Voltage Mk2

So, the bits arrived.

zx10r 2006 regulator is a FH010. The specification on this isn't clear but most people believe that these can generate 40-50amp. The great thing about these regulators is that instead of shunting spare volts to ground (and consequently wasting the power, and getting very hot in the process) I believe they switch the higher voltages on and off at a high frequency and the rapidly switched voltage looks like a smooth lower voltage. I think this is how it works (this is the same as a switched mode amplifier).

Anyway, on bikes that use alternators that include moving batteries the regulator/rectifier is an essential part. My alternator is more like a car one, the rotor carries a current (3 Amps, driven by the battery which I didn't include in my previous calculation) rather than magnets. The regulator in mine alters the rotor current so as to keep the voltage at 12v. This all makes good sense, but as noticed before it doesn't work so well for me.

So, I have removed the old regulator, and the old rectifier, wired up the rotor to 12v (non regulated), and fed the 3 phases of the alternator into the new FH010 regulator.

This is a new idea, but the logic is;

My problem isn't really different to the case with magnets, and these regulators work then.

This regulator should pump more current than the old one, as it can use the higher voltages from the alternator.

The old one didn't work, so hopefully this one will!

So far, it seems to re-charge the battery way quicker than the old alternator. At idle the battery charges EXCEPT when the lights and fans are both on. I tried dropping down to 1 fan (or 2 in series) and these were insufficiently powerful to cool the engine (it is 25deg C here at the moment, but that is no excuse).

More as it happens. The regulator runs nice and cool. I think this might turn out to be a great modification.

Secondary injectors working (sort of)

So, after all the wiring and plumbing the Secondary injection hardware was all in place.

I could have run this the other night, but I was a bit edgy about it all, so I left it for the weekend.

First run, much trepidation, set it to come in at 7.5k rpm. Drove along, and at 7.5k is just coughed and grumbled and coughed (note settings are 7.5k rpm, and manifold pressure (MAP) has to be above 80kPa). It felt like it was lean (rich was a possibility, but unlikely). Looking at the logs is turns out that the MS thinks everything is working, but the AFR (lambda probe) makes it clear that these is not enough fuel.

Tested polarity on injectors (doesn't seem to matter), tested injectors (fine), tested continuity from transistor in MS and the injector leads (FAIL). Basically I had wired the 37pin connector wrongly (dohhhh).

Test 2: drove out, and logged. I also put a video camera under the bonnet looking in through the window of the airbox. The idea is to see if the injectors are all working, and to see if the fuel is going in the right direction. Anyway it recorded for 10mins, and then shut itself down (damn just before the interesting bit!). Here is a log-file.

This log shows up a few things:

1) injector switch is working

2) injector switch doesn't happen over 25 cycles (as requested) but is instantaneous (this is a feature of the hi-res code, which I am investigating)

3) After both injectors come on-line the AFR jumps up by around 1 to 1.5units.

4) I need more fuel up the rev range (actually dangerously lean at 10k rpm).

re 3: I have told the MS that the primary injectors are 270cc/min (this seems to be a well-known number) and the secondaries are 294cc/min (cbr1000rr, number from internet), but there must be a problem. Could be:

a) injector values are not right

b) injector opening times are wrong (this comes into this in a complicated way)

c) the fuel isn't getting into the engine from the secondaries

Not sure what to do, but I will put VE table mapping points just before and just after the switch-over point and then the MLV tuning function can sort it out for me (cunning plan I thought).

The problem with fueling at the top end hopefully is because secondary injection really does buy power by pulling in more air, and causes good mixing. It may (on the other hand) be due to (a,b,c) or the fuel pump may have run out of capability.

All good fun.

Charging

One problem with putting fuel injection into a car with an old bike engine is that the fuel injection and the car all draw more current than the bike needed, and hence the alternator can struggle to deliver adequate current.

For example, the alternator in the GSXR1100 normally drives:

1 fan (5A)

1 cdi (low current draw, 4A)

and some lights including

1 headlight 5A

1 tail light 0.5A

In the car this increases to;

2 fans (10A)

2 headlights 10A

2 tail lights 1.0A

1 fuel pump 3A

Going fuel injected changes this further to:

1 Lambda probe with heater (3A)

1 set of fuel injectors (3A)

1 ECU unit 1A

1 ignition drivers (3A)

So, to cut a long story short. With lights on even at motorway speeds the result is a battery that is flattened. This doesn't stop the car, but means it won't restart.

A solution needed finding.

I swapped the lights to HID versions which draw less power, and changed lights to LED versions. So that is a start, but its not yet enough. After a drive the fans flatten the battery, so I need to do something about both the fans (probably try them in series instead of parallel, which will drop the current draw to 2.5A from 10A).

After that the only thing left is to improve the alternator. It is a ND (NipponDenso) 25Amp unit (I believe) which works like a car one. Basically the rotor current is regulated, and the outputs from the 3 windings are rectified. This is probably a good design, but by using a regulator/rectifier from a later model bike (in this case the FH012 unit, as promoted on the very impressive http://www.triumphrat.net site (see this posting for more help with this http://www.triumphrat.net/speed-triple-forum/104504-charging-system-diagnostics-rectifier-regulator-upgrade.html). It isn't clear whether this will work, as their modifications are for bikes with magnet based alternators, but its got to be worth a try. This RR unit seems to be used on the ZX10r (2006 model) which is where mine came from.

Air temperature problems

I noticed that the VE values (amount of fuel required for correct air-fuel-ratio) is different at night versus the day time. There are two possible reasons for this: voltage (lights on at night), or temperature. Generally when the weather is warm (extreme was the Abingdon track day where air inlet got to 52degC) the engine runs lean. This is the opposite of what might be expected and indicates that the ECU is correcting for temperature but over-doing it.

The view is that I have a problem with heat-soak into the sensor. Basically instead of the sensor measuring the air temperature it measures the air-box temperature. This isn't aided by the fact that my airbox is a chunk of aluminium behind the radiator. So, when the air-speed is high the air is much cooler than expected (and so denser) and hence the AFR is too high (which could cause engine damage if not sorted).

The gsxr600 sensor (from 2001-2003) is a solid brass affair which takes ages to warm/cool. This is not the best, and more recent sensors (especially for turboed engines) are designed to have low thermal mass.

I made a sensor based on a gsxr600 mechanicals, and maplin thermistor and some epoxy filler.

I recalibrated the ECU for this new sensor (resistance values in photo) and it seems to work fine. Note: don't try to upload to the Megasquirt unless your battery is putting out a decent number of volts.

I am keen to get tuning again and hopefully the AFR will be nice and stable.

RAM air seen

I have connected up the air inlet, and finally got around to trying the car out.

After initial messing around with TPS (different range on this one than my old one so) I had to re-write the tables.

Starts on new TBI's and runs. Very exciting. Ran around a bit collecting logging data. Updated tables and repeated a couple of times. I had done this already, but it hadn't been running so well. This time around I had changed the inlet rubbers from the short ones (the same as 4x of the 1&4 ones on a gsxr750) to the long ones from that same engine.

A few interesting results. Firstly, now the air hose is more rigid and the trumpets longer the VE values are a lot higher (maybe 10%). Not sure which is the source of this, but who cares.

The interesting detail in the plot above is in the MAP (manifold air pressure) in 2nd 3rd and then 4th. Note that as the speed increases the MAP increases (at the same RPM). The pressure increase is bugger-all, but that is the problem with RAM air, you don't get much at normal speeds. At least the little that is available is working for me.

New Throttle bodies in

Very excited, fitted new throttle bodies and air box, and went for a start-up.

It started first time (a bit of a first for me), and seemed to run on the old map reasonably well.

TPS settings were different from previous TBI's (that was a custom mount so as to get the ideal TPS range, which I haven't done with these ones).

Stuttering when accelerating at low RPM, probably need to balance the carbs (I only did a static balance, i.e. put a drillbit in the gap and adjust so they all fall out at the same time).

Also the air hose inlet was flopping around at the front of the car, so that needed sorting.

So I built an inlet ram pipe thing. From a Maplin woofer trumpet (£2.69) and a bit of bent ally. O.D. on this is just less than 3inch, so perfect fit to the 3inch hose.

Not much of a photo (taken with the built-in webcam on the netbook).

Now I need to rivet this on, and connect it up. This is slightly messy as the hinging bonnet makes this a bit messy as the hose changes length as the bonnet opens, but it should be alright.