MPVI Pro A/D inputs - averageing or instaneous?

[walt6076]

I'm using three of the A/D inputs on the MPVI Pro. I'm having good luck with the 0-5V signals from my WB and from my boost sensor. I'm having some strange results from the piggyback timing computer. It has a 0-5V signal line out I'm feeding into EIO 3. The output itself doesn't do any averaging on the signal. It might put out 0.33 V for two revolutions and then 0 V for a bunch of revolutions. It looks like the sample interval for the A/D inputs on the MPVI Pro is about 100 ms. If I'm at 3500 RPM, for example, that would be 34 ms at 0.33v and 66 ms at 0v.

What am I going to see in the datalog? Would I see 0v some of the time and 0.33v some of the time, proportional to how often the timing computer sent out 0.33v and how often it sent out 0v?

Or would the MPVI Pro A/D input average it and give me 0.1v?

(Apologizing in advance if this is covered somewhere--I did try and look. If it is explained elsewhere just share a link.

[schpenxel]

Going from memory but I believe it's taking a reading every millisecond then averaging over 10 miliseconds and reporting that, but don't quote me on that

One of the engineers would have to confirm

Could be off by an order of magnitude which would line up closer to what you're seeing... now I'm curious too

[Keith@HPTuners]

Data is collected over a 10ms period, averaged together, and then dispatched to VCM Scanner.

Could be anywhere from 10-100 data points collected over that 10ms period.

[schpenxel]

I was pretty close!

[walt6076]

So let me know if I've got this right: multiple EIO-3 v data points are collected over 10ms. Their average is sent to the VCM scanner. That 10ms value is reported in my 100ms data log event. The next 100ms data log event will have a new 10ms value. So if I've got an short, intermittent signal on EIO-3, statistically I'll capture something like 1 in 10 of them.

Are there any tricks to lowering the sample interval of the scanner?

[Keith@HPTuners]

No data is ever dropped.

The sample rate depends on the vehicle you're connected to.

[walt6076]

So is it correct that the averaging of the external sensor voltage reported will be over whatever the sample rate of my vehicle is? So my sample rate in the collected log is about 100ms. Then the average on the external sensor will also be over exactly that same 100ms? Thanks for helping me get this right.

[Keith@HPTuners]

Yes, if your vehicles sample rate is only 100ms, then that is the frequency you'll receive your AD inputs.

[sillycon]

Yes, if your vehicles sample rate is only 100ms, then that is the frequency you'll receive your AD inputs.

Is that also, then, a "Yes" to walts other question about the averaging of the data collected in that 100ms interval?

[Keith@HPTuners]

If there is other data collected in that 100ms interval, your AD input will poll faster.

AD input poll rate is directly tied to the data rate coming from the vehicle, and averaged over 10ms if this rate is shorter than 10ms spacing.

[sillycon]

Gotcha, so typically no averaging of values between polls; the ADs are just recorded as whatever they show when polled every polling interval (100ms in this example).

Different ballgame with sub 10ms polling intervals, but I expect few ECMs are pushing out data that quickly anyhow.

[Keith@HPTuners]

The time that matters is the time between messages the ECM pushes out.

Every CAN based ECM pushes data out that fast or faster.

The VPW vehicles are about that speed.

I don't know of a vehicle we support that is slower than 20ms.

[Keith@HPTuners]

fyi, this is a real easy thing to test. Connect a device to the AD inputs that changes voltage. Log it in VCM Scanner. Zoom all the way in to 1s. You should be able to see the data points.

[sillycon]

The time that matters is the time between messages the ECM pushes out.

Every CAN based ECM pushes data out that fast or faster.

The VPW vehicles are about that speed.

I don't know of a vehicle we support that is slower than 20ms.

I didnt realize that - I was going off the HPT documentation which indicated a max of about 40 samples per second (e.g. 25ms resolution), with most being around 20samples/50ms.

Or was that not an ECM-side limitation?