TIM-6 ignition > Arduino

[Tim Wescott]

... eg feed them into a ring buffer and take a weighted average, but that should be after trying to reduce noise electronically....

Aaaagh! My eyes, my eyes, I'll never un-see that.

Actually, it's a pretty common beginners mistake. There are some very good, very narrowly-focused reasons to put a moving-average (or other FIR) filter in a control loop, but this isn't one of them.

You should be able to make the temperature readings fairly quiet. If so, and if the pump control is on-off, just use hysteresis to prevent chatter (i.e., if the pump is on then turn off at 150 degrees, if it's off then turn on at 160 degrees).

If you're continually changing the speed of the pump then you're probably using a PI controller, in which case you may not need filtering at all.

If you do need low-pass filtering, use a first-order IIR:

temperature_state = temperature_state + some_gain * (temperature_measurement - temperature_state);

Just put the above line in your code after you read the temperature into temperature_measurement; it'll filter nicely, without using much resources, and the filter will be minimum phase, and thus won't contribute to loop instability. The number "some_gain" will need to be tuned, but it roughly has the same effect as the reciprocal of the length of your moving-average filter.

 

[Tim Wescott]

Hi Charles,

Texas Instruments make linear thermistors in surface mount packages. It would be easy to solder wires to them and then bond them to your radiator body with epoxy.

I don't see where a specific temperature sensor was mentioned, but if you're going to use a thermistor (which is easy to use with a micro if it's for a limited temperature range) then you can get really teeny ones. Here's an example, but I suggest looking on DigiKey and Mouser: https://www.digikey.com/product-detail/en/littelfuse-inc/GR103E1K/GR103E1K-ND/6009533.

A thermocouple would be a bit harder to apply electrically, but should be more rugged in small sizes than a thermistor.

 

[ICEpeter]

Hello Charles,
Regarding your radiator temperature monitoring, I build my own temperature sensors in miniature for monitoring radiator temperatures and described it in the post: A Radiator Story part 2
If you search for that post it may assist you in what you want to accomplish..

Peter J.

 

[bluejets]

Thanks for suggesting the DS18B20. I think it is much too big to be fitted on the engine. In a pocket by the radiator might be the answer, but even then, at 30mm long it would be nearly half the size of the 80mm fan and radiator!

Welcome to model engines where it is a fact of life that not all things can be miniaturised.
We use them in the return water line adjacent to the head.
The stainless tube type does not have to be installed in an upright position and there is also the option of using the to92 case device if space is a problem.
Averaging can work to a degree but not always .

 

[Charles Lamont]

Thanks for the input chaps, this is getting interesting. To clarify, the coolant pump will be driven continuously by the crankshaft. (There is enough going on here without thinking about miniature electromagnetic clutches!) The temperature input will be used to switch the fan on above a threshold temperature and then ramp its speed up with rising temperature.

If the fan speed hunts a little under a steady engine load it does not matter much. I do like Tim's IIR instead of a moving average and evidently have some reading to do.

I am moving things about on the website, and have expanded the ignition page to be the home for all the electrical jiggery pokerey:
http://www.charleslamont.me.uk/Seagull/electrics.html

 

[Tim Wescott]

I was stuck on thinking you were going to pump (or not pump) water, and then how in heck would you read block temperature from the water in the radiator?

You may be able to control temperature adequately with a simple integrator, although you may need a PI controller. If that's not in your background, you may find this article helpful: http://wescottdesign.com/articles/pid/pidWithoutAPhd.pdf

 

[BaronJ]

Hi Tim,
This book was one of those recommended when I was at Uni !

"Applied Control Theory for Embedded Systems"

Not that I remember that much of it nowadays

 

[Charles Lamont]

New question. What is the/a correct Hall sensor to use with the TIM-6 ingition?

 

[Nikhil Bhale]

New question to an old thread.

Is it possible/ advisable to use a proximity switch instead of hall sensor for the ignition circuit?
I just purchased a copy of MEB#34 for D Sage and Gadde ignition system and was wondering about it.

Edit I meant inductive proximity sensor like shown below. I believe they have more current carrying capacity.

Regards
Nikhil

 

[Charles Lamont]

Why do you want more current carrying capacity than a hall sensor?

 

[Nikhil Bhale]

Why do you want more current carrying capacity than a hall sensor?

I am just saying. I am just exploring the options. Also I have read somewhere that hall sensors may blow up.
I don't know anything about electronics but can these inductive sensors directly drive the coil?

 

[dsage]

You have a couple of questions in the last two posts.
a) I'm not familiar with the inductive sensors but I believe they require some intermediate circuitry to provide a switch action.
b) NO the sensor should not be used to drive the coil directly.

If you can give me the part number of the sensor I can check the data sheet to see if it can be a replacement for a hall sensor.
IMHO to try to arrange a sensor that large in proximity to the cam shaft of your model engine would be a lot more trouble than just using an old school set of points.
That said, the sensor blowing issue is usually caused by poor wiring practice. The most important point being that you need to have a DIRECT HEAVY ground wire from the engine block back to the battery negative. Every engine is different in the requirements.
Magnetic reed switches have been used as direct replacements for points or hall sensors. You know the kind in a small glass tube with a wire out each end and activated by a near passing magnet. Sometimes used as door sensors.

 

[Nikhil Bhale]

@dsage thanks for the reply. I have very little to absolutely no experience with electronics. As I said earlier I am exploring all the options. My ideas may sound totally whacky to anyone with electronics knowledge.

I have downloaded some electronic circuit simulation program on my mobile. I am trying to simulate all the ignition circuits that I can find on net. And maybe come out with my own.

I am attaching some datasheet of inductive proximity sensor.

As you said a sensor cannot be used for driving a coil directly. But what if the sensor output is used to turn ON/OFF a SSR, Solid state relay? Coil can be driven directly from SSR.
The system will have only two components, a sensor and a relay.

 

[dsage]

It appears the sensor could be used in place of a set of points or a hall sensor with a Sage/Gedde module. It's plenty fast enough. The part number appears to determine which direction it switches (high or low).
But I ask why bother? Becasue it's pretty huge compared to a hall sensor or a reed switch.
I can't find a good spec sheet for the relay so I don't know if it will be fast enough.
You'll need a relay that can switch at least 66 hz for a single cylinder engine and multiply that by 2,4,6,8 for more cylinders.
You are on your own in designing something new.
Let us know if you come up with something workable.