S2E18: We test a lean burn system on our Saturn . Will it get better MPG?

today's experiment is well sort of an experiment what i mean to say is the results we got were both good and good but not in the way you may think actually we're still a little bit puzzled ourselves so let's get right into it now way back at the turn of the century honda offered an odd little car called the insight these first generation cars were packed with all sorts of technology and were quite expensive during the six year production run honda only made 16 000 of these cars and that makes them kind of rare nowadays today we're going to focus on one of the features that was available on the manual transmission insights and that is the lean burn system now keep in mind lean burn technology was around long before the honda insight and this sort of fuel saving strategy has been discontinued due to the complexity of managing the emission control system you see the lean burn may use less gasoline but it also produces a byproduct in the exhaust called oxides of nitrogen as far as i know removing these oxides of nitrogen or nox from the exhaust is not easy to do so that's mainly why you don't see manufacturers offering this feature anymore well for our experiment we'll kind of roll back the clock a little bit and do a what if and what i mean by that is what if a similar lean burn system was available on a car like our second generation saturn oh that's ugly fortunately we'll just be using the technology from the honda an interesting thing to note is the components required to replicate a reliable lean burn system on a fuel-injected car back in the 90s was way too expensive for any manufacturer to even consider i reckon something like that would have bumped the price of the car up a few thousand dollars nowadays you can more or less buy everything you need from the jungle site for less than 200 bucks with that said this video is just for entertainment and educational purposes i'm in no way advocating anyone do this sort of modification we're doing it as a short term experiment the long-term and even the short-term effects of this experiment may damage the engine on our saturn and we're willing to take that risk so today we're going to build and test our experimental version of the lean burn system and see how it performs this should be an interesting video all right let's take a moment and discuss exactly how we're going to conduct today's experiment for this discussion i'm going to remove my hillbilly hat and put on my professor's cap this is the future jimbo here hey for the next eight minutes 45 seconds i go over the technical details that might not be something everybody wants to watch anyway we go from this diagram to this diagram it's a tedious discussion and if you're not up to it i recommend you jump forward to this time stamp you'll thank me later the first thing we're going to do is remove the narrow band oxygen sensor from the exhaust system on the saturn and then put it in the rubbish bin alright now we're going to install a wideband oxygen sensor in the same location that the narrowband sensor previously occupied now pay attention this wideband sensor produces a signal that's not compatible with the factory ecu that's used in the saturn and for that we have a solution so this wideband sensor that we're using for our initial testing comes in a super cheap kit that we got on the jungle site now this cheapo kit is of a questionable quality and more ways than one if we look at the instructions that came with the cheapo kit we can see that the output signal can be configured four different ways the first configuration is for use on an aftermarket efi computer like a mega squared or a speed duino and a bunch of others the second configuration is for folks who like to look at the mixture in lambda i'm not sure what the third configuration is for but it's likely for a 3.3 microprocessor the fourth configuration is to set the output of the wideband system to be the equivalent of a narrow band signal and that's a really cool feature if it worked and it doesn't work well it works occasionally but that makes it unreliable hey let's take a quick look so we have the system set up on the bench and the output of the wideband kit is connected to the multimeter now we're going to set the rotary switch in the gauge to the p3 position which is what we want to configure the output of the system to narrowband all right now let's power it up it goes through self-test that's good now right here we can see the gauge is confirming that it's configured for narrowband output by displaying the p3 mode setting which is great now we should see the voltage on the multimeter span between zero and one volt now keep in mind anything over 1 or 1.2 volts is too much voltage and can potentially damage the car's ecu even though the gauge is set up for narrowband output it's unfortunately generating a wideband output and that won't work now we did this multiple times and a few times we did see voltages that were appropriate for narrowband but this is really sketchy i definitely do not recommend this cheap kit from the jungle site alright well that sucks but we can work around that with some clever code later on anyway we'll take the zero to five volt signal from the wideband sensor and plug it into an arduino uno solid state microcomputer the arduino runs at 16 megahertz and has a 10 bit analog to digital converter the good news is the signal from our wide band sensor kit is a strong signal and we don't have to worry about the impedance so back to our story the zero to five volt signal is piped into the arduino and we do some math to convert the signal into the air fuel ratio now unfortunately like i mentioned before this signal is not compatible with the saturn ecu so we have to convert that signal into a narrow band signal and that can't be done with simple math nope for that we'll have to use lookup tables so here's a few graphs that will show the difference in signals that we're dealing with the first graph shows a narrow band signal and as you can see the operating range of this sensor is very narrow hence the name narrowband so on a narrowband o2 sensor the voltage generated by the sensor goes up when the engine runs rich and the voltage goes down when the engine runs lean the target air fuel ratio is right about here now this is a hard target to hit so generally the air fuel ratio bounces around quite a bit and its average value centers around the target kind of sloppy but it works okay on older cars like the saturn now the signal generated with a linear wideband sensor is both wider and it's linear the voltage generated by this kind of sensor is between 0 and 5 volts the lower the voltage the richer the mixture and of course the higher the voltage the leaner the mixture hopefully you notice the difference between the signals and the fact that the voltage swings are opposite of each other so what that means is you can't use a voltage divider to convert a wideband signal into a narrow band signal the reason i mentioned this is i've had questions in the past about this very same subject anyway this is just a brief overview on wideband vs narrowband if you want to learn more about this stuff the internet's full of knowledge but you gotta look once the wideband signal is converted to a narrowband signal via software on the arduino we need to further modify the narrowband signal with an offset now this offset will fool the ecu into running the engine leaner and theoretically burning less fuel so for the offset we really don't know how much to add and it's something that needs to be dialed in while the car is being driven too much offset the engine won't make any power and we risk damaging the engine anyway we're going to use a potentiometer to adjust the offset now this pots going to produce an analog signal that the microprocessor can read and we'll write some code to modify the narrow band signal in proportion to the position of the potentiometer next we need to know when the lean burn system should kick in we definitely don't want the engine to be in lean burn under heavy loads so again that's something we need to figure out while road testing the car so for this we'll use another potentiometer and some code oh and this is just a simple diagram and not a wiring schematic the last signal we need to look at is the map sensor on the engine so map stands for manifold absolute pressure the data the map sensor collects is used to calculate the air density and determines the engine's air mass flow rate so that's great but we can also use the map sensor data to determine the vacuum in the manifold high vacuum means the engine is not under much of a load and low vacuum means the engine is under a load so we want the lean burn to kick in when the engine is under a light to moderate load and the map sensor will provide that information we'll have to write a bit of code around the sensor data and the position of this pot and that will allow the lean burn to automatically kick in so the last part of this little experiment is send a 0 to 1 volt narrowband signal to the saturn's ecu well shucks the arduino has provisions to generate an analog signal but the method used to create that analog signal is questionable and we're not confident in the precision and quality of that signal so for the output signal we're going to use an adafruit 4725 digital to analog converter now this device has a 12-bit converter which is nice and will give us a great resolution in the zero to one volt range all right to complete the circuit we connected the 4725 to the ecu and the saturn more specifically we connected it to the oxygen sensor input keep in mind this device will now provide the 0 to 1 volt signal that the ecu is expecting from the oxygen sensor i reckon it would be handy to have something to indicate when the lean burn is active so for that we wired in a huge led indicator and that'll turn on when the lean burns triggered so let me give you two examples how this system works the first example i'll walk you through is how the system works in normal mode or in other words how the system works when the lean burn is not active the wideband sensor samples the exhaust and determines that the air fuel ratio is 14 to 1 and that voltage is sent to the arduino now the arduino will use a lookup table and match the 14 to 1 airfield ratio with the equivalent voltage that a narrow band sensor would have generated this information is passed on to the adafruit 4725 digital to analog converter and it will generate the required voltage that the ecm is expecting at that point the ecm can correct the air fuel ratio now the arduino is capable of doing this several thousand times per second and we slowed it down to about 12 times per second anyway that's an example of how the system is running in normal mode let me walk you through an example in lean burn mode okay let's say the wideband sensor samples the exhaust and determines that the air fuel ratio is perhaps 16 to 1 and that would be moderately lean once again the voltage is sent to the arduino and it looks up the air fuel ratio and finds the match for the correct narrowband signal however this time the arduino reads the position of the offset potentiometer and takes that offset value and adds it to the narrowband signal the modified signal is sent to the adafruit 4725 and this device produces a voltage that makes the ecu think the engine's operating at an ideal air fuel ratio and that keeps the ecu happy even though the engine is running lean the actual air fuel ratio is always going to be displayed on the gauge but through the magic of electronics and some code the ecu will be tricked into reducing the injector pulse width and thus allowing us to run the engine lean without generating any codes and that's the key to this trick if the ecu sees a fault it will disregard our signal and use backup data to correct for the air fuel ratio so here we have the completed and calibrated lean burn gizmo the software was fairly easy to write and the only thing that took a lot of time was confirming the calibration for the look up tables as you can see we have a box full of stuff not too pretty but this will certainly work for our experiment of course the arduino is easy to spot but the other stuff's buried deep inside the box i think we need to test this out in the car all right so now we have the engine all warmed up and the lean burn system is running but we have it set for zero offset and that's evident by looking at the air fuel ratio gauge the top knob is the offset dial and the bottom knob is the trigger dial now if i spin the dial over to this mark the system will engage the lean burn it's hard to tell in this lighting but the indicator is illuminated and that's cool but if you notice the air fuel ratio hasn't changed and that's great because we haven't dialed in any offset yet so let's do that now okay now we can see the air fuel ratio gauge is showing the engines running slightly leaner and here's the cool part the ecu has no clue because we're feeding it a fabricated signal and the ecu thinks the air fuel ratio is dead nuts on target i think we're almost ready for a road test now for the first test i'm going to monitor the scan gauge for instantaneous miles per gallon and average miles per gallon also on the first test we won't dial in any offset because we want to establish a new baseline for the miles per gallon previously we got 42 miles per gallon with no modifications done to the car and now we want to confirm that the car is still fine so before we head out let's take a look under the hood because off camera we made a few changes so first off we installed a fresh set of ngk spark plugs now the saturn s series engine was designed to use regular spark plugs well actually it's not necessarily the engine that cares it's the ignition system that cares and i recommend you research this if you don't believe me but iridium or platinum plugs can cause performance problems it's a saturn thing anyway while we were under the hood the non-functional air conditioner compressor was removed and just for the heck of it we removed the power steering pump the spark plugs well they may make a difference but the ac compressor and the power steering pump man i don't think we'll see much of an improvement in performance or fuel economy but that stuff has to be removed anyway when we do the diesel swap so now's a good time to take care of it i think we need to take the car out on the road and do some testing [Music] it's good to be back in a normal running car in the previous episode we did some pretty stupid stuff to this car and it got to the point where the car was basically undrivable poor saturn you know i really do like saturn's but this car is incredibly uncomfortable i think the coupes and the sedans are more or less the same except for where the seat belts are located this car wants to choke the driver every opportunity it gets and i guess that's fair play considering what we've done to the car well i guess it could be worse i could be driving a subaru now lately it's been extremely hot during the day and we found the best time to do our road tests is after six in the evening or early in the morning it really sucks not having air conditioning so today for our fuel economy test we'll be using the exclusive bigger hillbilly proving grounds each completed loop is about 30 miles give or take and we plan on making seven loops around the test track if you want to call it that anyway speeds range from 55 to 70 miles per hour on this trip i'll watch the scan gauge and try to maximize the fuel economy now previously we recorded 42 miles per gallon with pure expressway driving and that seems a bit high so on today's test we're hoping to see reasonable fuel economy with the lean burn system off and just driving the car in normal mode theoretically with the lean burn system off the fuel economy should be what a typical saturn would get with a single cam engine and a 5 speed manual transmission well we shall see anyway before we headed out for the fuel economy run we of course filled the fuel tank well it took several hours to complete the fuel economy run and i reckon it's time to fill the tank and see how much fuel we consumed for the entire trip the car ran perfect we didn't get any check engine lights and the air fuel ratio seemed to be right on target now keep in mind even though the lean burn system was not activated we were still processing the wide band signal through the arduino and converting it into a narrow band signal so at the very least it seems like the electronic gizmos were doing their job the lack of power steering was noticeable at first but i got used to it and the car seems to drive fine without it now this is interesting according to the scan gauge the car got 45 miles per gallon now i don't know how accurate that is and we'll obviously do the math but that's kind of interesting don't you think well there you have it 4.502 gallons okay now it's time to test out the lean burn system i'm really not looking forward to another several hours in this car but it has to be done so we have the offset already dialed in and the trigger point dialed in as well that's pretty much all we need to do so right away on the gauge i can see the engines running lean and the indicator lights on now due to an oversight the lean burn will actually engage at idle and this can be fixed with a few lines of code not a big deal i doubt we'll be doing any extended idling during our road test that indicator light being on really bugs me and like i said it's only a few lines of code to fix that now don't you just hate it when you buy a brand new gizmo of some sort and immediately right out of the box you have to plug it in for a software update yeah i know it's like the folks who write the code for a living ain't paying attention uh well in my case um well let's just change the subject so here we are just accelerating normally and as you can see when we put a load on the engine the lean burn indicator lamp turns off now when that happens the air fuel ratio automatically goes back to normal so what that means is we get full power on acceleration and that keeps the ecu happy and by happy i mean the ecu likely won't receive any abnormal signals from things like to knock sensor and whatnot so as you can see the bigger hillbilly proving grounds is more or less endless cornfields or whatever that stuff is i'm not really much of a farmer but the only interesting you see out here is an occasional horse or pony or whatever those things are sometimes you'll see a vulture-like creature eating roadkill and those things don't like to get out of the way i reckon kansas vultures ain't afraid to stand their ground so the air fuel ratio seems to be hovering around the low 17s and that's not really much of an offset one of the things we discovered with this cheapo wideband system is it won't work when the airfield ratio goes above 18 to 1. meh that's cool for now you see this is the first real road test and we're making notes on all the things we may need to improve on so after driving this car for a while i've noticed that it's a lot harder to get the scan gauge to show a decent instantaneous mile per gallon number previously on the first road test when the lean burn system was not active it was easy to hold the throttle just right and the scan gauge would show between 45 and 50 miles per gallon for the instantaneous reading now 40 seems to be the new normal it's hard to say what's really going on and the data will be interesting once we review it keep in mind we put a lot of effort into this gizmo but this is also its first real test i think if we can make it through seven laps of this fuel economy course without triggering the check engine light well that'll tell us we at least did something correct i guess i should also remind you folks that we're doing just the basics and today we're not going to fiddle around with the more advanced techniques that was used on the honda insight now when it comes to testing stuff out this ain't my first rodeo i think it would be unwise to try testing too much stuff all at once typically that makes troubleshooting very difficult and our best course of action is to do it in steps so as most of you folks know this car is scheduled for a diesel engine swap soon of course the sooner we can do the swap the better but as things go these days we've been experiencing a few delays i think nearly all the parts have arrived and we're working on the engine to transmission adapter plus the clutch and flywheel adapter so the engine that we'll be using is a 722 cc three cylinder kabuto diesel engine and we'll be mating it to a close ratio manual transmission out of an s series saturn so here's an update on our fuel injected supercharged intercooled 420 cc big block hemi-powered street legal go-kart well it's temporarily off the road for repairs and we also need to get a kansas license plate for it it'll be a few more weeks before that car is back in action and that's probably the best thing for it you see that car really doesn't behave well in hot climates that's difficult to test i think we've pushed that engine past its operating parameters by quite a bit anyway that's the story for another day so through the magic of video editing we're done with our seven laps around the fuel economy test track and we didn't get a check engine light i think i'll head over to the gas station and fill it up [Music] so i reckon it's time to look at the data this should be very interesting for the first road test we ran the car in normal mode or in other words the lean burn system was not activated so we had zero offset and the air fuel ratio was normal let's see okay we traveled a total of 211.6 miles and consumed 4.502 gallons of gasoline and that gives us 47 miles per gallon [Laughter] okay raise your hand if that doesn't make any sense all right i agree so let's go over the facts this is true this is true and during the road test the air fuel gauge indicated that the mixture was normal so yeah i don't know gnomes or something like that or and this is definitely possible something is not right with my setup anyway let's move on the second test was conducted with the lean burn system fully activated now all during the test i closely monitored the air fuel gauge and when the system was triggered the air fuel mixture went lean now we didn't go nuts and crank the system all the way up nope we kept it around 17.2 to 1. so given that information we traveled a total of 212.2 miles and consumed 4.782 gallons of fuel our fuel economy on the second trip was 44.4 miles to the gallon which is somewhat believable but i gotta tell you it doesn't feel right to me now during this test i closely watched the scan gauge for instantaneous and average miles per gallon and when the lean burn system was active well it was difficult to keep the instantaneous mile per gallon above 40. on the other hand in normal mode the car had no problem going above 50 miles per gallon for extended periods of time keep in mind we're talking about the scan gauge and the data that it reports is somewhat contrived but from what i gleaned off the internet the scan gauge pulls the injector pulse width from the data stream in order to calculate instantaneous miles per gallon oh yeah and for the entire 212 miles of the second trip we didn't trigger the check engine light so the ecm must have been somewhat satisfied with our fabricated signal now i don't care if an experiment doesn't produce the results that i was expecting that's the purpose of the experiment however i feel something went wrong with this experiment and even though we have interesting data i feel we need to do this over but differently don't worry i got this under control well i hope i do anyway let's take a look at what arrived at the po box this week so this package arrived a few days ago and i've been itching to open it but we got to set up cameras and all that stuff when we do an unboxing well the moment has finally arrived oh it comes with a note let me check this out real quick so this package is from one of our viewers anthony m he says he loves the channel and has sent something to help us out well let's check it out i don't know how this generous person saw into the future but this gift arrived just in time this is a really good quality wideband system i believe it's an innovative motorsports lc2 maybe i've used these in the past and never had problems with them i think this one even has a programmable offset and if it does that would be cool oh yeah this is awesome so for part two of the lean burn experiment guess what we'll be using this in our experiment this system will definitely get used on a lot of projects actually well like i said i've used these on previous projects and the one thing i like about this brand is it can be free air calibrated i guess the older ones can i reckon i should check the instructions to confirm that the el cheapo wideband system that we just used to make this video is of questionable quality now something like this i think i can trust it and that's important when you're trying to measure something like the air fuel ratio i'm actually pretty excited to start playing with this but i guess i have to finish this video first anyway with this gift i feel like we now have the proper tools to build more reliable lean burn system and it might actually be a lot simpler now like today anthony gave the channel a very generous gift and we certainly appreciate that in the past we received a set of turbochargers from aaron and we have plans for those don't worry about that now we can't forget stewart in michigan who donated the kabuto d722 diesel engine and before long that engine will be in the red saturn you know off camera we're doing a lot of work to mate the diesel engine to the spare saturn transmission i think the diesel saturn will be pretty epic when we get it finished so we recently passed 80 000 subscribers and thank you everybody for that it really helps out a lot and of course thanks again anthony aaron and stuart for your generous gifts to this channel if you want to help the channel out consider subscribing if you haven't already and of course the po box address is in the video description if you want to send us something so i have a lot of work ahead of me and i think it's time to sign off for now so we'll see you next time [Music]

Channel: Robot Cantina

Views: 74,182

Keywords: $99, 212 cc, 420 cc, 420 hemi, alternative transportation, custom, cvt transmission, diy, documentary, efi, electric car, electric vehicle, engine mods, engine swap, fabricate, fuel economy, fun, funny, go cart, megasquirt, gocart, gokart, NO2C, goonzquad, great idea, gx390, hack, hand made, harbor freight, hemi, hho, high performance, home made, homemade, hp, hybrid, lada, lawn mower, lawnmower, modification, modify, motor, motor swap, motorcycle, mpg, predator powered car, Fuel injection, speeduino

Id: SACCy4KBUKs

Length: 26min 47sec (1607 seconds)

Published: Sun Jul 31 2022