What is optimum AFR for peak power using LPG?

[CNG]

Reckon this has to be a Q for Simon.

I've been looking at numerous threads on t'internet about AFR for peak power. This against stoich. All the info I can find is broadly Yank which understandable but... it's petrol based too. And liquid gives a cooling effect. I looked to AFR for peak power running LPG and there's nothing. Let be for methane.

Reason I ask: I want to kid-on my Leonardo set-up, via microswitch, to alter my narrowband voltage at WOT. This at a lower rpm than Leo does now.

Thus richen mixture when required for my low-revving engine. Only:

1. I don't know if this is wise.
2. I don't know via a zirconium NB what voltage I'd aim for?
3. I don't know peak AFR for power w/methane.

Up my favourite hill my engine working WOT and CNG, seems happiest at about 0.98 Lambda.

Simon, pretty please - as distinct from stoich which I can see - what's peak power AFR with LPG, reckon that number makes a good start point.
And... do yo knwow what voltage a Zirc NB would show at that AFR?

AFR Lambda Table 19-02-20 01.png (153.06 KiB) Viewed 5402 times

[Gilbertd]

Why do you want to richen the mixture, surely the engine will run best and produce maximum power at Stoic?
You can't aim for a specific voltage with a Zirconia sensor, it only shows rich or lean and isn't a varying voltage. So you'll get the same voltage for slightly rich and very rich.
Peak power on LPG (Propane) will be at stoic, so 15.5 according to the chart you posted or 15.67 according to https://en.wikipedia.org/wiki/Stoichiometry

[LPGC]

Firstly, if you have an AFR meter designed for running on petrol (stoich 14.7:1) but you use it for LPG (stoich 15.4:1) then with stoich mixture on LPG it will show 14.7:1 not 15.4:1 and 15.4:1 would be a lean mixture on LPG, so you need to keep that in mind. Some AFR meters show mixture in terms of AFR, some show it in terms of lambda equivalence ratio (1 = stoich, less than 1 rich, higher than 1 lean), so if you've got an equivalence ratio of 0.95 it suggests the mixture is 5% rich, but again the equivalence ratio displayed might be geared for petrol and less accurate on LPG (but still more accurate than confusing AFR).

Engines run stoich during normal loads because stoich makes near the best use of the fuel and gives lowest emissions (at least when NOx are part of the consideration) but a slightly rich mixture does give a bit more power. You can put that to the test yourself by running an engine at idle and squirting some extra fuel (petrol or gas) into the intake, the rpm will want to increase because the engine makes more power, that is until you get to a bit too rich and then it makes less power so the rpm falls again. Nearly all engines are set up to go rich at high loads, mostly because that makes more power, also because it helps protect exhaust valves (less potetial for remaining oxygen to oxidise/burn the valves).

An obvious difference between a petrol fuel system and an LPG fuel system is that petrol enters the engine as a liquid where-as LPG enters as a gas. Liquids are much denser (take up less volume for same mass) than gasses so for a given flow rate (mass delivered / time) LPG displaces more room in the engine intake than petrol. Fuel is one part of what the engine needs to make power, air/oxygen is the other part, with liquid petrol this isn't really something we have to think about because we could easily put as much mass of petrol into the engine as we wanted right up to pointlessly rich mixtures (pointless because it would be so rich the engine couldn't even run because it would just misifre) without it displacing much air. But with gaseous LPG it's a bit different, if we're already making most power from a very slightly rich mixture, even if a slightly richer mixture would give more power the effect can be that going slightly richer actually gives less power because we are starting to decrease the amount of air that the engine can use by displacing it with LPG. Imagine we have an engine that can shift 100cfm of air, if we pumped 100cfm of gas into the intake manifold it wouldn't suck in any air. Example (with just example figures), that same engine that can shift 100cfm of air might make around 100bhp but if we're only putting 75cfm through and it's making around 75bhp we're probably restricting it by using the throttle/accelerator pedal and we could easily get back to 100cfm of airflow by just increasing the throttle opening, airflow doesn't need to be a consideration during throttled conditions because we can easily get more airflow... so at part engine loads it doesn't matter if we do displace some incoming air... so if we can go rich and get more torque we might want to do that, and if we can go richer still and get even more torque we might want to do that, because again it doesn't matter if we displace airflow. But at full engine load when the engine can still only shift 100cfm or air, if we go a bit rich we will see some displacement of air but we might still get more torque and bhp, but go richer still and we start displacing even more air and we'll probably see a drop in torque and bhp and there's no way to get more air without modifying the engine.

The way I set them up depends on the vehicle, some have narrow band pre-cat probes some have wide-band, some have post cat probes some don't, some have exhaust/cat temperature probes most don't. Nearly all with narrow band pre cat probes switch to open loop fuelling at high engine loads, most but not all with wide band probes will run closed loop under nearly all engine loads often including full engine load. If they stay in closed loop mode the petrol electronics will dictate mixture anyway, if not I'll always set mixture to go at least a little rich, at least rich enough that it's going to read as say 0.8v on the usual type 0>1v narrow band probe, but I'll also compare post cat probe readings and EGT readings if the vehicle has the means. There are even vehicles that have weak cats, can help protect the cats by running a slightly leaner mixture at high loads than usual (gives them less work to do, work that normally causes them to do chemical reactions that causes them to heat up) so on some vehicles properly set up it can be safer for the engine to run LPG than petrol.

CNG has a higher stoch than LPG and it is less dense than LPG, so the displacing air aspect will be more pronounced on CNG than on LPG, so we might expect peak torque and BHP to be at a leaner mixture than on LPG but you still want to go at least slightly rich for most power and for protecting engine bits from heat and oxidation.

[CNG]

Why do you want to richen the mixture, surely the engine will run best and produce maximum power at Stoic?

As previous thread, because a 175 does NOT aim for stoic. Not over 4000 rpm anyway, it maps rich. This it does for max-power. On most engines 4000 rpm is about 3/4 throttle and that's fine. Problem being that my set-up is so agricultural thus WOT won't see 4000rpm even flat out. Rarely anyway. Peak is 4200rpm.

I was going to add a voltage divider circuit - thus it richens at 3/4 throttle which is 3200rpm... ish up to WOT. My Q focuses in the wisdom of this.

Seems, Simon sees it as doable. Viable though? We'll see.

[CNG]

Firstly, if you have an AFR meter designed for running on petrol (stoich 14.7:1) but you use it for LPG (stoich 15.4:1) then with stoich mixture on LPG it will show 14.7:1 not 15.4:1 and 15.4:1 would be a lean mixture on LPG, so you need to keep that in mind. Some AFR meters show mixture in terms of AFR, some show it in terms of lambda equivalence ratio (1 = stoich, less than 1 rich, higher than 1 lean), so if you've got an equivalence ratio of 0.95 it suggests the mixture is 5% rich, but again the equivalence ratio displayed might be geared for petrol and less accurate on LPG (but still more accurate than confusing AFR).

I set my WB to display Lambda. Not sure of the value of AFR, more so with a BiFuel. Suppose if you only ever work with liquid it's OK. For me, 1.0 Lambda is 1.0 Lambda and all that.

....But with gaseous LPG it's a bit different, if we're already making most power from a very slightly rich mixture, even if a slightly richer mixture would give more power the effect can be that going slightly richer actually gives less power because we are starting to decrease the amount of air that the engine can use by displacing it with LPG. Imagine we have an engine that can shift 100cfm of air, if we pumped 100cfm of gas into the intake manifold it wouldn't suck in any air. Example (with just example figures), that same engine that can shift 100cfm of air might make around 100bhp but if we're only putting 75cfm through and it's making around 75bhp we're probably restricting it by using the throttle/accelerator pedal and we could easily get back to 100cfm of airflow by just increasing the throttle opening, airflow doesn't need to be a consideration during throttled conditions because we can easily get more airflow... so at part engine loads it doesn't matter if we do displace some incoming air... so if we can go rich and get more torque we might want to do that, and if we can go richer still and get even more torque we might want to do that, because again it doesn't matter if we displace airflow. But at full engine load when the engine can still only shift 100cfm or air, if we go a bit rich we will see some displacement of air but we might still get more torque and bhp, but go richer still and we start displacing even more air and we'll probably see a drop in torque and bhp and there's no way to get more air without modifying the engine.

Ah... I'm aware to get a gain I'm working with just about the worst fuel to find a gain. Something must give. It hadn't entered my head that at WOT there's no space left to play too hard with this.

[CNG]

Thanks for that Simon. Seems to me if 0.8V is rich, a voltage divider gets bunged-in, hence when the Zirconic Sender outputs 0.8V, we use a voltage-divider such the Leo is sent 0.5V (assuming 0.5V is stoich).

Thus, overall, at 3/4 to WOT, we switch in a circuit - Leo is sent 0.3V (ish) less than the zirconic actually puts out.

0.5 .png (182.61 KiB) Viewed 5369 times

Shouldn't be too hard

[LPGC]

I think you may be missing the point a bit. As said stoich makes the best use of the fuel and at anything much less than full throttle you might as well run stoich mixture. At full engine load (at least above 4000rpm) the 175 itself aims for rich mixture - There is no point fooling the 175 into running rich mixture because it will aim for rich mixture anyway at full engine load. Besides that, there's also the rest of what I said about fuel displacing a greater amount of air if you run a rich mixture, effectively lowering the engine's volumetric efficiency.

You will lose power running on CNG because it has lower calorific value than petrol or LPG but there isn't much way around that other than modifying or fitting a bigger engine. You''ll lose even more power if the mixer venturi is too small or too close to the carb/thorttle body.

[CNG]

Maybe I am missing the point here. 'Typical' engines in the era of the Leo produce peak power somewhere in the range 5500-6500rpm, thus Leo enriches at 4000rpm upwards making 4000rpm the upper 1/4 throttle - 75%-100% throttle.

For a Series LR they really don't rev. You're rarely over 3200rpm, and whilst peak power is at 4200, power levels off much past 3500rpm thus, we can rev at 4000+ yet there's more noise than power. 3500rpm gives 5-6bhp less than 4200rpm - and if we run over 3500, noise levels are horrendous. In reality you can, only you're 'persuaded' not to do this.

Faster noise soon gets tiresome. Hence my thinking goes - 3200rpm plus: We fool Leo to behave much as it would with a more modern design at 4000rpm. At 3200+ rpm the cylinders are not 'full'. There's room to add more fuel. Of course by 4200rpm (peak-power) there's no gain. In fact, less air gives power loss? Now I grasp your point, it'd be better, and easy enough to do, to remove the voltage-divider from the sensors output at WOT, having the voltage-divider in play only within 70-90% throttle. Correct?

Or have I still got this wrong?

[Gilbertd]

What makes you think the Leo enriches at 4,000 rpm? I can floor mine in Sport mode so it hangs onto the gears until the red line at 5,500 rpm and, assuming everything is working correctly, it stays with the lambda display flicking between red and green. If it en-richened it, the display would go to red and stay there.

[CNG]

What makes you think the Leo enriches at 4,000 rpm?

Because Simon says it does, and that'd make sense.

.... it stays with the lambda display flicking between red and green. If it en-richened it, the display would go to red and stay there.

Er, no... at least not as I grasp this; it enrichens slightly by changing the bias, thus it flickers all the same, yet spends more time on the red side of the display

[LPGC]

Yes I believe they generally do go rich upwards of 4000rpm with max load box indicated (as in lambda reading going into the red and staying there) but this may be depending on what firmware is installed, some may still aim for 0.5 on the lambda under such conditions. Setting the 'full throttle option' will disable this anyway because if you set the full throttle option you tell it where the stepper should move to under such conditions and it will move to that position regardless of lambda reading. By setting cut off or full throttle option you create open loop mode(s).

It's not just 175's that go rich at high load high rpms. all properly set up elctronic pulsed injection systems do it (slave and standalone), multipoint continuous injection standalone systems like Tartarini Etagas do it, most but not all of the factory converted standalone systems such as Nikon Kottec do it even though they don't need to to keep the petrol ECU happy... they all do it to create max power and to a lesser extent to protect the valves from higher exhaust temps and oxidation.

I think I have seen some 175's not do it, 4000+rpm with high (say 4.7v) TPS reading but still for some reason didn't highlight the 4th TPS box or aim for full throttle enrichment. Could be a label / year / firmware thing or options/settimgs related.

[Gilbertd]

Interesting. It may well be a firmware thing but I know mine doesn't go rich above 4,000 rpm, lambda stays flicking between the two extremes and actuator stays at default. I've just got my second P38 back in a state that it needs to be for a long run next week so intend checking the settings on that one too. That has a Millennium on it so I'll see what that does at full bore.

[CNG]

I'm pretty sure 4000rpm goes slightly rich on the 175 I tested. My current gauge-overkill helps see this. You'll not see it easily via software - or my narrowband gauge, too much flickering - my WB gauge shoves to 0.98<>0.99 Lambda for longer.

Out of my stack of 175 the only two that work are OMVL Millenium.

[robertXX]

0.9 lambada , is what i found best for power .

graphs on my threads .

that'll read 13.3 afr on a wb set for petrol .

[LPGC]

0.9 lambada , is what i found best for power .

graphs on my threads .

that'll read 13.3 afr on a wb set for petrol .

Do you mean lambda sensor reading of 0.9volts or 0.9 equivalence ratio Rob? I would expect you mean equivalence ratio because 13.3/14.7 = 0.9, while a leaner (still rich) mixture would still see a 0>1v lambda sensor more or less max out at around a 0.9v reading.

WIth 0>1v lambda sensors maxing out at around 0.9v there's no way a 175 could aim for 13.3afr with closed loop control, they normally aim for 0.5v, can aim for (say) 0.7v, but can't really aim for 0,9v because if the lambda sensor maxes out at 0.9v there's no way for the 175 to see richer mixture than the leanest mixture that gives 0.9v lambda reading... At least not with the normal setup with 175 wired directly to the signal output of a narrowband sensor, but would be possible if connected to a WB controller that outputted emulated 0>1v lambda sensor output below 4000rpm and emulated 0.5/0.7v output at 13.3AFR above 4000rpm with wide throttle TPS reading.

[robertXX]

Eq ratio simon , and yes a NB sensor would be a bit lost out of is range , and locked on .9 .

a wb like the lm1 innovate can be programmed to use any range of voltage for it's 9 to 20:1 range , so can be set to put out a 0-1v signal , or even i think a 1v to 0v signal , i am sure it could be tailored to fool the 175 some way .


i did many years ago map a NB sensor voltage to afr , and it as actually less switch like than i expected , it may be with a bit of mapping it's voltages against a wide band , possible that say .78352 volts would equate to the .9 lambda needed?



regards
robert

[LPGC]

i did many years ago map a NB sensor voltage to afr , and it as actually less switch like than i expected , it may be with a bit of mapping it's voltages against a wide band , possible that say .78352 volts would equate to the .9 lambda needed?

I know what you mean regards the possibility of mapping nb voltage to mixture but I think more likely in practice it would be too hit and miss due to the tolerances involved, mostly accuracy of the voltage reading. I just had a look at some online charts of nb voltage versus mixture, the pic below seems typical and shows that 0.9v = 11.78:1, 0.8v around 14:1, 0.85v around 13.4:1... so the voltage reading would need to be very accurate and the results repeatable over a wide range of temperatures, weather conditions (also affecting engine bay temps etc), slight swings in voltage bias. I don't know of any vehicle manufactures that try to make use of an nb lambda to hold a specific mixture other that stoch. But I would suppose it is viable for the 175's and similar to aim for a voltage of (say) 0.7v or 0.8v, just not 0.9v.

NB 0to1V lambda sensor response curve.jpg (191.64 KiB) Viewed 5186 times

[CNG]

I take your point. It'd be hard to achieve. I also see an issue with real world use. Power is literally startling in it’s complete absence

In the last week or so, I've made a mental note of how often I have my foot flat to the floor, this when I would not want enrichment. It's a LOT. So much so I think any tweak would come in as a net loss. This because there's so little power I'm forced to use the pedal more like a switch than a potentiometer. Use of throttle is more akin to the moped of my youth. ON or OFF. I'm simply not using throttle like you would in a vaguely more powerful vehicle. At 25-30mph I'm flat out to keep with standard traffic. I assure you, this isn't going fast.

The idea is kinda viable, yet a non-starter because of what this is, real-world, the juice is not worth the squeeze. This assuming it's easy to achieve, and the fact you can't trot out too much use of this trick tells me it probably isn't.

[CNG]

The bigger issue is that I can set 'Out of idle' limits - red-region; I suspect because we're not much over 3000rpm they won't come into play.

[Gilbertd]

The out of idle limits do what they say, they limit the maximum and minimum actuator positions but still following the lambda sensor signal. You can set it to 255 if you like but it won't ever go that high unless the mixture is really lean.