Defi and Syvecs 4bar disagreement!

[TimH]

I don't know that sensor, it's just that all that I have come across so far are nominally 0bar = 0.5V and full scale= 4.5V.

If I pull up SCal and use the Syvecs 4 Bar MAP linearisation, 4.5V is 3.5Bar, and 4 bar doesn't really exist...not what I would call a 4 bar MAP sensor!

[RICE RACING]

Thanks All - If the syvecs sensor is indeed OK (unfortunately I don't recall ever seeing a separate calibration sheet) then then the linearisation from the db must be out I guess - which I'm somewhat surprised by - I do note that there are two (or three) hint points that load up with the curve (straight line of course) - which also seemed a little odd.

I'll have to figure out a 'cost effective' method of calibrating manually - did have a look at the Fluke 700G - nice bit of kit no doubt - little out of my price range!!!

Funny thing is my Mightyvac gauge was not accurate and even less so after I tried to pull it apart and 'fix it' so I was looking to adapt a good replacement to it, well all of the 'good' mech ones I had here all showed different readings and I just cracked it and said I need a good stand alone one of certified instrument quality. Shopped around and got this one brand new for $250 (worth every cent now I have it) though I am pretty sure it was stolen? as the prices for them seem to be a fair bit more than that.

[stevieturbo]

As default high (4.5v) is 3.5bar?

Could be...best to test it

It is the only way to be sure

[dr_jones]

Cheers chaps - I will test...

[dr_jones]

Got some time to test / calibrate today.

Given that my leakdown gauge and DEFI pretty much match I'm going with them as being 'correct'

For anyone interested I've attached the cal containing the new data:

Syvecs 4bar (DEFI) is just the syvecs calibration against the measured (as accurately as I am able) data points.
Syvecs 4bar (DEFI CAL) is the sensor with its new curve interpolated through the measured data points.

Interestingly it looks awfully close now to the JTI 4bar - no doubt coincidence.

I've not started using this new sensor definition yet however as it means I need to adjust my fuelling table to suit - right now I've not quite got my head round how best to do that.

[TimH]

The CAL doesn't include any sensor linearisation files you've saved out to the sensor DB, and only shows the currently applied one.

But if the Syvecs 4 bar really is a 4 bar - and it ought to be!! - then I am not surprised if the linearisation is the same as a KA (JTi) one. As said, every pressure sensor I have come across has been 0.5V = 0 bar and 4.5V=full range. They are all "ratiometric" types.

[stevieturbo]

Got some time to test / calibrate today.

Given that my leakdown gauge and DEFI pretty much match I'm going with them as being 'correct'

For anyone interested I've attached the cal containing the new data:

Syvecs 4bar (DEFI) is just the syvecs calibration against the measured (as accurately as I am able) data points.
Syvecs 4bar (DEFI CAL) is the sensor with its new curve interpolated through the measured data points.

Interestingly it looks awfully close now to the JTI 4bar - no doubt coincidence.

I've not started using this new sensor definition yet however as it means I need to adjust my fuelling table to suit - right now I've not quite got my head round how best to do that.

So what you seem to be saying was the old sensor cal was incorrect ?

You'll need to adjust the tables manually based around previous voltage/map vs new.

ie, you'll have to compare old vs new and simply copy each row from the old table and apply it to the correct pressure row in the new table.

ie if old one had said 2000mbar at say 2v but you know know this was incorrect and was really 2500mbar....copy that row and apply it to the 2500mbar row in the new tables.
It will take a little time but should work ok for most tables.

[dr_jones]

The CAL doesn't include any sensor linearisation files you've saved out to the sensor DB, and only shows the currently applied one.

But if the Syvecs 4 bar really is a 4 bar - and it ought to be!! - then I am not surprised if the linearisation is the same as a KA (JTi) one. As said, every pressure sensor I have come across has been 0.5V = 0 bar and 4.5V=full range. They are all "ratiometric" types.

My mistake - have attached the relevant files.

Interestingly - although my curve fits the KA @ 4.5v (well it extrapolates to roughly the same point) the lower end differs - reviewing the logs shows readings as low as 189 which would mean a sensor voltage of less that 0.15v...

[dr_jones]

So what you seem to be saying was the old sensor cal was incorrect ?

You'll need to adjust the tables manually based around previous voltage/map vs new.

ie, you'll have to compare old vs new and simply copy each row from the old table and apply it to the correct pressure row in the new table.

ie if old one had said 2000mbar at say 2v but you know know this was incorrect and was really 2500mbar....copy that row and apply it to the 2500mbar row in the new tables.
It will take a little time but should work ok for most tables.

Yes precisely - I seems the sensor cal there in the Syvecs Suite does not match my Syvecs supplied 4 bar sensor. Somewhat surprising.

I think that approach may be the only way to go - the gradients of the two sensor cals are slightly different so as boost rises the reading error increases - which *I think* makes it less viable to just offset the fuel cal (as you may do with a change in injector size for example).

[TimH]

FYI the KA sensors have an accuracy (linearity) of .5% of full scale. A 4 bar MAP sensor will be accurate to 20mBar across the range. They have a calibration sheet that sets out the offset from 0.5V for 0, and the range.

This is why I have settled on these as the ONLY pressure sensors we offer. You can choose to calibrate each one using a pressure gauge but in most cases I would not think it was necessary. If you replace a KA with another, the error in pressure will be small enough that normal ECU corrections will sort it in all bu the most extreme of tunes. In my opinion!

I find it difficult to believe that the Syvecs one is non-linear, and that there isn't an accurate linearisation provided. Maybe one for Syvecs Tech Support to comment on?