Day on the dyno - R500

[Colin_T]

After I initially picked the car up and done the engine break in procedure the performance seemed a little underwhelming for a R500, power run on the rollers on its very best run showed peak of 206bhp and around 135lb ft. (On first run was only 201!?!) 
  Granted this was prior to me making up the cold air pipe work from the front of the carbon airbox down to the scoop under the radiator, so of course was effectively sucking in 80degree air off the rad - less than desirable - but I was doubtful it would be the 25*bhp I’m missing…. 

  Once I made the cold air feed it certainly felt a bit better on the road, and more consistent, but when driving the car I always felt there was room for improvement, certainly the power run showed fuel mixture being quite wavey on the top line, and a bit on the rich side in places, and cruising around it felt quite lean - surging and sometimes sneezing like a poorly set pair of Weber’s.

Very clear some work was needed, so I set about installing a serial comms port to the rather old MBE 967, obtained the correct firmware for the ecu version, and of course the PIN code… 

Back to the rollers today for some testing….

I began the day by setting a new baseline figure so i had a direct comparison with/without the cold air feed. Performance was immediately up 9bhp and around 10lb ft. So now 215bhp and 145lb ft. 
  Then for fairness of testing I optimised the fuel and ignition on the top line. The mixture line became flattened out, and I done some ignition swings, the results showed that the k series doesn’t seem to be particularly mixture sensitive to power, I tried several runs with quite a large difference in mixture (12-13.5  WOT) yet all yielded almost identical power throughout the rev range. I decided to leave it at 12.5 for a bit of combustion cooling with hopes of being able to extract something with the ignition timing.  
  The ignition timing as it turns out was not too far off, the maximum increase was found to be a little under 2 degrees, and no change in others cells, with no gains to be had trying more (or less) timing.
After optimisation results peak 216bhp and 148lb ft. But some larger gains throughout the rest of the rev range - and a smoother, less humpy curve - mostly due to the flatter, less wavey, AFR line. 

Next up I done some cam timing testing. I had discovered that the inlet timing was too far advanced - at least compared to the book figures. Meaning there would be little clearance under the valve! (Worrying as I’ve done 2 trackdays in it!) So I pulled the timing back to the Piper cams figure and performance improved a lot. I then done another run a further 0.5 camshaft degree retarded but performance significantly dropped off, so went back to the piper figure. 
  I then advanced the exhaust cam .5degree showing an increase. I then advanced a further 0.5 degree which showed a small gain at the low end, but loss in performance from mid range onwards. So back to the first “.5degree” position. After this ignition and fuel was revisited for any trims to suit the final position of the cam timing.
With the end result showing the cam timing optimisation giving a peak gain of +6bhp, 222bhp and 150lb ft with increases of 5-10bhp throughout majority of the rev range. Certainly worthwhile.
Seems as if this engine is not fussy with its fuel mixture, but sensitive to cam timing it seems… 

Red power / Blue torque /
Orange AFR. / Green driveline loss. 

SOLID LINE = FINAL run after cam timing and fuel/ignition optimised. (222bhp)
Note: flatter AFR line. (7k onwards they all about the same AFR) 
DOTTED LINE = BASE run prior to cam timing (215bhp)
DASHED LINE = BASE run before cold intake and any tuning (206bhp)
 

It may continue to make a few bhp more if I raised the rev limit and let it rev to 9 or more, but is the few bhp worth the risk? Probably not. 
Still not making the 230odd that caterham claim they make, however rolling roads, engine dynos, power figures… (let’s not disappear down that wormhole)
After all I could go on another rolling road and show 240… so I’m happy where it is. Plus I know this dyno shows proper “no nonsense” figures and is ultra repeatable making it perfect for testing - which is what you need when doing cam or fuel / ignition testing. 
 

Now all the top line stuff was complete, and I knew I wouldn’t be adding variables like swinging cam timing and having to do it all again, I then set about the rest of the map. Firstly I wanted to recheck idle balance and airflow as the cam timing had changed, and to fine tune the idle and idle fuelling.
I like to add to the ecu comments field for future reference any TPS voltage or value at idle, plus the idle airflow figure. E.g 5.5 kg/h 0.43v idle. 

The lambda control on some of these early MBEs is a bit questionable to put it politely, infact on my car although it has a lambda probe in one of the primaries, the lambda control, or any lambda target map wasn’t even active in the ecu! I was even advised not to use it. 
  So I actually just created a target map on a spreadsheet on my phone and worked to that and just left it open loop. 
  I like to break down the fuel mapping process into areas based on target mixture and map those areas as individual sections.
  I first set about mapping all the cruise areas of the map where I wanted to target 14.7 AFR. So in this instance anything below load site 8, and below 5000rpm. (Usually around 4000rpm is fine but given the R500s gear ratios I wanted to be sure that even at motorway speeds …and beyond… it would be economical.) I found the majority of this was actually too lean initially, 15 even 16s at higher rpm - explaining the horrible surging and sneezing I experienced. Once the fuelling in this area was all at target I then run through the ignition, optimising it where possible.

After this the same process is repeated for the higher load and higher rpm map targets. May 14, 13.5, 13, 12.7, 12.5 areas. Which taper and smooth off as the loads and speeds (where applicable) increase. I found that at higher loads the original map was very rich, and at high rpm mid to 3/4 throttle was excessively rich - enough to make it misfire, down in the high 9s at one point! 
  After the whole map is mapped to my target I run through the ignition making improvements where possible to find maximum brake torque (MBT) ((the minimum amount of timing required to achieve the most output))

Any smoothing and blending is carried out and the fuel map run through a final time to make sure it is all still on target. 

Final result - the car is a lot more drivable and smoother on the road, pulls much nicer everywhere. Plus of course a performance gain when the throttles get opened up. Looking forward to the summer.

It is my eventual intention to move away from the 20odd year old MBE and fit something a bit more modern. I will then get it on some proper lambda control, knock control and other engine protection strategies, plus mapped on some more modern injectors. But for now it should be a lot nicer. 
  I may even look at doing another cylinder head for it with different cams and further port and valve work. I’m confident with a more modern ecu and correct size injectors, properly mapped I could realise the power and torque benefits of the different cams without having to push the boundaries of the rev limiter, yet still retain road drivability. 

 

- Colin 

[PlastererPete]

Nice write up.

It seems you are just around the corner from me. I may well pop round with my CSR for a power run to see what shes actually putting out. Bog standard map and although perfectly fine to drive, she does run rich.