give the engine a modern-car smoothness, combined with performance. We have learned that just like with carbs, it's a trade off. We can either have a smooth engine, or performance. The FI manufacturer blames most of our problems on the hot cams, but I have my doubts.
I've set up two other high-performance engines with programmable FI, and have not had nearly this many problems. Both of those used the stock manifold with a single throttle plate, which makes things a lot simpler.
If we can't make it run well enough then we are considering going to 44IDF carbs.
Thanks,

Csaba
kpsig
Posts: 35
Joined: May 10th, 2008, 6:41 pm
Location: Greece

Re: Stay with FI, or go to Carbs?

Post by kpsig »

Keep in mind that you can have everyday drivability AND economy and then performance through switchable maps.
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Stay with FI, or go to Carbs?

Post by Guy Croft »

I don't know the cams and I don't know the FI system but certainly a reputable FI system will have no problem with any cam type if it's a proven profile and - provided (and it's a big proviso) - the lengths of header and inlet tract do not create adverse pressure wave effects in-culinder and up the inlet that upset the filling phase. The best FI in the world will not give 'more power' if things are 'fighting each other' there, though of course one can calibrate to make 'the best of it'. But in the latter case one could hardly argue that the unit was optimised. You can't do that with carbs of course, but why anyone should wish to do 'compensatory tuning' on a core engine that is way-off par - I can't imagine.

More details and photos pleasa!!

G
vandor
Posts: 108
Joined: November 30th, 2006, 3:24 pm
Location: Texas, USA
Contact:

Re: Stay with FI, or go to Carbs?

Post by vandor »

Here are the details: Fiat 2000cc 8 valve engine. 10:1 CR with forged pistons. 43/37mm in/ex valves, multi angle valve job, ported head.
"42/82 82/42" cams, 304º duration at running clearance, 260º at .050" (1.25mm) clearance. 84º overlap, but we have adjustable camgears, so overlap may be different once we arrive at our ideal setup. 10.6mm nominal lift, 110º lobe centers.

Exhaust is by the Allison 'snakepit' header, see pic. I can't remember if the exhaust is 2" or 2.25" (50 or 57mm).

Just put on a new intake manifold, which is custom made and puts the throttle bodies (TBs) at a 12º tilt from the vertical, so the air does not have to turn 90º under the TBs, only 78º. Also gives more clearance for the injectors. The TBs are made by TWM and have 40mm throttle plates. Injectors are 32 lb/hr (310 cc/min), just replaced 40 lb/hr (388 cc/min) ones.

The fuel injection system is made by SDS from Canada. It uses a crank position sensor and it controls both fuel and spark (two coil waste-spark system). It has the usual array of sensors: Engine Temp, Air Temp, MAP, TPS. It can be set to use mainly the MAP and RPM to calculate pulse width, and only use TPS for the 'accelerator pump' function, or not use MAP, and calculate the pulse width from TPS and RPM.

One thing that is a bit different about this FI system is that instead of filling in values in a chart (say, with RPM points on one axis, and MAP or TPS points on the other), one fills in values for RPM (every 250 rpm), and valus for MAP (or TPS, if switched to TPS mode). I think this gives much fewer data points, which may be our problem.

They told us to try and use the MAP, and only switch to TPS based fueling if absolutely needed. We were able to get a steady vacuum signal to the MAP sensor by placing a 1mm diameter restrictor in the hose, but the vacuum range of the engine did not give enough data points to give proper mapping. A stock engine may drive at partial load at 10 inHg (250 mmHg) vacuum, which gives about 15 data points between partial load and full throttle. However out engine only has ~4 inHg (100mm Hg) of vacuum at partial throttle, so we only had about 5 data point from quarter throttle to full throttle, which were not enough to fine tune part throttle operation. Mind you it took us a few trips to the dyno to figure that out, plus some ruined rings (read: engine rebuild) from running too rich.

So we switched to TPS/RPM based fueling and got the car running fairly well, still needed work on part throttle acceleration.
That is when I rebuilt the engine, had the new intake manifold made, and changed the injectors to the smaller size. I increased fuel pressure to partially make up for the smaller injectors, and started the new engine. The idle is erratic and air/fuel ratio bounces wildly between too rich and too lean. Sometimes it will lean out to the point the engine stalls. Other times AFR is 11:1. I can't figure out how can the injector pulse width be nearly the same, yet AFR vary so wildly. I am suspecting a vacuum leak, likely between the throttle bodies and the manifold. I tried spraying carburetor cleaner on that area to check for leaks, but the idle is too erratic to tell if it is changing by itself or because of the carb cleaner spray.
The throttle bodies seal with an o-ring, but maybe it is not thick enough? Yesterday I applied silicone gasket maker to the bottom of the throttle bodies, which requires 24 hours to dry. Today the roads are covered in ice and snow, so it will be a few days before I can get to it again.

I have 2 videos on Youtube showing the AFR gauge at idle, Guy, can I post the links?

There were other problems we had throughout the tuning process. The first was that at idle, and sometimes at other rpm ranges, there seemed to be no 'sweet-spot', meaning an AFR where the engine runs noticeably better. Seems like it was either too rich or too lean, but just making a little change would change the AFR a lot. This also happened on the dyno during full throttle runs. The FI system did not seem to react the way it is supposed to. A slight change in an RPM value would make a big change in the AFR at times, at other times it hardly made a change. This inconsistency drove us mad. We even went to a different dyno, but the AFR/RPM we got there was nearly identical to the one we got on the first dyno.

During the 4th or 5th (they all wash together now) trip to the dyno I was finally able to adjust the fuel values where AFR at full throttle was 13:1 throughout the RPM range. Which leads to another weirdness, even at that AFR, the car had some black smoke coming from the exhaust. Earlier when we tried adjusting the AFR to 12:1-12.5:1 at WOT, at two specific RPMs the exhaust would belch a lot of black smoke for an instant. From everything I read that is not so rich that a lot of smoke would be visible, and there were no spikes in the AFR/RPM chart, but the smoke was definitely there. And the piston rings definitely were ruined. We looked for possibilities for electromagnetic interference, watched the sensor outputs, but everything was fine.

Another 'anomaly' is that maximun Hp occurs at too low an rpm, around 6000, whereas I would expect such a setup to top out at close to 7000 rpm. I suspect it may be a poor joint from the header to the downpipe, where the exhaust shop slipped a smaller diameter pipe into the larger header collector, and the lip there may be killing airflow. First priority is to have that redone, but at this point I want to get the idle stable and be sure the AFR is not too rich, as I don't want to ruin another set of rings.

Just as an aside, the car belongs to my friend and customer Jeff, who is too busy working and traveling to do most of the work, but he is just as hands-on as I am. This car was going to be an experiment, and then I would duplicate the setup on my Spider, but seems like we are running into too many 'weird problems'! (one block leaked coolant from ALL the new freezeplugs!)
Any suggestions would be appreciated.
Thanks,

Csaba
Attachments
Headers
Headers
camcvrs.JPG (142.68 KiB) Viewed 13058 times
Headers from a different angle
Headers from a different angle
header8.jpg (126.87 KiB) Viewed 13058 times
Engine left side
Engine left side
engine1.JPG (97.24 KiB) Viewed 13058 times
kpsig
Posts: 35
Joined: May 10th, 2008, 6:41 pm
Location: Greece

Re: Stay with FI, or go to Carbs?

Post by kpsig »

OK, I will make some suggestions coming from my own experience
1. Check your wideband sensor. Calibrate it and compare its readings to a different one.
2. Reallocate you sensor. If it is located at the exhaust output, it might read wrong, although your compression ratio and other characteristics of your engine are not that peculiar in order to create "strange" pulses in the exhaust section. You should try making and insert point close to the beginning of your exhaust's one-tube section.
3. I think that you could have some problem with your ignition output hardware or software. Either your coils and/or cables have some problem OR the circuit of your ECU does not drive them well (most probable). TRY changing your Dwell settings. You could get the proper ones from the manufacturer of your coils. Wrong dwell settings can create the kind of problems you describe.
4. Injectors.... Make sure that they are also driven properly from your ECU. I mean voltage width, pulse width in relation to how many squirts you have per cycle. If you have many squirts (eg 4) it is most probable that your PW time is too short for the injectors to cope. also make sure that they operate properly.
5. When going to throttle bodies and wild cams, I strongly suggest your forget MAP and use only Alpha N (RPM/Load/TPS). You might use MAP as a secondary input from a certain RPM threshold upwards, a threshold to be determined by a lot of trials; However, such a setup is useable for turbocharged cars or motorcycles revving above 10000 rpm.
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Stay with FI, or go to Carbs?

Post by Guy Croft »

Model post!

well done Csaba, I hope we can help here.

I would be concerned about wildly fluctuating inlet pressure condition (due to waves) and the influence of that on the MAP sensor. I would not use it just as Kpsig suggests. Those things are best reserved for turbo units. If you have a brake booster installed (servo) make sure the vacuum line has a non-return valve.

Power too low might suggest insufficient ignition timing or mixture - too rich or lean will do that. I assume you do not have a back-pressure problem with the muffler (silencer) - it would need to be a 2.25" bore straight-thru large volume design.


I don't think you've got an inlet-ex mismatch problem. As for leaking core plugs (freeze plugs) I always use OE Fiat (and only them!) and with a bit of silicone sealant and use of the OE tool to fit them I have never had a leak. They are prone to leak if you don't use silicone or distort during fitting.

Post the link by all means, thanks for asking.

G
Rich Ellingham
Posts: 118
Joined: June 23rd, 2006, 6:54 am
Location: Glasgow, UK
Contact:

Re: Stay with FI, or go to Carbs?

Post by Rich Ellingham »

Csaba, a couple of points from my experience, MAP sensing on a N/A race engine is a NO go situation and be sure not to just ignore the MAP input but actually remove the MAP sensor, one of my fellow competitors had this issue on his DTA equipped MG ZR where the MAP was connected but the ECU was told to ignore the input however EM spikes were coming from the MAP and affecting the ECU calculations (tested on oscilloscope).

Make sure all the throttle plates are in line and opening to the same amounts and the throttle stop screws or bypass screws are the same for each body.

The next point is my most important one - how is the ECU controlling Idle? My Omex uses a Spark Scatter method to control bodies without an external electronic idle valve. I have to admit I did not fully understand the setting variables but I do know the idle speed is controlled by the ecu advancing and retarding the spark timing for the cylinders this will drastically alter an A/F ratio measured during idle scatter. I had to go to OMEX to get my unit re-flashed with a newer software, Andy Cornock who takes care of many programming functions there noticed the spark scatter variables were a long way from 'normal' range. I had always had a poor idle usually ending up in stalling, once he had put in some average variables the idle was significantly improved and stalling no longer occurred. I do not have the MAP and Software on the machine I am currently using to give more detail but I would estimate you may have a similar issue which maybe you could look at in your software.

Regards
Rich
Nobby
Posts: 87
Joined: August 24th, 2010, 11:58 am
Contact:

Re: Stay with FI, or go to Carbs?

Post by Nobby »

Thats interesting Rich. Sorry to go off topic - but I've been wondering how cars with individual TBs cope without having a single ICV to regulate things. That clears things up a little. One more question - how does the air actually get in then? Do the throttle's not fully close?
vandor
Posts: 108
Joined: November 30th, 2006, 3:24 pm
Location: Texas, USA
Contact:

Re: Stay with FI, or go to Carbs?

Post by vandor »

Hi Kpsig,

Thank you for your responses, I have some answers and questions:

1. We have used 4 different wideband sensors. Two at the dyno, and two for our own gauge, and all seem to agree. The one I have in the car now is brand new.

2. Actually the first header we had (same design) was of poor quality and cracked a lot, so we had to get a second one, and the oxygen sensors is now in a different place that on the first one. It is close to where the 4 tubes go into one, I can take a pic.

3. Do you know what is the proper dwell setting? I think the stock Fiat one for points ignition was 55º. I don't know how dwell would work on a lost-spark system (the electricity goes from the coil to the plug, to ground, to the other plug, and back to the coil, so there are always 2 plugs fired at once, 1 and 4, or 2 and 3).

4. Do you know how do I do those checks? The injectors are batch fired once a revolution.

5. We are not using the MAP sensor any more.

Guy,

I will look at the check valve for the booster, I actually want to disconnect all the vacuum lines to eliminate them as a possibility of leaks.

We had several dyno session and we experimented with a wide range of mixtures and ignition timing. We were surprised that mixture did not seem to have much effect on power, ie., no power change between 12:1 or 13:1 AFR.
We kept adding more ignition timing, but we actually lost power! I need to find my notes, but I think max power was with ~29 degrees of advance. I remember you saying that it's too little, but we lost power with more advance.

I need to look at the muffler, I do not think it is a straight through design, it's some kind of universal performance mullfer (Borla, or somesuch). I forget if there is a resonator in the system, but my experience is that a single straight-through muffler alone does not offer enough silencing for regular use. Not without earplugs :-)
I think we have a bottleneck or roughness inside the pipe at the header/downpipe junction, and I will have that redone as soon as I have the car running again.

Here are the Youtube links. The first one shows the mixture changing at idle. I made this for the FI system manufacturer, and you can see me scroll through screens on the controller. First it shows that Closed Loop is OFF, then I scroll through the GAUGE screens to show that none of the sensor inputs are changing.

http://www.youtube.com/watch?v=t8l3-4IqEi8

The second one shows the mixture changing (at the beginning I had to rev the engine to keep it from dying), and eventually the engine stalls.

http://www.youtube.com/watch?v=6ftkaqswBBc

The roads are still icy, but I may try and go look at the car later.
Thanks,

Csaba
Guy Croft
Site Admin
Posts: 5039
Joined: June 18th, 2006, 9:31 am
Location: Bedford, UK
Contact:

Re: Stay with FI, or go to Carbs?

Post by Guy Croft »

Forget 'resonators' as mid-position silencers are sometimes called, they can sometimes offer a small mid-range torque gain but will chop off the top-end. All you want is straight-thru and at the back of the car. The influence of joints and welds etc in the pipe, well, I'd be astonished if that was the problem, it would have to restrict the pipework very badly to cause what you're seeing, eg pipe crushed so it was half the internal area it should be, 1.75-2" ID from the end of the header collector to muffler...

You are wrong about noise and the influence of straight-thru mufflers. A good race design like the George Polley AX891 in my new book will chop off the harsh noise but at the same time not raise the static back-pressure. You can use 2 and be sure of flying thru any noise tests I can think of but at the same time see lower back-pressure in the tailpipe than even just one with static baffles or chambers of any design and a tortuous route.

I think your engine is struggling to inhale air to be honest. The insensitivity of the engine to mixture and ignition and low rpm @ peak power suggest very poor cylinder filling. You might give this some serious thought because the best FI in the world is not going to give super results ifthe core engine is to blame. As for ignition timing check out the ignition calibration on page 134 of my first book, sure there is much variation even at full throttle. Exactly what size are the throttle bodies?

I assume these cams are billet and of reputable origin and the supplier has dyno data to support their use & sale?

What is the CR, valve size? Has the head been gasflowed?

G
Post Reply

Who is online

Users browsing this forum: No registered users and 32 guests