sensors

FreeEMS topics that aren't specific to hardware development or firmware development.
wheeler_express
DIP8 - Involved
Posts: 28
Joined: Wed Apr 03, 2013 3:11 am

sensors

Post by wheeler_express »

1. What sensors is the code written around- OEM, digital, etc.?

2. Is an O2 sensor required? In the event I have to use Avgas from an airport, an O2 sensor will die in a matter of minutes. Avgas contains 2.5 times the lead of the old regular leaded auto fuel.

3. I would like to be able to manually (yes manually) adjust the A/F ratio to aid in cooling during max perf climb and to aggressively lean during long range cruise. I see you use tables, so here is my first thought-
Typically this A/F adjustment is limited to 20% either way, so use an analog input and either a pot or a VR through an ADC to determine which side of neutral the knob is set at. Are the A/F calcs and lookup tables in one file or several?

Without a total remake of the code, is this possible with the addition of a few lines of code? Assuming we use a pot, the voltage would be from vref (5v) to 0, so 2.5 would be neutral. If we use the ADC, 5v would be 255 (this is sufficient resolution), 2.5 would be 127.5, etc. and a simple formula could be inserted to multiply the results of the lookup table up or down by this figure.

Is this going to be a huge issue with the code as written? For airplanes, a few % lean of peak can really stretch your range at high altitudes.

Kevin Alderman
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sensors

Post by Fred »

1. Nothing digital just yet, however support for two different widebands has been considered/started on. SD algorithms require intake air temp, manifold pressure, and engine speed. Climatic change and cold engines after start require coolant or head temperature to be practical. TPS is not required, but is very useful to have in logs. Battery voltage is required, but comes essentially for free, so is seldom mentioned.

2. O2 is not required, however tuning without it isn't much better than tuning a carb with a needle 50 years ago. Look at Alan's Spartan units for value in terms of running N independent wideband sensors on your engine. http://14point7.com/

3. It'll be trivial to hack in a mixture adjust pot between 0 and 5V, however they're 10 bit, so the raw values are 0 - 1023, which can be used on any scale of +/- % that you choose.

Having said that, you'll tune each site for best torque or best economy, but not both. Both is discussed here, and would be a worthwhile feature on a plane more than anywhere else: viewtopic.php?f=8&t=1768

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
wheeler_express
DIP8 - Involved
Posts: 28
Joined: Wed Apr 03, 2013 3:11 am

Re: sensors

Post by wheeler_express »

https://www.sparkfun.com/products/11050

This is my temp sensor of choice at the moment for OAT, it takes very little code to get a calibrated output. I would use it for IAT, but it is good for 125c, and if my intercooler gets blocked, it would likely melt.

For MAP I have the same little unit you show in your earlier ECU pics- the MPX 4250AP. Does the code have a conversion formula in place for this unit, or is for another one? Also, I have 2 of them. One for MAP and the other for atmospheric for comparison.

I think an O2 sensor would be problematic at best with the likelyhood of having to run leaded avgas. I understand the issues of tuning, Lord knows I have spent many hours with that dang little screwdriver, replaced jets, etc. The mixture adjustment pot gives us a wide latitude for engine operation without extra switches, etc. For takeoff we richen the mixture, for cruise we lean it.

I am considering whether to install an O2 sensor for testing and tuning, and then remove it after this part is done. I think this is the way I will go. In the US we have to fly the plane for 40 hours within 40 miles of the airport for testing before we can just go anywhere, so I should be able to tune the engine during this period.

Thanks for all your help too- I know how much time I have just in my project and cant imagine how many hours you have in yours.

KA
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sensors

Post by Fred »

So your rotary is also turbo? Have you seen this? :-)

Image

I'd highly recommend a pair of normal GM IATs, you can run one to MAT, one to IAT, and figure out which one's lying, or use three, eliminate the outlier and average the other two.

Yep, there is a small library of MAP sensor configurations including that one, however anything linear will work fine.

Installing a wideband O2 purely for tuning and then always running open loop is pretty normal, many do it due to the cost of widebands.

NP, YW for the help. I'll try to get some others to pitch in and save my wrists some strain.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
HD13
DIP8 - Involved
Posts: 24
Joined: Mon Apr 08, 2013 8:01 pm

Re: sensors

Post by HD13 »

Wheeler_Expres,

The temp.sensors you are refering to are digital sensors.
I'am using an NTC (thermasgard HTF 50 NTC10K) 10 K -50 to 150 degrees celcius for cillinderhead temp of an aircooled engine.
I allso have more info on temp.sensors and willing to share.
Just let me know.
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sensors

Post by Fred »

Feel free to post it up, more info is always welcome, you never know who is lurking and reading that may benefit from it ;-)
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
wheeler_express
DIP8 - Involved
Posts: 28
Joined: Wed Apr 03, 2013 3:11 am

Re: sensors

Post by wheeler_express »

The digital sensor is a Dallas one wire. I downloaded the .h file from their website, assigned pin numbers and presto output. Then you just insert a couple of lines to convert the output into whatever you want to see- F, C, Kelvin, whatever. I assigned the output to a variable and read it through the program.

Regarding the IAT- the OEM sensors are heated filament type. The spreadsheet shows the sensor connected to an I/O, but I doubt the board will provide the current necessary to make the sensor work. Likewise, I have a concern that the board can't sink that much current to ground.

Dallas makes one that has a stainless tip which will tolerate higher temps, although it won't read higher than 125C. Also, each digital temp sensor has its unique serial number. The code is only required one time, as it reads each sensor by SN and provides the output by SN.

How are you handling the current feed to the OEM temp sensor?

And, I liked the graphic regarding the rotary. I spent a good bit of time with some rotary racers in Miami, as well as others that have rotaries in airplanes. I am not concerned about the seals. One rotary powered RV6 in Northern Florida has over 2000 hours on it. My $54,000 continental certified aircraft engine blew up at 1500 hours, and I take care of my engines. In contrast, my rotary has cost me around $1800 so far without the turbo. I also have a complete replacement in the building just in case of a problem. You see, one cylinder on my Connie cost more than a new rotary engine. So, instead of replacing a part here and a part there, if I have an issue I will swap engines and have a new one.

The aircraft rotaries have been running 1 quart of 2 cycle oil per 25 gallons of fuel with good results. Tracy, the guy in Northern Florida disassembled his rotary after 1200 hours and found no measurable wear inside the engine even on the apex seals. He put it back together and is still flying it. His RV 4 has a 2 rotor, his RV6 has a 3 rotor. www.rotaryeng.net has some more info.

Kevin Alderman
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sensors

Post by Fred »

Almost all OEM automotive temperature sensors, coolant, intake air, ambient air, etc are standard thermistors which simply require a low-current bias resistor of a suitable value from 5V and a ground, you simply take the centre-tap and do a lookup from a curve generated using this: https://en.wikipedia.org/wiki/Steinhart ... t_equation

Fair enough on the rotary front, they're not bad engines if well looked after and never subjected to certain abuses (some of which are not abuses to a normal engine). It's the type of owners/drivers that tend to get them a bad name in cars ;-) That and the fact that the early ones really were terrible things. My cousin had an RX-3 and he said he used to pull into the gas station, top up the oil and check the petrol ;-)

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
wheeler_express
DIP8 - Involved
Posts: 28
Joined: Wed Apr 03, 2013 3:11 am

Re: sensors

Post by wheeler_express »

1. In the main.c file at the very end is a reference to adjustPWM() and commented "PWM experimentation". Where does this adjustment refer back to? If this adjustment changes the injector PWM just before injector firing, this would be a good place to hack in my pot.

2. FAA requires a tachometer- is there currently a counter or output that does this, or will I need to add it based on the CAS teeth?

3. Where is the rev limiter code I read about in the forum? I definitely have to change that to 8500 RPM. I could use this counter and a serial out to provide tach info.

I am trying to get a grip on your code structure and the logical steps it takes. I want to ultimately push the code changes back to the group, along with the hardware that I am using so that others don't have to repeat the process, at least not all of it. If I just start hacking and inserting my temp sensors, pot, etc where ever, it might cause issues later on for other users. My engine is a rotary, but not a regular one. Its P ported, turbo, 4 injectors, no lead/trailing spark- both fire at once, etc., and optimized for 6000 rpm continuous. Not suitable for a car really unless it is a drag car, but the basic code should work for other rotary engines too.

KA
User avatar
Fred
Moderator
Posts: 15431
Joined: Tue Jan 15, 2008 2:31 pm
Location: Home sweet home!
Contact:

Re: sensors

Post by Fred »

1) No, that's for boost valves and idle valves, etc. Start a new thread in firmware about your requirements for the hack, then we can discuss the best approach.

2) There is a configurable general purpose timed output, that at least a couple of people are using or trying to use for a tacho, one is for sure working fine: http://www.youtube.com/watch?v=8ZEtEbMOvW8 "the night after I implemented it" or so he says, likely true :-)

3) Just adjust the configuration in the src/main/initialisers/ directory to achieve your desired rev limit and hysteresis ;-)

No hacking you do will make it back into the main code base for the time being, so do whatever it is that you want, without fear. Git's good like that. You can't disrupt things, I simply won't allow it. I don't even allow myself to do it. I have a dozen branches with various work on them. Nothing goes into the main branch until it's ready for prime time, so to speak.

Re your special rotary, I keep telling you, it's just not :-) The only thing that you really need to run it reasonably that isn't in the main code right now is staged injection. Sean hacked that in on his branch, so you could grab that patch, or I could do something more official, or a combination of the two. As far as getting it running, completely forget about all of the stuff it has to do in the end, and focus on the basics first. ONLY once you have it running should you start to worry about pots and dual fail-safe sensors and staged injection and so forth. If follow this path, you'll succeed more quickly, or at all :-)

Let's get on the phone and hash some of this out with voice, it's quicker, and it saves my wrist from the physical pain that each of these long posts is causing me.

PM me a phone number and/or skype account, and a time/date to aim for.

Fred.
DIYEFI.org - where Open Source means Open Source, and Free means Freedom
FreeEMS.org - the open source engine management system
FreeEMS dev diary and its comments thread and my turbo truck!
n00bs, do NOT PM or email tech questions! Use the forum!
The ever growing list of FreeEMS success stories!
Post Reply