Safe To Use ECU Connected Analog Sensors?

All sensor specific discussions! Temperature, pressure, steering angle, brake and throttle, etc. Post adaptations of OEM-style sensors, and also your clever DIY hacks and custom designs here too!

Moderators: JeffC, rdoherty, stieg

Post Reply
kbuckham
Posts: 8
Joined: Tue Feb 12, 2013 6:45 pm
Location: Victoria, BC, Canada

Safe To Use ECU Connected Analog Sensors?

Post by kbuckham »

I have several sensors connected to my car's ECU like MAP (manifold pressure), TPS (throttle position), AIT (manifold air temperature) that I would like to connect to Race Capture Pro. The ECU already supplies a 5V sensor reference line that all of these sensors use and of course these sensors are necessary for correct operation of the motor. How safe is it to tap these lines directly to also feed the Race Capture Pro? (I'm thinking about impedance issues or potential voltage offsets that could be introduced to the analog values seen by the ECU.)

brentp
Site Admin
Posts: 6282
Joined: Wed Jan 24, 2007 6:36 am

Post by brentp »

Hi kbuckham,

RaceCapture/Pro has an input impedance of 15.1K due to an internal voltage divider as part of the input protection / buffering / interface circuitry. Depending on the impedance of your sensors it may or may not interfere with your ECU. There's no voltage bias or internal pullup on the Analog inputs. You could probably do a test and tap in to one of the ECU sensors and see if that affects the sensor output.

You could directly measure the voltage at the sensor with and without RaceCapture/Pro connected, and see if the signal changes significantly.

Let us know what you find out!
Brent Picasso
CEO and Founder, Autosport Labs
Facebook | Twitter

kbuckham
Posts: 8
Joined: Tue Feb 12, 2013 6:45 pm
Location: Victoria, BC, Canada

Post by kbuckham »

The 15.1K ohm input impedance is considered low by the ECM vendor so I'll be testing all of the inputs. I think I'll build up some buffer circuits for the inputs using a TLV2375IN (http://ca.mouser.com/ProductDetail/Texa ... fWMBL0g%3d) as I'm now pretty sure I will see some voltage offset. Input impedances from 3K-10K ohm cause issues with several MAP and wideband sensors. (Most ECM input impedance values are normally close to a couple hundred kilo-ohm.)

When the expansion board comes it would be great to have even higher impedance inputs.

brentp
Site Admin
Posts: 6282
Joined: Wed Jan 24, 2007 6:36 am

Post by brentp »

thanks for the update.

If it's an active sensor that actually provide a current source (e.g. many MAP sensors like a MAP sensor or a wideband module) there's a good change there will be no effect providing the sensor can source enough current. 5v at 15.1K is less than 1mA, so it *should* be no problem.

tapping into resistive sensors like thermistors can certainly throw things off unless input impedance is a significantly higher - like 10 or 100x the source impedance.

There's an outside chance we could update RaceCapture/Pro to have an input impedance of 151K on the analog channels - we'd have to make sure that doesn't detrimentally affect the sample time on the processor.

please keep us updated on your findings!

dimondjack
Posts: 101
Joined: Tue Jan 15, 2013 1:37 pm

Post by dimondjack »

Could you add a resistor between the source and the racecapture pro? You would create a voltage divider that would result in an altered voltage but so long as you could scale the signal appropriately it wouldn't matter right?

For example, another 15.1k resistor would halve the voltage but you would have a total impedance of 30.2k. A 45k resistor would put a 0.25 multiplier on whatever you are measuring and would put the load impedeance all the way up to 60k, in the territory where most sensors start to not care. So long as you make sure that you have enough resolution in the measurement range I'd think that dividing the voltage like this would work ok.

An alternative is to whip up a high impedance opto-isolator circuit. I'm probably going to do that to make sure I never burn out the RCP with an overvoltage.

brentp
Site Admin
Posts: 6282
Joined: Wed Jan 24, 2007 6:36 am

Post by brentp »

dimondjack,

yes you could do that. it would affect the input range, but as you said, you can calibrate around that at the expense of resolution.

You wouldn't want to put an optoisolator on the analog inputs because you won't get the linear response necessary. Note, the inputs are protected to 40v. The only thing you don't want to connect to the inputs is the tach signal connected to the ignition coil as this exhibits 500v spikes. Logic level tach signals sourced from the ECU is fine, of course. in either case, you need to check the wiring of your car to know for sure.

For interfacing to the tach signal source from the ignition coil we deigned an opto-isolator board, designed to be installed between the high voltage tach signal and pulse input of RaceCapture/Pro.

https://github.com/autosportlabs/CoilX

@kbuckham we use the TLV274 in RaceCapture/Pro on the analog output section- very similar part to what you linked. It would also be a good choice for very high input impedance buffers.

kbuckham
Posts: 8
Joined: Tue Feb 12, 2013 6:45 pm
Location: Victoria, BC, Canada

Post by kbuckham »

Thanks, I'd love to say it was my idea but this is where I stole if from: http://www.dsmlink.com/wiki/ecuinputdata

brentp
Site Admin
Posts: 6282
Joined: Wed Jan 24, 2007 6:36 am

Post by brentp »

Recently whipped up this:

https://github.com/autosportlabs/analogBuffer

I am ordering a few test boards. If anyone wants one reply here (I will have 5 to spare)
Brent Picasso
CEO and Founder, Autosport Labs
Facebook | Twitter

kbuckham
Posts: 8
Joined: Tue Feb 12, 2013 6:45 pm
Location: Victoria, BC, Canada

Post by kbuckham »

Yes, please! That was fast! I just ordered two of the op-amps yesterday, but your circuit is much nicer than mine was going to be.

brentp
Site Admin
Posts: 6282
Joined: Wed Jan 24, 2007 6:36 am

Post by brentp »

cool, email me your address (again) to brent (at) autosportlabs.com. Note these won't be made as fast as your parts show up, so you may still want to wire in your version until I get one sent to you.
Brent Picasso
CEO and Founder, Autosport Labs
Facebook | Twitter

jakekooser
Posts: 64
Joined: Wed Apr 16, 2014 11:03 am
Location: Florida
Contact:

Post by jakekooser »

Are there any of these analog buffers left by any chance?

MichaelDrost
Posts: 30
Joined: Tue Aug 13, 2013 12:15 pm
Location: Amersfoort, the Netherlands

Post by MichaelDrost »

indeed yes, are there any left?
Might as well combine with the order I was planning for the CoilX x3 and expansion.

Thanks in advance for your efforts ...

brentp
Site Admin
Posts: 6282
Joined: Wed Jan 24, 2007 6:36 am

Post by brentp »

Hello,

So far we have made a sample set of buffers for testing purposes and haven't scheduled a formal run of adapter boards. We can look into it and make it available for a reasonable price. The form factor will be similar to the CoiX module / Tach adapter modules we have now.

What we need to do is make a few tweaks to the existing design and then get a small batch of boards made. Stay tuned!

-Brent
Brent Picasso
CEO and Founder, Autosport Labs
Facebook | Twitter

Post Reply