The VeeCU

[WesRagle]

Hi Guys,

Lots of distractions lately but ..., some progress.

I ported the software from the bread board to the prototype. I hooked up the prototype to Frankenpanel.



Frankenpanel allows me to drive either LEDs, for software development, or actual coils and injectors, for hardware testing.

I chose the enclosure I did largely to keep cost in check. One price paid for using a plastic enclosure is reduced heat dissipation. So, I've been worried about how hot it was going to get in the enclosure while operating. I couldn't resist running a quick check just to see. I ran the unit very hard to see if I had a problem or not.

14 Volts
3 Ohm Coils
Constant Dwell of 12 mSec.
4000 RPM
65.9 Injector Duty Cycle
10.16 GPH calculated Fuel Flow. (I had MAP and Mixture way up).
4.3 Amps Total.
20 Deg. C. Temperature rise internal.

Every component of the VeeCU is rated for at least 120 Deg. C operation. Seems like it's going to be just fine.

I have some thermally conductive glue on order. It should be here tomorrow. I'll glue a very small thermocouple to the ignition switching transistors and see how hot they are getting. They are rated for 175 Deg. C. operation but I don't want to see them anywhere near that.

Much more testing to follow.

Wes

[WesRagle]

Hi Guys,

I received the thermal conductive glue. A long time ago a guy showed me a method for measuring the temperature of an IC. First you get some high quality TC hookup wire. Then form a TC by stripping the wire and clipping off all but one strand of the wire. Next, twist the two strands together tightly and clip off all but the very tip. Put some thermally conductive "stuff" on the IC in question and stick the tip of the wire into it all the way to the surface of the IC. Make sure the strands do not touch outside the thermally conductive stuff.

That's all there is to it. A very responsive TC with very little heat sink effect.



I have an old EGT prototype gauge that I rescued from the trash after a lab clean up at work some 25 years ago. Still works and is very accurate.

To test the ignition transistor temperature rise I decided to go for broke. I hooked up a 3 Ohm Dynatek coil and set the RPMs to 60. At 60 RPM no dwell control is applied/possible so the coil is powered virtually 100% of the time. I intend to rate the VeeCU for operation over a voltage range of 6 to 20 volts so I set the voltage to 20.

I got some results I didn't expect.

At the beginning of the test:
IC Temperature 22 Deg. C.
Coil Resistance 3.0 Ohms.
RPM 60.
Voltage 20.
Current 6 Amps.

During the test peak IC temperature was 90 Deg. C.

At the end of the test:
IC Temperature 81 Deg. C.
Coil Resistance Approximately 5.0 Ohms
Current 3.4 Amps.

I didn't expect that! I have been assured by Dynatek that the coil does not contain a ballast resistor. So I asked Cpt. Google what the "thermal coefficient of resistance for copper" is. It's about 0.4 percent per centigrade degree. You can run the numbers but an increase from 3.0 to approximately 5.0 ohms requires an increase of the order of 200 Deg. C. I knew the outside of the coils got hot but I wouldn't have suspected the primary winding got that hot. I don't know, but that's what it seems like. I've duplicated the results with two different makes of 3 Ohm coils. The 5 Ohm is probably a little low. I couldn't get the power off and the meter leads on the coil fast enough. Once the leads were in place you could see the resistance of the primary drop at about 0.1 Ohm per second. Like I said, not what I expected.

Here are a some more "stabilized" measurements.

RPM 4000
Voltage 20
IC Temp 70
Current 2.15 Amps.

RPM 3000
Voltage 14
IC Temp 52
Current 1.3 Amps

I reduced the Voltage to 6 volts to check the spark with a hot coil. At 6 volts the spark duration was 0.75 mSec.

I also checked a hot coil at idle and 14 volts.

RPM 950
Voltage 14
IC Temp 36
Current 0.45 Amps
Injector was about 0.14 A (MAP 500 mBar, Duty 5.5, Calculated FF .816 GPH)

So for a hot coil at Idle total current draw = (2 x 0.45) + 0.14
or about 1.1 Amp.

I'll test a 5 Ohm coil soon.

Live and Learn,

Wes

[WesRagle]

Hi Guys,

In my last post I got so wrapped up in stressing the system I forgot to answer the basic questions that were on my mind. That is, under normal operating conditions how much power does the system require. So, I got up this morning and took some data under normal operating conditions with cool coils. I also got around to measuring the current requirements of the VeeCU itself.

Results:

0.05 Amps - VeeCU current requirements over 6 to 20 Volts input power range.

Power Requirements at Idle

14 Volts - System Voltage
20 Deg. C. - Initial transistor temperature
31 Deg. C. Final transistor temperature.
950 RPM - Engine Speed
0.5 Amps - Current draw per coil.
0.14 Amps - Injector Current (At FF of 0.825 GPH)
3.0 Ohms - Initial Coil Resistance
3.2 Ohms - Final Coil Resistance

Total Current (2 x 0.5) + 0.14 + 0.05 = 1.19

So, Idle current requirement is about 1.2 Amps + Fuel Pump Current.

Power Requirements at High Cruise

14 Volts - System Voltage
31 Deg. C. - Initial transistor temperature
52 Deg. C. Final transistor temperature.
3400 RPM - Engine Speed
1.7 Amps - Current draw per coil.
0.47 Amps - Injector Current (At FF of 4.15 GPH)
3.2 Ohms - Initial Coil Resistance
3.7 Ohms - Final Coil Resistance

Total Current (2 x 1.7) + 0.47 + 0.05 = 3.92

So, Cruise current requirement is about 4 Amps + Fuel Pump Current.

Edit: I tested the Dynatek 5 Ohm Coil.

Power Requirements at Idle

14 Volts - System Voltage
23 Deg. C. - Initial transistor temperature
30 Deg. C. Final transistor temperature.
950 RPM - Engine Speed
0.3 Amps - Current draw per coil.
0.14 Amps - Injector Current (At FF of 0.825 GPH)
6.0 Ohms - Initial Coil Resistance
6.0 Ohms - Final Coil Resistance

Total Current (2 x 0.3) + 0.14 + 0.05 = 0.79

So, Idle current requirement is about 0.8 Amps + Fuel Pump Current.

Power Requirements at High Cruise

14 Volts - System Voltage
30 Deg. C. - Initial transistor temperature
43 Deg. C. Final transistor temperature.
3400 RPM - Engine Speed
1.0 Amps - Current draw per coil.
0.47 Amps - Injector Current (At FF of 4.15 GPH)
6.0 Ohms - Initial Coil Resistance
6.3 Ohms - Final Coil Resistance

Total Current (2 x 1.0) + 0.47 + 0.05 = 2.52

So, Cruise current requirement is about 2.6 Amps + Fuel Pump Current.

I did look at the spark duration at three voltages for the 5 Ohm Coil.

14 Volts -> 1.25 mSec
8 Volts -> 0.75 mSec
6 Volts -> 0.50 mSec

I have the VeeCU power broken out separately from the Injector and Ignition power for testing. I could reduce the coil voltage below 5 volts before the plugs stopped firing. That is a couple of volts lower that what I measured on the bread board. That's probably due to better switching performance of the IGBTs (ignition transistors) used in the VeeCU vs the evaluation board used in the bread board.

Thanks,

Wes

[WesRagle]

Hi Guys,

I know I'm obsessing about component temperature, but I think for good reason.

I've tested three different coils. Two 3 Ohm and a 5 Ohm. I have one more 3 Ohm to test but I suspect it will be much like the other two.

I've run the system at 6 Volts, at 20 Volts, and everywhere in between. I've measured the switching transistors temperatures at cranking (no dwell control), at standard cruise, and at 4000 RPM at the afore mentioned voltages. I've driven the injectors at the same time from Idle to 80% duty ( > 10 GPH). I've had the case on and the case off. There were times when I was pushing over 100 Watts through the coils and injectors.

I've calculated and extrapolated until I'm blue in the face and I've come to the conclusion that it wouldn't be smart to mount the unit forward of the firewall. There were margins, but not the kind of margins I like with this type device. I'm a little disappointed, but not defeated.

There was one other factor favoring a cabin side mount. I used an automotive enclosure. The insulation for standard automotive wire is much thicker than tefzel wire. To get tefzel wire in the minimum diameter to guarantee a seal would require 16 Gage wire. That's not practical.

On the bright side, with the ambient operating temperature limited to 50 Deg. C. I have comfortable margins. Only under extreme conditions will the ignition transistors exceed 100 Deg. C. The transistors and the PC board ( FR4 tg rated to 170) are rated for 170 Deg. C., so I'm comfortable with that. And, I can use appropriately sized wire.

So, I won't consider this a defeat, I'll consider it a negotiation ;)

Now I can get on with my life, and software development.

Forever Forward,

Wes

[Bryan Cotton]

Wes,
What if you tapped the baffle for a cooling tube and gave the VCU a supply of fresh air?

[WesRagle]

HI Bryan,

I think that will work. However, there is the issue of heat soaking after shutdown followed by a restart. Having said that, that's exactly what I intend to do for the Onex installation. I have to have easy access to apply any software tweaks that might be needed.

I'll modify an enclosure so I can leave the programming cable hanging out. I'll cover the VeeCU with a small cabinet attached to the firewall with piano hinge and run a small blast tube to the cabinet.

That should get me through initial flight tests.

Wes

[Dave Wolfe]

Fwiw every aircraft ECU Im familiar with recommends placing the ECU in the cabin or aft of the firewall. There is at least one experimental that crashed (I think fatal!) due to placement ahead of firewall against manufacturer recommendations and this was with a rotax or UL power I believe.

I think you are talking about the ECU and not other items.

[WaiexB22]

I have an ecu. I bolted mine to the underside of the glare shield. It sits between the panel and fuel tank. I put a small fan there to hopefully keep it cool. The outer case of mine is a heat sink as well. Maybe consider attaching a heat sink to the enclosure?

[WesRagle]

Hey Robert,

Maybe consider attaching a heat sink to the enclosure?

The ECU generates right at 1 Watt of waste power/heat per amp of injector/ignition drive.

If you are interested, here is the data sheet for the IGBT used: https://www.mouser.com/datasheet/2/308/1/FGD3245G2_F085_D-3150312.pdf

If you look at the graphs starting at the bottom of page four you can see that V(CE) is about 1 Volt at the current required by the ignition coils. Luckily decreasing as the part heats up. That's where the waist heat is generated. So, about 1 Watt per Amp. If you look at the current at high cruise with the 3 Ohm coil you'll see it's about 4 amps. So, only 4 Watts to dissipate.

The Design Engineer working on the project did an excellent job of placing a lot of copper around the IGBTs and the "Engine Control IC" to use the PCB to draw heat away from the high power parts. With the PCB lying open on the bench it only heats up 10 Deg. C. under heavy load. With the case on, add another 10 Deg. All of this to say that with the VeeCU mounted in the cabin there is no real need for extra heat sinking. The case gets slightly warm to the touch, but that's about it.

Wes

[WesRagle]

Hi Guys,

Today was the day to invade the kitchen.



I glued a TC to one of the switching transistors and inserted a temperature probe in the rear of the enclosure to monitor internal temperature.



Unceremoniously the VeeCU was tossed into the oven.



The oven was turned to the lowest temperature it would accept, 170 Deg. F. I let the temperature stabilize for about 30 minutes to see what temperature the "bang bang" oven controller decided to control at.



Both the internal and transistor temp stabilized at about 82 Deg. C. I had checked the accuracy of the built in core temperature monitor of the MPU. It consistently reads higher than ambient at power on. It's read 89 Deg. C. while the other two temps read 82. That's what I have to work with so here I go.

I ran the VeeCU at several different power settings for 20 min. per test and recorded the data below.

I'm happy with the results. Under the extreme conditions of 20 Volts and 4000 RPM the transistors stayed below 150 Deg. C (device rated to 175 Deg. C) . All this while operating in an ambient environment well above 50 Deg. C (82 Deg. C).

BTW, I consider 150 Deg. C. as the ceiling for the transistor temp. It is considered good practice to
keep components at least 20 Deg. C. below the rating of the PCB (which is rated at 170).

So now I am actually comfortable that at 50 Deg. C. ambient I have enough safety margin.

I was also pleased with how well the plastic case shed heat. The internal temp of the box dropped rapidly when I opened the oven door.

Wes

=================================================

Test #1

Start

82 Deg C Internal Temp, 82 Deg C Transistor Temp, 89 Deg C CPU Temp
3.3 Amp Current
3400 RPM Engine Speed
14 V System Voltage
29.2 Injector Duty

End

87.2 Deg C Internal Temp, 119 Deg C Transistor Temp, 99 Deg C CPU Temp
3.1 Amp Current

Test #2

Increase Voltage to 20 Volts.

Start

87.2 Deg C Internal Temp, 119 Deg C Transistor Temp, 99 Deg C CPU Temp
4.5 Amp Current

End

90.3 Deg C Internal Temp, 136 Deg C Transistor Temp, 105 Deg C CPU Temp
4.2 Amp Current

Test #3

Increase RPM to 4000. Decrease Voltage to 14

Start

90.3 Deg C Internal Temp, 136 Deg C Transistor Temp, 105 Deg C CPU Temp
3.46 Amp Current
4000 RPM Engine Speed
14 V System Voltage
34 Injector Duty

End

89.3 Deg C Internal Temp, 124 Deg C Transistor Temp, 102 Deg C CPU Temp
3.51 Amp Current

Test #4

Increase voltage to 20 while keeping RPM at 4000.

Start

Initial current 5.4 amps.

End

Final current 4.74
91.2 Deg C Internal Temp, 141 Deg C Transistor Temp, 106 Deg C CPU Temp

=======================================================