Saturday, July 6, 2013

Tuning an 851 SP. Or, the things you don’t know you don’t know.

.
The first time I played with an 851 SP (SP3), I ran it on the dyno then had Duane make me an eprom as required.  But, it didn’t seem to work as I anticipated it would.
 
Next time it was an SP4, fitted with an FIM additional memory board in the P7 ecu.  Although I made many changes across the fuel mapping, and generally made the low speed running much nicer, the WOT fuelling didn’t change as I had expected.  Specifically, when I made zoned fuelling changes with the FIM hand held terminal the air/fuel ratio changed as expected.  But with the changes converted to actual mapping with the additional memory board (too much to explain), the air/fuel ratio didn’t change in some places.
 
Sometime later, when I had software to look at the fuel and associated maps themselves, I noticed a couple of things that explained much of this.  And introduced me to a concept I now appreciate well: it’s what you don’t know you don’t know that brings you undone.
 
First up, the fuel map on the SP2/3/4 shows the first issue.  Some background: the SP was the hot one, with the flash “race” bits such as dual injectors with staged operation.  This feature in particular really peaked people’s interest and really helped build the folklore.  Unfortunately, it didn’t quite work as expected, both literally and mythically.  As introduced on the 1988 851 Strada and Kit, the black or red (IW042) dual injectors were used up to SP4.  The same injectors were used on the factory race bikes (Lucchinelli and Roche Replicas, etc), but with the fuel pressure raised to 5 bar from the std 3 bar.  Which made complete sense once I’d seen the fuel map.  Incidentally, the P7 ecu only has one fuel map, not main and offset like the P8 and later ecus.
 
Below is the fuel map from the 888 SP4 037 eprom, in hexadecimal form.  For those unfamiliar with hexadecimal, it is a number system with a base of 16, not 10.  Like decimal, it has 0 to 9, but then above 9 there is A to F.  In 8 bit hexadecimal, you get a pair of digits for each number giving a total of 256 numbers.  But, you need to remember that 0 is a number, so the range is 0 to 255.  Like decimal is base 10, with a range of 0 to 9.  So 00 is 0, 10 is 16 and FF is 255.  The fuel maps in P7, P8, 1.6M and 1.5M are all 8 bit, although the rpm values are 16 bit on all from memory, meaning they are made up of two pairs, such as 3F4A.  The maps in the later 5.9M and 5AM on are 16 bit as well, giving a maximum map number of FFFF, or 65535.  The red numbers at the RH side are degrees of throttle opening, the blue number along the bottom rpm x 100.

 
The main point here is that the top line (WOT) you can see FF at 8,000 rpm, along with FA at 7,000 and FC at 9,000.  This means that the biggest number available is being used.  However, to clarify a point, this is not the same thing as duty cycle.  Duty cycle is the time the injectors are open expressed as a % of the total time available to open them.  The dual injector throttle body report deals with this more.  Going from FA to FF is only a 2.4% change, so there’s not much more at all available over the range in which the torque peaks.

What FF means is that the software can’t have a longer pulse width on the fuel map.  In the case of the Weber systems, FF means 17 milliseconds.  Why 17ms I don’t know, but that’s what it is.  And each increment is 17/255 or 0.0667ms.

In the case of the dual injector P7 and P8 bikes, once the pulse width duration goes over a certain time as defined in the software, the duration is cut in half and the second injector is opened as well.  Meaning each injector can be opened for a maximum of 8.5ms.  If you have a look at the dual injector report you can see the implications and benefits of that.

But, the issue the FF brings in this case is that, with std mufflers, etc, the SP2/3/4 is already using the maximum amount of fuel per cycle at the torque peak.  Meaning that when you open it up and decide you need more fuel, you can’t have any.  So, by definition, any std dual (black or red) injector, A cammed 851 kit to SP4 will be lean around 7 to 9,000 rpm.  As an aside, the 851SP of 1989 ran single green IW031 injectors, and would be fine.

How do you get around this?  Well, there are two obvious ways, and two not so:

1.       Fit larger injectors.
2.       Increase fuel pressure (ala race kit).
3.       Offset environmental trims.
4.       Fit an add on unit like a PC3, Bazzaz, Dobeck.

Why option 1 was never acted on until the change of throttle body design with the 888 I don’t know.  It’s obvious, as it was from the start: the 851Kit fuel map has FE on it.  And they were fitting the larger IW031 green injectors to the 1989 851 and later 907.

Option 2 was the route taken for the 851 Kit, and the race prep notes mention the 5 bar fuel pressure regulator and possibly also the special (read ‘it’s expensive because it’s race’) fuel pump to knock it out for long periods.  The road bikes run 3 bar, increasing to 3.5 bar gives you 8% more fuel, 4 bar gives you 15.5%.  But, you don’t need the blanket change that a fuel pressure increase gives you, you just need more in a certain range.  Although for many, especially back in the day when tuning options and knowhow were very limited, bumping up the fuel pressure would have worked just fine.

Option 3 I can’t confirm, but from what Duane Mitchell explained to me some time ago it should work.  Those familiar with 851 are most likely aware of the 008 and 009 eproms.  008 is for open airbox lid and std mufflers with baffles removed.  009 is for all std.  The difference is 6%, and is achieved not by richening the fuel map, but by increasing the ambient air temp trims by 6%.  This blanket change is much the same as increasing the fuel pressure from 3 to 3.4 bar.  Not very precise, but the 008 does seem to work quite well.  Of course, if you set the idle mixture via the idle trimmer with the 008, you will reduce the lower throttle enrichment considerably.

The further implication of option 3 is that you can use the environmental trims to overcome the FF issue.  This is why adding fuel via the FIM hand held terminal worked on the dyno.  The hand held terminal’s input to the pulse width is added well down the calculation stream, and the 17ms restriction is only on the actual map number.  So I believe what was initially 17ms can be trimmed up to (theoretically) the maximum allowed by the cycle time once the pulse width calculations begin for any given cycle.

The graph below shows the air/fuel ratio for the SP4 I ran back in August 2001, the first time I had used the Dynojet air/fuel ratio set up.  And, in hindsight, there was a very obvious reason why it confused me enough to pull the sensor tube out of the muffler after my first attempt at putting changes into the additional memory board.  I wish I hadn’t now.  But, anyway, blue is the base fuelling on the eprom fitted to that bike (BMM01), which I now know had FF at 6,500, 7,000, 8,000 and 9,000 rpm.  Red is with +10 via the hand held terminal.  Clearly, it’s a consistent +10 change across the range, so the hand held terminal is able to do what an eprom can’t.

 
I haven’t tried this though.  I was going to, but the time it takes to play with this sort of stuff is time I just don’t have these days.
 
Option 4 is simply because there is more time available (much more), so even though the PC3, etc, can only be connected to one of the two injectors per cylinder, it will help.  The point is that not a lot more fuel is needed, but it’s enough to be a bit of an issue.
 
Moving on, the environmental trim tables throw up some more issues though.  The SP3 ambient air and engine temperature trim tables are show below.
 

While the engine temp trims are as expected, and with the engine being water cooled, relatively stable and inert during use, the ambient air trims are a problem.  You can see that at 17 degrees, the correction is +3.1%.  At 29 degrees, it’s -3.1% and at 41 degrees it’s -6.3%.  As it turns out, these corrections are chemically correct for the change in air density due to temperature.  But, as ever, it’s the theory versus application thing that brings us undone.

If the bike is stationary or moving at low speed and hot, a lot of radiated heat comes up into the airbox.  Where is the ambient air temp sensor on an 851 and 888?  In the airbox.  So it will read the radiated heat and act on it.  If you pull up at a set of lights on a 17 degree day, by the time you take off again a couple of minutes later the airbox temp may certainly have hit 41 degrees or higher.  In which case, your fuelling when you take off will be 9.4% leaner than it was when you pulled up.  Once the bike is moving the airbox temp will quickly drop again, but by that point you’ll have stumbled or stalled and grabbed a handful of throttle to get the thing going and before you know it you’ve got a reputation for not liking traffic.

The 008 and 009 eproms don’t have this variation in ambient air temp trims (they have 3.1%), so I don’t know why it shows up on the SP eproms.  I moved my 851 air temp sensor from the airbox to up behind the LHF indicator, which helped, but didn’t totally eliminate the issue when using an eprom with the SP based ambient air temp trim table.

And, finally, we get to applying all the above.  I had an SP4 in recently which was running a bit odd after a restoration and also stalling.  During the resto it have been found that two of the injectors were dodgy, but they had been replaced with the green IW031 injectors as found on the 1989 on 851 Strada, 907 and 750/900SSie, not the red (IW042).  I supplied another pair of greens to make them all green, and on the injector cleaning company’s test bench the reds flowed 40 (of whatever units) and the greens 65 from memory.  So the greens flow quite a lot more.

I also went through the valve clearances, tightening the closing clearances and reset the cam timing to 98/102 centrelines.

With all the above in mind, I made a revised eprom with changes to the fuel (guessed) and spark mapping and the ambient air temp trims and then ran it on the dyno and road and adjusted the fuelling as required.  The fuel changes weren’t a blanket offset, and there’s definitely less change on the WOT line from 6,500 to 8,000 rpm.  The changes to the fuel mapping are shown below.
 

The changes I made to the spark advance mapping weren’t just dyno power related.  The eproms for the Kit and SP motors have a lot of advance at low rpm and throttle to try to overcome the cam duration, up to 59 degrees at 2,500 rpm and above.  But at idle they drop to 10 degrees, and it’s just not enough.  Generally, the more advance you can run the more stable the idle will be.  I was only able to add 5 degrees without causing the idle to increase markedly.  The issue you run into is that as you increase spark advance you increase idle speed.  You can lower the idle speed again on these P7 ecu bikes by winding out the idle stop, but this does two things: it resets your zero throttle line, changing the fuel mapping at low throttle and also make the idle much dirtier emissions wise for Hydrocarbons and Oxygen.  The long duration A cams make this much worse, so you have to strike a balance.  Once I was done, you could walk up to it cold, hit the start button and it would fire and idle without throttle.  Well, once I’d replaced the rooted starter clutch and fitted some of the Motolectric leads anyway.
 
Another way to improve cold idle is to add advance under the idle rpm.  The SP spark advance map has rpm break points at 1,000 and 1,500 rpm.  If I’d had the time to do all I wanted (i.e., if I was going to get paid for it instead of writing it off) I would have added a 1,200 rpm break.  This way, to have 15 degrees at hot idle, the 1,200 rpm point would be 15 degrees, as would the 1,500 rpm point, giving a nice stable idle.  However, the 1,000 rpm point could have 20 or even 25 degrees advance.  This way, when cold or if the idle drops, it picks up more advance so picks the idle up again without running away with itself above 1,200 rpm when the engine is hot.
.

1 comment:

  1. This comment has been removed by a blog administrator.

    ReplyDelete