Improving glow plug driver design for RC plane radial engine

[kikinou]

Here are the schematic and board I sent for PCB manufacture.

I moved on with a regulator (it adapts readily to various battery) though I give a try to the 555+MOSFET idea : sounds fun to me.

I chose to build two boards : one of the glow plug driver, and one for sensors. I am hoping to get smart enough so I can daisy chain the drivers and sensors so the same hardware self adapts to the targeted one, two, three and nine cylinders.


Circuit protection wondering
I hoped I got right the circuit protection with the D1 & D2 diodes and the PTC1 & PTC2 reversible fuse. This might be better than nothing, but I am wondering whether this could be improved.
The PTC3 is supposed to cut above 5A within 10ms when the user creates a short-circuit (it is often quite hard to plug the cable in the engine cowl, so short-circuit are regular). During normal operation, the glow plug (a ~200-500 mOhm resistor) drains about 2.5-3A. But, when short circuit, the wires should be about 50mOhm, so the current could peak at about 1.5V/0.05 Ohm ~ 30A. This could kill the expensive current sensor as well as the regulator.
How to improve this short circuit protection ?

Optocoupler wondering
I regret I did not put an optocoupler in between the ATTINY MCU (pin 6) and the regulator (pin 5). They are powered by two individual batteries.
- Would two individuals ground planes be better ?
- Is it better to separate high power and low power circuits ?

[kikinou]

Components and PCB just have been delivered. I am getting nervous.

[kikinou]

Yahhooooo .....


...... it doesn't work.

Not even the simple ON LED that is just behind the protection circuit and a filtering capacitor.

How to trouble shoot. My plan is to check voltages. For instance, the series resistor of that LED should be 5.2V, and it is around 3V. My plan is to use the multimeter to try to guess what's wrong, then pop components with the hot air gun. Remember, I had the circuit working on the breadboard. But I screwed up somewhere, somehow. But how many times ??

I uploaded everything on github :

[MrsR]

There is a commercial unit that will increase current/voltage to flooded Glow Plugs.
Try DESERT AIRCRAFT AUSTRALIA they had a number of units last time I looked and are very help full.

ALL THE BEST
Rachael

[kikinou]

Geeezz, that sounds like a fantastic idea.

I did not know about it, even though I noticed I was kind of doing it manually. Indeed, somedays, I applied up to 3A through the ~200mOhm glow plug resistor on an eventually flooded cylinder. Sometimes, in particular when it is not the first flight of the day, I just need a more conventional 2.5A. Also, to improve the idling RPM (which I really need/like on some of my 4 strokes engines), a slightly lower voltage is needed. So I might also improve glow plug life time, though I don't really have such problems.

How does their unit work ?

I could / will upgrade the current design (see github ) such that the attiny (that easily deciphers radio servo signals) can tweak the regulator voltage (with V_ADJ potentiometer on the schematic/board).

I don't know yet which kind of modification I need. I am going to do some research.

Many many thanks for that hint.

Using an on board glow driver without this idea made me loose this "starter" functionality. And that was a downside.

Many many thanks

[Buriedcode]

Yahhooooo .....

...... it doesn't work.

I really don't understand what you are trying to do on that schematic.  If you were driving them with constant current, why is the microcontroller there?  The comparator?  Now you've made boards, of course its best to use what you have so not to waste anything, but I fear you may have done what we've all done at some point - got lost on the little things, the details, and lost sight of the bigger picture.  I would have thought a small 'module' for each glow plug, perhaps with a control on/off signal, or simply an analogue voltage in that controls the current for each plug.  Using a dedicated current monitor seems overkill.

I'll re-read the thread to try and work out exactly what you're doing.  As for your boards - what *exactly* have you designed them to do?

[kikinou]

Please don't re-read the thread : let me put together a concise drawing tomorrow.

{" Yahooooo....." + nice picture +  "it doesn't work "} was meant to be funny.

You're right : I indeed moved on with little modules.

I will put together a pretty drawing instead of thousand words to improve the thread and my thoughts. But long story short : one big MCU and a few littles ones. You guess this right.

- The big one is for the screen, bluetooth, airspeed, barometer, thermometers, the SD card. That board is meant to be the same for the 1, 2, 3 and 9 cylinder engines. I started writing code for the AVR, but now I am playing around with a pi zero. A pi zero shield sounds like a more attractive way to go. It might be cheaper. Writing python code is faster, though I love C.

- Each little module has an tiny MCU (attiny85 - 8 pins) that deciphers the PWM signal from the receiver. Then, it switches ON/OFF the glow plug driver. It also digitizes the sensed current and sends it back to the main board with the big MCU (it seems like I2C should work). The comparator simply switches a LED ON when there is current (more than ~1A) through the glow plug : so I know the glow plug and the connectors are not totally dead. I should have chosen an attiny84 (14 pins) to get more pins : I did not know about this one.

But I will put together a drawing with functional blocs tomorrow. That will clarify my thoughts.

[kikinou]

Here are the ideas, as of now :

Before flight:
- cylinder running?  --> with cylinder temperature and glow plug resistor value [1]
- mixture rich?        --> RMP + temperature
- readings?             --> either smartphone, LCD screen, working fake instruments panels

After flight:
- SD card               --> use my laptop to plot the flight pattern (barometer height + airspeed) and check cylinder temperatures & RPM to improve or to make reliable the engine settings, trough the whole flight envelope.



That sounds both crazy and realistic to me. For sure, I am having fun with this project.


[1] For instance, with the radial 3, I have trouble to know wether 2 or 3 cylinders are running. When the engine cowl is screwed, I don't have access to cylinder tuemperature. My ears have hard time to recognize the sounds. The RPM is an indicator, but I don't have one. I stalled just after take off, at 2 meters high, because I mistakingly thought I had the power of al  3 cylinders.