Page 4 of 4

Re: TVR Chimaera LPG Conversion Advice

Posted: Sat May 20, 2017 9:23 am
by Gilbertd
Not having a pop Simon but you may be missing the point here a little. The OP is not advocating the Canems as an alternative to the OE ECU fitted by a manufacturer on a modern car but as a way forward for a car that originally was fitted with carbs, so no ECU at all or a very crude, early generation injection system. On the former, the only way to install LPG would be with a singlepoint running open loop unless a lambda sensor was added to give a bit more control. On the latter, on something like the 14CUX that the OP originally had, a conventional piggy back system would work but would never be perfect and it would still be necessary to retain the original distributor based ignition system with no control over the ignition timing to suit the different fuel characteristics. So in those cases, the Canems has been proved to be a viable way forward which gives a considerable improvement in both fuelling and ignition characteristics and control.

It still does beg the unanswered question on why he didn't upgrade the engine and control system to the vastly superior Motronic and start from there.

Re: TVR Chimaera LPG Conversion Advice

Posted: Sat May 20, 2017 1:35 pm
by classicswede
Europegas have ecu's where the timing can be altered and there are also timing advance processors available. There are also options like the 123tune or megajolt available for ignition only.

There is also the option of stand alone lpg injection systems where lpg can be electronic and petrol could stay carb or mechanical injection .

There are options out there and as always more than one way to get the result. The system used here is now show to work and is a option.

I have to agree converting to the later motronic engine and management would have been a better and more cost effective way of doing things. The down side would be if other engine upgrades are planned then the Motronic would not be ideal, a fully mappable system is really then the best way to go. There are lots of mappable systems to chose from. For the rover v8 there are plug and play mega squirt kits available.

Re: TVR Chimaera LPG Conversion Advice

Posted: Sat May 20, 2017 6:40 pm
Hehe... This is a thread where like Marek says it would be easy to talk at cross purposes and for confusion to be a factor but I think I see it fairly clearly...

Start of thread 4 years ago, OP's intention: Convert a carb'd Rover V8 to petrol injection, dual fuel with LPG injection.

Various ways of achieving this were suggested, the OP decided on Canems, maybe because it seemed to offer everything in one box, maybe because he was doing some engine tuning and an original map in a standard Rover V8 ECU wouldn't be suitable for the engine's new state, I can understand and agree with the thinking.

Canems ability at correct fuelling on LPG was doubted by others because it wasn't known at the time that it allows correction for gas temperature or pressure. I would agree with such doubts, so now I'm thinking a standalone petrol controller with a piggyback LPG ECU would be a better idea.

OP reckons Canems does allow those corrections. My doubts about correct fuelling on LPG are pretty much alleviated.

OP reckons some experimentation has been done with Canems and it was found that pressure and temperature correction are not necessary. My doubts are back, because this doesn't agree with laws of physics or with my extensive experience setting up and calibrating LPG systems, my results do agree with the laws of physics but there is another aspect that can more than offset any correction that might be necessary for density at short gas injector pulse duration such as at idle... I nearly didn't mention write the next bit of this paragraph but here I'm talking about injector response. At short pulse duration such as at idle, if you take 5% off injector pulse length then you'd expect to reduce fuelling by 5% but if pulse length was short enough to start with the -5% adjustment could effect fuelling by -10% or -20%. Someone who doesn't realise this might try dialling in density correction at -5% (which physics says should be correct), see that it has effected fuelling by -20% and conclude that with density correction enabled and set up according to laws of physics fuelling is incorrect by -15% but without density correction fuelling is only incorrect by +5%, leading to the next conclusion which would be that it's better not to use density correction. Bare with me on the figures because it is only the concept that I'm trying to put across here... I know that the OP reckons density correction isn't important, so implies that he got 0% difference with hot or cold gas. In this example if the same comparison was done with the injectors working at longer pulse duration it would be evident that density correction agreed with laws of physics, the exaggerated results at idle would be that the -5% correction pushed injectors towards such a short duration that they were operating below the minimum pulse duration at which they started to work linearly - the real fix in this example would be to reduce pressure or fit smaller nozzles, so that the injectors were working more linearly even with that 5% density correction subtracted from the pulse length. I can quite easily imagine someone who has fitted a completely different fuel system to their vehicle concentrating more on idle characteristics from cold start rather than starting the car from cold and immediately setting off to give it a boot full, and in the process missing the opportunity to check fuelling with gas that is colder than usual operating conditions with a boot full, and therefore discounting density correction, especially if they're not used to fitting LPG on more demanding engines and noting how necessary density correction is, and especially if running gas injectors at low pulse duration at idle because they hadn't thought about injector response.

Anyway, during the discussion it was pointed out that Canems set up to run both fuels would cost more than MS plus a piggy back LPG ECU, so now I'm all in favour of the MS with piggy back LPG ECU setup. MS with twin maps could have the same fuel map for LPG as for petrol, but it's LPG map could be set with ignition timing to better suit LPG. It doesn't matter that the MS LPG map isn't designed for LPG because the primary purpose of a piggyback LPG ECU is to take the petrol map and make it appropriate for LPG.

4 years pass.

The OP posts to tell us that the Canems has worked great on petrol and on LPG. I reckon this is a decent achievement, but it isn't clear if the engine mods (if any were done) would justify the extra expense of an aftermarket petrol injection system compared to say Gilbert's proposal of using a Rangerover ECU with the immobiliser function disabled and a piggyback LPG ECU. I still have reservations about the gas density compensation - the Rover V8 is a low demand engine and can seem to run well even if the driver is unaware fuel mixture isn't always quite correct, or if the fuelling ECU is having to swing fuelling (fuel trims) to keep mixture correct. If the engine wasn't modded then Gilbert's suggestion of an original ECU with a piggyback LPG system would make a lot of sense to me, If the engine was modded then Dai's second suggestion of using MS with a piggyback LPG ECU would make a lot of sense to me.

In one post OP equates adding a piggyback LPG ECU as cobbling two systems together Heath Robinson style. In another post (quote)..
ChimponGas wrote:LPG piggyback ECU designers know this and they understand their market aren't true engine mappers but guys trying to make a living out of price driven LPG installation that must work perfectly within 30 minutes or so of being fitted. This is why LPG piggyback ECU designers invest so much of their energies in self calibration algorithms and strategies designed to essentially use processor power to calibrate itself with little more required from the installer than a few clicks of his mouse and 30 minutes or so of driving the vehicle.
the OP misses the point that piggyback ECUs are piggyback because they have to be in order to fulfil the purpose at which they are aimed, and dismisses all LPG installers as people who rely on self tuning LPG systems. While I would agree that self calibration facilities that are built into many of the laptop tuning programs (and in part into firmware of the LPG ECUs) are aimed at amateur installers and done in a day installers, there are better installers who do a very thorough manual job of LPG system calibration and in the case of some vehicles manual calibration is the only way of achieving good results. I think of self tuning facilities as a gimmick, but some of the ECU's that have such gimmicks are extremely capable when the gimmicks are turned off.

Like Dai says there are plenty of options for adding an aftermarket petrol system and/or aftermarket LPG system.
But most of those options will make most sense on older vehicles and kit cars, not so much sense on modern vehicles where the original petrol injection ECU has to do all the other things I mentioned otherwise half of the systems/functions on the vehicle won't work. The older engine in this thread is low demand like most old engines, it will run without much in terms of symptoms if mixture isn't quite right which could be helping mask fuelling issues due to lack of density correction, it doesn't have VVT or Atkinson cycle like features and it isn't high revving, batch firing isn't a problem. If that wasn't true then recent MS running full sequential with features to control modern engines such as VVT, in conjunction with a piggyback LPG ECU, might highlight shortfalls with this type of setup, but the vehicle manufacturer's ECU (if it could be remapped to suit the state of the engine if any mods have been done) might highlight shortfalls with the MS setup.

The OP will have brought up piggyback ECUs in the quote above because most posters have mentioned them at some point on the thread, but his post doesn't refer to piggyback ECUs only in the context of what might be the best system for this application, it refers to them more generically, thus my last post also refers to them more generically. No bad feelings or big arguments to pick with the OP, I can see validity in a lot of what he's said, in fact my next post (this post) was going to be more along the lines of I think we're all getting closer to agreeing... Just wanted to get the points across about piggyback ECUs needing to be piggyback, installers mostly convert the type of vehicles that only piggyback systems would make sense on, for our purposes some of them are extremely good at what they do and the only option. If a good experienced installer was to go the whole hog on an older engine like the one in this thread then he'd probably go through all the same considerations discussed on this thread but might decide on a a different route to the OP... especially bearing in mind that any costs would have to be passed onto the customer if this were a customer vehicle, also relevant is potential for the customer to return with complaints about fuelling maybe after a change in the season.


Re: TVR Chimaera LPG Conversion Advice

Posted: Sun May 21, 2017 12:24 am
by marek
Dear Simon,

Both approaches mentioned factor in "injector response":-

The aftermarket stand alone ECUs such as Megasquirt or the Canems one have an injector calibration settings which dial in the injector speed or "dead time", per injector plus even a voltage correction for that. The "dead time" is the time for the injector to open minus the time for it to close*, so the actual injection pulse commanded is this "dead time" plus the time added for fuel and any adjustment (percentage or otherwise) added or multiplied up on top.

The equivalent for a piggy-back is done via the setting up of the graph which shows (perhaps) a 1.13 multiplier to the petrol pulse at low petrol injection pulse numbers through to (perhaps) a 1.03 multiplier at high petrol injection pulses.

It is of course a real chore to set this sort of thing up by comparison to the piggyback system's autocalibration.

Motronic have been mentioned a few times. Whilst robustness of the product and after sales support are a factor in choosing an aftermarket ECU, it needs to be pointed out that having the ability to alter settings yourself is also a consideration here. I don't believe that Motronic will offer you any software to do it yourself or give you a datalogger, whereas Megasquirt (and some others) will.

Whilst fuel temperature compensation is desirable, I can only speculate that the particular reducer location, its size and the ambient airflow under the bonnet mean it largely moves in tandem with what is reported for the plenum air temperature, which is conpensated for.

kind regards

*worked examples:-
injector opening time = 2.5 mSec
injector closing time = 1.0 mSec

1/injection pulse = 0.9 mSec => no fuel injected (injector didn't even open as commanded number was less than 2.5mSec)
2/injection pulse = 2.7 mSec => 1.2 mSec of fuel injected (none injected in first 2.5mSec, then 0.2 injected, injector commanded to close, but 1 mSec of further fuel gets through before it actually closes)
3/injection pulse =5mSec => 3.5mSec of fuel injected

There is even a short pulsewidth calibration graph, as it is recognised that very short pulsewidths don't follow a linear pattern.

Re: TVR Chimaera LPG Conversion Advice

Posted: Mon May 22, 2017 11:29 am
Hi Marek,

If you have a discrete stand alone map for LPG, why the need for a short pulse-width calibration graph? I mean you could just have one map which factors in injector response at low pulse duration. In fact, while injector response effects the map particularly at low pulse duration, the whole of the map has to take into account injector response.

Injectors don't work binary fashion, it isn't the case that they're either open or closed or that they take X ms to open and close regardless of conditions. Some injectors are slower to open and close when worked at a higher pressure. All can flow a fraction of fully open quantity when they are part way open and will be part way open during much of the time during both the opening and closing phases. For some pulse lengths they won't even get to fully open before they begin to close again - this implies that linearity can be effected by nozzle size too... an injector with 1mm nozzle may reach peak flow when the plunger is lifted 1mm from it's seat, but the same injector with 3mm nozzle will only flow 1/3rd peak flow when the plunger is lifted 1mm from it's seat. All of this is more relevant to some designs of injectors than others, because the operation of some injectors only partly involves the lifting of a plunger from a seat. It isn't the case that an injector goes straight from injecting zero fuel for a certain pulse length to injecting X ms worth of fuel for a pulse length that is only 0.1ms longer, in reality at such short pulse lengths around injector minimum pulse spec the injector doesn't dose fuel very accurately at all but some injectors are far more predictable than others, thus latency and accuracy are not always the same thing. Also, can forget manufacturers on paper specs and any hype, you'll only get to know real world performance through experience.

I think I've previously noticed you referring to injector response before on other threads, thought you were wrong but I didn't say anything... Something along the lines of gas duration shown on screen doesn't include the injector opening/closing time... Duration shown on screen is the actual pulse length of the injector including any opening closing time for any LPG system you can think of. The only difference between systems in anything like this regard is that some don't include the response time in the calibration multipliers. E.g., say Stag includes dead time (according to which injectors are selected) but KME doesn't, this would make a difference in the multiplier the installer must enter (particularly in the low pulse time range), because Stag seemingly wouldn't need as high a multiplier to calculate the same gas injector pulse duration, but KME shows a second multiplier line in which the dead time is included and the dead time used to calculate this depiction of multiplier is adjustable by the user separately to just selecting type of injectors. While as implied above, the only true dead time is at extremely short pulse lengths.
marek wrote: Whilst fuel temperature compensation is desirable, I can only speculate that the particular reducer location, its size and the ambient airflow under the bonnet mean it largely moves in tandem with what is reported for the plenum air temperature, which is conpensated for.
I don't agree that gas temp will move in tandem with intake air temp - Start the car (say start on gas) on a winter morning at say 5c outseide temp and gas temp will very quickly go way below outside temp, will then soon rise to something like normal range (say 30c), when all this time IAT is still close to outside temperature (5c). Go through McDonalds drive through on a hot summers day and gas temp may reach 75c while IAT is 25c. Go for a cruise on a Spring/Autumn day and gas temp may be 65c while IAT is 15c, but put your foot down and gas temp may fall to 40c while IAT stays at 15c. Then we could talk about pressure... and just to make things worse, when your reducer is cold (so gas temps are cold) pressure is likely to be a bit higher than when the reducer is warm, thus compounding the density point.


Re: TVR Chimaera LPG Conversion Advice

Posted: Mon May 22, 2017 1:39 pm
by marek
Dear Simon,

I'm not sure why you are taking the discussion in the direction of injector characteristics, as it doesn't really have too much relevance to the OP's thread.

Most people will chose injectors which are linear over their full operating range and/or adjust the default rail pressure such that they won't ever see small pulsewidths. My setup is no different and doesn't see low pulsewidths on either fuel.

If the oem petrol setup does see small pulsewidths, then that isn't necessarily going to translate across well to an emulated lpg pulsewidth, as the lpg injectors may require a different characteristic performance to the adjustment factored into the petrol pulse. Such a situation will typically only occur at low load overrun with very high vacuum in the plenum.

Injectors do indeed not work in a binary fashion. The example was illustrative so people would understand that dead time is made up not just from an opening characteristic, but also has a component related to the closing charcteristic, a different characteristic as supply voltage varies and finally, that its performance when commanded to give short pulsewidths may be radically different from the petrol injector it is copying. For example, two injectors (one lpg, one petrol) could have the same dead times, but the minimum amount of fuel delivered could vary because of the closing times being different to each other.

With regard to P and T corrections, it's proably best to restrict the discussion to how it works once the car is up to operating temperature. This is because the stand alone system and the piggyback vary in their approach during warmup:- the piggyback has to factor in that the new fuel has a reverse teperature gradient compared to the base fuel, whilst the stand alone is simply doesn't care and is calibrated totally differently for the two different fuels. The performance characteristics of petrol are of no interest to the stand alone ECU user.

What is interesting is to consider the heat transfer going on in relation to vapouristion:- if there were a heavy demand for fuel such that heat extracted from the coolant were to exceed the flow through the reducer, then I can easily see that temperature adjustment would need to be looked at. In normal operating conditions, a properly functioning cooling system will collar the operating temperature of the reducer to a range about 5'c above the thermostat opening temperature and the reducer is only producing gas on demand, so it is reasonable to focus on airflow and variation in under bonnet temperature.

Your example that wants to focus on warmup forgets that the standalone is already calibrated with a totally different coolant warmup curve for lpg anyway and that the subsequent gas temperature that results from coming a colder coolant is simply part of that stand alone coolant calibaration - there isn't a need to also factor in gas temperature variations prior to getting up to operating temperature because it has alreeady been done once. You can trade the two tables against each other in pretty much a 1:1 ratio so temperature adjustments prior to reaching operating temperature are pretty much a dead rubber. That is the reason for my wanting to direct attetion to under bonnet temperature variation and inlet air temperature under ordinary running rather than simply focus on gas temperature in isolation.

kind regards

Re: TVR Chimaera LPG Conversion Advice

Posted: Mon May 22, 2017 1:57 pm
by marek
LPGC wrote:Hi Marek,

If you have a discrete stand alone map for LPG, why the need for a short pulse-width calibration graph?
The stand alone software lets you set up different injectors with different dead times and different low pulsewidths charcteristics. It then lets you set up basic fuelling for all cylinders for each fuel type. Then it lets you, should you wish, to alter have a trim map (not just a flat percentage) per injector as trim may vary with duty cycle or manifold pressure. You can then even have another map per cylinder for ignition trim.

That's slightly more sophisticated than just throwing it all into a new fuel map as how each injector performs at different voltages or pressures or duty cycles or low pulse widths may vary.

It's a racing certainty that an lpg injector will not only have a different characteristic performance to the petrol injector it may be asked to emulate, but that there any be variation even in a batch of the same injectors. This may not be a factor for a four cylinder engine but may be a concern for a twelve cylinder engine.

You don't really want to have a petrol injector's adjustments made to a different type of injector unless that injector also shares the same characteristics.

kind regards

Re: TVR Chimaera LPG Conversion Advice

Posted: Tue May 23, 2017 11:34 am

Reason I went in the direction of injector characteristics is because at low pulse duration there is the potential for characteristics to mask need for temp correction.

No injector with any latency (or dead time), which means all injectors (though latency of petrol injectors is very small) is ever linear across it's full operating range, the opening/closing time must always be factored in (which flies in the face of the term linear) but the opening/closing time represents a much bigger fraction of total pulse time for short pulses than it does for long pulses. Latency is one thing, accuracy is another, both affect linearity. If injectors are slow to open and close but take exactly the same time to open and close every time then accuracy doesn't have to suffer even at low pulse times. Suppose injector A takes 2ms to open and close while injector B takes 3ms to open/close. If both have the same 10% margin of error for open/close time then injector A has 0.2ms margin of error while inj B has 0.3ms margin of error, so injector A in this case would be more accurate at low pulse times... but while quicker open/close injectors are usually more accurate at low pulse times this isn't necessarily always the case, because while injector A has that 10% margin injector B might only have 5% margin. Linearity can be more concerned with speed of open and closing than outright accuracy because linear implies if you pulse the injector for twice as long you get twice as much gas dose and if we factor in the open/close time injector A is going to more closely match that definition than injector B (and have a flatter map) even if injector B has that 5% margin for error while inj A has 10% margin for error.. But all 3 factors effect each other in application and, as you've implied, the best way to ensure good results is to fit injectors of a good spec and have them pulse for long enough at low loads that any margin of error for open/close time is only a small fraction of total injector pulse time. If we fit very accurate injectors but they have a lot of latency then we have less resolution (for want of a better term), by which I mean slow injectors at idle might need to pulse for 4.5ms where fast injectors at idle might need to pulse for 3.5ms, but regardless idle pulse time we are restricted to a certain maximum pulse time (which might be 16ms)... an injector with pulse width between 3.5ms and 16ms has more scope for adjustment of fuelling than one that pulses within a range of 4.5ms and 16ms.

Quote 'with regard to P and T corrections', no mate I don't agree, I described a few scenarios that highlight why I don't agree.

I don't agree with your gas temp assumptions either... I'll name a few variables: C coolant temp, R reducer body temp, T tank temp, W weather (external) temp, Z cooling effect of gas changing phase in the reducer, V gas vapour temp, U under bonnet temp. The design of the reducer tries to bring V up to R, which would be the same as C if the engine used a negligible amount of gas, so standing gas in the reducer would reach R which would equal C. But when running on gas V is affected by all of the variables. If the reducer wasn't plumbed to hot water then both R and V would eventually equal T - Z, very cold. Of course the reducer is plumbed to hot water so R conducts heat to V but there is less time for this heat transfer to occur in the reducer when gas flow (engine demand) is higher because gas will be flowing through the reducer faster, not only that but Z is also greater when the reducer is flowing more gas, not only that but Z starts to lower R too (because R also relies on heat being conducted from C to R). When vapour is in the rubber hose from reducer to injectors there is another opportunity for V to be steered towards U, again this steering will have more effect when gas flow is small and vice/versa, while U itself is effected by many things such as exhaust heat, radiator flow, vehicle speed and W. If the car has been parked up engine off for a while T will also equal W but T also decreases below W when gas is drawn from the tank, the faster gas is drawn from the tank the greater the cooling effect on the tank. If you've just been booting it and then you slow down, or if you've been parked up in the sun for a while with engine running U is likely to be higher. A sudden change of throttle position won't have much immediate effect on any of the variables but continued use of that throttle position will effect all variables over time and you cannot reliably predict what any of those variables will be. Past occurrences (recent driving conditions) affect current V. To say V simply rises and falls with U is a big assumption that takes hardly any of the variables into account, doesn't agree with physics and doesn't agree with my findings. It would be possible to make the assumption closer to reality by running V through an intercooler at the front of the car but you wouldn't want V to equal W in very cold conditions... That point leads me onto this next point - Especially in cold W conditions you wouldn't want V to be close to W, but you reckon V and engine intake air temp rise and fall closely together anyway even without a V intercooler?


Re: TVR Chimaera LPG Conversion Advice

Posted: Tue May 23, 2017 9:20 pm
by marek
Dear Simon,

One reason why you may be seeing larger temperature variations than the original poster or myself could be in how installations are configured. In my case, the car has 20 litres of coolant and ALL of it runs through the reducers. This may account for why I see relatively little movement in temperature once warmed up.

For a car which has much less coolant and is teed at the heater matrix pipes, the same heat has to be extracted using much less water flow. Temperature movement is leveraged if only a small amount of coolant is used to service the reducer. I've seen this effect on my car:- when turning the heater on, the reducer on that side (which now shares its water with the heater matrix) drops 3'c. With less coolant and/or a lower proportion of the coolant running through the reducer to start with, you may well be building in such variability in your conversions. Indeed, if insufficient heat can be extracted from coolant flow, then the reducer takes it from the environment and ultimately tries to freeze up.

The original poster reported that variations were easily accommodated within his closed loop running.

The whole point about stand alone custom installations is that you design a system from scratch as well as can be done, rather than shoehorn in the third party product which is available, into somewhere it probably doesn't quite fit in insufficient time and to a cut-throat budget.

I only joined this discussion because I couldn't see why people were knocking it. The OP is happy with his conversion and reports that it runs well.

kind regards

Re: TVR Chimaera LPG Conversion Advice

Posted: Wed May 24, 2017 12:11 pm

I have configured top side of a thousand vehicles myself and seen thousands of other configurations, not all use T's for reducer plumbing. Not sure what you mean by all coolant runs through your reducer, how did you plumb it in? We must agree that another reason for not seeing changes in reducer and/or gas temperature would be if not monitoring them...

Your point about greater flow of water through the reducer being more likely to maintain reducer temperature cuts both ways - Like I said above, more gas flow through the reducer, which has a greater cooling effect than a smaller gas flow, combined with the greater gas flow moving more quickly through the reducer and therefore spending less time being heated in the reducer... will lead to cooler gas flow out of the reducer even if reducer body temp remains the same (so regardless of how much hot water you pump through the reducer)? Greater gas flow through the reducer also implies greater gas take-off from the tank, which lowers tank temperature, which lowers temperature of gas flowing even into the input side of the reducer, and the tank temp issue will continue to have an effect long after an event such as booting it (or filling up with gas)?

There is variability in all conversions, the variability will effect fuelling if it is not accounted for with aspects such as temp and pressure correction addressed. Indeed the more advanced vehicles become the greater the number of variable aspects are addressed.

Nobody here has made a fuel system from scratch, your fuel system uses all of the same components that could be fitted as part of any LPG installation but uses an all in one ECU that controls both the petrol fuelling and the LPG fuelling directly instead of an OEM petrol ECU and piggyback LPG ECU, nobody here designed the all in one ECU either, all components anyone here fits are either OEM or 3rd party including your LPG injectors and LPG reducer. I allow plenty of time for conversions, this is usually a few days, I certainly don't need weeks to get everything spot on or for it to remain spot on for years to come - You may note I have that reputation, a reputation I probably wouldn't have if my results didn't speak for themselves, my results couldn't be so good if I didn't very thoroughly understand what I am doing, knowing what I am doing involves all of the tech aspects we are discussing here.

If you're taking much longer than a few days to get decent results and you were to offer the all in one route on a commercial basis you would have to charge appropriately for your time, it also turns out that the all in one route is more expensive in parts - Your point on insufficient time and cut throat budgets doesn't make much sense when applied to installers such as myself but could be turned around to show how the extra money and time the all in one route takes could be wasted. Whether or not the extra money and time are wasteful will depend on if there are enough advantages to the all in one route, I can see some advantages but I also see disadvantages, we have been having that discussion where on the one side you and the OP seem to think it worth the extra money and time, on the other side almost everyone else disagrees and some of us are putting over some pretty convincing reasons for disagreeing, not least references to laws of physics that are indisputable (yet seem to be disputed by some). I don't believe my question on whether anyone here would offer the all in one route as a commercial proposition to paying customers with warranty was answered. If we're talking about shoehorning it is just as likely anyone taking your approach would shoehorn than anyone taking the more usual approach... plenty threads and pics of my installs on this forum, people could make their own minds up as to whether I shoehorn.

Some dodgy piggyback tuning is easily accommodated with closed loop running, but that doesn't mean it is ideal.

I got the impression early on that you were in some way involved with the OPs project, maybe because you don't often post on this forum or mention much about many LPG conversions you've done. I think the OP cleared that up for me by being first to mention that you'd done a similar project. My thoughts on this discussion are the same now as they were early on - If anyone doesn't have much experience with piggyback systems how can they know that the standalone route is superior? You could turn that around and say how do I know the piggyback route is superior - I started from the position that all in one could only ever equal piggyback not better it but then when further info on all in one came to light that steered me to thinking piggyback was the better route. Other standalone petrol controllers are available (that are not all in one) such as MS, piggyback systems can piggyback onto any batch fired or sequential petrol system including any standalone, a piggyback system on the back of standalone would compensate for pressure and temperature... compensations which you consider unnecessary or compensated for by default, but which I know through experience are necessary and cannot be compensated for by default. I even went off at a bit of a tangent to partly explain a scenario that can mask the need for density correction at low loads, I might not have expected someone without a lot of experience with piggyback systems to have already anticipated the effects I mentioned, but apparently they have and they still don't see the need for temp correction.


Re: TVR Chimaera LPG Conversion Advice

Posted: Fri Jun 02, 2017 7:13 pm
by ChimponGas
So, on the subject of GEMS being better than my batch fired Canems duel fuel stand alone ECU I think it's important to accept while fully sequential injection may indeed give a slightly more refined idle and lower emissions at idle it really isn't the holy grail many will have you believe. There's good reason why most stand alone ECU systems continue to stick with batch fired injection and it's really all down to a basic balance between ease of mapping vs the benefits going fully sequential would genuinely deliver in the real world.

I think it's fair to say it'll be blindingly obvious to anyone with a basic foundation in fuel injection that injector opening times over 2,000rpm are so fast any benefit of implementing fully sequential injection over batch are largely lost. Now lets look at the patient we're working with here, our old friend the Rover V8 offers extremely poor levels of combustion efficiency by modern standards so the margin of gain is diminished still further. Drop old Rover in a 1000kg highly involving analogue sports car who's owners care little for the much hyped environmental concerns of the tree hugging fraternity and is highly unlikely to be using said V8 as a limousine or urban carriage.... one can start to see why the tiny idle quality and idle emission benefits of fully sequential injection fell right to the bottom of this TVR owner's design brief.

All this adds up to a neat little equation that shouted at me not to bother with the additional cost and complexities of going fully sequential, the truth is the real world benefits just aren't there, so while some may assume rather short shortsightedly that all this wasn't looked at and studied very carefully before this project was started is really showing themselves to be rather ignorant and disrespectful to the highly skilled team that delivered what has already proved itself to have delivered a highly successful outcome.

Back to the GEMS idea, while I'd agree this is an interesting and sophisticated OEM quality ECU that can be had for very little money, it is not what you call user friendly. Lets remind ourselves of the patient again, its a TVR not a Range Rover and while the engine is basically the same power plant there are some very important differences that mean you cant just lob a GEMS on it and get up the road. For starters the TVR will invariably run a much more progressive camshaft in both lift and duration, most sport heavily ported cylinder heads, different pistons and a raised compression ratio to go with them... larger valve sizes, a lighter flywheel the list goes on not to mention the use of the Borg Warner T5 manual gearbox over the Range Rover's auto and the requirement for a cam position sensor on an engine never designed for one because all TVRs ran a distributor and intermediate timing cover to suit. Mapping the petrol side alone represents a big challenge with the GEMS compared with the laughably easy software and user friendly process of creating the perfect calibration with my Canems ECU.

And lets forget my objective was always to implement LPG, well with the GEMS fitted it really would have only left me with one option, slap an LPG piggy back on it. Now while I appreciate some of these piggybacks are getting quite smart these days, it's still a slave ECU that does nothing with ignition which is a crying shame if your running a performance car (or any car for that matter) on LPG, because not only is the fuel 110Ron but it's burn speed is very very different to petrol too. Being able to create bespoke ignition calibrations from scratch on a rolling road to suit each fuel type really shouldn't be underestimated, even firing my injectors using the ancient but tried and tested batch method. This is not hearsay or speculation either, sure the car ran just fine on LPG with the petrol ignition map, we only did this to prove a point, using that as a control by the time we'd finally built the perfect bespoke LPG ignition calibration we'd picked up 12 ft/lbs of torque at peak and roughly 7% everywhere else, and trust me the dyno does not lie. The gas map on the fuel side is obviously going to look different, but you'd be amazed at the differences between the ignition calibrations, and as we know what you do with the ignition will influence what to need to do with the fuel and vice versa, this is really where the Canems system really starts to surge ahead of the traditional piggyback LPG ECU sliced to a petrol ECU that you can't even calibrate, and I'm very much including the fully sequential GEMS petrol ECU in that.

Hopefully you can see now why the oversimplified idea of putting a GEMS engine management system on a Rover V8 TVR isn't quite the walk in the park some may think it is, and why it's not the brilliant idea LPG installers clutching their faithful LPG piggybacks will have you believe, and I haven't even got onto the blocks TVR were using either. The biggest hurdle for most TVR GEMS conversions would be whether the block is the 38A cross bolted type fitted in the P38 range rovers as these have the lugs and fittings for knock sensors etc. For many TVR owners this is going to be the deciding factor on whether a GEMS install is a viable option or not, all TVR Wedge models used the earlier pre-serp engine as did many Chimaeras and Griffs. The mighty 5.0 lite engine was somewhat of a TVR hybrid where TVR Power of Coventry who built all TVRs engines started from scratch with a very very bare un-machined block, for example while it has the bosses in the casting for cross bolting TVR never drilled and tapped them. Fortunately (in a way) I have a 4.0 litre serpentine, with this model TVR did indeed use the large journal cross bolted block so although I rejected the GEMS install idea from the outset I did at least have the option of considering it as unlike a lot of RV8 TVR owners I am at least blessed with a later type 38A block.

However to be fair to the GEMS idea there are a few in the TVR world (with Mark Adams in support) who have been beavering away on the idea for a few years now, but it has to be said progress has been painfully slow with only one car actually up and running on it after all this time, and I'm sure he has no intention of running anything other than petrol. As I understand it if you include Mark Adams, there are three guys collaborating on the project with just some of the challenges delaying the project being designing a suitable trigger disc to fit the dizzy cam nose with a JP timing gear, but the challenges as I've already pointed out go way beyond this, but I believe much legwork and 'banging head against the wall' on all the other stuff to make it all work has now been done, these guys really do deserve a medal for all their efforts.

So, and given I started all this over five years ago when the TVR GEMS development team hadn't even started their work, and given I wanted something that I could get working quickly with all the benefits of a stand alone ECU (on both fuels and ignition) the batch fired Canems Dual Fuel made for the perfect choice.

GEMS my arse, and chucking an old Rangie lump in my 150mph+ TVR clearly aint the answer either, I hope that puts an end to all this GEMS nonsense because for all the above reasons and many more besides, especially when my original objectives are factored it, the whole GEMS idea made for and still makes for.... a very poor idea indeed.