If you'll be changing the Maf anyway you might as well do that and see if it makes any changes to fuel trims.

For the benefit of others who might be interested in reading this and for my own notes...

Had a Whatsapp chat with Khemitude today, showed me around the engine bay during the video call. It's got a (relatively recent) AEB badged AEB2568 with old Bigas reducers and Valtek type30 injectors. The 2 injector rails are each fed from a separate reducer, no balance pipe. The gas inlet points on the are in the 'old skool' position on the Thor plenum so quite a distance from petrol injectors and engine inlet valves. This also means that each rail feeds 2 cylinders on one cylinder bank and 2 cylinders on the opposite cylinder bank, and without a balance pipe this means it's likely 2 cylinders on each bank get higher pressure gas and hence richer mixture than the other 2 cylinders on the same bank.

Plan is for the OP to come to me maybe next week. I'll replace the 2 old reducers with a new MJ one and calibrate it as well as can be.

Just had an old mk2 (facelift) Jeep Grand Cherokee 4L straight six here with a very similar problem... The Bigas reducer was leaking gas not to the cooling system but to the gas output, so when the system was turned off gas pressure rose to >3bar and was ballooning gas pipes. I replaced the Bigas single reducer with a MJ reducer. One of the Valtek type34 injectors wasn't working at all (coil OK but no output of gas) on number 1 cylinder, so I left it set to run on 5 cylinders on gas with number 1 on petrol until the owner has replaced that injector himself. It also has an occasional lambda signal shorted to heater circuit problem, so the owner is going to replace the pre-cat lambda sensor.

Turns out it's one I converted in 2008 but I originally fitted a closed loop mixer system (back then I'd offer customers the choice of a mixer system or sequential system and the mixer system would be a bit cheaper), so the front end components have been changed since I originally converted it.15 years ago. Heh, first thing I said when looking under the bonnet was "Wow your ECU is just hanging and resting against the engine". He says "Well you fitted it". I said "I didn't fit this mate lol". My original receipt (which he brought with him) showed I'd fitted the mixer system but it now had the sequential system fitted.

Rough idle could be due to any or a combination of worn injectors, points of gas entry in plenum too far from engine valves, calibration.

But your immediate problem is the reducer(s) leaking LPG into engine coolant.

Most reducers have 12mm gas outlets, Bigas reducers normally have 10mm gas outlets.
Most injectors have 12mm gas inlets but Bigas injectors normally have 10mm inlets and 10mm inlets are available for V30 injectors.
10mm to 12mm inline pipe connectors are available.
If your install has 10mm reducer outlet(s) and 10mm injector inlets you may need to either change injector inlets to 12mm or use a 12mm to 10mm inline connector when fitting a reducer that has 12mm outlet. Since 12mm pipe is better available than 10mm pipe (and you may need to reroute piping to use just 1 reducer) and since most injectors are only available with 12mm inlets it would be better to change piping to 12mm if you're changing the reducer.

I think I understand the pipe routing you've described. The installer may have intended the pipe between injector rails (with the T to the AEB025) to act as a pressure balance pipe. But if that pipe is narrow diameter it may not be able to flow enough enough to balance pressure between rails in some conditions (e.g. big difference in pressure output of reducer, certain engine loads).

Gilbertd wrote:

Simon, LPGC on here (and on www.lpgforum.co.uk) is just off the A1 near Doncaster. He will no doubt see this and reply shortly. Sounds as if you have correctly identified the problem and it may be due to a bodged attempt at rebuilding the reducers (or they are simply knackered). There will no doubt be a single reducer that Simon will recommend. Are we talking GEMS (pre-99) or Thor (99 onwards) motor?

I'm here ;-)

I agree it's going to be one or both of the reducers leaking gas into coolant, most likely just one,

My go-to reducers are MagicJet (Compact version and Extra version), those produced over the last few years all have a very wide range of pressure adjustability, either version (Compact or Extra) would be a good choice of reducer for a P38 in terms of both physical and technical and they're reliable.

If 2 reducers are fitted the gas outlets need to be T'd together. There's only one pressure sensor so if one reducer feeds one bank and the other reducer feeds the other bank the pressure sensor will only read pressure from one bank and the other bank could be getting very different pressure. It wouldn't cause any leaks or coolant pressurisation but would cause different fuelling to individual banks.

Does the system have a Bigas ECU (probably AEB2568?) and Bigas injectors?

I'm in South Elmsall, between Wakefield and Doncaster.

Hehe yes cheating in a way but I have to cut all injector negative pulsing wires anyway. Even if I'd looked up wiring colours on a diagram there'd still be the need to confirm which colours went to which injectors, with my method there's no need to confirm.

Might as well have the data logger connected. In fact you could use it to make sure Stft's don't hang at extremes for too long, if they're hanging at an extreme change the throttle position and the Stft's will momentarily return to zero which resets the timer for Stft hanging at extremes and prevents throwing a MIL code before Ltft's are learned.

Before because if it works it could fix it, then you wouldn't need to buy any parts.

I'd do it with a cold engine, that way you're doing the same as I did when I first found out it could reset the underlying trim. When I'm cutting petrol injector wires to wire in an LPG system, sometimes the easiest way of finding which colour wire leads to which petrol injector is just to cut a wire and see which cylinder the petrol ECU reports has the misfire and open circuit petrol injector.

Not cutting you off though, we could keep going with testing etc if you like. But it does seem there's not much more to learn from testing and it's time to start thinking about which components the results of those tests could point to. Components/things those results could point to include the maf, fuel pressure sensor, lambda sensors (whether or not they're correct spec), fuel spec, false (unmetered) air leak between maf and throttle body. I probably forgot something!

But it's occurred to me there's one more thing I'd try before replacing any parts. At the rear of each fuel rail (near the firewall on the engine) there's a fuel injector harness connector... Disconnect one of those connectors and start it up, so it only runs on 4 cylinders because 4 cylinders have fuel injectors disconnected. Let it run until the engine warning light comes on (and probably flashes). Use your OBD tool to confirm that it's logged faults for at least one disconnected fuel injector. Then turn the engine off, reconnect the loom, turn ignition on, clear the engine error codes. Turn ignition off, restart the engine and immediately do some mixed style driving (like I advised before, don't just stick to a steady cruise or steady engine load). This doesn't seem intuitive I know, but there are good reasons for this advice. Back in the days when 4.2 SC RR's were relatively new, lots of people in my business (LPG converting vehicles) reckoned these vehicles had 2 fuel maps that the ECU's switched between, which they blamed for poor LPG conversion results. My take on it is that there's only one fuel map but there's a kind of underlying/initial adjustment/map and that adjustment/map is learned during early engine running from new and after and severe error codes are cleared with the idea being to have the 'usual' map have fuel Ltft's of around +7% when learned. But if that initial/base adjustment/map is learned wrong then even if all mechanicals and sensors are in good order and well within spec the Ltft's may be well outside of that average +7% when learned. You can't read that base/initial/underlying map on any scan tool but the concept isn't really that unusual, for example a lot of vehicles that run wide band probes kind of do much the same but in reverse, on those vehicles the fuel trims for idle are usually not anything like what would usually be expected (for an old skool vehicle with narrow band probes) until the fuel trims for under-load driving are learned, and then the base/underlying map for idle conditions is based on the under-load trims and the idle trims start to read as would normally be expected.

Not really relevant but I did have a look at the last couple of scan tool PDF's you posted, thing is in those data sets you'd selected S2 (post cat sensor) Stft's which are less relevant than S1 Stft's so the overall data sets were not as useful as the earlier data sets I based some of my advise on. Still they do confirm consistency with earlier data sets so they were useful in that respect.

Aren't they usually 7/8 hex?

Agreed with Gilbert Stilsons are great for rusted hex's.

You can buy a socket from Halfords specifically designed for removing lambda probes... Obviously deep, it has a slot along the length, the wire protrudes through the slot. The top of the socket has it's own hex, so even if there isn't enough space to use the socket with a ratchet/bar/extension you'll probably still be able to get on it with a spanner. I've got a couple, very good.

Tried all sorts of tools over the years for getting lambdas out of many models of car, made my own special tools for some models of car (usually just the cut-off end of a ring spanner welded to an extension, a slot cut in the ring for the wire to go through, but different angle between ring and extension for different cars).

p73990 wrote:

So why the Maf? Due to age? Or simply to hard to determine if it is reading out of spec. (sort of a preventative measure).

Because you've disproven a lot of things already, Mafs do fail in time / use, you'd probably spend as much time and money replacing lambdas as you'd spend on a maf, a maf can be changed with very little effort compared to welding (or maybe even bodging) pinholes in lambda sensor bosses. Vehicle work doesn't get much easier than changing a maf. If we're wrong and a new maf doesn't fix it, keep it as a spare (in case you intend to keep the vehicle as they do tend to fail in time anyway) or sell it on as a part bought in error. I don't often advise swapping a part that hasn't been proven faulty but it seems a sensible next step in this case. I've spoken to owners of SC RR's who have had the same problems as you fixed when they changed the maf.

I'm not saying the maf might have completely failed, it definitely hasn't completely failed... but it could be reading out of spec (outputting a signal that's around as linear as spec but under-reading across the entire air flow range). They can do that with age, use, or if they get dirty. I used to see a lot of under-reading Maf's on Lexus vehicles, some of the air filters owners were fitting imparted a yellow dust onto the maf sensor which insulated the innards from flowing air and caused the mafs to go out of spec. Often when people think they're cleaning the maf sensor they're actually just cleaning the IAT intake air temp sensor that is built into the same unit as the maf. The maf itself is far more delicate than the IAT and it can be difficult to access so attempts to clean it can cause problems in themselves.

Most mafs output a voltage signal that represents airflow, some mafs output a pwm or digital signal. Of course there's a wide range of mafs of each of those types and you need the specific correct maf for the vehicle. Without re-reading the thread I don't know if you've said how long you've owned it, maybe if you've only owned it a short while it's possible a previous owner fitted an incorrect model maf. I'd only but a genuine maf for a RR SC, or in your situation I might check to make sure the maf is the voltage output type and if it is then make a project circuit that reads maf voltage and outputs a different voltage (in your case a higher voltage). There's a guy/thread on this forum where someone (I forget who or I'd link to it but I expect someone else reading this will link to it) made such circuit/Arduino project that does just that.. He fitted a completely different spec maf designed for a different model vehicle to his P38 which usually wouldn't work on a P38 but the Arduino project he made allows (within reason) any Maf to be fitted to any vehicle.

Holes in the exhaust do have potential to throw readings from O2 sensors, usually by holes acting as a venturi which causes air to be sucked in. But I'd only expect that to happen in an exhaust that has very little back pressure. Of course, a normal exhaust running at part engine load would have very little back pressure. Put all that together and it means that holes in exhausts effect mixture readings more at part engine loads, while your fuel trims are consistently high across the full range of engine loads.

Interpretation of post cat readings can offer insight into whether pre cat probes are working properly, but it's easier to detect front probes that are reading too rich (so leaning mixture off) than reading too lean (so richening mixture) using the post cat probe reading method.

Since front probe bosses (screw in points for probes) seem to have been tampered with, perhaps the 2 front probes have also been replaced with incorrect spec probes for the vehicle. Wide band lambda sensors are available in a wider range of specs than narrow band probes and are not nearly as intercompatible as narrow band probes.

But I'd be changing the Maf before any probe changing or boss welding.

The format is good thanks. I have one browser tab open showing the definitions/scales, another tab open for scrolling sideways through the data (just looking at the top chart which shows all the fields).

Both traces (before cleaning the MAF and after) are very similar.

The long term trims are generally high positive but the short term trims are not far from zero. If we had a high positive trim at a certain engine load (when running closed loop) with a high positive short term trim and with equivalence ratio holding above 1 it would point to a lean problem at that engine load but we're not seeing that. The engine seems to be getting correct mixture during all closed loop conditions and mixture goes suitably rich during high load conditions (like it's supposed to).

Like I said early in one of my early posts on this thread, supercharged Rangerovers can throw a few codes for mixture if/after fuel trims have been reset (and you said you reset fuel trims?). If the code doesn't recur it could be that you've done enough drive cycles with various engine loads for fuel trims to have been relearned... No codes because mixture doesn't hang lean for long enough to produce an error code now because trims have been relearned.

If code(s) do recur I'd expect it would be because Ltft's (particularly bank1 Ltft) is quite high peaking at around 17%. If I remember correctly most SC RR's I've worked on have Ltft's of around 7%. There could be various reasons / faults causing consistently high Ltft's including under-reading Maf (yes they can be contaminated with oil / dust from air filters etc, or can just go bad with age and use). Can probably rule out a vacuum leak because trims are around the same at low engine loads (when there's most vacuum and a vacuum leak would be most revealing in terms of trims) as they are at higher engine loads (when a vacuum leak wouldn't make much difference), though that still doesn't disprove a none-vacuum false air leak (a leak in ducting between the Maf and throttle body), a leak there would see high trims across all engine loads. There could be a problem with the fuel pressure sensor, if that were to read above actual pressure the effects would be to both reduce petrol injector pulse lengths and cause the fuel pump system to deliver lower than normal fuel pressure. What fuel is in the tank (high ethanol content etc)? High ethanol content fuel has a different stochiometric ratio to normal petrol (it needs a different amount of air to burn the same weight of fuel) which can be reflected in fuel trims. But I think at this point I'd be wanting to try a new Maf on it.

P0135, not P1035.

P0135 does point to the lambda sensor's heater circuit. With this error code the lambda sensor may output correct signal or not when the lambda sensor is hot enough (the heater normally makes it hot, the error says the heater won't be working) and the ECU may run in closed loop mode (so attempt to correct mixture based on lambda sensor signal) or not (may run in open loop mode and ignore lambda sensor signal). But it does all point to you needing to change that lambda sensor or fix it's wiring.

Sometimes a lambda sensor with failed heater circuit will output correct signal when hot enough but it might not be hot enough when the engine is just idling.

If there's a problem with the heater circuit the heater circuit could be shorted to the signal circuit, which will mess up signal.

When you first start it from stone cold it'll be running 'open loop' mode ignoring all the equivalence ratios etc. But it'll switch to closed loop mode after maybe 30 seconds to a minute or so depending on how cold it (engine, weather) is, yes it would be a good idea to monitor readings during that time too.

If you can only monitor 8 fields go for rpm, load, ltftb1, ltftb2, stftb1, stftb2, equivalence ratio b1, fuel pressure.

It's fuel pressure on demand and fuel pressure does rise under high load conditions, there will always be some fuel pressure (even with engine off there should be some pressure retained in the system), pressure will probably be at around it's lowest during very low load and over-run conditions (downhill cruise on like you've said).

The same stretch of road / conditions compares like with like conditions but probably doesn't see a very wide range of conditions. The problem might not occur during some conditions. I'd test under the same conditions but also test under a wider range of conditions. For example you might be able to pull over and let it idle for a while beside that stretch of road, then set off and accelerate quickly putting the engine under boost loads.

I've had a look at the 2 traces.

On both traces long term trims (Ltft's) are a little higher than what I'd expect but I dunno if the 14% would be high enough to cause an OBD code in itself. There's nothing to show engine load in the traces (the tool could be set to show RPM / throttle position / calculated load / etc). Were the driving conditions during the traces the same conditions that cause the codes? Also Ltft's by themselves don't indicate mixture or what the fuel system is doing to try to correct mixture, to see that info we need Ltft's, Stft's and lambda equivalence ratios all on the same trace. Stft's normally steer the Ltft's over time but Stft's reach either extreme (say +25% / -25%) the ECU can interpreted that as a potential error condition during which the Ltft's are not steered by Stft's. So, say you've got a Ltft of +10% and a Stft of + 4%, over time that would steer the Ltft to +14% and the Stft would be zero. But if you've got Ltft of +10% and Stft of +25% the Ltft will remain at +10% and eventually the ECU will generate an error code. It might be a good idea, if possible with your scan tool, to do some traces showing all of the fields I've mentioned here.

Fuel pressure also has command pressure but I don't remember if command pressure is a field you can read with a scan tool. I'd normally expect fuel pressure to rise during high engine load on-boost conditions. But if fuel pressure is close to within command pressure I wouldn't expect much of a problem (at least not .. You reckon it should be 45psi during the trace conditions but if so, and if it only falls to 41, I wouldn't expect that to cause much of a problem (at least not during low engine load conditions). Behind the scenes the ECU compensates fuelling for fuel pressure, I don't know if the ECU has a table for fuel pressure petrol injector pulse duration compensation or calculates the compensation but the amount of necessary compensation for lower or higher than expected fuel pressure is quite simple - compare the square root of expected command pressure (45 is 6.708) and the square root of actual (read) pressure (41 is 6.403) and there's only a 4.8% difference of what a petrol injector would flow for any given pulse length with the lower than intended pressure. Even if there were no petrol injector pulse duration compensation for fuel pressure, with expected pressure of 45 and actual of 41 we'd only expect to see a 4.8% change in fuel trims. But behind the scenes the ECU will add that 4.8% onto petrol injector pulse duration and in theory (because of this compensation) the lower fuel pressure should therefore not be reflected in fuel trims at all. Normally on these engines if there is a fuel pressure issue the ECU will come up with specific error codes pointing to the fuel pressure problem.

You've done smoke / plenum leak tests.
Done a couple of traces already,
I'd advise doing some more traces with the extra fields as mentioned above and in various driving conditions.
Maybe do a compression test.
If all seems good from the above I'd reset adapted values (fuel trims), start the car and immediately go for a drive with mixed driving conditions whilst monitoring Ltft's, Stft's and equivalence ratios (like I mentioned in an earlier post).
Still no joy I'd be looking at the Maf, maybe compare voltage readings at warm idle and under other easily comparable conditions (such as 60mph cruise on level road) with someone else who has the same model vehicle.
I still wouldn't rule out the evap system despite the tests you've already done.

I'll say again - Supercharged Rangerovers have wide band pre-cat lambda sensors fitted. With wide band sensors voltage readings are meaningless in themselves. The only way to read mixture from a wide band probe wired to a vehicle ECU is by using OBD live data to read the equivalence ratio. ***

Lambda probes including wide band sensors need to be hot to work properly. Heater circuit readings just tell you about the heating of the probe, there's little useful info about actual mixture from heater sensor duty cycle / voltage / etc readings.

Just because you usually get an error code for mixture on one (either) cylinder bank doesn't mean that the other cylinder bank isn't also sailing very close to the wind in terms of having correct mixture (or fuel trims reaching an extreme) and if fuel trims are similar on both cylinder banks over a wide range of conditions the other bank probably is close to also flagging an error code.

I haven't yet looked at the video but I had no need to see the video to say the above. I will watch the video and see if I can offer any insight after watching it.

*** A good way to understand wide band probes is to think of them as a narrow band probe but with a built in air-pump that blows atmospheric air close to the narrow band sensor. The more air that is blown past the sensor at any given mixture the lower the voltage reading from the narrow band sensor will be. So the voltage reading in itself from the narrow band probe gives very little information about actual mixture unless you also know how hard the pump is blowing. You can only have one control/reading system connected to the wide band probe at any one time because the system has to simultaneously read the voltage from the sensor whilst controlling the pump, only with both sets of information (sensor voltage and air pump control) can the system know the actual mixture... It would be no good having one control unit read the sensor voltage while another unit controls the pump. If mixture strays far from anticipated mixture it can take a short while for a wide band lambda sensor control system to 'home in' on correct mixture, in fact even at close to stochiometric ratios they can take longer to reflect a true mixture reading than a narrow band probe because it takes time for the control system to do it's calculations and effect pump output. Because it blows air close to the sensor a wide band probe can accurately read much richer mixtures than a narrow band probe but they're not much better than narrow band probes at accurately reading lean mixtures.

You've twice mentioned the fault being ambient temperature related, no faults seem to occur when the weather is warmer.

What readings are you seeing for intake air temperature and do they agree with actual air temperature?

Really IAT shouldn't make much difference because MAF sensors are self compensating for intake air temperature. Yet all MAFs have built-in IAT's...

Gilbertd wrote:

CNG conversions are a common alternative to LPG (Propane) in many European countries, particularly Italy. There's a number of trucks running on it in the UK now too. Energy density of Methane is slightly better than petrol or diesel (see https://en.wikipedia.org/wiki/Energy_density) although you will use slightly more of it due to the stoichiometric ratio (see https://en.wikipedia.org/wiki/Stoichiometry).

The other downside is the tank, Propane needs a tank capable of storing a liquid at around 10 bar (145 psi), so fairly simple to construct, but a CNG tank needs to be able to store the gas at around 10 times that, LNG needs an even greater pressure.

There's little difference in peak torque or bhp between LPG and petrol but engines run on NG make less power than engines run on LPG.

I should define a few terms to make sure the rest of what I write makes sense. CNG is compressed natural gas (NG), LNG is liquified NG.

One downside to LNG is that it has to be kept at cryogenic temperature... If you needed to store it (store could mean just park a vehicle for a week without using it) there'd need to be some measures taken to keep it cold or the increase in temperature would mean an increase in pressure that could burst any tank. I believe that in practice they have a vent valve which allows a little NG to escape to atmosphere and the evaporation keeps it at cryogenic temperature during such storage. But that's not environmentally friendly because NG is one of the most potent greenhouse gasses, much worse than CO2.

LNG is used on trucks etc but CNG is usually better suited to cars (mostly due to the above reasons). If we have NG delivered to our homes (which we do if we have piped gas to the home) we could easily buy a CNG compressor and fill CNG converted vehicles at home. I have looked into this. CNG tanks are difficult to source in the UK, in fact it seems easier to buy a CNG compressor than it is to buy a CNG tank in the UK! Running a vehicle on CNG is probably the cheapest way of running a vehicle... It's just that with the lack of CNG filling forecourts the only place we'd be able to refuel with CNG is at home using our ow (expensive to buy) CNG compressors. I would have already done it (fitted a CNG tank and CNG pressure reducer to one of my cars and bought/fitted a CNG compressor at home) if tanks were more readily available.

Morat wrote:

300 miles from a Tesla 3 is ambitious. You'd need a full battery. IIRC even the long range version tops out at 350 miles (in ideal conditions of course). Tesla recommend that you keep the battery between 20 and 80% whenever possible to increase the battery life, so most of the time the range is much less. Of course, most of the time you don't need 300 miles but it can't be a nice feeling to imagine you're pushing such an expensive component when you do.

I personally would like to check out a Tesla, but I now WFH so the only time I need to drive is usually a dash into York or a road trip 3 or 5 up for a long way. So, I think the Simon's home LPG pump is FAR more interesting!

Maybe worth looking into yourself Miles? Contact a few local-ish gas suppliers, tell them how much you'd expect to use, ask them how much for 47kg bottles delivered to your door, see if you can haggle the price down. Then you just need a pump... You could buy a pump tomorrow from a well known supplier (MIck at LPGAutosupplies, ex TinleyTech) for £600+VAT. If you made the calls on Monday you could be set up ready to refill at home by Tuesday or Wednesday and for a similar cost per litre to refilling at Morrisons.

There's a guy on LPGforum who reckons he uses a £30 car electric petrol pump to pump LPG from bottles into his car LPG tank, I'm not sure if that really works or if it does work how reliable it would be but I've thought about trying the cheap pump myself.

If electric vehicles really do take-off then we could expect forecourts that sell petrol and diesel to decrease in number, we may even see the day when petrol/diesel are difficult to buy and people with petrol/diesel cars have range anxiety. But even then, or even if we couldn't buy petrol or diesel at all, as long as people use LPG for heating and it isn't illegal to run vehicles with engines we would still be able to run our vehicles on LPG if we refuel at home.

With an extra LPG tank fitted my car will do 700 miles on a single fill of LPG. I made my extra tank quick fit / quick remove, I removed it today because I needed the luggage space so at the moment my car only has 350 miles range on LPG. The extra tank is now sat in the yard and it's full. If next week I needed to refuel but all LPG forecourts were out of action I could still put the extra tank in the back, connect it up and drive 350 miles on LPG. In fact I've got loads of tanks in the yard, mostly new ones, any of them could be used as the extra tank and I could carry several of them in my car. I've also got a full 70L forklift tank and a full 18kg propane (caravan) tank - I could easily make a short (3ft) pipe with correct fitting on one end to connect to any of those tanks and quick-fit (hanson coupling) fitting on the other end and run my car from any of those tanks.

Years ago I used to fix local Calor depot's fleet of LPG converted vehicles and got to know a couple of their delivery drivers. The delivery drivers sold me LPG in red bottle forklift tanks, 10 tanks at a time at a price that worked out at around 11p per litre. They'd park on the top road outside my place and we'd simply roll the full bottles downhill 20yards down the driveway into my yard, carry the empty bottles back up to the lorry. I didn't bother trying to pump from the red tanks into my car tank, I just rigged a pipe into the luggage area of my car and carried 2 red forklift bottles in the luggage area. Around 35litres in each of those (small) forklift bottles.

Bolt wrote:

If we start with a natural gas fired power plant, and 100 cu litres of CNG.
Burn that in a furnace to boil water in a heat exchanger, make steam, turn the turbine to generate electricity.
Best case is 60% efficiency and only if you can use the waste heat to do something. (as low as 33%)
https://energyeducation.ca/encyclopedia/Natural_gas_power_plant
So, 40 of our Cu litres gone to heat at the plant.
Now we have electricity at about 500kv which has been stepped up there using transformers.
These are pretty good, however we go through no fewer than 4 to go from the plant to the rural grid, to the local grid and to your house. we will through in transmission losses here as well so lets call that 10%, but it can be much higher.......So, kiss off some more CNG..
https://blog.se.com/energy-management-energy-efficiency/2013/03/25/how-big-are-power-line-losses/
Now, we have a battery charger coming in at about 80% on average, so more heat and more wasted energy.

Batteries themselves are not 100% efficient and degrade as they are cycled, so average over a 6 year life cycle? call it 80% with 20 % being turned to heat in the battery.
Electric motor and controller? 90% is generous, so another 10% loss to heat.
I have used the average losses here and leaned toward very conservative numbers.
Even so, you are kissing off more than 50% of the electrical energy produced at the plant before
it even sees the EV. other types of plant, like coal are far less efficient.
Not my idea of an environmentally friendly solution.
Question for LPGC: How much better would it be to just burn the CNG in a converted vehicle?

As much as I love engines they're only around 30% efficient and the fuel has to be refined and transported to forecourts. But the infrastructure is already there, no need to build more power stations, upgrade the electrical grid or source lithium.

Gilbertd wrote:

From what I've seen the idea behind it being a renewable energy source is that by producing it by hydrolysing water using electricity from renewable sources, you've got a zero carbon fuel.

Yes, so if we ever have abundant clean electricity we can also have abundant clean burning hydrogen. Then maybe we could all go back to ICE cars and run them on hydrogen hehe.