FYI Went to Dynajets website and in the FAQ section they state that it is possible to keep your O2 sensors connected with the PCIII installed if fuel economy is a factor, Has anyone tried this?

 

Glens on the Touring forum tried it, but I think reinstalled the eliminators after a short test. The PC site says:

On bikes that use an O2 sensor in the stock exhaust have a closed loop area that can not be adjusted by the Power Commander alone. If you make adjustments with the PCIII in the closed loop area the O2 sensor will sense these changes and the ECM will alter the fuel curve accordingly. If you want to have full fuel control over the entire RPM range of your bike then you will need to bypass the closed loop area which the O2 eliminators will accomplish. If you are looking for the best fuel economy and not worried about making changed to the closed loop area you can leave the stock O2 sensors connected and still make changes outside of the closed loop area with the PCIII.

The problem is determining what areas are affected by closed-loop, which they don't mention. PC tech support once told me it went as high as 50% TP, so I don't know how you would ensure the PCIII isn't interfering in closed-loop mode without more specific instructions. I know that when I lean 20% or higher too much, I get a flat spot at roll-on. I would also assume that once the unit is shut down it reverts to the default map, so any "learning" that takes place would be temporary. I would consider doing this, but 14.7:1 is very lean and I probably would not want to run it there. I would rather find the right A/F for cruise (probably around 14.0-14.2:1) and adjust the PCIII accordingly with the eliminators connected.

The problem is determining A/F. The Wideband Commander will do this, as well as other A/F gauge kits, but all that I've seen are expensive. Using a wide-band O2 sensor you could measure the voltage and determine A/F based on existing conversion tables. I've thought about doing this if I could find a small LED display or analog gauge that could be mounted somewhere on the bike. As a less-desirable alternative, I could connect a conventional voltmeter and carry it around temporarily for testing certain areas of the map.






[/align]

 

OK I guess without knowing just exactly where the o2s take over it will be best to just leave eliminators in place. If we knew for sure that they would take over only at idle or light cruise they would be good at 14.7...

 

The reason for the test was to verify suspicion one or both injectors where slightly fouled [bad gas]. Hypothesized that the ECU would, using feedback from the sensors, even the cylinders up a at least a little and smooth the engine. It did make a difference that way. Once the tank-full of mostly good gas with a lot of cleaner in it passed through everything's okay again.

Next time I'm out riding for a whole day I'll try it again to see if/how it impacts fuel economy.

$289 for a wide-band sensor, the necessary controller, and a nice, pretty gauge. The controller is the same unit than can have an analog output tweaked to fake that it's a narrow-band sensor with the "14.7" being any desired AFR. I'm wondering if it would be possible to use a couple of resistors for channel isolation (if they're even required) to feed both front and rear inputs to the Delphi from one unit.

I've got to put some more thought into that last statement. What we'd want would be to have the resultant injector duty-cycle from the "faked-out" input to be mirrored on the other cylinder. I'll get back when I've fleshed it out a little more...

 

Lots of good information and fixes at this link, including how to cool engine without losing a lot of gas miliege.:

http://www.nightrider.com/biketech/h...erformance.htm

 

I wouldn't mind trying that too, and assume that any learning that takes place will be lost after shutting the motor down. The problem is that stoich is just too lean for these engines, but if we could get it down to 14.2:1 or so....

You lost me. What's the result of that?

As for the kit above, it sounds good, but where would we put the gauge, and how big is it?

If you're contemplating fooling Dephi into producing less than 14.7:1, I don't think that can be done. That must be changed by proprietary software, like SERT. If it could be done, it would make it very feasible to reconnect the O2 sensors.

 

It's easy to get it to 14.2 with a voltage-divider network. Two resistors per cylinder is all it takes. I'd rather get a full set of both male and female connectors for the sensors before I do it, so I can just plug them in without breaking into the wiring harness.

I'd bet good money the Delphi "don't remember" the changes. It doesn't have "auto-tune" and "rememberin' the changes" is really all "auto-tune" is.

I was thinking (obviously not loud enough) that one wide-band sensor with it's (necessary) controller (which has an "adjustable" narrow-band-compatible output), as in the linked kit(s), might be put-into-service-able for both cylinders' worth of Delphi-feeding. Maybe that wasn't much clearer of an expressed thought...

Dobeck's site(s) are woefully sucky in many respects. The gauge at the bottom of the page is referenced as being 52mm so I'd assume the unspecified one is probably the same size.

I'm thinking some double-sided tape or taped velcro strips and it would ride on the dash quite nicely. But you've got fairing-dash space that I don't, so your options would be better.

Again, a couple of 25¢ resistors per cylinder would get to 14.2:1.

The wideband controller is adjustable on an analog output to mimic any reasonable value you want as being "14.7:1". But they're a couple hundred bux apiece (with sensors). Add $400 to a PC-111 and you're getting right close to a Twin-Tech or Thundermax...





[/quote]

 

I should flesh that out a little. The narrow-band sensors have a range of just under one volt DC at the "fat" end of the range and pretty much zero volts DC at the "lean" end. If you use a voltage-dividing network of appropriately-sized (in resistance values) resistors you can drop the almost-one-volt to the half-volt the sensor produces at 14.7:1. That way, the ECU will enrich the mixture until the sensor is putting out almost one volt, which is decreased by the resistor network to a half volt, and the ECU is "happy" in its ignorance.

You really can't get very close to even 14.0:1, though, because the output of the sensor isn't linear. Once the sensor gets to a saturated value (anything lower than about 14.1:1 is the almost-one-volt) the output will be at a steady voltage which could represent anything below ~14:1, or maybe 14:1, who can say? That's why you've got to keep it within the (more) linear portion of the sensor output, so it'll actually represent something tangible.

This is basically what the "bias tables" do in SERT. I'm quite sure you'll find that with them you're limited to about 14.2:1. I think in "reality" what that function is for is to correct a misbehaving O2 sensor, but we all know what it usually gets used for...

I've had a couple beers so if this isn't making sense ask again and maybe I'll do better on a less-cloudy brain.

 

I'm missing something here. I thought the O2 sensors were mostly switches that told the ECM whether or not to go into closed-loop mode, and the A/F for closed-loop is imbedded in the innards of the ECM. Only SERT or other proprietary software can induce it to change. If that's the case, then how can resistors placed in-line between the O2's and ECM make any change in A/F?

The last PC tech I spoke with said it had limited learning ability, and could alter its own map by ~5% either way.

I've been looking for some handlebar-mounts for 52mm gauges, but have yet to find any. This size gauge is really too big for a MC anyway, IMO.

I looked at the LC-1 on the Nightrider site--which they claim can work with PCIII, etc.--but I'm having trouble figuring out how it works based on their description. It's <$200 and sounds like it might be a solution.

 

I didn't look just now but think I'd explained that. I'll try it again anyway.

The O2 sensor produces DC voltage based on the amount of O2 present at its tip. The voltage is centered at about 0.5 VDC for 14.7:1. When there's less O2 the voltage output heads toward 1.0 VDC and when there's more, it heads toward 0 VDC. Hook your DVM up to the output of one once and watch it work.

By using a couple of resistors with appropriate values you can decrease the voltage output in a proportional manner. This will bias the output of the sensor, with the effect that it will be reporting a higher AFR than what it should be reporting. The ECU will say "Whoa, let's add some fuel to get that back down." Net result will be a lower AFR in truth but the ECU will be happy with it.

I know I discussed this earlier. Look at that again. Just remember that no matter what you're letting out from the sensor, it can only actually report to about 14.0:1 so don't go thinking you can turn a narrow-band sensor into a wide-band one, or alter the actual range under which the NB will operate.

I don't doubt that for a minute. I've experience it first hand with my seat-of-pants dyno and gas sniffers. What I'm saying, however, is that the alteration to the map is transitory. It only remains in effect for the session during which the computer is booted. It doesn't have a system of retaining that information across sessions. If you can determine authoritatively that the case is otherwise I'd really appreciate a pointer to the information.

Agreed. But not too terrible for temporary usage.

That very LC-1 is the one which comes in that kit with the gauge. The way it works is, aside from having the necessary controller which makes the wide-band sensor function, it has an extra 0-1 VDC output for feeding the ECU in the place where the stock one would be feeding it. You can bias the 0.5 VDC "14.7:1 AFR" output to any actual AFR within range of the device. Instead of 0 VDC being, say 15.4:1 and 1 VDC being, say 14.0:1, you could set it for 0 VCD output at 14.0:1 and 1 VDC at 12.6:1. The ECU will never know the difference. It will "think" it's targeting 14.7:1 all the while achieving 13.3:1 (as would be the case for this example).

Naturally, you'd need to use something like SERT or PC-III to cover the open-loop areas of operation. Also, note that the duty cycle of the injectors will be out of range from what's programmed into the (stock) Delphi. Although it will accommodate the changes (so long as they're within the ±5%, as you report) it will take a few moments to "learn" that offset the next time you boot up the ECU (again, as I currently understand the situation).

But at a couple hundred per, the LC-1 units are really not the best solution IMO. Why not just get a TMAT or DTT for about the same amount of money (as adding two LC-1s to a SERT, or even PC-III) and be able to go closed-loop everywhere?

If it were possible (and I haven't pondered it further yet) to feed both Delphi O2 channels from just one LC-1 then a couple hundr