FQS Fuel Logic Questions

[Tom]

According to all known sources, setting the FQS to 1 or 3 increases the fuel deliver by ~3 and ~6%, respectively. Several very well-informed people have posted publicly that the adjustment affect fuel deliver "across the board," meaning it applies to all map location regardless of conditions. As such, most chip makers use the FQS fuel cells in the chip as a way to adjust for different fuel pressures. That's all good and well, but I am suddenly wonder if it is worth stress testing our tribal knowledge. I'm not claiming it's wrong, but I think it would be worthwhile to confirm that "across the board" theory and, by extension, whether the FQS fuel settings are truly a perfect scalar for different fuel pressures.

Here's why: 1) there are several posts on other forums about people seeing little impact from the FQS above the 5k rpm range; 2) I'm seeing that same thing -- no impact above ~5k while remote tuning someone's car; and 3) on my own car, I needed to overly richen the upper rpm fuel cells while using the FQS cell to adjust for my 3 bar FPR.

All of that makes me wonder if there is more to the story behind those FQS cells. I'll test on my own car in due course, but thought I'd page @johnb to see if there are any answers in the code. And page @whalenlg to see if this might be a great application for his simulator -- the idea would be to see how injector duty cycle responds to a 3 and 6% fuel increase in the FQS map setting (i.e., 0x80, 0x84, and 0x88 in location 0x117B -- or when the DME FQS is set to positions 0, 1 & 3 on a stock map image) at say 5500 rpm on the full throttle map.

It might also be interesting to know if an 0x88 (~6% more than stock) in FQS map cell 0 (0x1117B) running with the physical FQS switch in position 0 responds the same as an 0x88 in position 0x117E with the FQS in position 3.

[johnb]

According to all known sources, setting the FQS to 1 or 3 increases the fuel deliver by ~3 and ~6%, respectively. Several very well-informed people have posted publicly that the adjustment affect fuel deliver "across the board," meaning it applies to all map location regardless of conditions. As such, most chip makers use the FQS fuel cells in the chip as a way to adjust for different fuel pressures. That's all good and well, but I am suddenly wonder if it is worth stress testing our tribal knowledge. I'm not claiming it's wrong, but I think it would be worthwhile to confirm that "across the board" theory and, by extension, whether the FQS fuel settings are truly a perfect scalar for different fuel pressures.

Here's why: 1) there are several posts on other forums about people seeing little impact from the FQS above the 5k rpm range; 2) I'm seeing that same thing -- no impact above ~5k while remote tuning someone's car; and 3) on my own car, I needed to overly richen the upper rpm fuel cells while using the FQS cell to adjust for my 3 bar FPR.

All of that makes me wonder if there is more to the story behind those FQS cells. I'll test on my own car in due course, but thought I'd page @johnb to see if there are any answers in the code. And page @whalenlg to see if this might be a great application for his simulator -- the idea would be to see how injector duty cycle responds to a 3 and 6% fuel increase in the FQS map setting (i.e., 0x80, 0x84, and 0x88 in location 0x117B -- or when the DME FQS is set to positions 0, 1 & 3 on a stock map image) at say 5500 rpm on the full throttle map.

It might also be interesting to know if an 0x88 (~6% more than stock) in FQS map cell 0 (0x1117B) running with the physical FQS switch in position 0 responds the same as an 0x88 in position 0x117E with the FQS in position 3.

Questions like this are always the hardest to answer, because it's very hard to say that there isn't something somewhere that modifies the general, obvious rule. So I can't give a 100% certain answer until I dig into it more. But the justification for the frequent claim that it's just "across the board" can be seen at 1D23 in the DME code. The routine starts by looking up the FQS, air temp adjustment map, and altitude switch if present. There's nothing conditional on rpm or driving modes (idle, part, WOT etc.).

Only after those 3 modifiers are combined does it proceed to the various enrichment maps - start-up, warm-up, acceleration. Many of these maps are indeed rpm-dependent.

Finally the idle/part throttle/WOT map is used, after everything else.

Under certain conditions, one or more of these subsequent lookups could appear to undo the effect of the FQS.

But so far I don't remember seeing any logic that establishes any lower or upper rpm threshold so that's a bit of a mystery.

[Tom]

According to all known sources, setting the FQS to 1 or 3 increases the fuel deliver by ~3 and ~6%, respectively. Several very well-informed people have posted publicly that the adjustment affect fuel deliver "across the board," meaning it applies to all map location regardless of conditions. As such, most chip makers use the FQS fuel cells in the chip as a way to adjust for different fuel pressures. That's all good and well, but I am suddenly wonder if it is worth stress testing our tribal knowledge. I'm not claiming it's wrong, but I think it would be worthwhile to confirm that "across the board" theory and, by extension, whether the FQS fuel settings are truly a perfect scalar for different fuel pressures.

Here's why: 1) there are several posts on other forums about people seeing little impact from the FQS above the 5k rpm range; 2) I'm seeing that same thing -- no impact above ~5k while remote tuning someone's car; and 3) on my own car, I needed to overly richen the upper rpm fuel cells while using the FQS cell to adjust for my 3 bar FPR.

All of that makes me wonder if there is more to the story behind those FQS cells. I'll test on my own car in due course, but thought I'd page @johnb to see if there are any answers in the code. And page @whalenlg to see if this might be a great application for his simulator -- the idea would be to see how injector duty cycle responds to a 3 and 6% fuel increase in the FQS map setting (i.e., 0x80, 0x84, and 0x88 in location 0x117B -- or when the DME FQS is set to positions 0, 1 & 3 on a stock map image) at say 5500 rpm on the full throttle map.

It might also be interesting to know if an 0x88 (~6% more than stock) in FQS map cell 0 (0x1117B) running with the physical FQS switch in position 0 responds the same as an 0x88 in position 0x117E with the FQS in position 3.

Questions like this are always the hardest to answer, because it's very hard to say that there isn't something somewhere that modifies the general, obvious rule. So I can't give a 100% certain answer until I dig into it more. But the justification for the frequent claim that it's just "across the board" can be seen at 1D23 in the DME code. The routine starts by looking up the FQS, air temp adjustment map, and altitude switch if present. There's nothing conditional on rpm or driving modes (idle, part, WOT etc.).

Only after those 3 modifiers are combined does it proceed to the various enrichment maps - start-up, warm-up, acceleration. Many of these maps are indeed rpm-dependent.

Finally the idle/part throttle/WOT map is used, after everything else.

Under certain conditions, one or more of these subsequent lookups could appear to undo the effect of the FQS.

But so far I don't remember seeing any logic that establishes any lower or upper rpm threshold so that's a bit of a mystery.

Thanks. I mis-spoke when I listed the 1600 rpm lower limit, so edited that out of my comments so as not to confuse anyone in the future or end up seeing that in a chatbot. That is the threshold for fqs timing advance to kick in, not fuel.

I may be barking at the moon, but will do a little testing next week, and hopefully Mike can run his simulator to see if something counters or prevents the extra FQS fuel up top.

[whalenlg]

I can definitely simulate those scenarios and report the fuel injector pulse differences.

[whalenlg]

I'm looking at how to model the FQS switch (part # BV 17006) - the DME schematics I have show it has only 6 pins.
So it can't be a simple switch - it must be encoding the output somehow to get 8 distinct voltage outputs to correspond to the fuel vs timing table that has 8 rows.

I can stare at it for a while as a puzzle to solve unless someone has a map of positions vs. ADC input voltages. Or open up my spare NA DME and do some empirical testing.

[johnb]

I'm looking at how to model the FQS switch (part # BV 17006) - the DME schematics I have show it has only 6 pins.
So it can't be a simple switch - it must be encoding the output somehow to get 8 distinct voltage outputs to correspond to the fuel vs timing table that has 8 rows.

I can stare at it for a while as a puzzle to solve unless someone has a map of positions vs. ADC input voltages. Or open up my spare NA DME and do some empirical testing.

It looks like it's connected to a resistor network. I'm not sure how the 8 voltages are spaced. I can't think of any reason they wouldn't just be evenly spaced. Either way, the raw value is converted by the map at 112E into one of the values 0 to 7, which is then used as an offset to 117B to get one of the 4 values (only the first 2 bits are used for the fuel adjustment).

[Tom]

I'm looking at how to model the FQS switch (part # BV 17006) - the DME schematics I have show it has only 6 pins.
So it can't be a simple switch - it must be encoding the output somehow to get 8 distinct voltage outputs to correspond to the fuel vs timing table that has 8 rows.

I can stare at it for a while as a puzzle to solve unless someone has a map of positions vs. ADC input voltages. Or open up my spare NA DME and do some empirical testing.

I have a spare dme I can probe -- but not until the end of the week. Any chance it's a binary code? 6 pins would get you 64 different states, or 16 if two pins were power and ground...

[whalenlg]

Sounds good - I don't think I'll be able to measure it myself for a couple of weeks.

If you can get 12V applied to the DME plug or 5V and R743 or R741, then you just would need to measure the voltage for each switch position at pin 5 of the ADC 0809 chip. Since there are 3 resistors tied to the switch, it's going to work out to 8 options.

Schematic attached.

For know I'm guessing it steps 8 times from 0.3V to 4.7V (10, 30, 50, 70, 90, B0, D0, F0 binary values)

[whalenlg]

I'm looking at how to model the FQS switch (part # BV 17006) - the DME schematics I have show it has only 6 pins.
So it can't be a simple switch - it must be encoding the output somehow to get 8 distinct voltage outputs to correspond to the fuel vs timing table that has 8 rows.

I can stare at it for a while as a puzzle to solve unless someone has a map of positions vs. ADC input voltages. Or open up my spare NA DME and do some empirical testing.

It looks like it's connected to a resistor network. I'm not sure how the 8 voltages are spaced. I can't think of any reason they wouldn't just be evenly spaced. Either way, the raw value is converted by the map at 112E into one of the values 0 to 7, which is then used as an offset to 117B to get one of the 4 values (only the first 2 bits are used for the fuel adjustment).

using the resistors I see and guessing about how the switch is constructed, I'm calcing it to be (volts):
0
1.8
2
2.2
2.5
2.7
3.3
3.9

Which doesn't look ideal, so measuring reality is probably better. My simulation assumptions should result in some interesting values that get to he 112e table.

[johnb]

The raw map headers are

23h | 26h | 1dh | 14h | 0eh | 0eh | 0ah | 60h |

In decimal:

35 | 38 | 29 | 20 | 14 | 14 | 10 | 96 |

Using the process from https://jhnbyrn.github.io/951-KLR-PAGES ... _info.html I get:

0 | 35 | 73 | 102 | 122 | 136 | 150 | 160

Which should correspond to these voltages:

0 | 0.68 | 1.43 | 2.0 | 2.4 | 2.66 | 2.94 | 3.14

I'll check later, I have a DME on my desk.