thanks GraemeD - connectors with a back door, learn something every day!
Stock NA1 motor...who is running anything OTHER than OEM fuel injectors?
- Thread starter Speedmaster
- Start date
I just had 410cc RDX injectors installed in my NSX and a chip custom programmed on a dyno using SR5guy's definition.
Using SR5guy's definition, it's impressive to see all the data stored in the PGM-FI's chip. My tuner said it is way beyond what European manufacturers were doing in the early 1990's. There are five fuel maps, three ignition maps, and loads of other variables that can be set. My tuner also said that no one other than a manufacturer can spend the months required to tune every cell of every table to optimize the data set for a certain engine type. The configuration of that engine type includes its injectors. The injectors flow a certain volume of fuel per minute, have a certain lag time after they receive the signal but before they open, spray the fuel and atomize it in a certain way, and the OEM chip is programmed to take all those factors into account. If you put in a different injector, you should modify all data points that are impacted by the injector's different characteristics.
The RDX injectors have a greater lag time than the stock NSX injectors and flow more fuel per minute. My tuner said the NSX's PGM-FI sets some limits regarding injector duration at idle and if you take the 410cc RDX injector lag time into account (which helps the engine run well), getting a stable idle in an NSX is borderline. If you don't take the lag time into account, it's easier to get a good idle but there are some other tradeoffs. If I would do it again and wanted to use RDX injectors, my tuner recommended using the smaller injectors from the later, naturally-aspirated RDXs, assuming they have same spray pattern, atomization, and lag time as the larger injectors in the earlier turbo. It would be easier to get an NSX to idle well with the smaller injectors and they would still be easily large enough at full throttle.
But his overall recommendation was just to stick with the OEM NSX injectors. They're also large enough for naturally-aspirated NSXs (at least with stock displacement) and every cell in the PGM-FI's maps has been optimized to run with them.
Using SR5guy's definition, it's impressive to see all the data stored in the PGM-FI's chip. My tuner said it is way beyond what European manufacturers were doing in the early 1990's. There are five fuel maps, three ignition maps, and loads of other variables that can be set. My tuner also said that no one other than a manufacturer can spend the months required to tune every cell of every table to optimize the data set for a certain engine type. The configuration of that engine type includes its injectors. The injectors flow a certain volume of fuel per minute, have a certain lag time after they receive the signal but before they open, spray the fuel and atomize it in a certain way, and the OEM chip is programmed to take all those factors into account. If you put in a different injector, you should modify all data points that are impacted by the injector's different characteristics.
The RDX injectors have a greater lag time than the stock NSX injectors and flow more fuel per minute. My tuner said the NSX's PGM-FI sets some limits regarding injector duration at idle and if you take the 410cc RDX injector lag time into account (which helps the engine run well), getting a stable idle in an NSX is borderline. If you don't take the lag time into account, it's easier to get a good idle but there are some other tradeoffs. If I would do it again and wanted to use RDX injectors, my tuner recommended using the smaller injectors from the later, naturally-aspirated RDXs, assuming they have same spray pattern, atomization, and lag time as the larger injectors in the earlier turbo. It would be easier to get an NSX to idle well with the smaller injectors and they would still be easily large enough at full throttle.
But his overall recommendation was just to stick with the OEM NSX injectors. They're also large enough for naturally-aspirated NSXs (at least with stock displacement) and every cell in the PGM-FI's maps has been optimized to run with them.
Nice work Andreas, that's exactly what I'm about to do with my local tuner guy. (I recently purchased and installed the Ostrich2 emulator from Moates, and also their BURN2 PROM burner). Are your 410cc RDX injectors the same type as being sold by prospeed with his RDX tune?.
Yes, I bought my RDX injectors from Prospeed along with the top hat adapters for the injectors, the pigtail adapters for the electrical connections, the additional spacers for the fuel rails, and the injector resistor bypass plug. I also bought the Moates Ostrich2, which emulated the chip in the PGM-FI while my tuner was adjusting the maps. My tuner has his own chip burning hardware so I didn't buy the Moates Burn2 PROM burner and maybe I should have. It wasn't easy to get the PGM-FI to accept a newly-burned chip. The only one that worked was a 27C256 PROM.
Hi! ...and what are the gains? Are there any? Sounds like you are going back to stock? Do we have a part number on the NA RDX? I am thinking of doing a AEM piggy back computer (F/IC) to go along with the oem ECU. Those setted partial throttle parameters playing along with O2 sensors, etc. like you said, cause nothing is more perfect than an oem developed ECU (for economy).
The gains after tuning the PGM-FI and installing the RDX injectors were 14 hp but my tuner guessed that almost none of that came from the injectors. During the baseline run with the stock injectors, the PGM-FI wasn't tuned for the engine's current mechanical configuration. Then the RDX injectors were installed and the PGM-FI was tuned. How much of the 14 hp gain came from the injectors and how much from the tune is impossible to say. If someone spent the money to tune the PGM-FI for the stock injectors, then installed RDX injectors without changing anything else, and re-tuned the PGM-FI, then we'd have a good idea what gains the RDX injectors alone bring.
A friend of mine came along with his NSX while mine was on the dyno and my tuner recommended against his installing 410cc RDX injectors. I like the idea of a better spray pattern and better atomization but the turbo RDX injectors do seem to have some downsides. With the turbo RDX injectors installed and their latency entered into the PGM-FI, my RPM at idle isn't quite as solid as it was with the stock injectors. When the RDX latencies were reduced a bit in the PGM-FI, the idle speed was completely solid but I don't know what knock-on effects entering the wrong latency has so they're now back up where they probably should be. Furthermore, while driving today, my engine threw a CEL code 44 - "Rear Fuel Supply System". I'll have to see what's causing that now.
Given the issues I've experienced, I'm toying with the idea of going back to stock NSX injectors. If I do that, I'd have the PGM-FI reprogrammed and then we'd have a comparison between an optimized RDX tune and an optimized NSX tune so we could see how many horsepower the injectors really bring. On the other hand, modern 310cc naturally aspirated RDX injectors (part number 16450-R70-A01) may be worth a shot as well after having their flow rate and latency / dead time measured.
If you want to keep the OEM PGM-FI and add an AEM F/IC piggyback, what do you expect that can do better than just reprogramming the maps in the PGM-FI? Honda probably tested every cell in every map of the PGM-FI and we're not going to be able to duplicate that effort regardless what route we go. Of the options we have - an AEM F/IC, a standalone engine management system, or just reprogramming the maps in the PGM-FI - reprogramming the maps in the PGM-FI seems to be the most elegant solution to me and I don't see any downsides to that route for naturally-aspirated engines. If you're not happy, you can always just put the OEM maps back on the chip.
A friend of mine came along with his NSX while mine was on the dyno and my tuner recommended against his installing 410cc RDX injectors. I like the idea of a better spray pattern and better atomization but the turbo RDX injectors do seem to have some downsides. With the turbo RDX injectors installed and their latency entered into the PGM-FI, my RPM at idle isn't quite as solid as it was with the stock injectors. When the RDX latencies were reduced a bit in the PGM-FI, the idle speed was completely solid but I don't know what knock-on effects entering the wrong latency has so they're now back up where they probably should be. Furthermore, while driving today, my engine threw a CEL code 44 - "Rear Fuel Supply System". I'll have to see what's causing that now.
Given the issues I've experienced, I'm toying with the idea of going back to stock NSX injectors. If I do that, I'd have the PGM-FI reprogrammed and then we'd have a comparison between an optimized RDX tune and an optimized NSX tune so we could see how many horsepower the injectors really bring. On the other hand, modern 310cc naturally aspirated RDX injectors (part number 16450-R70-A01) may be worth a shot as well after having their flow rate and latency / dead time measured.
If you want to keep the OEM PGM-FI and add an AEM F/IC piggyback, what do you expect that can do better than just reprogramming the maps in the PGM-FI? Honda probably tested every cell in every map of the PGM-FI and we're not going to be able to duplicate that effort regardless what route we go. Of the options we have - an AEM F/IC, a standalone engine management system, or just reprogramming the maps in the PGM-FI - reprogramming the maps in the PGM-FI seems to be the most elegant solution to me and I don't see any downsides to that route for naturally-aspirated engines. If you're not happy, you can always just put the OEM maps back on the chip.
Thank you for the feedback on your experiences. I assume the peak HP gain was 14hp, but what did the before/after torque curves look like? Was it a solid gain all across the RPM range? Were the same timing maps used? Did your tuner try a lower ignition timing and then increase from there?
As you know, besides a bigger injector for more fuel flow with modified engines, the next biggest benefit of these was the modern spray pattern and therefore combustion efficiency increase. The engine may respond better with reduced ignition timing than OEM.
Furthermore, some people thought it may be beneficial to run the injectors at a higher fuel pressure for even better fuel atomization. The injector engineers may cringe, but any changes would most likely be lost in data scatter with the pressure increase (45 to 65 psi for example).
Finally, I know you have a modified engine. What was your previous NSX OEM maximum injector duty? Close to 90%? With your cam timing changes and so on, are you going to increase fuel pressure to add injector duty margin? It's a fine line of balancing injector pressure and injector spring float when looking at injector longevity. The early CT supercharger kits ran the OEM NSX injectors to near 90 psi though.
Dave
Last edited:
Hard to believe you run on stock injectors with stock fuel press your setup. How much hp do you have in total? 340hp I guess with comptech cams and headwork. You compression is oem, right? I have on my racecar 360hp with comptech cams, RDX 410 and 11.5:1 CR, Taitec exhaust and headers, CT adj cam gears and AEM EMS2 tuned.
Last edited:
Yes, the peak gain was 14 hp but there were solid gains everywhere. The torque curve was basically just shifted upwards a bit.
My tuner didn't experiment with various ignition maps. He just copied in the ignition maps from a 1994 JDM NSX. Honda strongly recommended those NSXs be run on Japanese super premium gasoline, which supposedly usually has close to 100 RON. 100 RON is what I routinely fill my NSX up with so it seemed like a good match. Your point about less ignition timing is a good one, though. Perhaps the better atomization of the RDX injectors results in a faster flame front so if maximum cylinder pressure is developed significantly earlier, JDM super premium ignition maps may actually be counterproductive. Hmmm… Have you seen any research regarding how much optimal ignition advance changes based on atomization if knock were not an issue? Does improved atomization increase a gasoline engine's propensity to knock at a given ignition timing? I could always go back to the USDM ignition maps I was running before and see how those impact power.
Without looking into injector duty much, at 8000 rpm at full throttle the value in the appropriate fuel map is about 9870, which I assume is microseconds. In the OEM map, the value is about 13230. At 8000 rpm, the engine completes one complete cycle (two revolutions) every 15000 microseconds. So at full throttle the injectors probably spray at the back of closed intake valves a good percentage of the time. Is that correct?
By the way, my tuner reduced the injector latency settings in my chip so that they are now close to the settings for stock NSX injectors. Like that, the CEL light doesn't come on anymore and my idle is much more stable. The only issue I have with the idle now is that the rpm swing up and down a couple of times if I force the air conditioner to turn on at idle when the engine is stone cold. Other than that, all is good. It seems my PGM-FI really didn't like the RDX latency settings posted here: http://honda-tech.com/engine-management-tuning-124/rdx-injector-dead-time-test-results-2742347/. I wonder what the correct latency settings would really be for the turbo RDX injectors. If someone has the latency measured again, it would be good to measure the latency of the stock NSX injectors as well.
Yes, my compression ratio is pretty much stock. Comptech milled my heads a bit when they ported and polished them back in 1995 but only increased the compression ratio to something like 10.6:1. I don't trust the absolute numbers dynos spit out so I asked a friend with a bone stock NA1 five speed to come along and have his NSX tested on the same dyno within half an hour of mine. For what it's worth, my car supposedly put 26.1% more power to the rollers than his. If his car generates 270 hp at the crank, mine should generate something like 340, as you guessed.
My tuner didn't experiment with various ignition maps. He just copied in the ignition maps from a 1994 JDM NSX. Honda strongly recommended those NSXs be run on Japanese super premium gasoline, which supposedly usually has close to 100 RON. 100 RON is what I routinely fill my NSX up with so it seemed like a good match. Your point about less ignition timing is a good one, though. Perhaps the better atomization of the RDX injectors results in a faster flame front so if maximum cylinder pressure is developed significantly earlier, JDM super premium ignition maps may actually be counterproductive. Hmmm… Have you seen any research regarding how much optimal ignition advance changes based on atomization if knock were not an issue? Does improved atomization increase a gasoline engine's propensity to knock at a given ignition timing? I could always go back to the USDM ignition maps I was running before and see how those impact power.
Without looking into injector duty much, at 8000 rpm at full throttle the value in the appropriate fuel map is about 9870, which I assume is microseconds. In the OEM map, the value is about 13230. At 8000 rpm, the engine completes one complete cycle (two revolutions) every 15000 microseconds. So at full throttle the injectors probably spray at the back of closed intake valves a good percentage of the time. Is that correct?
By the way, my tuner reduced the injector latency settings in my chip so that they are now close to the settings for stock NSX injectors. Like that, the CEL light doesn't come on anymore and my idle is much more stable. The only issue I have with the idle now is that the rpm swing up and down a couple of times if I force the air conditioner to turn on at idle when the engine is stone cold. Other than that, all is good. It seems my PGM-FI really didn't like the RDX latency settings posted here: http://honda-tech.com/engine-management-tuning-124/rdx-injector-dead-time-test-results-2742347/. I wonder what the correct latency settings would really be for the turbo RDX injectors. If someone has the latency measured again, it would be good to measure the latency of the stock NSX injectors as well.
Yes, my compression ratio is pretty much stock. Comptech milled my heads a bit when they ported and polished them back in 1995 but only increased the compression ratio to something like 10.6:1. I don't trust the absolute numbers dynos spit out so I asked a friend with a bone stock NA1 five speed to come along and have his NSX tested on the same dyno within half an hour of mine. For what it's worth, my car supposedly put 26.1% more power to the rollers than his. If his car generates 270 hp at the crank, mine should generate something like 340, as you guessed.
There's a thread on here somewhere about what timing people are using that I posted some of what I am doing. Basically, I tune it myself with an AEM Series 2 and a TFX Engine Analyzer that I can monitor cylinder pressure. During my engine rebuild I rewelded the heads, dimpled the piston domes and combustion chambers in specific locations based on all the old NSX combustion patterns I witnessed on engine rebuilds, ceramic coated the combustion chambers, indexed about 5 sets of plugs, and had custom MLS HG made for a my desired piston-head (squish) clearance. Unfortunately, I didn't have the pressure logger before the rebuild and all these changes. It's neat now since I could watch the piston rings seat during the break-in, and also what happens when I make my header/exhaust manifolds and turbo changes.
I do believe the OEM timing tables being extended for higher-octane fuel (JDM AND US-spec). About 7 years ago when I went to the dragstrip, I first mixed up my own batch of fuel with toluene as the primary ingredient with a little MMO. I calculated the equivalent octane to be ~100. The engine was a lot stronger and smoother.
Finally, yes, the OEM injectors do spray at the backs of the closed intake valves quite a bit at high RPM. Actually, it is desirable for them to run at 100% duty at redline, but of course, that leaves no margin for error during tuning.
I do believe the OEM timing tables being extended for higher-octane fuel (JDM AND US-spec). About 7 years ago when I went to the dragstrip, I first mixed up my own batch of fuel with toluene as the primary ingredient with a little MMO. I calculated the equivalent octane to be ~100. The engine was a lot stronger and smoother.
Finally, yes, the OEM injectors do spray at the backs of the closed intake valves quite a bit at high RPM. Actually, it is desirable for them to run at 100% duty at redline, but of course, that leaves no margin for error during tuning.
Last edited:
I've seen your posts and must say, dimpling the tops of my pistons and brewing my own fuel is way beyond what I want to do!
If your NSX ran stronger and more smoothly with a home brewed fuel based on toluene, do you think that was because of the higher energy density per liter or because of the fuel's higher effective octane? Honda recommended that USDM NSXs be fuelled with 91 (R+M)/2 octane gasoline or higher. If your car ran stronger because of your fuel's greater resistance to knocking, that would mean Honda's USDM ignition maps were optimized for octane ratings above the official recommendation, right?
I know the 1994 JDM NSX ignition maps have more advance than the 1991 USDM maps. For example, at 8000 rpm and full throttle the 1994 JDM map has 28.75° of advance compared to the 23.00° in the 1991 USDM map. If a USDM NSX knocks with 91 (R+M)/2 gasoline (96 RON?) and develops more power with a higher octane fuel, I wonder what the optimal RON rating would be for NSXs running stock injectors and Japanese ignition maps.
If my engine would currently like to knock with the 1994 JDM ignition maps, 100 RON gasoline, and turbo RDX injectors, maybe dialing the timing back to the 1991 USDM specs would reduce how much timing the PGM-FI pulls to prevent knocking. Since I can't measure cylinder pressures (and most likely never will be able to) and can't yet do datalogging, maybe empirical tests would be my best bet for now. It would be really nice, with sr5guy's help, to be able to datalog and monitor my OBD1 NSX's ignition timing in real time.
If your NSX ran stronger and more smoothly with a home brewed fuel based on toluene, do you think that was because of the higher energy density per liter or because of the fuel's higher effective octane? Honda recommended that USDM NSXs be fuelled with 91 (R+M)/2 octane gasoline or higher. If your car ran stronger because of your fuel's greater resistance to knocking, that would mean Honda's USDM ignition maps were optimized for octane ratings above the official recommendation, right?
I know the 1994 JDM NSX ignition maps have more advance than the 1991 USDM maps. For example, at 8000 rpm and full throttle the 1994 JDM map has 28.75° of advance compared to the 23.00° in the 1991 USDM map. If a USDM NSX knocks with 91 (R+M)/2 gasoline (96 RON?) and develops more power with a higher octane fuel, I wonder what the optimal RON rating would be for NSXs running stock injectors and Japanese ignition maps.
If my engine would currently like to knock with the 1994 JDM ignition maps, 100 RON gasoline, and turbo RDX injectors, maybe dialing the timing back to the 1991 USDM specs would reduce how much timing the PGM-FI pulls to prevent knocking. Since I can't measure cylinder pressures (and most likely never will be able to) and can't yet do datalogging, maybe empirical tests would be my best bet for now. It would be really nice, with sr5guy's help, to be able to datalog and monitor my OBD1 NSX's ignition timing in real time.
Last edited:
I thought I had documented my findings on here:
http://www.nsxprime.com/forum/showthread.php/153746-Anybody-Else-Play-With-Fuel-Concoctions?
I guess an equivalent 96 octane took full advantage of the OEM USDM timing curves.
http://www.nsxprime.com/forum/showthread.php/153746-Anybody-Else-Play-With-Fuel-Concoctions?
I guess an equivalent 96 octane took full advantage of the OEM USDM timing curves.
I missed that thread!
If a stock NSX with USDM ignition maps knocks, pulls timing, and reduces power unless the fuel has at least 96 (R+M)/2 octane (about 101 RON), Honda should have mentioned that in the Owner's Manual. Given that 1994 JDM NSXs run more ignition timing than USDM NSXs, that would mean they should be powered by a gasoline with even more than 101 RON. As far as I know, such high octane fuels are not available at public filling stations in Japan. Maybe your NSX ran better on toluene for other reasons.
Datalogging the knock events in an NSX with stock injectors could clear up the car's gasoline octane requirement. Hopefully soon…
If a stock NSX with USDM ignition maps knocks, pulls timing, and reduces power unless the fuel has at least 96 (R+M)/2 octane (about 101 RON), Honda should have mentioned that in the Owner's Manual. Given that 1994 JDM NSXs run more ignition timing than USDM NSXs, that would mean they should be powered by a gasoline with even more than 101 RON. As far as I know, such high octane fuels are not available at public filling stations in Japan. Maybe your NSX ran better on toluene for other reasons.
Datalogging the knock events in an NSX with stock injectors could clear up the car's gasoline octane requirement. Hopefully soon…
Interesting. Then my car running better with the ~96 octane toluene blend must have been because of other reasons (energy density, flammability, or just compensating for crappy gas in my old area). Too bad I didn't have the knowledge and tools I do now back then!
In the future I may work with ethanol fuel and implement an ethanol fuel sensor, but that is low on my priority list. I have everything I need to do it except for the flex sensor... and those are ~$25 for a relatively new one at a junkyard. Almost all new domestic vehicles here are flex-fuel compatible so there is no shortage of supply.
In the future I may work with ethanol fuel and implement an ethanol fuel sensor, but that is low on my priority list. I have everything I need to do it except for the flex sensor... and those are ~$25 for a relatively new one at a junkyard. Almost all new domestic vehicles here are flex-fuel compatible so there is no shortage of supply.
1. Myself and DDozier did a scientific before and after RDX injector power test. The car we tested on was running extremely well on the stock injectors before we started. It was mostly stock with headers and it dyno'd 276whp IIRC. The only change made was installation of RDX injectors and a mathematical re-scale of the 5 stock fuel maps. There was a noticeable, across the board gain of 6ft/lbs of torque that could only be attributed to the injector atomization improvement. We weren't able to squeeze anything else significant out of it with timing changes leading us to believe that there was a flow bottleneck and/or some heat soak issues.
2. The injector latency settings are not correct in the link you posted.
3. I believe the fuel flow at low pulse widths with the RDX injectors doesn't follow a linear path like the stock injectors do. I haven't had the time to work out the details but in experimentation, leveling out the low load, low RPM areas of the fuel maps as well as adjusting timing accordingly has resulted in stable idle. I believe whats causing the unstable idle is a variation of manifold pressure caused by "curve" in the low pulsewidth fuel flow which causes a feedback loop (the idle surging). Simply scaling the factory fuel maps isn't enough to compensate.
4. If your tuner tuned the fuel maps with stock-like injector latency settings I can assure you the tune will need to be completely re-done once the correct latency settings are entered.
5. The NSX uses a dynamic injector timing methodology that varies injector timing based on desired pulsewidth and RPM. Some settings have been defined in the public definition but they're mostly useless without a detailed description which I don't have time to write up unfortunately
-Matt
Last edited:
I can totally confirm that the turbo RDX injector latencies that WitchHunter measured did not work well in my NSX's PGM-FI. Hopefully we'll get better latency data to work with relatively soon.
Thankfully, using your definition, I can look into the fuel maps my tuner programmed over the course of several hours on the dyno. As you said, the fuel maps my tuner came up with are not simply 58.5% of the OEM NA1 injector duration maps (240 cc/min NA1 NSX injector / 410 cc/min turbo RDX injector). In my current maps for the turbo RDX injectors, at low throttle openings my injector durations are a lower percentage of the OEM durations than they are at high throttle openings. So if you install turbo RDX injectors, the injector duration in the PGM-FI should be determined on a dyno, not just on the back of an envelope.
If you weren't able to squeeze out any more horsepower with changes to the ignition timing, did you try reducing the timing? As Mac Attack indicated, that may be worth a try. The better the atomization, the faster the flame front so if peak cylinder pressure occurs sooner, maybe less ignition advance will result in more crank horsepower. Right now my NSX is running more ignition advance than OEM USDM NSX specs but with the turbo RDX injectors, maybe less timing would increase my crank horsepower. I'll have to check.
And datalogging would really help. If there's anything I can do to help that along (like becoming a paying customer of yours), just let me know.
Thankfully, using your definition, I can look into the fuel maps my tuner programmed over the course of several hours on the dyno. As you said, the fuel maps my tuner came up with are not simply 58.5% of the OEM NA1 injector duration maps (240 cc/min NA1 NSX injector / 410 cc/min turbo RDX injector). In my current maps for the turbo RDX injectors, at low throttle openings my injector durations are a lower percentage of the OEM durations than they are at high throttle openings. So if you install turbo RDX injectors, the injector duration in the PGM-FI should be determined on a dyno, not just on the back of an envelope.
If you weren't able to squeeze out any more horsepower with changes to the ignition timing, did you try reducing the timing? As Mac Attack indicated, that may be worth a try. The better the atomization, the faster the flame front so if peak cylinder pressure occurs sooner, maybe less ignition advance will result in more crank horsepower. Right now my NSX is running more ignition advance than OEM USDM NSX specs but with the turbo RDX injectors, maybe less timing would increase my crank horsepower. I'll have to check.
And datalogging would really help. If there's anything I can do to help that along (like becoming a paying customer of yours), just let me know.
Coming into this late. I see SOS/RC offers stock 250cc direct replacement injectors. Can anyone yay or nay if these are as good as OEM.
** updated the charts with some minor corrections, and reduced to three RDX data sets for clarity **
View attachment ECU latency tuning (f).pdf
The “elephant in the room” is whether the RDX injectors are unnecessarily large (and thus unnecessarily slow) for those who prefer their NSX naturally aspirated. All we really want is the advanced spray pattern offered by the RDX injectors, and perhaps the SOS/RC 250cc/min injectors are a better choice, (although I haven't seen any info on spray or latencies for them).
The RDX 410cc/min is a lot more than 240cc/min but we don’t need that much more fuel for N/A. Lower flow seems be associated with less latency (presumably because the mechanics are smaller/lighter) and ideally we should choose one that is almost as fast as OEM.
I've just had a couple of RDX injectors and an OEM injector tested for latency (slowness) by KPM Motorsport here in Adelaide. I've attached a summary chart comparing 5 injector data sets (2 x OEM, 3 x RDX):
1) OEM ECU - these are the latency vs voltage values embedded by the OEM (Honda) in the ECU ROM (extracted using TunerProRT and sr5guy's definition)
2) OEM_KPM - latencies of my OEM injectors measured by KPM, a bit odd at low voltages < 10V
3) RDX ACS 1.2 - greenberet's latest trial values
4) RDX-KPM - latencies of my RDX injectors measured by KPM
5) RDX-HondaTech - the Which-hunter Performance table for RDX injectors at Honda-Tech.com
Ignoring the lower latency readings by KPM for the OEM injectors below 10V, I think the data is quite conclusive. The two injector types (OEM and RDX) have produced reasonably consistent results above 10volts.
AFAIK the only time the voltage will fall below 10V should be when starting the car. Once the car is moving the voltage will be in the range 12 – 14.5 volts. Occasionally if sitting at idle with A/C, heater, lights, blnkers and wipers etc all on, the battery drain my drop volts a bit, but typically to only to 12 volts, worst case perhaps 11.5V.
So 99.9% of the time the volts will be > 12.5V and that’s where we now have good data… for the voltages > 10V the three RDX readings are very close, differing by only 0.15 to 0.20 ms. This suggests the Honda-Tech / Whichhunter was actually correct. It would be helpful if a tuning expert could chime in about whether 0.2ms is significant when it comes to tuning and latency tables.
I asked the KPM engineer on how they test latency when I collected the injectors. As I understood his explanation: their test device can apply a train of short pulses to the injectors where the spaces between each pulse is much longer (i.e. a very low mark-space ratio). The short pulse width (mark) is under control of the operator. If the pulses are less than latency time no fuel can come out of the injector. As the pulse width is gradually increased eventually the latency is overcome and a drip will start to form on injector outlet. The pulse width when fuel first appears is the latency period. (I have sent this paragraph to KPM seeking confirmation I have explained this correctly).
View attachment ECU latency tuning (f).pdf
The “elephant in the room” is whether the RDX injectors are unnecessarily large (and thus unnecessarily slow) for those who prefer their NSX naturally aspirated. All we really want is the advanced spray pattern offered by the RDX injectors, and perhaps the SOS/RC 250cc/min injectors are a better choice, (although I haven't seen any info on spray or latencies for them).
The RDX 410cc/min is a lot more than 240cc/min but we don’t need that much more fuel for N/A. Lower flow seems be associated with less latency (presumably because the mechanics are smaller/lighter) and ideally we should choose one that is almost as fast as OEM.
I've just had a couple of RDX injectors and an OEM injector tested for latency (slowness) by KPM Motorsport here in Adelaide. I've attached a summary chart comparing 5 injector data sets (2 x OEM, 3 x RDX):
1) OEM ECU - these are the latency vs voltage values embedded by the OEM (Honda) in the ECU ROM (extracted using TunerProRT and sr5guy's definition)
2) OEM_KPM - latencies of my OEM injectors measured by KPM, a bit odd at low voltages < 10V
3) RDX ACS 1.2 - greenberet's latest trial values
4) RDX-KPM - latencies of my RDX injectors measured by KPM
5) RDX-HondaTech - the Which-hunter Performance table for RDX injectors at Honda-Tech.com
Ignoring the lower latency readings by KPM for the OEM injectors below 10V, I think the data is quite conclusive. The two injector types (OEM and RDX) have produced reasonably consistent results above 10volts.
AFAIK the only time the voltage will fall below 10V should be when starting the car. Once the car is moving the voltage will be in the range 12 – 14.5 volts. Occasionally if sitting at idle with A/C, heater, lights, blnkers and wipers etc all on, the battery drain my drop volts a bit, but typically to only to 12 volts, worst case perhaps 11.5V.
So 99.9% of the time the volts will be > 12.5V and that’s where we now have good data… for the voltages > 10V the three RDX readings are very close, differing by only 0.15 to 0.20 ms. This suggests the Honda-Tech / Whichhunter was actually correct. It would be helpful if a tuning expert could chime in about whether 0.2ms is significant when it comes to tuning and latency tables.
I asked the KPM engineer on how they test latency when I collected the injectors. As I understood his explanation: their test device can apply a train of short pulses to the injectors where the spaces between each pulse is much longer (i.e. a very low mark-space ratio). The short pulse width (mark) is under control of the operator. If the pulses are less than latency time no fuel can come out of the injector. As the pulse width is gradually increased eventually the latency is overcome and a drip will start to form on injector outlet. The pulse width when fuel first appears is the latency period. (I have sent this paragraph to KPM seeking confirmation I have explained this correctly).
Last edited:
I just wanted to report that my NSX has been running very well with turbo RDX injectors and PGM-FI latency values based off of the tests sparky had carried out by the company KPM in Australia. Thank you sparky!
In the PGM-FI, the latency values for stock NSX injectors are a bit lower than KPM's measurements above 13 volts. Below 13 volts, the values in the PGM-FI are above KPM's measurements. Given that, I entered turbo RDX latency values into my PGM-FI which are also a bit lower than KPM's measurements above 13 volts and are higher below 13 volts. Specifically, I entered 3.469 at 7.03V, 1.936 at 9.04V, 1.242 at 12.06V, 0.915 at 14.07V, and 0.726 at 16.01V. Like that my idle has been as stable as OEM in all conditions for several weeks now and I haven't gotten a Check Engine Light.
In the PGM-FI, the latency values for stock NSX injectors are a bit lower than KPM's measurements above 13 volts. Below 13 volts, the values in the PGM-FI are above KPM's measurements. Given that, I entered turbo RDX latency values into my PGM-FI which are also a bit lower than KPM's measurements above 13 volts and are higher below 13 volts. Specifically, I entered 3.469 at 7.03V, 1.936 at 9.04V, 1.242 at 12.06V, 0.915 at 14.07V, and 0.726 at 16.01V. Like that my idle has been as stable as OEM in all conditions for several weeks now and I haven't gotten a Check Engine Light.
That's great news greenberet.
I'm no injector expert, but it surprises me how such small differences in your latency setting make the difference.
It would be interesting to hear form an experienced tuner on why the tune is so sensitive to these values.
Also what is smallest latency change that would make a noticeable difference ? e.g.... 0.1ms or 0.01ms or even smaller? Another way of asking same question: what precision / significant digits are meaningful in latency values: for example say a particular latency was measured, it could be reported as 1.1ms or 1.06ms or 1.064 or 1.0642. These are potentially all the same, just different precision.
I'm no injector expert, but it surprises me how such small differences in your latency setting make the difference.
It would be interesting to hear form an experienced tuner on why the tune is so sensitive to these values.
Also what is smallest latency change that would make a noticeable difference ? e.g.... 0.1ms or 0.01ms or even smaller? Another way of asking same question: what precision / significant digits are meaningful in latency values: for example say a particular latency was measured, it could be reported as 1.1ms or 1.06ms or 1.064 or 1.0642. These are potentially all the same, just different precision.
Thanks for doing the testing and putting that summary chart together Sparky!
Latency, or more correctly injector dead time, is important for idle and low-load low-RPM conditions when the injector dead time is a significant fraction (25-40%) of the total injector pulse time. It's importance diminishes when injector pulse time increases such that the injector dead time only accounts for maybe 5-10% of this period.
Especially for idle conditions with closed-loop feedback from the O2 sensors, if your injector dead time is 25-40% of the total pulse time and your O2 sensor is determining a lean condition, then the ECU is using the interpolated fuel map with interpolated latency to determine the needed increase in injector pulse width. At the low injector flow rates, it gets a little screwy due to injector non-linearity.
As far as adjusting latency, tenths of a millisecond are OK for initial tuning since we are talking about a total latency period of ~1 milliseconds at 13.5 Volts. Since we are talking about idle quality, low RPM driveability and fuel economy, your feedback on further refinement isn't going to be very precise!
Oh, if you want to bump up your fuel rail pressure as has been talked about with this mod, then the latency and fuel injector flow curve will need to be adjusted further. Latency should increase, and the fuel injector non-linear flow curve will become even more apparent.
Dave
Latency, or more correctly injector dead time, is important for idle and low-load low-RPM conditions when the injector dead time is a significant fraction (25-40%) of the total injector pulse time. It's importance diminishes when injector pulse time increases such that the injector dead time only accounts for maybe 5-10% of this period.
Especially for idle conditions with closed-loop feedback from the O2 sensors, if your injector dead time is 25-40% of the total pulse time and your O2 sensor is determining a lean condition, then the ECU is using the interpolated fuel map with interpolated latency to determine the needed increase in injector pulse width. At the low injector flow rates, it gets a little screwy due to injector non-linearity.
As far as adjusting latency, tenths of a millisecond are OK for initial tuning since we are talking about a total latency period of ~1 milliseconds at 13.5 Volts. Since we are talking about idle quality, low RPM driveability and fuel economy, your feedback on further refinement isn't going to be very precise!
Oh, if you want to bump up your fuel rail pressure as has been talked about with this mod, then the latency and fuel injector flow curve will need to be adjusted further. Latency should increase, and the fuel injector non-linear flow curve will become even more apparent.
Dave
Last edited:
Thanks for the explanation Mac Attack. I guess that's the awkward truth. If the car doesn't idle well we should try different values but it's hard to predict what values to go to !
Interesting point. The curve differences in my charts are most noticeable at low voltages (i.e. starting / idling). I wonder if this is due to different test fuel pressures? I asked KPM to use 50psi for my tests.
Last edited:
Resurrecting this thread as I plan ahead for my "Stage 2" process for my 92 auto. In this stage, I will convert the car from an auto to a 6-speed. I'll also be adding Comptech cams and new OEM valvesprings. The fact that [MENTION=12723]greenberet[/MENTION] was able to run this setup for 17 years without trouble has convinced me it is the way to go, well, before Stage 3, which is ITB and full engine management. 
Question though- it appears from the thread that the OEM 240cc injectors are sufficient to handle even the increased lift from the Comptech cams. Do we know what the Honda 240 duty cycle is at 8,000 rpm? From my GT-R tuning experience, I always understood that 80% or lower is ideal, but of course that is when you are dealing with 20 psi of boost! Still, I can't imagine the 240 cc injector is running below 80% at WOT using the Comptech cams. Am I wrong? I would love to put the RDX injector money toward ITB instead and keep my OEM 240's, but don't want to put the engine in danger either.
Question though- it appears from the thread that the OEM 240cc injectors are sufficient to handle even the increased lift from the Comptech cams. Do we know what the Honda 240 duty cycle is at 8,000 rpm? From my GT-R tuning experience, I always understood that 80% or lower is ideal, but of course that is when you are dealing with 20 psi of boost! Still, I can't imagine the 240 cc injector is running below 80% at WOT using the Comptech cams. Am I wrong? I would love to put the RDX injector money toward ITB instead and keep my OEM 240's, but don't want to put the engine in danger either.
Similar threads
