My GEMS Story
From the moment I first climbed
into my beloved Chimaera on a freezing winter’s night in 2006 I had already
visualised its future. Setting off from the home of the (visibly upset) former
owner onto a pitch-black motorway with the ice warning light emblazoned before
me, worrying that I had just bought far too much car for my level of driving
experience, when I really should have been thinking about simply enjoying my
new car; I confess that at this point I was already plotting my first
modification. I am going to take the time right now to place the blame for this
squarely on one person: Mr. Chris Manley.
Unfortunately for me, Chris was
and still is very good friends with Tim Lamont of ACT and even more
unfortunately around 2004, had recently sold his S3 and bought Tim’s own highly
modified pre-cat Griffith – a lovely car that he still owns today. One evening
Chris kindly invited to take me out for a short ride in his new car. I
gleefully agreed – knowing nothing about what he was about to do to my future aspirations
of having disposable income.
After jumping in and cruising
gently out of the carpark to a quiet T-junction; Chris calmly proceeded to slam
his foot on the accelerator - violently introducing me to the world of TVR. If
you don’t know this car, it is very quick and VERY loud. I was hooked - Cheers
Chris!
A month or two after buying my
own (completely standard and far too quiet) 1997 450 Chimaera, I had already
visited Tim for the first time to pick up a set of his beautiful stainless,
de-catted sports manifolds and a full stainless cherry-bomb exhaust together
with the obligatory induction kit and carbon trumpets in order to get the noise
that I so craved. There has been a fairly steady flow of projects, plans and
large injections of cash ever since, always needing to make the car just a
little bit faster/prettier/better; I love modifying it almost as much as I like
driving it, despite quite often having very little idea what I’m doing when I
set out.
Around 8 years ago, I decided
that I should probably make something of a plan of how I was going to modify
the car to avoid doing things in the wrong order or altering things that might
not be compatible with future plans. Knowing that I would definitely want to
make more changes to the engine, I knew I would need to have the ECU re-mapped
at some point to give the engine the extra fuel it would need. At the time
there seemed to be two options; re-map the standard Lucas 14CUX ECU via Mark
Adams each time I changed something at quite a high cost (and at the mercy of
Mark’s famously busy schedule), or replace the ECU with something more modern
which could be adjusted by any engine tuner worth his salt (more on this
later).
After doing quite a lot of
research into the costs and various aftermarket ECU offerings on the market, I
opted for the (then) new and popular Emerald K3 ECU – a great and well-priced
piece of kit which among several other things had an excellent and easy to use
software interface; the nerdy engineer within me particularly liked this as I could
see and learn exactly what was going on and how the whole thing worked - even
tweak the fuel map myself to my liking. So after asking lots of questions on
the Pistonheads forums, studying the manuals and the Steve Heath bible, I
rather naively ordered all of the bits and then commenced the disassembly of my
pride and joy with a detailed step by step plan that I’d written with an
estimate to finish the job in a couple of days – how hard could it be?
About 4-6 weeks later I managed
to get the engine to start - luckily for me my naivety was outweighed by a
steely determination - I shall leave out the details of the debacle and process
of finding out that the new cam had been installed 14 degrees out of time by a
reputable garage. Finally however it was time to trailer the car up to Norfolk
for setting up on the rolling road by Emerald – they did a lovely job and all
was good. This is when everyone on the forums or at car meets says that they
live happily ever after and the car is perfect. These stories are most
definitely lies.
Fast forward 7 years and I think
I have had the car on dynos around 6 or 7 more times at a cost of probably £300
a go. Some of these visits have been to address areas of the map which don’t
feel right after running the car for a while, or some were to re-tune to cater
for new engine modifications. Some were to try and diagnose issues with the
engine (usually after me fiddling). I noticed a few things through this
experience:
●
No one can map a car in a day. There is a reason that
car manufacturers spend months on this process, and test in all sorts of
temperatures and conditions (especially out on the open road).
●
There are so many variables in an ECU map, and all of
them have an effect on the way a car drives, or the ability to start when cold
or hot, or the ability to not produce a very subtle shunt at 74 miles per hour
on a very specific piece of downhill motorway when the weather’s cold.
●
The map can completely change the character of the car;
I also noticed that different mappers can imprint different characteristics in
terms of the response and feel, and getting it wrong can make you absolutely
hate driving the car. Put simply, it is just as critical as having the
suspension set up correctly – probably even more so.
Last year I was driving the car
one day and for seemingly no reason, the car started running strangely – it
would cruise very lean, then idle very rich and generally sound rather poorly.
Plugging in my trusty laptop and firing up the software revealed that nothing
was amiss; all sensors seemed to be reading ok, but things weren’t running well
at all. Calls to various specialists were made, tests were done and heads were
scratched and the problem didn’t reveal itself – a suggestion was made to have
it put on the dyno to do some more tests for another £300, plus whatever extra
labour might be required to remedy the issue. That was it; there must be a
better way of doing this.
Having heard about his reputation
many times within the TVR world, but never actually managing to speak with him,
I decided to try calling Mark Adams. I have met all kinds of characters in the
TVR tuning world, and I can safely say that Mark is the most accommodating and
most friendly of them all. What an absolute gentleman. Being at the end of my
tether with ECUs, and being a mechanical engineer by trade, I asked Mark to run
me through all options including severe engineering detail. Over the next 4
hours on the phone, he walked me though the reasons behind the
drawbacks of aftermarket ECUs, and the limitations of the standard Lucas 14CUX,
and various options on the market right up to the extremely expensive offerings
from the likes of Motec and others. None of them he explained; come close to
the capabilities of the 1994 Lucas-Sagem GEMS ECU. I was sceptical but
intrigued.
The GEMS ECU is like no other ECU
before or since. It was designed by Lucas-Sagem to be the OEM ECU to beat all
others, and underwent a long period of design and development, with a budget of
millions of pounds back in the 1990s. Utilising a dual core processor, fully
waterproof and pressure sealed casing, OBD communication and extremely hard
wearing design, the ECU is tough, and quick enough to do what it needs to do
even when paired against modern contenders. Its special skill however is its
programming – while many ECUs have a main fuel map to work from, and a few
adjustment tables for things like air temperature and air fuel ratio, GEMS
takes things far beyond this and has extremely complicated maps for a level of
detail simply not seen in the aftermarket world, and all of these were set up
and tested over a period of years attached to a Rover V8 engine. Among other
things, these engines were run at conditions not achievable by a rolling road
using a large and very expensive electric engine dyno which could simulate any
load on the engine, they were also run in all conditions around the world in a
heavy testing programme.
During this long gestation
period, the engineers at Lucas-Sagem managed to bake in a level of
self-learning unlike anything else before or since. As well as being able to
map itself completely autonomously to a level of refinement simply not
achievable by a human tuner (Mark’s own admission), the GEMS ECU can also still
get the car along the road after many of its sensors have failed; its ability
to keep calm and carry on despite having nearly all of its electrical arms and
legs cut off is uncanny – and to the point that the driver may not even notice.
It also has extremely complicated diagnostics such that it doesn’t need to go
to a garage if something goes wrong – it will tell you exactly what ailments
are affecting it. They should really have called it Skynet.
This then is what really
attracted me to the GEMS system, having worked with a lot of ECUs over the
years, Mark explained that its this uncanny ability to work against all odds,
and refine itself permanently with no human intervention is why GEMS is so
good. Additional benefits of this are that it is extremely economical with
fuel, fully Euro 3 compliant if plumbed in in OEM guise, and can be diagnosed
by any garage with an OBD computer. I decided this was the only option
(particularly as my car can be tricky to map due to having a supercharger
fitted), despite finding out rather alarmingly that it had never been done
before on a TVR (except 2 factory specials from TVR themselves destined for the
emission-crazy Switzerland). Sadly for Lucas, GEMS never made it to serious
mass market due to its massive over-engineering driving the cost too high
against its rival, Bosch with a cheaper and less clever alternative so it was
only ever really seen in the GEMS Range Rover and a few Morgans.
So
why had it never been done on a TVR before? This was my next question to Mark,
and his answer centred around three main points; the cam position sensor, crank
trigger and knock sensors could present quite a large pain in the arse. The
latter two could be solved quite easily using some luck; the later block type
used by TVR for the 4.6L engines (which my Chimaera has) was the same as used
by the GEMS Range Rovers and as such has the mounting points for the two knock
sensors already in the block. Happily, it also has the attachment point for the
GEMS crank sensor; GEMS uses a crank trigger mounted on the flywheel (out of
the way of any debris that may foul it), so the sensor is mounted on the
opposite side of the engine to the starter motor.
At
the same time as the gearbox housing was done, I had my trusted mechanical
genius – Dave Batty, fit the flywheel, crank sensor, knock sensors, a GEMS
Airflow meter, and also a GEMS coil-pack with leads – I got a second hand but
new-looking one from Ebay with all leads – this bolts neatly behind the plenum
to existing lugs in the top of the inlet manifold.
I
then drove my still sick sounding Emerald brained car back home, disconnected
the battery, found my super sharp wire cutters, and proceeded to chop out the
entire engine wiring loom, including everything in the foot-well and the engine
bay. God that was satisfying! I wish I could do it again…
For convenience, here are the
parts I needed to add for my conversion with prices I paid:
Part
|
Part Number
|
Price (ea)
|
New/Second H
|
Coil Pack
|
ERR6269
|
£35
|
SH
|
Air Flow Meter
|
ERR5595
|
£35
|
SH
|
Crank Position S
|
ERR6357
|
£30
|
SH
|
Cam Position S
|
ERR6169
|
£20
|
N
|
Knock Sensors
|
NSC100650G
|
£70
|
N
|
Lambda Sensors
|
ERR1834
|
£45
|
N
|
ECU
|
N/A
|
£20
|
SH
|
Tornado Chips
|
N/A
|
£500
|
N
|
You will note that I haven’t yet mentioned
the first major reason why no one has made the conversion before; the cam
sensor. To add some background to the problem, while the later TVR block is the
same as the GEMS engine, the front cover is not; and neither is the sump. Since
the GEMS engine doesn’t have a distributor, it does not have the hole in to
which it is mounted. It also uses a long-nose camshaft with a very different
cam sprocket mounted to the end; the “spokes” of which are the trigger pattern
for the cam position sensor mounted in a hole in the front cover just above the
crank pulley. This front cover then has a different sealing arrangement with
the sump below, so it would require a GEMS sump to be fitted. There are a
number of problems going down this path.
●
It is bloody expensive
●
None of the nice, fast road camshafts available
for TVRs are available in the long nose guise (or if they were then I’m sure
they would be with a price hike)
●
The GEMS cam sprocket uses a simplex chain, and
isn’t compatible with the stronger duplex chain that TVRs commonly use for the
higher revs used on the engine.
●
The big off-road GEMS sump won’t fit in the TVR
chassis and you can’t buy a standard one with any baffles in for track use.
As you can see then: this is not
a path to be walked, and has deterred many, but being optimistic I was sure
that it could be done another way. The first idea discussed with Mark was to
design a bespoke sensor based on some existing components on the market which
would drop into the distributor hole and translate the rotation of the camshaft
up into a small rotating trigger in a sensor housing machined to emulate that
of the GEMS. This was deemed perfectly feasible, but expensive and would still
see a lump sticking out of the block where the distributor once was which
sounded a bit ugly. This should be a neat solution.
About this time, Mark put me in
touch with Mike Timm; a guy who had been planning a GEMS conversion for some
time on his Griff, but on a much more relaxed schedule. I contacted him and we
discussed the concept of a trigger disc similar to one that David Blades had
built for his VEMS powered Griffith some years earlier. Mike, also an engineer
had taken the measurements and knocked up a CAD drawing in no time at all and
we both set upon checking the measurements independently. Satisfied that we had
them correct, I popped down to my local laser cutting shop and had a few
trigger discs run out in 3mm plate; look how pretty they are!
This is where things get a bit
scary (for me anyway). After draining the oil and coolant, the swirl pot,
coolant pipes, power steering pump, the supercharger pipework, exhaust
downpipes, chassis cross-member and crank pulley, I could finally remove the
front cover – there’s actually nothing hard here, it’s perfectly doable by
anyone at home with a pair of axle stands, a jack, a basic tool kit and some
Swarfega.
Note; Holding a
drill loaded with such a large drill bit gives quite an awesome sense of power
– this has got to be one of the more pleasurable steps of the conversion.
After degreasing the cover in the
kitchen sink, and remembering to centre punch the hole as I was taught at
school, I proceeded to be amazed at how easily such a large hole can be bored
into metal using a hand drill (while covering the kitchen worktop in aluminium
shavings). After an extensive wash (again in the sink) I returned to the car to
offer up the trigger disc and see if my sensor would fit correctly. It did not;
while the position of the sensor was correct relative to the disc, the gap
between the trigger disc and sensor was far too big. Through some kitchen-bench
testing earlier, I had worked out that the sensor reacts correctly when it is
around 1mm from the disc; in this case I reckoned I had around 5ish mm - my
hopes for an easy day’s work were instantly dashed.
My first cunning plan was
to add another 3mm trigger disc in front of the first to reduce the gap to
around 2ish mm, when I tried this though, I found that I could no-longer re-fit
the front cover; a little examination revealed that there was a small lump in
the casting which I assume is there to stop the cam sprocket walking forward in
older engines. This was fairly easily dealt with
using my favourite new giant drill-bit, allowing me to then fit everything back
together again with the extra trigger disc (now 6mm thick) – this brought the
gap right down to manageable levels.
The final piece of the
puzzle was to re-assemble the rest of the arrangement, deleting the distributor
drive gear and replacing with a thinner washer to account for the thickness of
the new trigger disc. To save anyone else time in doing this, I have had
several of these discs made in 6mm thickness with washers, and I will be
selling them through Mark Adams – please do get in touch with him if you would
like to consider taking on one of these conversions!
With the hardware now
appearing to work, I needed to fit the sensor. It needed to work first time and
every time, as it is extremely difficult to access once everything is back
together. So after measuring the air gap several times, I realised that adding
the gasket and any sealant to the front cover when I fit it finally may affect
the sensor-disc air gap by enough to make it not read correctly. So after
speaking with Mark at length, I decided to opt for more of a “TVR” approach and
use a - technically sound – bodge. But a bodge nonetheless. To be absolutely
assured of the sensor gap being correct, I refit the front cover with gasket
and strolled down to the shop to buy a bench grinder with which I ground off
the mounting lug until the sensor was a perfect cylinder, creating far more
mess with black plastic dust in the kitchen than the aluminium shavings – eek!
The sensor was then fitted with a Sikaflex jacket and slid very tightly into
the front cover until it touched the trigger wheel before being pulled
back fraction. What a neat bodge! Mark
had done this on several previous GEMS engines with good results so I decided I
was fine with it. As it happened, this worked flawlessly and has ever since.
After a celebratory beer or two,
engine bay was re-assembled and all fluids renewed: Hardware. Done!
The next bit was the wiring loom.
Because I’m perhaps slightly crazy; I was looking forward to this part as I
have always enjoyed the simple logic of how circuits work and have played
around with car electrics since I built a windows PC and sound system into my
Ford Fiesta as an enthusiastic 18 year old.
Having played with the Chimaera
wiring loom before to put in my Emerald system some years ago, I had a good
working knowledge of where everything was, but it really is quite simple once
you get your head around the wiring diagrams. Unfortunately though, historic
Nick had cut several corners, and chopped and changed the loom over the years
to add an engine start button two different aftermarket alarms, extra gauges
for the supercharger and all sorts. This added to my workload significantly, so
my advice here for future Nick and anyone reading this is: do not cut corners.
Oh, and: write down anything you have changed which is not on the standard
wiring looms! This will save quite a lot of head scratching and swearing at
yourself in future years.
Thankfully, Mark had such
a loom on his shelf new from the factory which I duly purchased and laid out on
the lounge floor in what I thought was the correct shape. Armed with the Land Rover Defender RAVE
manual and lots of wiring diagrams, I began to mark up the loom (nothing was
labelled), and work out what I could cut out safely in terms of gearbox control
wiring, air con and things like that. It became clear about this point that due
to the left hand drive affliction of US cars, and the layout of the Defender
engine bay, I would struggle to make this fit in the Chimaera as it was due to
the position of the battery, fusebox and lack of a firewall bulkhead (the
Defender has everything in under the bonnet). Despite my best intentions of
using the Land Rover loom as a base, I ended up pretty much cutting the loom to
pieces and tailoring it to fit my engine bay exactly. For this reason I would
recommend anyone trying this themselves to just start from scratch, it’s
probably simpler. Or, if this sounds too much like a pain in the arse to you, I
will be providing either measurements or my entire loom to Mark and his team to
clone so that he can provide pre-made looms specifically for TVRs at varying
degrees of quality (OEM/race spec) and/or completeness (he can supply wiring
pre-terminated with the correct plugs if you’d like to have a go yourself).
As you can imagine; this took AGES. And as
with all wiring, there can be several hurdles along the way. After a few weeks
of the odd hour here or there however, it was ready to fit and forget. Or so I
thought. I thought I would be clever and install a military spec 62-pin bulkhead
connector so I could easily remove the engine for future power mods. This took hours to solder
up and pin test, so you can imagine the volume of the swearing when I went to
connect the battery and realised (after lots of strange readings at each ECU
pin) that I had held the soldering iron on one of the pins slightly too long
and the insulator had melted inside one of the connectors; shorting several
pins together. Both male and female connectors were immediately cut out and
hurled into the nearest bin, and then proceeded the long task of soldering 62
pairs of wires together in order to test the car. Not a good day that one, much
beer was needed to console myself.
Thankfully, my workaround worked
and correct voltages were now seen at all pins. Success! Nearly.
The next point to note for anyone
trying this is that GEMS learns the very first throttle position that it sees
as base, and cannot adjust this base downward - only up (no idea why).
Unfortunately for me, when I connected my GEMS up for the first time, I had the
throttle springs removed and the throttle sat half open. The first inkling I
had that this may be the case was when I started the car and it shot straight
to around 4k rpm – with no silencers on the car this was *quite* a shock.
Although it did start first time so I can’t complain too much! Once I re-fitted
the springs and set the throttle however, the car simply would not idle and
sounded terrible – this one took an age to diagnose. It was GEMS that finally
told me however; on Mark’s recommendation I fitted a WIFI CAN-bus dongle to the
OBD port that I’d put in, and connected this to my Ipad (not bad for 1994
technology!) where it told me lots of things about how it was running,
including that it was at 60% throttle on idle – doh! Resetting this via Mark
sending me a new chip in the post (as I didn’t have the proper GEMS computer to
do re-set the throttle), I fired it up again and this time, the car ran like a
champ.
To say I’m pleased with it then
would be an understatement; although it took time and a lot of understanding to
fit to the car, it is an extremely good ECU now that it’s up and running, and I
would thoroughly recommend it to anyone with a RV8. It is a fantastically
capable ECU and comes in at a fraction of the cost of anything which could come
close in terms of capability.
Before I wrap this up, I would
like to take some time to say a very big thanks to Mark. He has been fantastic throughout
and has taken far more time than I could possibly have hoped for to help me
understand how everything works with regular and in-depth discussions over the
phone; not always in business hours! As a small return favour to him, I have
agreed to write a conversion guide in order for anyone else to be able to
undertake this conversion themselves with a little more ease then I, and will
have the parts produced to sort out the cam position problem. If there is
enough interest then Mark may also put together a kit of parts with the
installation guide, and possibly even a standard loom. If you are interested in
doing this, please drop either myself or Mark an email and we can both answer
any questions that you have.
So what’s next for my car? Well
there are a few niggles to sort out still; while I was replacing all of the
dashboard wiring, I have managed to stop the headlights working which is a bit
of a problem at this time of year. I also managed to accidently cut out and
break the ECU speedo box so the ECU currently doesn’t know when the car is
moving and when it is not which causes GEMS to raise the idle a little high to
prevent any stalling. I have a fix planned for both of these things but I have decided
to wait until spring to finish them up as its far too cold to work on the car
outside right now!
Beyond that, I think I might
drive it for a while, but driving it is only half of the fun of owning a TVR in
my opinion; there’s always the re-trim that it needs, and maybe another
supercharger, and maybe a 5.x engine, and maybe a chassis refurb, and maybe a
bit of paintwork, and…
Thanks for reading!
Nick Simpson
April addendum:
Since writing this article, the
car was kept in hibernation over the winter period (as it was far too cold to
work on outside!). I managed to find the fault with the lights; a rather
embarrassing one I might add – I’d simply put the ribbon cable on the switch
upside down. Which I realised only after dismantling the entire dashboard. Doh!
The speedo box ended up
being a little more involved. GEMS it turns out, requires a greatly different
rate of pulses from the speedo sensor than is used by the speedo gauge. For
this reason I had to delve a little into the world of integrated circuits,
GULP! What I came up with (with some helpful advice from an electronics
engineer) is this beautiful little creation, which I’m clearly very proud of.
It replaces the existing box behind the dashboard, and splits the signal
provided by the sensor on the diff (which I also upgraded so it actually reads
below 10mph). One output runs to the speedo, and the other to GEMS, which
actually sees half the speed the car is actually travelling which neatly gets
around the GEMS speed limiter that Mark warned me of. If anyone would like one
of these, I’d be happy to make a few more.
Once this final problem was
solved, I ran the car up to Cheltenham to at last meet Mark and have him check
the map over. As previously mentioned, my car is a bit of a funny one due to
the supercharger and giant injectors, so Mark wanted to see it in person to
ensure GEMS was adapting correctly.
It turned out to be a good
decision, as Mark almost immediately identified a broken throttle position
sensor; he still has no idea how the car drove so well without it working!
Luckily he had a spare in the car so we changed that out and got the car
running on the rollers. Expecting a full day as per usual, I stocked up on
biscuits and drinks and brimmed the tank, expecting to have to again before the
day was done. How wrong I was though… Once set up with a couple of ancient
laptops used to talk to the GEMS chip directly (which was hugely impressive to
watch by the way), Mark set about a very short tuning session. He simply scaled
GEMS’s map to cover the injector range, tweaked the idle and some of the ECU’s
targets before moving onto power runs. It was all over in a couple of hours and
less than ¼ tank of fuel – a definite first for me at a rolling road!
Unfortunately the dyno only read
up to 300bhp, which thankfully the car reached comfortably - even with
wheelspin despite myself sat on the back, a boot full of weights, a driver and
a passenger helping on the friction front! What we did learn though was that
the power curve is just about a dead straight line, and that the car absolutely
sang all the way through. An unbelievable result; the car is now immensely
smooth through any driving condition, very, very predictable and all around an
absolute joy to drive.
It’s been a long old road to get here, but I don’t think I’ve ever enjoyed driving the car as much as I have now – I can definitively say that it drives much better than OEM, which was the original point of doing all of this.
It’s been a long old road to get here, but I don’t think I’ve ever enjoyed driving the car as much as I have now – I can definitively say that it drives much better than OEM, which was the original point of doing all of this.
Comments
Post a Comment