Tear Down and Rebuild of My R400 Duratec Engine

[aerobod - near CYYC]

9am on the 2nd of May is the appointment time for my rolling road day.

Of the 2 rolling road operators who weren’t phased by ECUs they haven’t dealt with before, both having experience of at least 10 different ECU manufacturers, one tapped out today after talking with his friend in Australia who described the MBE ECU as ‘a nightmare to tune, better to swap in a new standalone, mate”. This operator wanted to learn the MBE and do the tuning with his computer gear.

The company I picked is not phased by me bringing my computer gear along, then working with me on tuning. They currently have a Dynojet 248E rolling road that is good up to 1200bhp and fairly common here in North America for tuners.

[Colin_T]

I wouldn’t go as far as calling it nightmare but some of the way it goes about doing things is more long winded than it needs to be. So can take longer to do some tasks than it would on other systems, but ultimately it does do a good job of running an engine once it’s done. 
 

im sure between the two of you you will work through it. 
 

Assuming the guy won’t have a pot box for the MBE and will be using the keyboard, familiarise yourself with the keys and shortcuts.
use F9 to highlight cells you’ve adjusted if not in “mapping mode” mapping mode is ok if your doing fine tuning and able to be perfectly in the centre of a site, the issue is with mapping mode you can’t select and adjust a group of cells or a nearby cell. It locks you into one cell only until you “stop mapping” then you have to adjust what you wanted to adjust, then “start mapping” again (have to “enable mapping” first !!) And if you’re not in mapping mode and just adjusting tables you have to “send” before the changes update…. to be honest that’s how I usually do it, change the value, click send, check the change, if ok, highlight it F9 so I know I’ve done it, move to next site.
initially you can quickly and roughly map it by skipping load and speed sites and just calculating between sites ((note: can only do 2 rows/columns at a time or else it tries to interpolate a square corner to corner)) then you can go back after and fine tune it after. 

oh and save your work regularly !

Colin 

Edited April 10 by Colin_T

[aerobod - near CYYC]

4 hours ago, Colin_T said:

I wouldn’t go as far as calling it nightmare but some of the way it goes about doing things is more long winded than it needs to be. So can take longer to do some tasks than it would on other systems, but ultimately it does do a good job of running an engine once it’s done. 
 

im sure between the two of you you will work through it. 
 

Assuming the guy won’t have a pot box for the MBE and will be using the keyboard, familiarise yourself with the keys and shortcuts.
use F9 to highlight cells you’ve adjusted if not in “mapping mode” mapping mode is ok if your doing fine tuning and able to be perfectly in the centre of a site, the issue is with mapping mode you can’t select and adjust a group of cells or a nearby cell. It locks you into one cell only until you “stop mapping” then you have to adjust what you wanted to adjust, then “start mapping” again (have to “enable mapping” first !!) And if you’re not in mapping mode and just adjusting tables you have to “send” before the changes update…. to be honest that’s how I usually do it, change the value, click send, check the change, if ok, highlight it F9 so I know I’ve done it, move to next site.
initially you can quickly and roughly map it by skipping load and speed sites and just calculating between sites ((note: can only do 2 rows/columns at a time or else it tries to interpolate a square corner to corner)) then you can go back after and fine tune it after. 

oh and save your work regularly !

Colin 

Thanks for that Colin. Without the mapping box, I will be restricted to Live Mapping mode only. I think R79 has a few enhancements over earlier versions to smooth the mapping process, such as colour intensity in the active map indicating the cell that is closest to the engine speed and throttle position, so the injection time or ignition advance can be quickly jogged to the modified value with either the course or fine change keys.

One question I don’t know if you can answer (maybe @7 wonders of the world or someone else will also know the best answer) is the best place to tie down the Caterham while on the rolling road. I think the Dynojet 248E is a 2-wheel dynamometer, so I was planning to strap through the stationary front wheels, but other than around the tubes where the strengthening plates for the rear jacking points are, I can’t picture anywhere else to tie the back down to.

[7 wonders of the world]

Over the tyres rather then thro. The wheel. To reduce the chance of rim. Damage. 

Strops and ratchets from the cage or roll bar there's not a lot of weight on them.. You can always add ballast to the boot and passenger side. 

Single rollers to can strop around the A frame to Dedion but this will pull off the drive roller. On a twin set up. 

[aerobod - near CYYC]

10 minutes ago, 7 wonders of the world said:

Over the tyres rather then thro. The wheel. To reduce the chance of rim. Damage. 

Strops and ratchets from the cage or roll bar there's not a lot of weight on them.. You can always add ballast to the boot and passenger side. 

Single rollers to can strop around the A frame to Dedion but this will pull off the drive roller. On a twin set up. 

Sounds good using the roll bar instead of the jacking points for tie down to get that extra load on the rear tyres, Neil. I will see what straps they have for the tyres, I'm OK going through the wheel rims if they can't easily fit under the front wings, as that is what I do on my track wheels when the car is on the trailer due to awkward strap angles if I go over the tyres, the scuffing is already well established!

[Geoff Brown]

And/or as in my case when car on RR sit on the boot lip for an extra 85 Kgs !

Edited April 12 by Geoff Brown

[aerobod - near CYYC]

With Sat being start day, I wanted to just double check the fuel injector timings I had built into the new map. Compared with conventional high output naturally aspirated engines, I keep coming back to what seems to be very efficient BSFC figures for the Duratec in R400/420R and R500 variants.

Looking at logs I have taken over the years and checking BSFC at peak specified torque and power, I end up with 235g/kWh at peak torque (with Lambda=0.86) and 240g/kWh at peak power. I’m also adjusting for my altitude with typical 0.88 bar pressure, normalizing back to 1 standard atmosphere. One would expect 250 to 300 as a typical value.

I will have to keep a close eye on Lambda values before taking the engine to high loads, but so far all the calculations seem to be lining up.

[aerobod - near CYYC]

Parameter	R400 @torque peak	R400 @power peak	R500A @torque peak	R500A @power peak
Power output (bhp)	180	210	240	258
Power output (kW)	134	157	179	192
RPM	6300	7600	7165	8200
Torque (Nm)	203	197	238	224
Observed injector time (ms)	12.94	12.50	-	-
Observed ambient pressure (bar)	0.891	0.889	-	-
Injector time at standard atmosphere (ms)	14.71	14.25	9.60	9.36
Injector flow at 100% duty (g/min)	172.6	172.6	310.0	310.0
Lambda	0.86	0.82	0.86	0.82
Injector duty cycle (%)	77.2	90.2	57.2	63.9
BSFC (g/kwh)	239	238	239	248
Total air mass flow (g/min)	6651	7407	8857	9432
Volumetric Efficiency @ 40C intake temp(%)	89	87	108	101
Thermal Efficiency (%)	35	35	35	33

Using measured parameters on the old engine and projected parameters on the new engine based on other people's info and projections from the existing engine, this is my best guess at the engine parameters at the moment. My general conclusion is that if BSFC is a little optimistic, then the published Caterham info is over starting the power output of the standard engines (which has generally been the caes). Also, the Caterham peak torque RPM for the R400 plenum engine seems to be a bit low, my previous data logging suggests it should be more like 6,700RPM instead of 6,300RPM, which would change the BSFC to 237g/kWh at that engine speed, logically making the BSFC lower at peak torque than peak power, which is the way it should be.

[7 wonders of the world]

Caterham over quoting on the power output.... that's almost like saying their cars are heavier than published weights.... 

Go stand in the corner and say 200 hail. Mary's........ 🤣🤣

[7 wonders of the world]

Though not famoar with the MBE I assume. There is a reference target the fuelling works from... So as long as these are in the ball park it should learn in a form of adaptive tuning....? 

Do you run a CID sensor...? 

[aerobod - near CYYC]

Still lots of mapping adjustments to do, also need to determine if my new Lambda sensor is faulty as it would suddenly go from rich to maximum lean, but I could still smell excess fuel. Only one small leak requiring the clip on the main hose at the back of the head to be tightened more.

After the first start I shut the engine down quickly from what I thought was excessive top-end noise, but it turned out the throttle bodies allow quite a banging noise out of the intake at low throttle openings, probably due to the late valve closing.

Took about 5 brief runs for about 20 seconds at a time to fiddle around with the throttle mapping at sites 0 and 1, as it would not rev initially due to being either too rich or lean just off idle. Did the 20 minute initial cam and ring running-in as specified by the manufacturers, then did all the throttle air-flow balancing (only needed to use th3 bleed screw on #4 to make 0.5kg/hr adjustment, setting all to 9kg/hr after setting mechanical idle at 1,200RPM, then re-activating ECU controlled idle to 1,000RPM. Did a 6km brief drive to log some data with medium load up to 4,500RPM.

Did the first oil change and cut the oil filter open, a few bits of old RTV debris and tiny specks of what looked to be machining debris. Probably the amount expected, typically like this every couple of filter pleats:

The finger filter also trapped a few bits of larger debris:

Overall on track, need to have a look at the logs and make adjustments before more road use and the upcoming rolling road session.

 

[aerobod - near CYYC]

8 hours ago, 7 wonders of the world said:

Though not famoar with the MBE I assume. There is a reference target the fuelling works from... So as long as these are in the ball park it should learn in a form of adaptive tuning....? 

Do you run a CID sensor...? 

Hi Neil,

I can do adaptive tuning with the MBE, but I need to determine if the new Lambda sensor is giving correct results first. I do have a cam sensor fitted, haven’t checked the data yet to see if cam sync was achieved on target.

[7 wonders of the world]

Glad to here she is back to life.. 😃

Only reason I asked about the CID if not your injectors would be paired and the injector timing then makes little difference... I don't run one and mine are paired. 

I emailed. My old and current data over for reference providing your initial lambda map is reasonable your will be OK and update after a brief run after checking your adaptive data. 

[aerobod - near CYYC]

2 minutes ago, 7 wonders of the world said:

Glad to here she is back to life.. 😃

Only reason I asked about the CID if not your injectors would be paired and the injector timing then makes little difference... I don't run one and mine are paired. 

I emailed. My old and current data over for reference providing your initial lambda map is reasonable your will be OK and update after a brief run after checking your adaptive data. 

Thanks for the emails Neil, I was going to compare them with my logged data tomorrow. The weather would be good for putting some running in km on the car tomorrow, but I don’t want to rush the analysis and also promised that SWMBO and I would go out somewhere, too (but not in the car!).

I’m currently set for sequential injection once cam sync is achieved, batch (paired) before sync when the engine starts.

[7 wonders of the world]

We had dry weather here yesterday too so I had my first blat since October and the first on the new engine after mapping.... Very very impressive the torque is wonderful and much more linear than previous and bloody fast.. 😁

Light load mapping is superb I only ran on the open loop map to try in order to highlight and coughs etc... Which I will. Correct with adaptive... 

The noise on WOT is enough to wake most of Herefordshire... 😅... I no longer need a horn... 

[aerobod - near CYYC]

Last weekend when doing the throttle body balancing and adjustment, I noticed the air temp data showed it rising to 60C due to the proximity of the air filter backing plate to the engine. Even with rubber isolation washers and 30 mins of driving around, the temp only dropped down to 48C.

I decided today not to be lazy about sorting a better location for the sensor, especially with more mapping and the rolling road session coming up, I want as accurate temperature input as I can get. I picked up some PVC threaded 1” water pipe fittings, machined them to make a suitable stand-off tube and cut a 32mm hole in the air filter backplate:

The end plugs were drilled, the one that was shortened for the sensor end, the other slotted to pass the cable through with the minitimer connector left in place. The main body was enlarged inside to allow space for the connector and a spacer was created to optimize the distance from the backing plate:

The position is very close to the trumpet inlet, but not too close to affect the bellmouth flow:

The PVC should provide good thermal isolation:

 

I’ve recalibrated by Lambda system, updated the maps from data logged last week and fuel data provided by Neil from his engine, created a new target Lambda map that should be close to the needs of this engine and switched on adaptive Lambda. Hopefully I will get at least 200km of moderate load running-in tomorrow to populate the adaptive maps, then more analysis, followed by another few hundred km over the next week or so, before another oil change to synthetic 5w50 for the rolling road session.

[7 wonders of the world]

Hi James, 

this can be an issue as an heat soak on the inlet, I place my sensor between 1 & 2 on the upper portion of the backplate to catch the airstream isolated on thick fibre washers in an oversized holed, I have also used a 7mm phenolic spacer between the head and the manifold which makes a significant difference to the temperature of the entire inlet tract.

When i get a new infa red thermometer i will give you some numbers both running and after heat soak when stood

[aerobod - near CYYC]

The repositioning of the air temp sensor worked well. When it was in the backing plate it would be about 45C above ambient with the car idling for 10mins or so, then only slowly drop to 30C above ambient while driving with the engine up to temperature.

Moving the sensor so it is actually outside the bonnet but inside the filter protrusion next to the trumpet bellmouth, dropped the temp to 30C above ambient for an extended idle with an up-to-temperature engine, but it drops to about 18C above ambient after a couple of minutes of driving. I think the value now it is quite representative of the average air temperature in the throttle body inlet tract.

[aerobod - near CYYC]

Dyno Day at RCTS Autoworx was successful today, 7 hours of setup and tuning to cover all but cold starting. I was going to try and log a cold start as it was sleeting and -1C when I got to the rollers, but needed to get the car inside instead of trying to log info while the laptop was getting wet. The car starts and idles well at 10C and above, but needs throttle and coaxing at 0C for a minute or so.

Initially I was disappointed with the engine output, but they are geared towards measuring wheel horsepower at sea level (rollers are at 1050m altitude). They dynoed a Hellcat with a quoted 717 crank horsepower last week and measured 594bhp. They measured just over 212bhp on my car, that would equate to 256bhp at just under 8000RPM with a similar ratio as the Hellcat, right were I was hoping to be, if that is the case.

Just waiting for them to email me the Dynojet plot.

Here is the full throttle run once the mapping was generally complete: 

 

A few other photos of the rolling road setup. Without the nosecone and bonnet on, coolant temps hit 100C and needed a cooldown in extended sessions. With the nosecone and bonnet in place the normal 90C was seen during extended sessions.

We spent a couple of hours investigating a couple of issues.

The first was that the dynamometer Lambda sensor only agreed with the wideband sensor in my exhaust at medium loads and was about 0.08 different at high loads and 0.10 different at low loads. Due to the inflexible probe inserted into the exhaust, it only makes it to the bend, we concluded by reviewing plugs, smell and peak power when nudging fuel up and down that the exhaust collector was allowing reverse flow to cause the dyno probe to give false readings, tuning was much more predictable using my wideband sensor.

The second problem was my knock sensor system kept activating around 6500RPM (it used to do this at about 7200RPM with the plenum). We pulled timing by 4 degrees and it still activated at the same point, pushed timing up a degree and power increased 3bhp, another degree and didn’t increase at all. We concluded that after multiple inspections of the plugs with a loupe that there was no detonation, the knock sensor filter is just picking up mechanical noise at a slightly lower RPM than before, probably the more aggressive cams and throttle bodies to blame.

Lots of maps saved from different scenarios and logging to match them, so I can go back and look for any anomalies that we may have missed during the session.

I’ve also now settled on 1200RPM for idle due to the sound through the throttle bodies being very tappety at 1000RPM, almost like a valve is about to drop into the engine - gave me flashbacks of the original engine failure 😀

The proof of how good the engine is now will be in the first track day on 12th May.

[John Vine]

Great stuff, James.  I'll peruse the video tomorrow.

Interesting rear plate!

JV

[aerobod - near CYYC]

19 minutes ago, John Vine said:

Great stuff, James.  I'll peruse the video tomorrow.

Interesting rear plate!

JV

Hi John,

We are lucky here that the number plates can be chosen for about $200 as long as it is no more than 7 characters, doesn't have a reserved name or special format used by, for example, the Government and has no offensive words or connotations.

It is a pity that we can't have 8 characters, so I had to go with Caterum instead of Caterham, but it does cause the North Americans to pronounce the name close to correctly, instead of the normal emphasis they would have on the 'ham' part of the name!

[aerobod - near CYYC]

The graph from the rolling road:

Values are measured wheel horsepower and torque corrected to sea-level values, compared with the rolling road altitude of 1050m / 3,400ft.

I'm very pleased that we got the AFR / Lambda value so close to a straight line across the whole range, with it slightly decreasing to the torque and power peaks.

The actual revs measured may be slightly lagging on the dyno, as we did hit the rev limiter at 8200RPM, so peak power is effectively at the redline. The torque peaks are influenced by the throttle body resonances, likely 5th harmonic lifting the higher torque peak and 7th harmonic lifting the lower torque peak.

The revs reached at throttle cut on the Easimap plot show 8202RPM, with Lambda just where I wanted it to be for safe operation.

Applying a 20% drivetrain loss factor in 4th gear to the wheel measurements lead to 255bhp at nominally 8200RPM and 188lbft of torque at 6500RPM, compared with the quoted 263bhp@8500RPM and 176lbft@7165 for an "official" R500, so I think it is close enough to job done to call it an R500A!

Edited Friday at 17:47 by aerobod - near CYYC

[aerobod - near CYYC]

From some of the info put out by Dynojet, the manufacturers of the rolling road I used yesterday, I can see why they don't give the tools to convert to crank horsepower and why they emphasise wheel horsepower measurement: https://www.dynojet.com/blog/whp-or-chp-which-should-you-trust/. They also tend to lease the rolling roads to the operators and provide the calibration and maintenance services to keep them in good order, discouraging them from doing any adjustments. Mitch, the RCTS owner, says that they don't provide any documentation on re-calibration, just setting the system up for the local atmospheric conditions and checking accuracy during scheduled maintenance.

[7 wonders of the world]

 

Calibration and certification by the manufacturer and not the user is pretty standard.. It's in there interest to be accurate. No manufacturer wants to gain a reputation for telling porkies

WHP is what you see however you can accurately measure the losses on run out in order to gain accurate flywheel figures useful for comparing to the figures provided by vehicle manufacturers for instance it takes the guesswork out of comparisons

Gear choice and tyre pressure can always influence readings though single rollers are less susceptible to variations.  

If you leave with better numbers and better driveabilty than when you arrived you've had a good day... 

 

[aerobod - near CYYC]

12 minutes ago, 7 wonders of the world said:

If you leave with better numbers and better driveabilty than when you arrived you've had a good day... 

That was my aim yesterday which seemed to be achieved, with full verification next weekend on the track. Entered with very good mid range, paid virtually no attention to it during tuning as the first run showed little change needed. Was all about the top end (ensuring safety of mixture under high load), bottom end from low revs with high throttle (Caterham is too light to get meaningful readings from road logging) and playing with the timing to get it to the right position.