Printable View
RB20/25/26 & CA18det, info relates to the SR20 aswell,
they mention belt deflection on big power lumpy camed RB's suffer about 8 degrees of timing retard and power lost :cry:
do CA18's suffer the same loss is power? https://www.youtube.com/watch?v=mh4hxnfMfGA
Strange thing is that if the stock ECU detects detonation or the sensor is short circuit (unless replaced by high ohm resistor) it retards the timing. I don't know by how much but around 10 degrees was claimed. Video says "high power" engines run near the detonation limit. Video also doesn't say that some det events can't be heard by ear but a piezo sensor will hear them.
The CAS isn't to blame it's the "rubber band" cam belt. The core of the belt isn't a "rubber band" it's KEVLAR and it doesn't stretch every time the engine is revved hard. If the belt "stretch" causes retard then the cams also retard (Haltech don't seem to be leaping in to fix that). 7.5 degrees is 1/2 tooth, 1/2 a tooth in 48 teeth between crank and cam is 0.01 (1%) strain in the belt. I think this claimed elastic [1] "stretch" is highly improbable. It's well beyond the limit of elastic proportionality of KEVLAR - belt will be permanently stretched, engine would be retarded and stay retarded. At the limit of elastic proportionality 0.005 (0.5%) there would less than 4 degrees of timing change and I'm fairly sure that no one runs timing belts anywhere close to that. l'd hope about 1 degree and 25% of the limit of proportionality = factor of safety of 4. It may stretch a bit as it ages over 60K miles / 5 years.
https://www2.lbl.gov/ritchie/Library...nDependent.pdf
The condition they describe will apply to ALL belt driven cams.
CAS has 4 "home windows" one for each cylinder, the windows are different sizes and the ECU counts 1 degree pulses during window "open" period, this tells it which cylinder it's on. Engine only has to turn 1/2 revolution for a "window" to move past the sensor and ECU to start running engine. After market missing tooth from crank sensor wheel may have to turn a full crank revolution to get the reference and start running engine. A single cam "home" signal with fully toothed crank sensor may have to turn 2 revolutions before ECU can start running engine. (Though the camshaft magnet in the video appears to have 2 teeth, one turn of crank but this means they can only run in bank fire mode and not sequential.)
Nearly all after market ECU have much lower resolution of engine position. At best they have 36 teeth on crank, that's 72 on/off signals for each revolution, 5 degree resolution. Video 12 teeth on crank is 24 on/off for each revolution and a (frankly laughable) 15 degree resolution. They then have to "estimate" based on engine speed where it really is. But actual engine rotation speed slows as it comes up on compression and speeds up on first 1/2 of power stoke so they have to "guess" [2] where the engine is, 15 degrees is a lot of rotation time to be guessing. Nissan CAS has 360 slots, 720 on/off signals for each cam revolution and ONE degree resolution of engine position. I'm not sure if it does any guessing or if timing changes in 1 degree steps (I think the latter). The Nissan CA18DET ECU runs off a "timing chip" (fiendishly cunning and complex device) and the CPU only determines the position that it programs the timing chip with.
[1] Elastic - the material recovers it's original length when stretched and released. If the "elastic limit" is exceeded the deformation is permanent and the material has "yielded".
[2] "guess" requires a fast and powerful CPU and a 4D acceleration curve v's rpm v's load.
Quote:
Originally Posted by skyshack
hmm, even better approach than mentioned yout video...well done
I clicked on the video as I had endless problems with the CAS splines wearing many years and thought they would mention that and the timing wander issue
when they went onto the dreaded power loss I was a bit surprised as not read or heard of it before on the rb or any belt driven car engine enough to cause a loss
many years back this came up in the world of ford pinto tuning and later the cossie yb200 and tests if I remember were similr to a chain
as in both had near zero strectch at operating loads ,so I think they are leading us astray a bit saying the belt stretches
it was the belt deflection bit they mention, I'm not convinced it can deflect or move at all when its going up the rev range
thinking how fast the thing is spinning,it's under load can it really move and alter cam timing?
read about lots of rb builds and not something thats come up when chasing high power,
the RB has many issues tbh but at most a swap to a Greedy or Gates racing Belt is only thing done regards the belt
and it's changed to a more reinforced belt as a 'hold it together because it cost a lot' mod not because the stock belt stretches
so I wont be designing some sort of super close fitting belt guard just yet, I might see if it's come up as an issue on RB's or other engines
Come on skyshack you've been on here long enough to know all about CAS issues with worn splines on exhaust cams...
In an ideal world blah blah, but 30 year old engines...
Exhaust cam CAS drive spline wear is a given on the CA,the fix is just a camshaft with good splines and an unworn CAS and lube the splines
and a good tight fitting 'O' ring on the CAS drive to possibly further resist it trying to move on the splines,
I'm still curious if anyone else has a take on the claimed power loss from belt deflection
and it retarding cam timing as the belt (or chain) is rapidly moving
if the rapid belt/chain movement alters timing,
then the same rapid movements are rapidly altering the timing back and forth between perfect timing
its not one constant timing swing like they sort of suggest causing a loss in power,
it would be multiple times every stroke
if you had a calibrated crank pick up/sensor and a cam signal pick up upon each cam it could be montored/data logged
to actually confirm if it is actually swinging the timing at any point of the rev range,
I haven't found any testing like that so far
They swapped the CAS. I ass u me ed that they would have been working with a fairly fresh build. "high power" suggests new high lift cam so will have new cam splines. I would hope they checked the crank keyway, must have seen it when they put the 12 tooth wheel on unless they did it with their eyes shut.
The power the engine makes doesn't affect the cam timing or belt tension. It's simply down to cam inertia "I", acceleration "a" (I^2a) and cam profile lift "x", engine speed "v", valve spring rate "K" (Kxv). So they will have high lift cams with a more aggressive profile and possibly stronger springs but that's not going to create this sort of issue.
Having an inertia dyno that doesn't load the engine enough or having slip on rollers so it accelerates much faster than it will when tyres have grip on track would affect belt load. (But pulling the timing so engine makes less power is called "traction control".)
So I'm pretty sure that they have observed OEM ECU detonation detect page EF&EC-17.
This says it pulls the timing additional 5 degrees retarded from knock map (LOW octane map on CA). IF they have knock map that is 5 degrees retard then they get 10 degrees.
http://forum.nistune.com/viewtopic.p...8aeaf&start=30
Not RB as this is CA specific. "These days we remove knock maps from this vehicle with feature pack updates, so just one set of maps (Nissan did this for Europe). ECU will still retard timing on knock count, but nothing beats properly tuning the vehicle to avoid needing this backup feature"
http://forum.nistune.com/viewtopic.p...=2027&start=30
On a JDM CA18DET ECU and loom there is a signal for the turbo boost solenoid which could be used to indicate that it had switched to the LOW octane map. Euro Hi and Low octane maps are the same, JDM are for 96 Ron and 89 Ron, based on increasing knock count ECU will shift timing to lo map. If that knocks as well it pulls the timing, don't know where that is but it's separate to the 2 ign maps or factored off knock count. Knock count gets forgotten every time ECU is switched off so it re-calibrates to the fuel mix every time it runs.
It may be 8 bit 1 MHz but it's still quite clever (as the timing chip does the grunt).
I understand if it detects knock the ecu signal to cas pulls 5 and that = 10 at the crank,thing is they describe the belt is deflecting/oscilating/wobbling very fast
and that is retarding the cam timing, I see were one was going with that with the knock issue,a lot will be on stand alone ECU as well
then the thing is they don't cure the belt deflection with extra tensioners and some close fitting guard so the belt has zero slack & no long runs
they fix the electronic timing with a crank sensor/toothed wheel , if the belt is still deflecting and retarding timing in a mechanical sense
they said =10% power lost
corrected ignition/injector timing is not really the 100% cure imo as it must be loosing power from being retarded up to 8 degrees from optimum timing
if the belt deflection issue is left (10% power lost) how much is gained back from just correcting ign/fuel fire points 5%?
the engine is still down 5% , why do they not do something about the belt deflecting to get back the full 10% the 8 degree retard causes
As I said up ^^^ there. Agreed they don't seem to be leaping in to fix the retarded valve timing which must be hurting power.Quote:
Originally Posted by ANDY black s13
Haltec have a another video on RB blocks. They admit it's a contentious claim but say they have heavier springs and lumpy cams.
But chains can have problems as well.
The only time I've know an OEM fit gears in place of a chain was Honda's VF750 when it became the VFR750. They had run into their age old problem of incorrect design of a 180 degree twin (V4 is 2x 180 degree twins at 90 degrees). Twin cylinder CB250K4 snapped cams as unsupported in middle, fitted a middle bearing on CB250G5 and that wrecked the chain tensioner. All the bumps are in 180 degrees round the cam and it runs unloaded on base circle for that period. Then it hits the first cam and the tensioner goes from unloaded to full load. An inline 4 has lumps spaced evenly all the way round the cam and never unloads fully so tension change is reduced. CX500 V twin had chain tensioner issues for same reason, all the lumps in one 1/2 of cam rotation and tensioner that couldn't cope with fully unloaded to loaded cycles. They had to recall and fit a new tensioner with bigger bolt in place of the skinny one that sheared.
sorry missed the bit were you mentioned they skip over the belt issue,didn't see the other video and seems like more an idea/theory than an actual fault
so as I suspected as not mentioned in countless builds chasing big numbers not something to really worry about
agreed chains can be no better,years back I ran a motorbike workshop with another chap, the cx500 was one bike we see often (dispatch riders hack)
VFR750 was a bike in its day, then they did a few with gear driven cams rc40 vfr400 they sound amazing as well,
I have a vtr1000 sc36 that has a hydro chain tensioner issue were the front cylinder tensioner
on decel can loose tension and the chain skips,swap the hydro's for manual tensioner fixes that
the rc51 vtr1000 (the base bike for the british superbike series) got more power 136hp and changed to gear driven
the Moriwaki squeezed over 180hp out of that motor
funny how first thing I thought of when they started the belt deflection thing was gear driven cams from years back and the honda bike engines
Having had my water pump fail last week I'm now in possession of a 30K mile used cam belt.
Distance between marks for crank and exhaust sprockets = 384 mm (measured with tape so +/-1mm).
Teeth = 48.
Pitch = 8.00000 mm.
So at 1/2 life there is no measurable [1] permanent stretch on a stock engine. [1] Using DIY measuring devices found in home AKA a tape measure.
Crank sprocket has 24 teeth. Error 1 mm = 360/(24x8) = 1.875 degrees.
If Haltech's claims are real a 4 cylinder will have less than 2/3 the stretch under dynamic conditions. So 5.3 degrees retard. A 83.5 mm stroke CA engine revs lower than the RB26 73.7 mm RB so that will also reduce the load. Then there is the lumpy cams and valve springs to consider (excessively over-rated? possibly double - you know those RB guys).
Here's another angle on it.
https://youtu.be/_sEDuNGnRKA
quite good, the bit were he says testing the stock cas v's another trigger was the same further suggests is the belt actually causing a faultQuote:
Originally Posted by immy21
has anyone run an engine with a camera mounted to veiw what the belt is doing,is the cam timing actually wandering?
I'm not convinced there is actually a measurable loss at all,
we have skyshacks used belt showing near zero stretch,and previous post about testing chains vs belts were the same
so the strectch idea is done imo
I think the haltech video was missleading like we are all loosing power,
I have yet to find anything testing the theory and anyone trying to cure the belt/chain from deflecting like they found it was a real power loosing fault
not on about the crank trigger method correcting the timing/firing points, I have yet to find anyone else claiming power loss from chains or belts
but if anyone has something please post it up as I'm curious to learn if it's worth chasing,can you really pick up a 10% gain
Not many other makers run ignition timing off cam. Not even Mitsubishi who made the CAS.
Toyota 2JZ has a toothed wheel on crank. Initially 12 teeth, then they went to 36-2 teeth.
No one has 15 degrees advance under load or at high speed, 35 degrees advance is nearer the mark.
Nissan CA's 1 MHz "BBC model B" ECU (I kid you not) doesn't have to count the 360° pulses, they can drive the fiendish timing chip directly. The ECU can get speed from the 4/6 cylinder windows or by counting while just one 360 degree window is "ON".
Haltech can't use the NISSAN's "too fast" 360 degree CAS signal, 9 min in to this video (then hear the rest).
https://youtu.be/XqUDHwquH-w
And another cat is out of the bag. All high power engines are run to higher than stock rpm. 25 m/s mean piston speed is not unreasonable ~10,700 rpm for RB and 9,700 for CA, Toyota VVTi engine runs that stock. At some point the stock ECU and many after market ones will give up. Stock CA ECU gives up on sequential injection at 3200 rpm, then it batch fires all injectors together.
Replacement CAS with any other number of slots is just to replicate the crank trigger wheel of an after market ECU without the problem of mounting it on crank and fitting a pick up. 24 slots is 12 tooth crank wheel and single slot is cam sensor input to after market ECU.
that bit about the haltech ecu not reading is confirmed in something I found here https://www.sau.com.au/forums/topic/...ons-in-rb2630/
one tuner found the aftermarket ecu's couldn't read the nissan parts, that was one common pattern with cars having the timing scatter at certain parts of the rev range
one had a 5500 rpms scatter issue,that was possibly a CAS drive wear fault, another posted a data log with 3 rpm points were the timing became random
As one mentioned RB's were making 1K whp from the start on the OE ecu/cas/coilpacks etc
another constant was people using the BLUE uprated Gates timing belt,
one the tooth profile dont match perfectly the pulleys tooth shape
possibly allowing it to move on the pulleys a few degree either way
and the belt stiffness is some how creating more belt deflection at certain rpms,
three cars having the scatter/missfires power loss were fixed by going back to the OE Nissan belt
I was going to get the blue belt thinking it would be better/less likely to fail on a tuned engine, but I ain't now after reading that
I was also going to splash out on aftermarket ECU, thinking nistune board in the OE ecu now as a cheaper and less faff to work option
it seems people swap the ecu then have to design or buy and make parts fit like in the video to make it work
the Apexi power ecu works like the stocker they work fine as well