0 - 5 Volt sensors - Loss of resolution with 0 - 12 Volt ip?

SAD_Man
Posts: 7
Joined: Wed Jun 03, 2009 11:17 am

0 - 5 Volt sensors - Loss of resolution with 0 - 12 Volt ip?

Postby SAD_Man » Wed Jun 03, 2009 11:26 am

Hi,

Looking at the specifications for the analogue inputs on say the DL1 and the Dash2 I notice they have 0 - 12 volt analogue inputs.

Most of the sensors I use are 0 - 5 volt output, is there anyway to re-scale the 0 - 12 volt inputs to 0 - 5 volt?

By re-scale I mean that the internal ADC's now expect a 0 - 5 volt input to the analogue channel/s. In doing so using the full 12 bits.

I ask this because if I were to use my 0 - 5 volt sensors I would effectively be losing more half of the available 12 bit resolution that the Race Tech units have?

Just wondered what you guys do about this?

Cheers

faraday
Posts: 267
Joined: Sun Feb 25, 2007 1:18 am

Postby faraday » Wed Jun 03, 2009 3:03 pm

I hope your SADs are improving with some increased sun (northern hemishere, I presume, unlike myself). My mood slipped somewhat when we lost the 0-5 V inputs the early DL1s had. An economy measure, no doubt, but perhaps justifiable when the price of a DL1 is considered.

:idea: Look at your problem from the other end...
Is it possible to build amplifiers to scale the analogue signals :?:
I imagine that as well as the design and construction costs, there will be some distortion.

Perhaps an electron herder out there can comment on this whimsy.

SAD_Man
Posts: 7
Joined: Wed Jun 03, 2009 11:17 am

Postby SAD_Man » Wed Jun 03, 2009 4:02 pm

Cheers Faraday,

How can you not be an "Electron chaser" with a name like Faraday!

Thanks for your reply, I only occasionally chase electrons (So my understanding could and probably is flawed) but this is my understanding:

When you buy an ADC (Analogue to Digital Convertor) you buy them in various flavours 8 bit / 10 bit / 12 bit etc... And it is the number of bits (amongst other features) that determine the cost of the ADC.

So RT has a 12 bit ADC already, after this (this is the bit I am not sure about) the ADC is supplied a stable reference voltage, in the case of the current DL1 this is 12v but could just as easily be 5v as it was (from your comments) - The ADC then 'knows' that the incoming analogue voltage is say 5 volts and will convert it in to the 4096 separate bytes for use in the digital domain accordingly.

So my understanding it wouldn't be a cost issue?

My guess is this was changed to 0 - 12v as this can cope with almost any sensor and they (we) live with the fact that anyone using 0 - 5 v output sensors just gets less resolution?

Expensive data systems like PI Research have the ability to change the input range 0 - 1v, 0 - 5v, 0 - 10v etc... to some of the analogue inputs to cater for this.

But I do appreciate that DL1 / Dash2 products are cheaper and are built to a cost in comparison.

SAD are my initials :cry:

I am in Birmingham (Not Alabama) but West Midlands, England.

We have just had our two days of sun for this Summer :roll:

Cheers

osborni
Posts: 497
Joined: Fri Dec 07, 2007 9:08 pm
Location: USA, Michigan

Postby osborni » Wed Jun 03, 2009 4:51 pm

The 12 bit ADC gives you a resolution of roughly 0.003 volts. On a 0-5 VDC sensor this gives you about 1666 divisions or 333 divisions per volt. That should be enough for most purposes.

I use LM19 one wire sensors for temperature. The effective output range is about 1 volt for ~60-210* F range - IIRC. So I do see some "stair-stepping", but it is still accurate enough for measuring temperature close enough to a +/- 1 degree.

Realistically, pressure and temperature only need to be good to a few degrees or PSI. Stuff like brake pressure, steering, TPS and manifold pressure are more for reference of my behavior as a driver then for literal translation of system response. I just want to know how smooth I am (not) and am only peripherally concerned with the actual data numbers. I can see that stuff like AFR needs to be more accurate, but a 12 bit ADC is good enough for that too, even over a 4 or 5 volt effective range.

It's not like some industrial DAQ systems that I use at work that need 16 bit resolution on a 0-5 volt range. That sort of accuracy isn't mission critical for monitoring an engine or the other typical sensors on a race car.
BMW 2000 M Coupe

faraday
Posts: 267
Joined: Sun Feb 25, 2007 1:18 am

Postby faraday » Wed Jun 03, 2009 5:07 pm

Perhaps all I share with Faraday is his christian name, but we should all aspire to achieve as much as he did. As a student more from the sciences than humanities and to my detriment now a mechanical engineer, I was aware of Faraday and "his" cage.
Thirteen years ago, I adopted our current ginger moggie from the RSPCA animal shelter. The tiny kitten was as you'd expect in a cage, but hanging inverted from a wire cage protected incandescent light fitting. I'm not sure that this was the best way to keep his tummy warm, but it was certainly a great trick.
For some reason I thought Faraday was Irish and associated ginger colouring more with that island, so electricity + cage + ginger = Faraday -> the name I chose.
When my wife (humanities background) stopped abusing me for bringing home another cat, she challenged me as to his silly name. Subsequently, I thought I should do a little biographical research to arm myself for any future inquisition.

But back to ADC.
My cost argument presumes that initially R-T saw the need for 0-12V as well as the usual 0-5V channels. I guess that ADCs come with multiple channels. The early DL1s would thus have needed two, assuming that members of each gang shared the reference voltage.
If all the above applies and cost savings were sought when a major revision was happening, I expect that a single ADC with more channels would achieve this.

My home town of Melbourne is sometimes thought of as more English than most Australian locations, particularly with regard to climate. We will nevertheless probably see more hours of sun over the next wintry months than you do in a year, but it often still feels as cold as a snowy alpine region because of strong and damp winds.
Despite being grey today, we had light winds and 18 degrees. Most of us enjoyed the decent rain that is too infrequent these days.

However glum the "Old Country" might be, I hope I can experience it someday soon.

SAD_Man
Posts: 7
Joined: Wed Jun 03, 2009 11:17 am

Postby SAD_Man » Wed Jun 03, 2009 6:19 pm

Thanks for the replies guys,

Faraday,

The last time I was in Melbourne was 1995, I stayed for two weeks visited (Race circuits) Philip Island, Calder Park and Winton? (Tasmania). I stayed in a hotel in the middle of Melbourne, around the corner from a Harley Davidson dealer, sorry that's like saying "Oh you live in Melbourne, you must know Mr. Smith, he lives in Melbourne"

My time with the Australians and the Tasmanians seemed to hint at a fair amount of banter (to the detriment of Tasmanians) which inferred that they were all related! :lol:

And another piece of useless information is that on my flight to Tasmania I sat next to Perfect Peter Brock, he was racing at Winton.

Yes I am a Mechanical engineer for my sins as well, but seem to spend more of my time setting up, looking at, using and analyzing data and data logging equipment.


Back to ADC's

Yes I see that RT's decision to make the analogue inputs 0 - 12v meant it would cater for a wider range of sensors and for most circumstances this was a greater benefit to its users than the loss of resolution, even with a 0 - 5 volt sensor you still nearly achieve 11 bit resolution. And the facility to be able to re-scale the analogue inputs would add expense.

So I guess the answer to my question is that the inputs are 0 - 12v and are not able to be re-scaled. To be honest I suspected they wouldn't be.

Osborni,

Yes I agree entirely that 12 bit is more than enough for most applications even with a 0 - 5v input and the 1700 odd bytes this results in.

It sounds like we both use similar DAQ equipment at work, my questions are directed more towards some of the type of situations I need for them at work in which case sometimes I need this resolution for a one off rig I might need to quickly put together.

I am also interested in putting the Dash2 / DL1 in my road car having only just become aware of RT products.

I have been looking for a few years at some of the cheaper DAQ systems as we all too often end up spending thousands on LMS / National Instruments / P.I. Research type equipment because people are comfortable with them and scared to risk looking else where.

Cheers

EDIT:

To further confuse myself I have just been looking through the knowledge base for the DL1 pin out and and it clearly states that the 8 analogue inputs for the DL1 MK2 are all 16 bit?

http://www.race-technology.com/wiki/ind ... onnections

Were they all once 16 bit or is it a mistake in the documentation, what is the documentation like generally?

faraday
Posts: 267
Joined: Sun Feb 25, 2007 1:18 am

Postby faraday » Thu Jun 04, 2009 3:23 am

Many young engineers think in black and white and lust for a squillion significant figures.
Real life, which includes inanimate machines and the measurement there, of just ain't like that. Lawyers are not the only chaps dealing with grey areas.

The same applies to the documentation :wink:

Nevertheless, as time goes on, things have improved. Unfortunately, it seems to me that the volume of words has increased faster than the removal of sections that either no longer apply, or never did apply.
R-T's proprietors are creative and imaginative chaps who have planned some features that have not eventuated.
More conservative and established motorsport DAQ firms have documentation shortfalls too. It has been a problem for Motec, with whom I am close.

This brings me to Aussies.
Not all Taswegians are inbred, but then Tasmania really is the closest thing we have to England.
Winton is actually in Victoria; I should be testing there tomorrow.

I met Peter Brock, but never sat next to him. One of my closer friends actually did, in (Motec proprietor) Richard Bendell's Daytona, which was designed by my best friend.
Unfortunately, Peter's perfection deserted him.
Peter and the Daytona did not survive.
The tree and my navigating neighbour, Mick Hone did. Mick was a very successful bike racer and has the local Suzuki dealership. He still races cars as an amateur.

Life goes on, but not for Brock.
:cry:

SAD_Man
Posts: 7
Joined: Wed Jun 03, 2009 11:17 am

Postby SAD_Man » Thu Jun 04, 2009 7:10 am

You are indeed correct, I had a feeling my memory was deserting me when I made the association between Tasmania and Winton (hence the ?) it was of course Symmons Plain. Interesting place none the less.

My "Tour guide" and entertainer for the visit and chief instigator of the Tasmanian banter was a guy called Dick Johnson, John Bowe was usually on the receiving end of it. I was working with the Shell sponsored Falcons run by the Stone brothers.

The V8 Supercars were very "entertaining".

Sadly my earlier visit to Philip Island coincided with the accident which claimed Gregg Hansford and I was only introduced to him for the first time just prior to the race start to discuss some handling issues.

Alas, working in Research and Development, trying to find the minute differences between strategy A and B or component X vs Y all too regularly has me chasing too many significant figures. But I am all too regularly presented with test results unnecessarily quoted to 5 decimal places from measurements made only to 1 d.p.

OT:

As you seem to "know your way around" the DL1 what is your opinion of the GPS speed information it gives and your view on the 5Hz vs the 20Hz option? I only have experience of the CORRSYS Datron Microsat system. Especially with regard to the speed at which the DL1 acquires satellites on start up and the number if any of signal drop-outs / glitches in use.

Also how well does the engine ECU serial input work, is it robust for a variety of ECU data streams.

Good luck in Winton (I'm left feeling unsure whether the third venue I visited was actually Winton or Calder Park, I'm sure I have notes tucked away somwhere which will enlighten me)

jakk2809
Posts: 1
Joined: Thu Jul 30, 2009 1:40 pm

Postby jakk2809 » Thu Jul 30, 2009 1:43 pm

I'm having the same problem. I got current sensors that work on 5V but I want to take advantage of the 12V range. One option is to use op-amps to scale the 5V to a 12V. A simple gain circuit will work but I'm also trying more complicated subtractor circuits to fine tune things.

osborni
Posts: 497
Joined: Fri Dec 07, 2007 9:08 pm
Location: USA, Michigan

Postby osborni » Sat Aug 01, 2009 10:12 pm

Why? I have had zero issues running 0.5-4.5 volt range sensors. Plenty of resolution for a race car.

I do run some temp sensors that have an effective 0.5vdc range, and you can see some lack of resolution in that, but it still works out to less then 1/2 of a degree F where the data is sort of flaky.
BMW 2000 M Coupe


Return to “Sensor-related questions”

Who is online

Users browsing this forum: No registered users and 35 guests