Author Topic: Improving glow plug driver design for RC plane radial engine  (Read 21765 times)

0 Members and 1 Guest are viewing this topic.

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Improving glow plug driver design for RC plane radial engine
« on: February 23, 2016, 09:33:09 am »
Hi there,
This is my first post here, aiming at showing some fun stuff, and getting some help - yep, I have some issues.

The following image shows the ugly design that I am aiming at improving:


It looks ugly but it works well enough for me. The following picture shows the plane flying, and the 3 cylinder radial engine.



But I would like to add three functionalities (3 current sensors, 3 thermometers, and voltage regulators) because I wish to troubleshoot the engine. He (it) is moody. Sometimes, the glow plug filament is broken. Sometimes, it is not warm enough, or the opposite. Sometimes, the engine is too hot (mixture lean ?). Most of the time, one of the cylinders doesn't want to cooperate, and it is hard to know, even though I am now familiar with its roaring. Hence, I am willing to do my own clean PCB design. This is a first for me. I also get some motivation because that effort would also benefit the in-line twin engine that tracts a P51 mustang and the radial 9 a friend of mine just got. Another motivation to get a PCB done is to redo my ugly on-board glow driver : you might have notice the hot glue blob on the DC/DC converter in the above picture. The capacitors got unsoldered after about 50 flights. Now is a good time to invest some time in this project.

So, I am giving a shot to the following PCB design project :

A preliminary schematic draft is attached as a pdf file.

The idea is simple :
1. an attiny converts the receiver PWM in ON/OFF (0V or 5V) signals
2. that triggers a relay to power up the 3 glow plugs (~3A each)
3. an ATMEGA 328 is electrically independant and does the clever work :
  - 3.a - measure each cylinder temperature
  - 3.b - measure each glow plug current
  - 3.c - report the 3 glow plug currents on a 3 color LED
  - 3.d - report the 3 cylinder temperatures on a screen (on board / hot plug, don't know yet)

Here is the idea behind the current measurement.
When the cylinder is OFF, the glow plug current is about ~2.5A.
When the cylinder is ON, the glow plug current is about ~2.3A.
When the glow plug filament is broken, the current is exactly 0A.
--> so I can make sure this when my engine is moody.

Here is the idea behind the temperature measurement.
- I can double check if a particular cylinder is ON or OFF (warm or cold)
- I can make sure the fuel/air mixture is about right ( warm=lean , cold = rich)

Here is the idea behind the UART and ISP connection
- I use quite randomly expensive OS F plug (4 strokes) and very cheap hobbyking 2 strokes glow (it works fine for a third the price). So I need to adjust the firmware to account for this.
- I will need to toubleshoot before it works.


Here are the ideas behind using regulator :
- LiFe (safer than LiPo) battery are by far lighter for the same energy storage, but their voltage is about 3.3V, or some multiple. So I would need to get the voltage down to 1.2V - 1.5V, depending on the glow plug brand / batch.
- Three glow plug is up to 9 A @ 1.5V. The DC/DC converter is specified for 6A. I should be pulling about 7-9A through it (never measured it actually). It is warm but it works. There is a potentiometer to adjust the output voltage : that is necessary : some plugs likes 1.2V, some other 1.5V, and I like switching in between the cheapo and the expensive one.







So far so good, but I am stuck :
I have hard time to find a simple switched regulator that gives me 9A @ 1.2V-1.5V adujstable from a 3.3V or 6.6V battery.  Would you recommend some ?

- The LM1771 sounds great because it outputs low voltage (down to 0.8V) from 2.8V-5.5V (1S LiFe is 3.3V +/- 0.3V). The TI datasheet section 8.2.3 indicates it can delivers up to 5A @ 3.3V from a 5V battery.
--> Can I get 9A @ 1.2-1.5V ?
--> Can I put a couple of regulators in parallel ?
--> Should I simply harve 3 regulators, one for each glow plug ?
--> Can I simply design the circuit with figure 6 BOM : sounds complicated to me (2 FET, 3 caps, 1 inductor, 1 resistor, 1 pot).

- The LM2576 datasheet suggests my life would be easier with this guy. Indeed, when I look at figure 33 of the TI datasheet, I readily find the circuit I need, and it is simple (2 caps, 1 diode, 1 inductor, 1 pot). It is less efficient but simpler : I like it. But it outputs max 3A. I would need 3 either in parallel, or one on each glow plug.
--> Can I connect this guy in parallel ?

- The LM338 is linear, even less efficient, but that is fine with me. Doubling the LiFe battery size is not a problem on that plane. According to the datasheet page 9 (1.2V–25V Adjustable Regulator figure), the circuit is fairly simple (2 caps, 1 potar). That's sweet for my PCB area, soldering skills, though I guess I can handle more. But the may point here is that I may be able to two of these regulators in parallel to output up to 10A.
-->But can I put such a regulator (linear) in parallel ?

Many thanks for your hints for theses issues. I hope there are quite interesting to you.

May be I should get these various guys on a breadboard, and see how it goes. Would that be the path moving forward ?

- kikinou


PS : english is not my native language, so I welcome notification on sentences that should be clarified.
« Last Edit: March 05, 2016, 08:33:29 am by kikinou »
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug drive design for RC plane radial engine
« Reply #1 on: February 23, 2016, 10:03:05 am »
Why dc/dc 1,2V ?
Nominal voltages of glow plug's is not 1,5V and 2V ?
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 8725
  • Country: de
Re: Improving glow plug driver design for RC plane radial engine
« Reply #2 on: February 23, 2016, 11:42:48 am »
I would avoid using a relays - a MOSFET should be smaller, lower power and less trouble with variable contact resistance. Some DCDC converters may also have an enable pin, so that no extra power switch is needed.

For just 6 inputs there is no real need for a 16 input MUX chips. The µC has enough analog inputs, and there are smaller MUX chips around with 4 or 8 inputs.
 
The following users thanked this post: kikinou

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug drive design for RC plane radial engine
« Reply #3 on: February 23, 2016, 01:13:05 pm »
Why dc/dc 1,2V ?
Because NiMh works fine with OS F and the Hobbyking#3 glow plug.

Nominal voltages of glow plug's is not 1,5V and 2V ?
My cox 0.06 and 0.02 sometimes needs higher. I don't really now the right voltage actually : I turn a knob or plug NiMh till it works.

To mitigate the issue you are bring up, I like the pot I have on the DC converter shown in the above picture. I set it to the glowing red I like.
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #4 on: February 23, 2016, 01:20:05 pm »
I would avoid using a relays - a MOSFET should be smaller, lower power and less trouble with variable contact resistance. Some DCDC converters may also have an enable pin, so that no extra power switch is needed.
I'll do the extra effort to get it that way. Using a relay with an attiny was intellectually easier for me. I am going to use a breadboard this time and get the MOSFET right.

For just 6 inputs there is no real need for a 16 input MUX chips. The µC has enough analog inputs, and there are smaller MUX chips around with 4 or 8 inputs.
Thanks for the catch. That is much better. I'll update the schematic in the next few days.
The truth is that I initially planned to do the board for the radial 9 (9 current sensors, 9 temperature sensors), so I needed the MUX. But that idea was too ambitious. So I got more modest now, and I am doing a first step with the 3,2 and 1 cyl first.
« Last Edit: February 23, 2016, 01:21:36 pm by kikinou »
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #5 on: February 23, 2016, 01:21:18 pm »
I'm a former amateur of model aircraft and have significant experience with 2T glow plug and diesel engines.

I do not understand where is your problem.

Your 3-cylinders engine is a 2T or 4T engine ?

With 2T engines, you need no longer supply the glow plug once the engine starts, so your project does not make sense.

Regarding multi-cylinder engines 4T, it seems to me that the glow plug principle is not good and that you need to prefer a gasoline engine.

In my opinion, the real solution would be to change the engine if it is a 4 strokes engine.
« Last Edit: February 23, 2016, 01:22:55 pm by oldway »
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #6 on: February 23, 2016, 01:42:41 pm »
It is a four strokes with a gorgeous sound, and edgy power/weight ratio. It barely allows for looping : but that's quite realistic. Nevertheless, the whole point is that I love its roaring.

Having an on-board glow driver has 3 advantages.
1 - Flip a switch on the radio, flip the prop and here we go.
2 - Much lower idle RPM, which is nice/required for taxing and landing.
3 - In flight, there is always a cylinder that switches OFF below 25% power. It should not but it does. So I used to turn it back ON. Now I have the driver automatically ON when throttle is below 30%.

Changing the engine, is a solution. But that would mean storing/selling this 520 EUR nitro engine and buying a 840EUR gasoline engine. The power is 20-30% lower and the plane would have hard time to take-off, and there is an additionnal 100g for the CDI (sparking ignition). I tried a compromise. I purchased a CDI ignition and use it with methanol/nitro (not gasoline). In principle, I should have los only 5% power (not 20-30%). But I failed getting all 3 cylinders running. Then the sunset came, and I put back my glow plug back the next day to have fun on the field right away.

So, now I have this engine, which is a quite excellent value for the money, I stick with it, despite the few solid drawbacks your are pointing out.

And, here I am with the strong desire to get my glow plug current monitored and cylinder temperature measured.
« Last Edit: February 23, 2016, 01:48:33 pm by kikinou »
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #7 on: February 23, 2016, 02:38:59 pm »
To function properly, especially at low rpm's, the glow plug must remain sufficiently hot.

This is easily obtained with the 2 stroke engines, but not with 4T engines.

If you seek the maximum power / weight ratio, choose a 2T engine, not a 4T engine, there is no match at all.
(NB: for exemple, OS 35AX 2 strokes with muffler: 3.52HP/Kg, OS FS95 4 strokes with muffler: 2.58HP/kg)

In my opinion, you must have a reliable engine for the sake of your plane.

For this reason, you should reconsider the choice of your engine even if it cost money.

What you try to do is in french, "une emplâtre sur une jambe de bois", and that's never a good choice !
« Last Edit: February 23, 2016, 03:21:02 pm by oldway »
 

Offline promacjoe

  • Regular Contributor
  • *
  • Posts: 65
  • Country: us
Re: Improving glow plug driver design for RC plane radial engine
« Reply #8 on: February 24, 2016, 03:50:52 am »
have you considered running your Glow plugs in series. Then you could use a regulated power supply at 4 to 5V, 3A. to know which glow plug is not working, a simple LED and resistor across each glow plug, will work. when a glow plug is burned out, the LED associated with that glow plug will light up. the advantage to this is, if one glow plug is burnt out, the engine will not crank. also you reduce your component list. Only one power supply and current sensor is needed.
 
The following users thanked this post: kikinou

Offline BurningTantalum

  • Frequent Contributor
  • **
  • Posts: 321
  • Country: au
Re: Improving glow plug driver design for RC plane radial engine
« Reply #9 on: February 24, 2016, 07:04:57 am »
Kikinou-
The Cox engines, being American and of old design, use 1.5 Volt glow plugs/heads as they were designed for zinc carbon cells.
Most modern plugs are designed for 1.2 Volts as NiCad and NiMH are commonly used. I think that other voltages have been used in the past but I'm not sure.
Regards,  BT
 
The following users thanked this post: kikinou

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #10 on: February 24, 2016, 07:08:07 am »
Kikinou let me know by PM that he wants to use his 4T engine.

We will than solve his problem, but not in the way he recommends.

We are on an electronic forum and most of the participants do not know anything upon glow plug engines.
It is therefore necessary to give some explanations about it.

What is a "glow plug?
It is a coiled rhodium platinum wire which is heated to red at start by an electric current and which is maintained at high temperature by the heat of explosions.

The supply voltage of the glow plug is 1.5 or 2V according to brand and model.

There are different types of glow plug from hot to cold depending on the engine type, fuel type, and so on...
More details here:
http://www.osengines.com/glowplugs/

Why we need a glow plug?

The glow plug engine is running on diesel engine principle but the compression rate is insufficient to ensure ignition of methanol.

We need a hotspot to ensure ignition of the fuel mixture, this is why whe need a glow plug.
Note that the diesel vehicle engines also use a hot spot for easier cold starting.

The glow plug has another important function: to automatically adjust the timing of the engine.
Running at Higher rpm makes the plug hotter and "fire" the fuel-air mix sooner. At lower rpm, the filament cools and the plug fires less frequently.
But at low rpm, especially with 4T engines, the filament may became too cool to fire the mixture: that's the problem of kikinou.
« Last Edit: February 24, 2016, 07:12:41 am by oldway »
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #11 on: February 24, 2016, 07:33:35 am »
How to solve this problem ?

You should know that we can not constantly feeding the filament when the engine is running in full speed because the filament too hot, it becomes brittle and can break.

That is why we remove the connection of the glow plug as soon as the engine is started.

Whe also feeds the filament with constant voltage and not constant current because platinum has a positive temperature coeficient and the resistance of the filament increases as the temperature increases.

This has a temperature regulating effect when feeding the filament with constant voltage.

We can not connect the glow plugs in series as has been suggested because the hottest filament will have the higher voltage, what can dammage it.

The solution proposed by kikinou is complicated, inefficient and dangerous.
To detect the extinction of a cylinder by measuring its temperature is slow and complicated.
If one of the cylinders is switched off, there is a lot of risk than the 2 other also risk to switch off and the engine can stop before the temperature detection does act.

The solution of the problem is much simpler:
Simply measure the engine rpm and energize the glow plugs if the rotation falls below a specified value.
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #12 on: February 24, 2016, 07:43:25 am »
How to measure rpm simply?

By sensing the sound of the exhaust by a microphone.
The microphone must be protected from the wind (protection with foam, placed in the fuselage).
It needs an amplifier and a low pass filter.
If there is no sound or if there is a low frequency sound, the glow plugs should be supplied at reduced voltage (1V, for example), just to help keeping the filaments at a sufficient temperature.
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #13 on: February 24, 2016, 07:58:15 am »
It seems to me that my project description is poorly written. The topic is unfortunately drifting toward rc model stuff, as opposed to EE. Let me try to get back on track : circuit & PCB design.

Basically, I am adding functionalities to that widely used thingy:

http://www.hobbyking.com/hobbyking/store/__15517__Hobby_King_Onboard_Auto_Glow_Plug_Driver.html


Let me briefly summarize their marketing.
- The red and black leads connects to the glow plug (2 leads device), that itself is connected to the filament : a ~100 mOhm resistor that gets hot and glowing :

- The three lead connectors (white/red/black = signal / 5V / ground) is plugged to a channel of the RC receiver. The signal is a PWM (20 ms period) with a duty cycle that commony varies in between 5% and 10% (1ms to 2ms).
- When one flips a switch on the radio, this thingy powers the glow plug one can starts the engine by flipping the propeller.
- When the radio throttle is low (below a value one adjusts to the engine behavior), this thingy also powers the glow pug and the engine can idle at lower RPM. That desirable when landing and taxiing. 

That works sweetly, and I already improved this design for the sake of my plane safety (the above corsair is an ~ 800 hours build). On this picture (of some other setup I built),

Here is the schematic with an ATTINY85 that translates the PWM signal S from the receiver into 0V-5V on a MOSFET gate G.


one may see that :
- the glow plug battery is different from the battery of the receiver and servos
---> the receiver battery is the 5V connector on the little above schematic
---> the glow plug battery is the (drain / 10V) connector and the DC/DC converter (10V to adj. 1.2V) is hooked to the (GND / 10V) connector
- the glow plug voltage is adjustable

I would avoid using a relays - a MOSFET should be smaller, lower power and less trouble with variable contact resistance. Some DCDC converters may also have an enable pin, so that no extra power switch is needed.
There I used a MOSFET because I got smart enough that day. And indeed, this particular board has an enable pin, which I located, but my guts told me not to hack a board that works.

There are more details on rcgroups, which is a sweet spot to talk about RC related subject such as model engine.
http://www.rcgroups.com/forums/showthread.php?t=2228980
 

My current project is to add temperature and current monitoring to my previous builts, which work wonderfully on 4 planes of mine, and two of friends. That sounds like an interesting project to me, but the circuit design challenges me.
« Last Edit: February 24, 2016, 11:15:30 am by kikinou »
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #14 on: February 24, 2016, 09:43:26 am »
Kikinou-
The Cox engines, being American and of old design, use 1.5 Volt glow plugs/heads as they were designed for zinc carbon cells.
Most modern plugs are designed for 1.2 Volts as NiCad and NiMH are commonly used. I think that other voltages have been used in the past but I'm not sure.
Regards,  BT
Make sense. I always wondered why I had to crank up the voltage on my cox engines.
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #15 on: February 24, 2016, 09:50:34 am »
have you considered running your Glow plugs in series. Then you could use a regulated power supply at 4 to 5V, 3A. to know which glow plug is not working, a simple LED and resistor across each glow plug, will work. when a glow plug is burned out, the LED associated with that glow plug will light up. the advantage to this is, if one glow plug is burnt out, the engine will not crank. also you reduce your component list. Only one power supply and current sensor is needed.
I don't really have the choice though : the plugs have a common lead that is the engine chassis itself. Please have a look on the following image of my engine :
« Last Edit: February 24, 2016, 11:16:41 am by kikinou »
 

Offline BloodyCactus

  • Frequent Contributor
  • **
  • Posts: 480
  • Country: us
    • Kråketær
Re: Improving glow plug driver design for RC plane radial engine
« Reply #16 on: February 24, 2016, 05:26:53 pm »
thats a sweet looking engine. the sound must be amazing! i have some nitro engines but they are single cylinder, you dont get that awesome sound from those ones!
-- Aussie living in the USA --
 
The following users thanked this post: kikinou

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #17 on: February 24, 2016, 06:01:14 pm »
Seems to be a Saito engine.
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #18 on: February 24, 2016, 06:40:04 pm »
http://www.sonictronics.com/xcart/home.php?cat=285
Seems to be what you are looking for.
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #19 on: February 24, 2016, 07:20:17 pm »
thats a sweet looking engine. the sound must be amazing! i have some nitro engines but they are single cylinder, you dont get that awesome sound from those ones!
Christmas is 10 months ahead, so 52 bucks a month are worth it. These are rare and keep their value for second hand sale, I won't loose as much money as with common engine. I hard very hard time to get a saito twin. Go for it ! Go for it !
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #20 on: February 24, 2016, 07:43:13 pm »
http://www.sonictronics.com/xcart/home.php?cat=285
Seems to be what you are looking for.
Yes and no. I built this device for a third the price, and plenty of fun. This is the message I was trying to convey in my very first post. Here is a picture of my on board glow driver on my corsair (same picture as in post# 1).


Now, moving forward, I have two goals :
- put all these dangling electronics on a clean PCB ;
- add temperature and current sensors to monitor each cylinder health (no plan for an IOS Icylinder app though)







Thanks to Kleinstein, I am now in the process of improving the schematic posted above.

And I think I made my mind regarding the regulator issue. I am going to use three LM2576 to get up to 3A at adjustable 1.2V-1.5V from either a 3.3V or 6.6V LiFe battery.


But this LM2576 solds in various packages and output voltage (fixed 1.2V,5V,... and Adjustable). That is want I can understand from here :
http://www.ti.com/product/LM2576/samplebuy

Hence, the one I believe I need is :
- voltage adjustable
- TO-220                [  --> sounds good for both breadboard and routing pcb 3A traces]
- TO-220 (NDH)      [  --> sounds good to save PCB real estate]
=> TI part# LM2576T-ADJ/LF03 is my guy !




But I don't know where I can buy this guy in France / Europe. Sounds to me I am way too demanding.


« Last Edit: February 24, 2016, 08:14:00 pm by kikinou »
 

Offline oldway

  • Super Contributor
  • ***
  • !
  • Posts: 2174
Re: Improving glow plug driver design for RC plane radial engine
« Reply #21 on: February 24, 2016, 09:23:53 pm »
The glow plug is a simple resistor, no need of filtering, you can feed it with a PWM signal.
Use a logic level mosfet.
 

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #22 on: February 24, 2016, 09:32:20 pm »
Yep.

What would be the best way to get the components? I am planning to get these on ebay but one by one is a pain. Do you know of any ebay store that is large enough so I can get all very common diode, regs, caps, etc.... within a single order ?

 

Online TheSteve

  • Supporter
  • ****
  • Posts: 3271
  • Country: ca
  • Living the Dream
Re: Improving glow plug driver design for RC plane radial engine
« Reply #23 on: February 25, 2016, 06:52:08 am »
You can always check Digikey or Mouser and see what shipping would be.

btw, beautiful engine!
VE7FM
 
The following users thanked this post: kikinou

Offline kikinou

  • Contributor
  • Posts: 31
  • Country: fr
Re: Improving glow plug driver design for RC plane radial engine
« Reply #24 on: February 27, 2016, 10:11:30 pm »
Thank you very much. Digikey shipping is free over 65 EUR, and the cart is 112 EUR for now. Tax might be another hidden cost though, I don't know. Thanks TheSteve for letting me check.

DIY is surprisingly very expensive. Am I doing something wrong ?

Attached is the updated schematic. I used the ON/OFF pin of the relay, as suggested.

I plan to order all components and get the circuit working on a breadboard before routing a PCB. I feel like it is the right thing to do, but I am just guessing.
 
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf