TDV8 auto into a defender 110

[estu1708]

Hi, I am looking at building a spectre replica, and was thinking of putting a TDV8 auto in to it(if possible) the only problem I am coming across is finding a gear box to work with it. 

Any suggestions?

 

thanks 

 

[Carloz]

"the only problem"?? 😁

Maybe use the autobox which is normally behind the TDV8 and then make a outputshaft (maybe based on parts from the original transfer case) to a remote LT230 which you move backwards.

[miketomcat]

If memory serves me correctly it has been done and they used RR engine and gearbox mated to an lt230.

Mike

[L19MUD]

36 minutes ago, Carloz said:

"the only problem"?? 😁

 

My thoughts exactly!! A TDV8 in a Defender would be great, putting aside the electronics I would imagine the physical size could be an issue? 

 

Is there space between the chassis rails? The TDV8 is also very tall which is overcome in a RR by bolting the diff to the side of the sump with the driveshaft running through the middle of the sump, you won't be able to do that with a beam axle Defender

[Maverik]

TDV8 Into a 100"defender. - International Forum - LR4x4 - The Land Rover Forum

[L19MUD]

1 hour ago, Maverik said:

TDV8 Into a 100"defender. - International Forum - LR4x4 - The Land Rover Forum

wonder if that ever came to anything?

[RedLineMike]

On 7/2/2021 at 1:48 PM, L19MUD said:

wonder if that ever came to anything?

i believe the project was canned due to issues & delays on the fabrication side , the rolling chassis came up on ebay about 18 months ago for sale,

BAS & Nene did put the TDv6 engine into several defender at one point but the cost of the ecu & its finickyness to run the ecu outside the donor vehicle is what stopped the project, 

 

https://www.bodylogicuk.com/index.php/10/327154/

[Chicken Drumstick]

JE Engineering did a few TDV6 Defenders I believe. I think they had to notch or modify the front cross member or something.

Does anyone know what arrangement Land Rover used on the 70th Anni models using the AJV8 and ZF 8 speed auto? I'd assume that would be the easiest approach and copy what they did. As I'm guessing they would still be subject to the Radically Modified vehicle regs and nobody paying £150,000+ on one would be happy with a Q plate vehicle.

[landroversforever]

I had a long chat with Ian at IRB regarding the 3L TDV6 as an option and in the end it was doable - but from memory it was £15-20K to get it running properly outside the rest of the original car.

Anything is doable... but its how much.

[FridgeFreezer]

8 hours ago, landroversforever said:

I had a long chat with Ian at IRB regarding the 3L TDV6 as an option and in the end it was doable - but from memory it was £15-20K to get it running properly outside the rest of the original car.

Anything is doable... but its how much.

That's a lot of cash for a rattly lump with average power considering you could throw a Rover V8 in for a grand and have 15k spare to spend on fuel, or throw a very fruity LS in for that budget and have a hell of a lot more to smile about.

[Ed Poore]

On 7/14/2021 at 12:49 PM, landroversforever said:

but from memory it was £15-20K to get it running properly outside the rest of the original car.

Is that £12-15k per vehicle or was that factoring in development costs? If the former that's frankly ridiculous! I'm sure @FridgeFreezer and I could knock together some custom ECUs for less than that to run the engine outside of the vehicle for a fraction of that.

[landroversforever]

I'm sure a bit of that would have been development costs. I appreciate yours and Fridge's backgrounds, but its a bit different to throwing MS on a Rover V8. The bodylogic one wasn't spoken kindly of regarding how well it would run the 2.7 or the 3.0. I'd trust Ian's work in that respect as he spent a long time at JLR as a powertrain engineer so has no doubt got some background knowledge of them.

[landroversforever]

To add to that, if you chaps could make one work well... then get on it!  There must be a good supply of TDV6 and 8 engines from crash damaged cars that aren't viable to repair. 

[Ed Poore]

I did actually look into it and there's no way I'd start from an existing ECU. I did look into it a while back when I had the 3.6 and actually controlling the sensors and everything else looked fairly simple and straightforward starting from scratch.

What I didn't know was how to tune / map the vehicle so if there's someone who knows that I can build an ECU to control it.

[FridgeFreezer]

All I know is modern diesel control is waaaay harder than petrol where you just have to squirt roughly the right amount of fuel in and light it at the right moment. Also modern direct diesel injectors are way harder to drive (=expensive electronics) than the simple 12v solenoid that is a petrol injector.

You'd need a very good understanding of the exact ins & outs of it all (timings etc.) and probably bank on damaging/exploding at least one engine for R&D purposes when it goes wrong.

[garrycol]

In Australia we have a company that puts the 2.7TDV6 (and the 5 cylinder 3.2 Diesel out of a Ford Ranger) into modern Puma Defenders and another company has just put a LS crate engine in a Puma Defender.  I dont know the configurations but they all work very well but are expensive.

Garry

[Ed Poore]

3 minutes ago, FridgeFreezer said:

Also modern direct diesel injectors are way harder to drive

If I remember correctly all you need to do is "send" a pulse of current to inject fuel so not massively difficult... All you've got to do is send roughly the right amount of diesel at roughly the right time - you don't even need to faff with this whole lighting it at the right time either . I mean fundamentally a diesel is a diesel - it needs fuel and air (okay and maybe a bit of compression).

The high pressure fuel rail takes care of itself if I remember correctly, so crank the engine and start squirting fuel with respect to cam shaft and crank shaft positions and away you go. Admittedly the voltages and currents are a wee bit higher but I mean that might bump up the cost to a few pence for a suitable MOSFET... 

[Chicken Drumstick]

1 hour ago, Ed Poore said:

I did actually look into it and there's no way I'd start from an existing ECU. I did look into it a while back when I had the 3.6 and actually controlling the sensors and everything else looked fairly simple and straightforward starting from scratch.

What I didn't know was how to tune / map the vehicle so if there's someone who knows that I can build an ECU to control it.

How easy would the AJV8 be to control? Thinking 4.2 suprcharged and ZF-6 speed from an L320 RRS or L322.

[FridgeFreezer]

41 minutes ago, Chicken Drumstick said:

How easy would the AJV8 be to control? Thinking 4.2 suprcharged and ZF-6 speed from an L320 RRS or L322.

If it's petrol, take yor pick of aftermarket ECU's from Megasquirt 1 upwards, MS2 would be my choice.

 

55 minutes ago, Ed Poore said:

If I remember correctly all you need to do is "send" a pulse of current to inject fuel so not massively difficult... All you've got to do is send roughly the right amount of diesel at roughly the right time - you don't even need to faff with this whole lighting it at the right time either . I mean fundamentally a diesel is a diesel - it needs fuel and air (okay and maybe a bit of compression).

The high pressure fuel rail takes care of itself if I remember correctly, so crank the engine and start squirting fuel with respect to cam shaft and crank shaft positions and away you go. Admittedly the voltages and currents are a wee bit higher but I mean that might bump up the cost to a few pence for a suitable MOSFET... 

Ed - generating a good solid 120v rail inside the ECU adds a problem / cost, and as I understand it anything that doesn't have a mechanical injection pump is likely to be doing multiple carefully-timed squirts per "event", that's how modern lumps are so much better and smoother than older ones. With spark you just need to fire the thing at the right moment (and can do it wasted-spark too), with diesel you need to fire the thing at the right moment AND for the right period AND potentially do that multiple times within each event and do it per-cylinder not banked/batched as you can with petrol.

Put it this way - to run a V8 pretty well an MS1 has 2x 12v-capable injector drivers and 4x 5v logic coil drivers and can run well enough on a very slow 8-bit micro with ~0.1ms resolution. For a TDV8 you'd need 8 injector drivers that can do 120v into a seriously chunky inductive load and I'd guess you're going to need <0.01ms resolution for decent control.

I'd be tempted to look to the states where they have a ton of aftermarket / tunable diesel ECU's and see if one can be made to play nicely with a TDV8.

Anything's possible but there's a reason the diesel Megasquirt project never really got off the ground. Again, I'm certain it's possible if someone wanted to pay the R&D costs, I'd quite like to give it a go TBH just for the curiosity.

[Ed Poore]

19 minutes ago, FridgeFreezer said:

and can run well enough on a very slow 8-bit micro with ~0.1ms resolution

And there in lies my real bug bear with things like Megasquirt etc., yes I appreciate it was started decades ago and people struggled with soldering anything smaller than a battery cable but technology has moved on so far since those days. I had high hopes of the Speeduino project but the moment I saw the fact that they had "highly optimised the code to run on an Arduino" immediately meant to me they'd chosen the wrong micro in the first instance.

My inclination for a rpelacement Megasquirt / Dieselsquirt whatever you want to call it would be scrap the micro, put a tiny little ARM core in there as a supervisor and a small little FPGA. V6, V8, V10, V16, W16 who cares that's what FPGAs are superb at doing multiple things in parallel. Again <10us resolution is perfectly doable chosing the right components. I guess my background is in timing things down in the picosecond range and having to process things in nanoseconds so when you've got microseconds to achieve something you can do soooo much in that time . Plus you don't really have to react to something in an engine - it's all fairly predictive (yes you do need to adapt to things but that's in slow time - hence the ARM).

Generating a 120V rail that can withstand inductive loads is tricky but not difficult. Generating a 500kW 1.2kV DC rail from 3-phase whilst retaining a 0.99 power factor and >98% efficiency more or less irrespective of loading, now that's a challenge but achievable.

Like you I too am very tempted to give it a go just for curiosity. I've got a bunch of systems in my desk that are more than capable of handling the timing / complexity requirements, once it's developed could easily be shrunk down on device size to reduce cost. I mean you don't need a quad core 1.2GHz SoC to run an engine but it'll be handy to develop on .

Just need to find someone to donate us a 4.2 Supercharged and a TDV8 that we are allowed to dismantle / blow up.

[Bowie69]

40 minutes ago, Ed Poore said:

"highly optimised the code to run on an Arduino"

As a developer, it would be stupid not to optimise it to be honest., last time I looked, the Arduino used for speeduino is streets ahead of the MS1/2 chip.... not sure about the V3 though, but that is commercial.

there's only so much power you need to run an engine, I suppose

[Ed Poore]

9 minutes ago, Bowie69 said:

As a developer, it would be stupid not to optimise it to be honest., last time I looked, the Arduino used for speeduino is streets ahead of the MS1/2 chip.... not sure about the V3 though, but that is commercial.

there's only so much power you need to run an engine, I suppose

There's a time and a place for optimisation trust me. Usually right at the end when you've got things working reliably and want to improve performance / efficiency / space. You shouldn't need to optimise just to get the thing running. I had a look through the code and they had (at least when I looked) had a load of assembly mixed in with the C. So what started out as a nice potentially cross platform bit of code quickly got shot down.

In one of my past projects there we were heavily looking into optimisations both on a micro level but also in my instance on a 72 core 256GB RAM monster with 64MB (I think) cache. There was only ever once that I managed to beat the compiler at optimising a bit of code and that was because I knew the incoming data structure so could swap a few things around to keep it in cache more often.

Usually what the optimisation does is make the code unintelligible for anyone but the original author unless it's done well. In a commercial environment that's a big drain and usually not required. I don't have against it but requiring optimisation to make it meet timing requirements to make it run screams to be you've messed up the design big time.

[FridgeFreezer]

Ed - you can do anything especially with time & budget but making something that's as complicated (or more complicated) than the original doesn't really appeal to me and will limit the number of folks who would feel able or willing to contribute. In the early days of MS loads of folks chipped in because the codebase was small, simple, & fairly well-documented / easily understood and you could compile it with free tools.

Also TBH you can achieve incredibly fine timings on a bog standard £5 STM32 with the built-in timers/interrupts... the issue is arriving at the timings in the first place and making a bulletproof (and fail-safe) codebase & hardware that you'd trust not to grenade your engine because you plugged the USB cable in at the wrong moment and caused an interrupt... making an ECU that works on the bench is a lot easier than making one that will survive in an actual car & work reliably.

I assume the Speeduino brag about optimisation just means they've re-written a few bits to avoid using the noddy standard Arduino libraries that aren't up to mission-critical applications beyond blinking an LED, and perhaps made use of hardware peripherals rather than bad habits like bit-banging with blocking delays and the like.

Supercharged V8 is no issue to run with standard MS or whatever, it's only diesels that are difficult.

[smallfry]

Wibble wibble wibble. You guys have lost me with all this speaking in tongues, and now Ive got a headache. I couldnt learn any of this stuff due to lack of Gorm.

With respect, If it really  is that easy, why has none of the great and good done it ? Surely it would be popular conversion ?

Im sure you would all be better off cobbling up a gearbox controller for my ZF 4HP24 😆

[Ed Poore]

2 hours ago, smallfry said:

If it really  is that easy, why has none of the great and good done it ?

Simple answer - no one would be willing to fund the development so given we'd have to source at least one or two donor vehicles means it's a back-burner kind of project. Electronically it's reasonably simple to design but actually getting it reliable will simply take a lot of time and I for one have other work on the go both higher priority and paying (that's definitely higher priority) so whist I'd love to I just simply can't devote the time at the moment.