NTK L2H2 New Linearisation

[pat]

Tim,

NTK manufacture a lot of sensors that the general public don't know about or get to see. The best ones are laboratory grade and last I enquired Horiba had an exclusive on those, ie you cannot buy one from NTK and use it in a lab grade application.... but then considering the price you probably wouldn't want to anyway There are some lesser sensors that are available for around UKP 350 plus the VAT, which Horiba did use in their lower end kit, and which you could get from NTK so long as you were not going to use it in a lab-type environment... but there was a requirement to prove that you were using it for "something else", in-car lambda control in motorsport being an acceptable "something else". Then there are the mass produced sensors like the LZA09-E1 which are OEM / pattern part and have no restriction on application, but they are no match for the lab grade ones so Horiba didn't care: if you wanted lab grade measurements then you couldn't do it with an LZA09-E1 anyway so you were back to square one - buy from Horiba.

The motorsport ones I have looked at / used are different to the L1H1 and L2H2 both in construction and markings. I've not come across them through any of the motor factor type sources. We can supply them, but don't tend to hold stock because we don't get asked for them that frequently (not surprising considering the cost!).

Cheers,

Pat.

[TimH]

So, which one was the S6 designed to use?

[pat]

The one that's about 350 quid and not available through the usual channels

[Benkku]

Purpose of the calibration resistor is to 'tell' to the circuitry/electronics, what is the ion cell current at the free air condition. Pump cell current level is different between these sensors. Altought pump cell current does not differentiate nor similarize sensors, when this resistance 'information' is not used in a first place. So the idea is/was to use sensors which produce similar pump cell current - or sensors which were/are calibrated to produce similar pump cell current through the resistor.

What comes to sensors accuracy, 'ordinary' sensors (ie. other than laboratory or 'motorsport' grade), fall typically inside 1-3 percent at lambda 1. Rich side is bit more accurate, typically 2 percent difference at most, because the pump cell current slope is more steeper as a function of the lambda.

But... There are caveats.
All of these sensors, have an temperature and pressure dependency, which is far more greater, than 'basic accuracy' between 'ordinary' and/or laboratory/so called 'motorsport' sensor grades. Pressure dependency is main affecting factor for the lesser accuracy - be it 'ordinary', laboratory or 'motorsport' grade sensor. Reason not to mount the sensor upstream of the turbocharger, in forced induction applications for example.

In my opinion 'ordinary' (read - monetary wise cheap) NTK sensors, are suitable/useable, when calibrated - ie. sensor look up table voltage values defined.

Most used 'motorsport' sensor seems to be LZA03-E1-P. I do not know, if this is the sensor which Syvecs ECU's are designed to use or not. But all sensors could be 'similarized', by adjusting the pump cell current through the resistor, to the same value as the 'design' sensor puts out at the certain oxygen level (be it free air or some other percentage).

[CT9A]

Is the NTK output linear like the bosch sensor? So I could do a 2 point calibration on the dyno, one from 10 afr and one from 18 afr and interpolate in the software correct?

[TimH]

NTK's don't give a linear output - nor do the Bosch sensors (unless pre-linearised to a 0-5V output with an external box).

[CT9A]

Tim I bought the sensor from you, the linearisation found in the software is way off what the dyno wideband reads. What do you suggest? Take measurements from 10-17AFR in steps of 1 afr and adjust?
How do others calibrate this with the LZA sensor?

[TimH]

Which linearisation are you using? I see that there are now 3 that could be the right one!!

[stevieturbo]

Tim I bought the sensor from you, the linearisation found in the software is way off what the dyno wideband reads. What do you suggest? Take measurements from 10-17AFR in steps of 1 afr and adjust?
How do others calibrate this with the LZA sensor?

get a 3rd opinion ? lol

What wideband does the dyno use ? and where is the sensor located ?

[CT9A]

Dyno uses a NTK wideband sensor, I think it is the same as a AFR500, it was measured at the back, I know not the best place to get accurate readings, especially at low airflow.

I welded another bung next to the LZA09 and mounted a older LSU4.2 sensor, I calibrated the wideband in the syvecs to read almost the same. It was off by 0.5 AFR. Still think there must be a better way to calibrate the LZA09. I will try again soon with a friends Autronic model B wideband. That one has a lab grade NTK sensor.