Thanks for the info Scott (winders). I will increase the fuel pressure. Regarding a fuel damper, I am still running the fuel accumulator from the the CIS system - I think the accumulator is intended to smooth out pressure fluctuations?

Thanks for the info Dave (montauk). Yes, the MegaSquirt does have a screen to input the injector dead time. Not knowing any better, I have been using the default value of 1 ms. But I will look into this further and adjust as necessary....

http://forums.pelicanparts.com/uploa...1750511910.jpg

I looked at the last data log I recorded and yes, it looks like I will be running out of pulse width by the time I get to max rpm as the commanded pulse width at is 13.2 ms (1.0 ms dead time + 12.2 ms of fuel) at less than 6000 rpm.

http://forums.pelicanparts.com/uploa...1750511998.jpg

I will be making the fuel pressure adjustment, then make a testing run with auto tune. Then I will schedule another dyno session so Gary (Bains Tuning) can dial in the VE table.

From what I understand, the accumulator dampens pulses from the fuel pump but would not dampen pulses from fuel injectors. Are you using the stock CIS fuel pump? If you are not, I would probably ditch the accumulator and get a damper somewhere near the fuel rails.

Scott,

I'm still running the OE CIS fuel pump, accumulator, and fuel filter set up. My current fuel system (based on the installation diagrams with my kit) is a dead leg configuration. My current FPR has 2 in/out ports plus a return port at the bottom.

I know from reading some of your posts on the subject of fuel system design that you recommend a flow through system as the dead leg system can be prone to vapor lock, etc.

I think once I get the fuel pressure turned up to the 55-60 psi range and see the effect on VE table cell values and injector pulse widths, then I'll start planning the next step with my fuel system set up using a flow through design with the Radium FPRD-RA fuel pressure regulator with dampening.

Dan,

This is the guide/ sheet that I used to calculate injector dead times. Not sure if Megasquirt has the same sort of hardware test page.

https://docs.google.com/spreadsheets/d/1YOHLYERVBvFJWPd9v-q5CJqnmSxO_cNkOPxhmKbs7Vc/edit?gid=692277831#gid=692277831

VE Table Cell Values After Fuel Pressure Adjustment
 

This afternoon I adjusted the fuel pressure with the engine running from 38 psi to 55 psi. Almost immediately. I observed that the AFR at idle dropped from 14:1 to 12.7:1. On the VE table, I had to adjust the cell values in the idle region from 35 to 29 in order to bring the AFR back to my 14:1 target.

http://forums.pelicanparts.com/uploa...1750550820.jpg

Next, I took the car out for a testing run with TunerStudio running "Tune Analyze Live" (auto tune). I drove the car north out of Milpitas on I-680 running in 4 gear and accelerating up to near 5000 rpm whenever traffic would permit until I got to the top of the Sunol grade where I burned the auto tune changes to the ECU, saved the changes to the tune file, shut down my laptop and drove back down the hill and home.

Here is a close up of the VE table difference report: the left table is 55 psi fuel pressure and the right table is 38 psi fuel pressure (blue numbers on the right table are values that decreased, black values did not change as I was mostly running 3000-5000 rpm with 10-30 % TPS load).

The changes to the VE table cell values to achieve the same target AFR after the fuel pressure increase are significant, there are deltas of 10 or more.

http://forums.pelicanparts.com/uploa...1750551390.jpg

Next stop on this journey is to the dyno shop. I have already paid the reservation deposit and am waiting for an appointment date / time.

So in my last post I had turned up the fuel pressure from 38 psi to 55 psi, saw a reduction in AFR and adjusted for it by decreasing the cell values in the VE table.

Today, I went out and turned the fuel pressure up from 55 psi to 60 psi as Scott (winders) had recommended.

Happily, the AFR did NOT change so now I can have the benefit of an improved spray pattern and know that variations in the fuel pressure well above the 43.5 psi max injector flow pressure are not effecting the amount of fuel delivered per injector cycle.

I have a dyno tune session scheduled for next Tuesday and I will report more test results then.

Living and learning....

Dan,

Just a quick heads-up about your injectors—dead times are usually rated at 43.5 PSI, which is the pressure they're designed to operate at. In fact, the Maxxecu data lists 43.5PSI with their dead times. When you run them at 55 PSI, the injectors have to work harder to open, and that can increase latency. It might not seem like much, but it could affect your tune, especially at idle or during transient conditions.

Unless there's a specific reason you're targeting higher pressure, it might be worth asking your tuner about it. If the injectors are designed for 43.5 PSI, sticking with that baseline keeps things predictable—and makes published specs like dead time more accurate. Just something to consider while you're at the tuners.

Dave

Check the injector’s required pressure as too high of a pressure can also cause increased stress and wear on the injectors, seals, etc which can cause premature failure.

Also, if the pressure is too high (above 55-60) or too low (below 30-35) the spray pattern may be compromised.

As always, it’s injector specific as some are rated for high psi (up to 100psi) but most are not.


Sent from my iPhone using Tapatalk

Dave (montauk) & Julian (Showdown) thank you for your comments.

When I think back yesterday and I changed the fuel pressure from 55 psi to 60 psi, the AFR DID change. The AFR at idle went up from 14.0:1 to 14.5:1 which had to be from an increase in dead time (latency) / decrease in fuel flow time due to the higher fuel pressure as the ECU commands a PW time that is dead time + fuel flow time.

I think I'll turn down the fuel pressure back to 43.5 psi (the gauge dial is course so probably 45 psi will be as close as possible) and see where the AFR comes in, adjust the VE table to achieve my desired 14.0:1 AFR and use the injector dead time Dave (montauk) found for the 0280155831 injectors and leave it alone.

I'll talk with Gary (the dyno tuner) about it on Tuesday before he runs the car on the dyno.

Sigh...

The Bosch 0280155831 is a Bosch EV6EL style injector that can run on systems pressures as high as 8 bar (116 psi). Running them at 60 psi (just over 4 bar) is not a problem in any way shape or form.

Again, the spray pattern is improved at 60 psi AND this reduces the duty cycle which was said to be at issue.

https://www.bosch-motorsport.com/media/downloads/injection_valve_ev_6_datasheet_51_en_2775912459.pd f

I'm done....

Thank you Scott (winders). I have been flip flopping. Yes, the 0280155831 injector is a Bosch EV6EL. I am staying at 60 psi. Now I'm done.

Scott,
The info sheet you posted lists flow rates at 3 bar. The dead time data I found is at 3 bar. If you run at a higher pressure, how do you set the dead times?
Dave

There are tables out there. I don't have specifics for the Bosch EV6 injectors.

I used Injector Dynamics ID1050x injectors with my MoTec M150 and Injector Dynamics supplied characteristics tables that automatically adjusted for changes in fuel pressure.

Here is some Bosch injector info:

https://www.bosch-motorsport.de/content/downloads/Raceparts/Resources/pdf/Basic%20Information_67843467_Injection_Valve_EV_14 .pdf


http://forums.pelicanparts.com/uploa...1750994397.jpg

For the 0280155831 injector: 31.4 gal/hr @ 3 bar (43.5 psi) becomes - FR2 = (4 bar / 3 bar)^1/2 x 31.4 = 36.2 gal/hr at 4 bar (58 psi), and
FR2 = (37 psi / 43.5 psi)^1/2 x 31.4 = 29.0 gal/hr at 37 psi

Note that with pressures higher than 3 bar, voltage to the injectors must be increased. So you’ll have to do something to drive more (more than 12v?) to the injectors if you raise the pressure above 3 bar…. That makes sense but also seems quite complicated and I don’t know if MegaSquirt has that capability- not sure if any ECU does… dunno.


Sent from my iPhone using Tapatalk

It makes sense. Higher pressure requires more work by the solenoid to open. Does running at 4 or 5 bar offer a measurable advantage for naturally aspirated engines?

Plus, you'll still need to create test bed to figure out the dead times if you can't find them published somewhere. One tuner told me that they prefer using the actual ECU to develop dead times not just a voltage range. That seems a bit over the top but I guess with high end stuff, you can't be too accurate.

I'm running with my ECU monitoring fuel pressure. My pressure regulator has the MAP connection to adjust fuel pressure based on manifold pressure but I decided not to connect it. If it turns out that I need it, I'll hook it up. So far, I haven't seen any issues. My logs show fuel pressure at 3 bar with extremely little variation regardless of operating conditions.

Excellent find.

On a speed density (a true volumetric efficiency fuel model) system you have to tell the ecu what the fuel pressure is either by inserting that pressure manually or having a fuel pressure transducer to feed the pressure in real time to the ecu, along with as complete as possible injector characterization, fuel choice chemistry, etc so that it ultimately knows the exact fuel mass being injected into the motor per injector pw.

The VE table in conjunction with environmental data and engine size is to tell the ecu how much air mass is in the combustion chamber.

And the Lambda table ultimately dictates what the end PW will be to meet the target air/fuel mixture.

I say all this in simplistic form as there are plenty of resources that explain in greater detail by people that are blessed with gifts in communication.

If you adjust the fuel pressure without telling the ecu that the operating fuel pressure has changed then it will have a direct effect on lambda (air/fuel mixture) that would require a VE table change to meet target since the ecu has no idea what the actual fuel mass is anymore. If you had a transducer on the fuel rail feeding the pressure in real time into the fuel calculation OR update the fuel pressure value to what guage pressure is then the adjustment in pressure will not have an effect on the air fuel mixture or the value in the VE table as the calibration will change the PW to meet the target lambda in the background.

With that being said, increase in fuel pressure from a semi typical 3 bar relative pressure will have greater atomization generally but that does depend on the injector design as well and when you get to the point of diminishing returns as an increase in pressure will also heat the fuel more, increase amperage at the pump and decrease flow at the pump which is fine if you have enough head room flow wise with your pump choice and you have wiring appropriate for the amperage the pump runs.

I highly recommend a fuel pulse damper on a dead headed rail and still recommend them for a flow through rail. You can tune around the injector pulse width "hammering" in most cases but the VE table will end up ugly as all hell and will not represent the true air mass in the motor at smaller pulse widths.

And another small point with regards to the injectors themselves. A good injector company will provide data that will encompass dead times and flow at various fuel pressures and voltages if that injector was designed for aftermarket use (i.e Injector Dynamics). If the injector was originally spec'd for an OEM manufacture then that injector data is proprietary and will not be shared so you have to either go by data provided by the internet which is highly suspect usually, provided by a vendor, or send them out to be characterized. Also this data changes depending on the hardware of your chosen ECU as not all injector drivers operate the same, this is a big selling point to Injector Dynamics as they provide ECU specific characterization data.

Thanks for posting this. I appreciate it.
Dave

Here is some more data from an injector supplier who has worked with Bosch Motorsports.

The data is for 2 different product series, and are tables of injector dead times for pressure versus battery voltage. I am putting this out here for information only - the data shouldn't be used as is, it is just to illustrate some of the discussion points that have been made in this thread regarding fuel injectors and their dead time (or latency or offset).

Discussion points:
1. Higher battery voltage results in shorter dead times.
2. Higher fuel pressure results in longer dead times.

http://forums.pelicanparts.com/uploa...1751074794.jpg

http://forums.pelicanparts.com/uploa...1751074794.jpg

What jumps out at me from this data is that fuel pressure changes (40 to 100 psi) at lower battery voltages (8 V) result in much larger changes in dead time (delta of approx 1.0 ms) than the same fuel pressure changes (40 to 100 psi) at higher battery voltages (16 V) (delta of approx 0.1 ms).

Interesting.

The bottomline is that at normal street battery voltages the difference in dead time when going from 43.5 psi to 60 psi is quite small and nothing to get excited about.

this is a little off topic here as I am doing carbs but should have some cross over info

1978 ROW 911 SC with large port heads (39mm) - planning to use PMO intake headers (PM-O911-0) which are the 46/39mm Tall for CIS.

looking at gaskets and heat insulators - for CIS heads there appears to be only 38mm or 40mm - which is the preferred size for my application - go undersize or go oversize

my apologies for being a bit off topic but this thread is a lot more active than the one I started.

I would go with the 40mm gaskets and keep them out of the air flow.

would there be any concern with the .5mm exposed portion of the head...

maybe a stupid question but there it is

I'd either go 40mm and keep that .5mm out of the airflow as Dan suggested-it's just going to create turbulence where you don't want it, or have custom gaskets made. A laser cutter and a scanner and it's a piece of cake. I'm sure you could have it done online too.

Dyno Tune After Fuel Pressure Increase
 

I went to the dyno tuning shop and got my car tuned after I increased the fuel pressure from 37 psi to 50 psi. Max injector flow has gone from 29.0 lbs/hr to 33.7 lbs/hr.

The injector % duty cycle at 7200 rpm was 89% with 37 psi, now the % duty cycle at 7200 rpm is 72% with 50 psi.

Difference report for the VE table, left side graph is 50 psi & right side graph is 37 psi.

http://forums.pelicanparts.com/uploa...1751430782.jpg

The engine made a little bit more hp at the top with 1 degree more advance at 6500 - 7500 rpm as shown in the difference report for the Ignition table, left side graph is 50 psi & right side graph is 37 psi.

http://forums.pelicanparts.com/uploa...1751430988.jpg

The dyno graph shows max wheel torque unchanged at 206.5 ft-lbs while max wheel hp increased from 232.8 hp @ 6600 rpm to 238.1 hp @ 7100 rpm.

http://forums.pelicanparts.com/uploa...1751431155.jpg

Thank you Dave (montauk) for asking me why I was running with the fuel pressure so low. From the resulting discussion on this thread and observations made while working on my tune, I have learned a bunch and gotten a better performing engine as a result.

Dan,
That's great news. Thanks for posting!!
Dave

Dan
Did you end up installing a fuel pulse damper on your setup?
I’ve been chasing some odd behavior in my long-term fuel trim table—huge swings between adjacent cells, like the ones in the screenshot I attached. From what I understand, that kind of variation really shouldn’t be happening under steady-state conditions.
After reading through some of what Rivet and Winders shared, I started wondering whether those big jumps could be due to slight fluctuations in fuel pressure. During driving, my gauge reads a pretty consistent 45 PSI, but when I reviewed the fuel pressure graph, there are some subtle oscillations riding on top of that flat line.
I decided to order a Holley 12-1004 AN6 fuel damper and plan to install it this weekend. One gripe though—the fitting uses a male AN6 ORB connection, so it relies on an O-ring rather than the usual 37° flare. Not ideal, but I’m hoping it takes the edge off those pressure ripples and smooths out the trims.
http://forums.pelicanparts.com/uploa...1751542030.png

No, I am running the CIS Fuel Accumulator between the fuel pump and the FPR, then the fuel supply deadheads to the fuel rails.

Regarding your long term fuel trim screenshot, do those dips move around from one run to another (that is moving to different rpm / TPS conditions)?

My experience with OEM emissions systems is that long term trim changes more slowly. If the LTFT dip persists, then it would indicate the base fuel trim map is too rich in those areas - 1500 to 2500 rpm at low TPS is like pulling away from a stop with light throttle then cruising around the neighborhood?

Anyway, back to my fuel set up and observed fuel pressure changes. While Gary Bains was running the car on the dyno, I stood at the back of the car and watched the fuel pressure readings (my MegaSquirt 2 doesn't have an input for fuel pressure).

When Gary floored the gas (WOT) at 2000 rpm, the fuel pressure would drop from 50 to 48 psi then recover to 50 psi after 1 second or so and then stay steady at 50 psi through the rest of the WOT ramp run until after 7000 rpm and the fuel pressure would start to drop slightly (not sure of the amount, only that I started to see the gauge slowly moving lower), but then at 7300 rpm Gary ended the ramp run and the fuel pressure went back to 50 psi. That indicates to me that my fuel pump may be getting flow limited above 7000 rpm. If I was a racer, I would be thinking about a higher flow rate fuel pump, but for canyon carving and cars & coffee duty I am okay with what I have.

The graph represents the long term trim after running the car quite a bit and applying those long term corrections to the base map. At this point, I shouldn't be seeing dips like that. You can see that the adjustments around the big dip area are modest. More to play with but I'm enjoying the process.

BTW, I finally drove the car, about 100 miles, on Monday without logging data. The laptop was with me just in case but it wasn't needed.

Results May Vary Esspecially When You're Working On Somebody Else's Car
 

Here is a tale of caution for those of you considering a CIS to ITB EFI conversion.

Earlier this year I met Andy at a local PCA cars & coffee in Soquel. Andy has a 74 911 and he added a wide body kit, rebuilt and swapped in a 3.0L SC motor which he had converted the CIS to ITB EFI. So, given that I also have similar motor work on my 75 Carrera, we hit it off and became friends.

Andy has a friend, Todd, that owns a 79 911 SC 3.0L and Todd saw my ITB EFI set up and decided he wanted that too. Well, Todd paints cars and Andy has a 64 356C that he wanted to get painted (Todd had painted Andy's 74 911 previously and was happy with the result), so Todd and Andy worked out a deal: Todd paints Andy's 356 and Andy converts Todd's SC to ITB EFI.

Andy knows Al Kosmal at X-Faktory.com and decides to buy Al's Option 3 conversion kit (MegaSquirt3 + MS3X expansion board, etc...) because Andy and Todd decide they want to run a coil on plug (COP) set up instead of using the distributor. Andy also decides not to install the Clewett Engineering cam position sensor so this will be a semi-sequential FI with COP waste ignition system.

Since Andy's 74 911 is running an old AEM Infinity ECU, he is not familiar with MegaSquirt and asks me to help with the MegaSquirt tune file set up and VE table tuning duties. Little did I know what I was signing myself up for.

On the first page of Al's installation document is a clear statement about how the target vehicle must be in good running condition or else the ITB EFI installation will become an "advanced" installation requiring more work.

http://forums.pelicanparts.com/uploa...1754964777.jpg

Well, I had never seen Todd's car until I made my first trip down to Andy's house in Salinas. This 79 911 SC looks nice on the outside with a fresh coat of GP White paint and Turbo whale tail spoiler, but beneath that exterior is a car that had obviously been run hard and not well cared for. The exhaust was pretty rusted, the CV prop shafts had inner rubber boots that weren't secured so all the inner CV joint grease was gone, the brakes are completely used up and the shifter bushings and coupler are completely worn out such that the gear shift lever just flops around. Andy corrected the exhaust and CV shaft issues by installing SSI's + a new muffler and new CV prop shafts.

Andy's installation of the ITB's and the EFI + COP hardware and wiring is first rate. He mounted the ECU + a relay board to an aluminum plate and mounted that to the floor pan under the passenger seat. All of the wiring connections are located in the rear shock upper mount cross member area and he used the ground lug on the left rear inner fender (behind where the old fuel accumulator and fuel filter were located) for the EFI ground terminations. Switched 12V is coming from the ignition switch and main relay power (connected to terminal #30 on the relays) is via a 4 gauge cable connected to the battery + terminal at the starter.

The existing wiring is messy as a car alarm system and who knows what else had been added then removed in the past. There are stray wires in multiple places under the front hood near the main fuse panel and under the dash. There is also a ground shut off switch on the battery ground cable with about 30% of the cable strands cut away to allow installation of the shut off switch.

There is also a problem with the alternator which is a replacement unit with an integrated voltage regulator as the original voltage regulator had been removed. The problem is that the generator charge warning light in the dash gauge stays lit even when the starter key switch is in the off position. This turns out to be the reason the battery ground shut off switch was installed.

You probably know where this story is headed....

So before hand, I had set up a tune file using the MS3 sample project in TunerStudio and used the my VE table, AFR table and ignition table values to create those tables for Todd's set up - I rescaled the rpm to 6500 rpm max and I converted the ignition table from 12 x 12 (MS2 standard) to 16 x 16 (MS3 standard). Load sense is Alpha-n using the TPS. The FI is using the MS3X board outputs and is set for semi-sequential. The ignition is using a 36-1 trigger wheel/crankshaft pulley with a Hall effect sensor and a plug to fill the distributor shaft opening in the crankcase (from Clewett Engineering) and Denso style COP units. The ignition outputs are set to COP waste.

I loaded my tune program into the ECU and entered calibration data for the CLT (coolant temp), IAT (manifold air temp) and wideband O2 sensors. We counted the trigger wheel spaces from the case parting line to the crank position sensor, multiplied by 10 and entered the result in the "Tooth #1 Angle(deg BTDC)" field.

Then we opened up the ITB's a bit. Powered up the fuel pump and performed a leak check and regulator pressure adjustment (I set the fuel pressure to 50 psi) without problems.

Then we tried starting the engine - nothing. We spent the next 3 hours checking and retrying. Got a few back fires; otherwise, nothing. By now the battery was low, so we stopped, hooked up the battery charger and called it a day.

I decided to reload / update the ECU firmware so I downloaded the current firmware revision from DIYAutoTune.com to my laptop. I also spent an hour practicing how to perform the Inj/Spark tests in TunerStudio using my car as a test mule.

Several days later, I made the trip back down to Andy's house. I was able to successfully update the MS3 firmware from rev 1.5.2 to rev 1.6.1 and verified the tune file was correct. I was able to successfully test all 6 injectors and all 6 COP units so I knew the injector and COP units were wired in the firing order correctly.

Then we tried starting the engine. This time it started but ran horribly and the TunerStudio dashboard was showing the rpm read out jumping all around and the sync error status light was blinking - it was a mess.

We took a break and headed out to get some lunch. As we went over what we were seeing, we kept coming back to that generator charging light problem. Was the alternator output noisy? I didn't have my oscilloscope with me, but could we isolate the EFI electrical system from the charging system? Then Andy said, why don't we remove the fan belt so the alternator doesn't turn and try starting the engine?

So we went back to Andy's and did that, and the engine started and ran more or less normally, given we still need to verify the timing, sync the ITB's and tune the VE table.

Now a new alternator, new battery ground cable, and a new engine/transaxle ground strap are on order and once received and installed, we can try again to get this CIS to ITB EFI conversion up and running so we can then get Todd's car on the road and use the auto tune function in TunerStudio to dial in the VE table.

Stay tuned....

Alternator info
 

Dan,
I am glad that you and Andy are working together on this one.

Here is good info regarding the install of an alternator with integrated regulator

http://forums.pelicanparts.com/uploa...1754975268.jpg

regards,
al

Yikes!

I was at a show talking to a fella about efi and he seemed convinced that “the computer will fix it” if there were problems as if the computer was an all knowing entity capable of miracles.

I told him that efi will amplify any issues that CIS can hide.


Sent from my iPhone using Tapatalk

CIS can't "hide" anything. CIS is far less flexible than EFI as the ECU, the "computer", that controls the EFI system can make adjustments based on far more variables and, depending on the ECU, can even "autotune" to an extent.

So no, EFI will not "amplify" any issues.

Perhaps "hide" was the wrong word.... Tolerate, compensate, accept...

CIS is a pretty forgiving system and can run with all sorts of problems. Case in point; the hundreds of CIS Porsches out there that run and drive with little quirks, nuisances or deficiencies, but run nonetheless. Air leaks, pressure issues, WUR problems... the cars my not run well but they run and CIS tolerates it. Take that same car and throw EFI on it and the small issues that CIS can tolerate become major issues for EFI. Sure, you can modify the tune to compensate but that's crappy practice and not the correct way of going about things.

EFI depends on the tune which in turn depends on the tuner. There are countless options and features and tweaks that can be done in an EFI system (regardless of the ECU) and knowing how to work all of those is a lot and it's very easy to start chasing one's tail.

Ask any EFI ECU manufacturer, any tuner, any shop and to a T some version of "you have to have a well sorted car before going to EFI" will be their gospel.

But what do I know, feel free to take a crap funning CIS car and throw an EFI system on it and have at it.

CIS is not better than EFI in any way, shape, or form. CIS is just an electromechanical injection system. It does not tolerate or compensate for issues better than EFI. It just doesn't know the issues are there to begin with. EFI systems are far more flexible and make it easier to diagnose problems when they occur.

One thing I’ve come to appreciate is how fault-tolerant modern EFI systems can be. Even with a rough initial tune, the car can run surprisingly well—mine certainly did. And as the tune improves, so does drivability and performance. Dan’s setup is a great example: his car was running well even with low fuel pressure. Modern cars even have a limp home mode to compensate for failing components.

But this isn’t unique to EFI. CIS is also a tuned system, and every component was carefully selected to make the engine run optimally. The Warm-Up Regulator (WUR) is a perfect example—swap in the wrong one, and the engine might still run, but not as it was designed to.

Whether you’re running EFI, CIS, or carbs, the underlying engine condition matters. A well-tuned system can compensate for a lot, but to get the most out of any setup, the engine itself needs to be in great shape. That’s the foundation everything else builds on.