EDF Rocketry

[Richard Dierking]

there are a lot better fans out there. Also could have used NACA ducts for the inlets.

Yes, I like the idea of using ducts, but it does add a bit of complexity and mass. I think lightweight fiberglass ducts would be a great idea on a future project. Ducts on EDF aircraft like scale fighter jets are commonly used.
Not sure about better fans. I know there are lots of manufacturers and suppliers of EDF's for RC aircraft. I needed something that was relatively cheap and easily available. I chose these because they were in stock on Amazon and after testing the first one, they preformed as advertised. Please keep in mind that I also have several other projects using EDF's and I've purchased a bunch of them - more than 10 now, along with ESC's and batteries.

 

[Richard Dierking]

I mistook the large item near the top of the tube for the battery. Would you identify some components to help discussion?

Yes, the battery is shown in the photo of the nose. The dark area at the very tip is just where I epoxied the aluminum all thread that secures the battery holder to the nose.
In the photo of the forward end of the airframe, you see the parachute deployment compartment (half-cylinder). The parachute deploys out a side hatch. Then, there's a small 1 cell LiPo for the BP charge. The control is a RC switch seen on the side of the deployment compartment and the other component you thought may be the ESC is actually the 4 channel RC receiver. I'll create a simple drawing with some additional photos.

Hey, form follows function.

Yeah, but I want it to still resemble a traditional amateur rocket. Someone on YouTube made a "rocket" mostly to land like Falcon 9, but it's a large tube with an open forward end. It lands pretty good, but it's really not a rocket.
I would like to have parallel motors on a airframe with an EDF. Launch with the solid motors and land with the EDF. We'll see.
More about roll torque a bit later, but I'm probably going to actually hold the EDF rocket on the rail during the "launch."
I would ask the guy in your photo for advice, but didn't Kirk kill him?

 

[Rob Campbell]

I would ask the guy in your photo for advice, but didn't Kirk kill him?

And the "probe" used in the movie was built by Public Missiles.

 

[Richard Dierking]

I'd be very concerned about flying something with such a low thrust to weight ratio without some sort of active stability.

I need to make my final case before the jury goes out on that one. Again, I'm going to run Rocksim on the design. First I need some static test numbers to create a RASP file.

But, please consider that the motor is controlled by the RC transmitter (me). If it started to go south, I could cut the throttle (switch). Also, the parachute deployment is RC controlled, so I could activate the deployment at any time during the flight.
The RC control has a safety feature that if the signal is lost, the motor cuts.
Btw, I could even restart the EDF on the final part of the descent to slow the landing.

But, lots of questions and concerns to follow. The idea is to hash out some of the problems for something like this. And, it's a classical engineering approach.

 

[jqavins]

I would ask the guy in your photo for advice, but didn't Kirk kill him?

No, Kirk kept him busy while Picard killed him.

And the "probe" used in the movie was built by Public Missiles.

And Paramount won't let let them kit it (rasin fracka rat bastards).

 

[Richard Dierking]

No, Kirk kept him busy while Picard killed him.

See, no matter what happens with the EDF rocket, you will learn something in this thread.
I did some testing this morning. Still don't have the 5 Kg load cell in, but the 50 Kg works OK.
First, some photos... I'm going to split this into 2 posts.

The 80 amp ESC and forward of that, the parachute deployment bay.



Static test of the EDF rocket. If you do something like this, wear safety glasses! It's like running a leaf blower in your garage. Also, must be careful of not allowing any debris into the intake.

 

[Rob Campbell]

And Paramount won't let let them kit it (rasin fracka rat bastards).

Paramount's Fan Film Guidelines essentially shut down fan made Star Trek. While most fan made Star Trek is unwatchable, it improved to the point Paramount felt threatened. And some of the newest fan Star Trek is better than what Paramount is making today. The guidelines essentially killed Axanar, which was their intent, and Star Trek Continues cut two episodes and ended their series early to avoid a lawsuit.

 

[Richard Dierking]

Here's some info on the static tests. The first two used a 64 mm EDF (same I'm using on the lifting body). The fan is in a simple housing with an exhaust reduction (photo).

D2836, 64 mm EDF, 3900 Kv, 3S, 11 blade fan.
Using Turnigy 3 cell 2200 mAh 40-50C battery: 0.93 Kg force
Using Avian Smart 3 cell 1300 mAh 30C battery: 0.90 Kg force

Next, tested the D2842, 70 mm EDF, 3400 Kv, 4S, 12 blade fan. In the airframe.
Using Ovonic 4 cell 1300 mAh 80C battery: 1.12 Kg force

I don't like the results I'm seeing from the 70mm EDF in the rocket. I'm suspicious of the E-Flite ESC. I've tried several times to calibrate it and I'm still getting some dead band when I start it up. Also, I'm not sure the lower thrust than advertised for this EDF (1.81 Kg thrust) is an issue with the intake "dirty air." I'm pulling the motor and ESC and I'll test it outside the airframe.

 

[neil_w]

If you're willing to share, I would be very interested to know what all the EDF componentry costs for a system like this.

 

[Richard Dierking]

f you're willing to share, I would be very interested to know what all the EDF componentry costs for a system like this.

Of course, I would be happy to share all the info. I'm planning on actually providing some conclusions after the maiden flight and will provide all the components with suppliers and cost.
Even if it's not this rocket, I think something like this has a lot of potential. Maybe a 3D printed design? Then, you could obtain a motor and fan blade kit and install it in the 3D mount. This would save mass and you could switch out different motors and fan blades. The battery and ESC conductors could be built into the airframe that would have ducts. What a cool thing that would be!

I think of all the events I participated in like Space Day and Sally Ride festivals to name a few and was never able to launch a rocket; only static test motors. Before turning down my permit, I actually had a fire inspector ask me "would fire come out of the rocket?" I would have loved to say, "well, no."
Btw, the Academy of Model Aeronautics (AMA) that I'm also a member of, covers rocketry. But, this is new territory. We will see.

 

[jqavins]

(I'm using wire data from here.)

So, I just did a little math based on the 14 AWG wire (2.525 mΩ/ft) from the ESC to the motor, the motor's K_V (3400 RPM/V) and the motor's current (75A). That wire is costing you about 1288 RPM per foot, or just over 100 RPM per inch between the ESC and the motor one way. So I can see why you're sweating every inch.

But, for 75A continuous current, you should be using 8 AWG (actually good only for 73A, but who's counting?) and that would make the loss only about 27 RPM/in one way. Now, I don't know how much thrust that added fan speed gets you, but I bet you a dollar it's more than the added wire weight, by quite a bit. That's without considering the air impingement, that might be hurting the thrust, and certainly could be reduced by allowing a little more length to keep the middle of the tube clear.

 

[Richard Dierking]

Thank you! Wow!
I also wonder how extra length of wire between the ESC and motor could affect the timing?

Funny, this started out as just a "quick - fun" project because I was working on the EDF for the lifting body project. LOL

I just pulled the motor and ESC and it went fairly well. When I static test, I'm going to keep the wiring the same. I'll also get some temperatures.
The complete airframe (including the nose) minus the EDF, ESC, and battery weigh 266 grams.

 

[jqavins]

Timing? The signal speed down the wires will be a substantial fraction of the speed of light, so better (i.e. less) than 10 ns/m. So never mind the timing.

Also, between posts I calculated what I could about the weight. The copper weight for the 8 AWG wire is about 46 grams per foot one way. (Yes, I know, grams per foot. So, I'm a sinner.) That's small if not tiny compared to your 266 gram airframe, but when the EDF, ESC, and battery are added, the copper weight will be tiny indeed. I don't have enough information about the insulation to go on.

 

[Richard Dierking]

It's cool about the units. Originally, the rocket was going to be longer but I started counting grams and determined the 3" airframe was 7 grams per inch (see, I do it too).

I think the key to the best performance will be a 3S/4S motor and 4 cell battery with a lower current ESC. That way, the mass of the motor and ESC will be less but the higher voltage will drive the RPM higher resulting in more thrust. Because the current will be less, the battery doesn't need the high C factor and that will lower its mass too. So, it will be like 50 amps not 75. Hope that makes sense.

For anyone else trying this, it's important to static test the system first. If there's an issue with the current capacity anywhere, the motor will surge at maximum throttle. I should have done that first with the 80 amp ESC and 70 mm EDF.

 

[georgegassaway]

As far as roll torque, it's not just a momentary thing for a throttle change. Full size helicopters with a single set of rotors, will spin around and around without the tail rotor operating. So, the rocket will spin even if you go max throttle for liftoff and never change throttle. Could try to reduce that spin by using some angled vanes downstream of the turbine blades, to produce an opposing roll force. But that would reduce thrust. Adding roll tabs to the fins would not affect the thrust, but the effectiveness would vary greatly with the model's flight velocity.

Let me ask this. What kind of flight profile are you planning for? To climb under thrust for say 10 seconds? I think it will start getting off-vertical sort of quickly, although the spinning would help reduce that a bit. So, it will get horizontal under thrust after awhile.

I figure you plan for R/C ejection or altimeter-sensing ejection. Not something based on timing since "liftoff", as the flight path may get horizontal quickly, or it may stay keep pointed "up" for a good while. And also, to shut the EDF off.

 

[Richard Dierking]

I figure you plan for R/C ejection or altimeter-sensing ejection. Not something based on timing since "liftoff", as the flight path may get horizontal quickly, or it may stay keep pointed "up" for a good while. And also, to shut the EDF off.

The EDF will either shutdown when I trip the throttle cut switch, or if the receiver losses signal. This is a RC safety feature.
Deployment uses a RC switch. I can also create a failsafe so the receiver will activate the switch if the signal is lost.
Not sure which ESC and battery I'll be using for the flight. If it's the Spektrum "Smart" equipment I'll be able to monitor battery condition via telemetry to the transmitter.

I'm waiting for the static test data and the possibility of Rocksim flight simulation before developing a flight plan. The 1300 mAh battery could develop full thrust for close to 45 seconds and hopefully not draw the battery down too much. This btw is the reason I have a separate battery for the deployment charge.

As far as roll torque, it's not just a momentary thing for a throttle change.

Since it would be useful to know something about the rotation before the flight, I could rig my spin table up used mostly for axial balance. So, no power to it, and have the EDF thrust up with the ESC and battery counter-balanced on the table. I can read the RPM of the table.
Sound like a good plan?

 

[jqavins]

I don't see how this would work unless you know the friction characteristics of the table's bearing. You're looking for the steady state torque of the motor at full speed, so the table's speed would be steady but that doesn't tell you the reaction torque unless you know the torque in the bearing that cancels it.

Another way would be to hold the table still and measure the torque needed to do that. Attach a fish scale to the rim of the table and the motor to the table's center. (Dead center isn't important, but keeping the scale level is.) Spin up the motor while holding the other end of the scale. Once the scale's reading settles, multiply the weight by the table's radius and that's the torque.

 

[Richard Dierking]

Sorry if I gave the impression that I wished to quantify the spin torque. I was thinking that perhaps if we could get an idea of the potential to spin, and that would be adequate.
How about this; I set the table on its side and it took 3 grams to rotate the 42 cm diameter plate. So, not much. I got the bearing from a recycle place and removed it from an industrial sorting machine. Anyway, it's a very high-quality bearing, and the downward force (about 2 Kg) shouldn't affect it too much. I'll run a "simple" test this afternoon.

Here's some interesting data from a couple of static tests I just completed. I used the same EDF, ESC, and battery that was in the rocket. I tried two of the exhaust tail cones with the same diameters, just different lengths (photo).

Both tests 31 seconds long and going to full throttle immediately. Sampling every second.

1" long exhaust (this tail cone was used on the rocket)
Avg 1460 grams
Min 1387 grams
Max 1657 grams

1.5" long exhaust
Avg 1433 grams
Min 1356 grams
Max 1627 grams

Interesting! I want to try the second test with the 1.5" again with the same battery that I used for the 1" tail cone this afternoon. First, it looks like it's loosing about 0.36 Kg thrust in the rocket.
Next, it looks like the longer tail cone doesn't improve the thrust.

 

[jqavins]

So what about a 1/2" tail cone, or none at all?

 

[Richard Dierking]

That's so funny, I was afraid you would ask that. Darn control stuff. Remember, I called this classical engineering but it's really casual engineering or perhaps hobby engineering. ;-) This is when you work on stuff while you are having a beer.

OK, I'll run without a tail cone this afternoon, this time.
Also, please remember that the tail cone looks good and that's a good reason to have it.

Oh boy, just started thinking about base drag. I need to get this stuff out of my head! EDF rocketry could be a hobby in itself.

 

[georgegassaway]

I wrote:
>>>
As far as roll torque, it's not just a momentary thing for a throttle change.
<<<<

Since it would be useful to know something about the rotation before the flight, I could rig my spin table up used mostly for axial balance. So, no power to it, and have the EDF thrust up with the ESC and battery counter-balanced on the table. I can read the RPM of the table.
Sound like a good plan?

As the turbine blades rotate into the air, and deflect the air, there is a torque created from that alone. The faster the RPM, the greater the torque that is created by the blade's "twisting" into the air (Crude, I cant think of a better term for it). As I referred to before, a helicopter with a single rotor system, has that aerodynamically created torque which requires a tail rotor to create a counter-torque effect. You have probably seen videos of copters crashing after the tail rotor failed, causing the copter to spin around and around.

Your spin table would show you something, but unless it has air bearings, it won't show you how bad the spin can be due to rotational friction. Would be more effective, I think, to suspend the rocket by the tip of the nose, attached to a ball bearing fishing swivel. No issues of a spin table mass, or thrust impingement on to the table affecting it. Although if the rocket is not laterally balanced well, it might have a wobble rather than spin as smooth at a lathe.

 

[jqavins]

As the turbine blades rotate into the air, and deflect the air, there is a torque created from that alone. The faster the RPM, the greater the torque that is created by the blade's "twisting" into the air (Crude, I cant think of a better term for it).

Yes, good points that I did not consider.

I think, to suspend the rocket by the tip of the nose, attached to a ball bearing fishing swivel. No issues of a spin table mass, or thrust impingement on to the table affecting it. Although if the rocket is not laterally balanced well, it might have a wobble rather than spin as smooth at a lathe.

To partially deal with the imbalance issue, one might suspend the rocket between to such swivels, one attached by a short string to the tip of the nose and the other to the center of the aft end (somehow). Another method would be to attach the rocket to the spin table, supported above the table on a simple open truss to allow air flow through the EDF, then use the fish scale setup.

And this is naught but an interesting exercise in the mechanics of the vehicle and the test, as I don't think Richard has any real need to - on intention of - going to such lengths.

 

[Richard Dierking]

And this is naught but an interesting exercise in the mechanics of the vehicle and the test, as I don't think Richard has any real need to - on intention of - going to such lengths.

Thank you. Perhaps if someone was really pursuing a EDF rocket design they would need some additional study. I'm more interested in judging the safety of launching my project. Again, please keep in mind that I can cut the throttle at any time including immediately after launch.

There was a bit of a delay in conducting another static test of the tail cones. I just tested the longer (1.5") tailbone again but this time with the 5Kg load cell I received yesterday afternoon. (It takes some work to connect and calibrate a new load cell.)

Again, a 31 second long test at full throttle. Sampling every second.
Average 1.47 Kg
Maximum 1.65 Kg
Minimum 1.44 Kg

Compares favorably with the original results:

1.5" long exhaust
Avg 1433 grams
Min 1356 grams
Max 1627 grams

I'll run a test with no tail cone later. I want to move the EDF over to the spin table. Depending on what happens, I might create a YouTube video of the test.

 

[Richard Dierking]

Here's the spin table set up. The video link follows.

 

[Richard Dierking]

I'm working on addressing this roll issue. I already conducted one this this morning, but I believe the control surfaces (just forward of the tail cone on the inside) were too small. So, I got a negative result. I'll increase the surfaces and will try again this afternoon.

I might even have to increase the control surfaces from two to four. That would be adding two more 3.7 gram servos and the associated mass.

If I can RC control the servos and their control surfaces, then I will work on active roll control. Then, I will static test to see how much adding the control surfaces cost in thrust.

 

[Richard Dierking]

Successful 2nd test using two internal control surfaces. Here's the video:


Next, I need to create some thinner surfaces and try these before conducting a static thrust test to see how much thrust this cost.

Finally, I want to pursue an active control system to automatically control roll.

 

[JohnCoker]

Very cool project; I love the experimentation and look forward to seeing where you go with it.

 

[Richard Dierking]

Thank you John. Wish I had your knowledge and skill at building stuff.

There's a lot to try with this project. I'm doing some more testing today.
I don't think the light-weight servos I received yesterday will do. They are 3.7 gram with plastic gears. I'm going to a 4.8 gram digital servo with metal gears.
I will probably post some thoughts on what I have learned so far after the static tests with the new control surfaces.

 

[Richard Dierking]

Just completed a test with the control surfaces starting in a position to control (stop) the roll. Going to full thrust quickly it rolled on the spin table less than 1/2 rotation.



I need to create another version of the control surfaces. The current version only uses epoxy to hold the surface on the servo arm. I don't trust it, so I need to make something that is secured with 0-80 bolts.

Just a thought from today, I think when I get the rocket complete, I will test launch it at the local park. I need to see how fast I can get it up to full throttle and test the throttle cut first. Just curious how it will launch off the pad now that roll will be controlled.

 

[Richard Dierking]

Did some static testing today with the 70 mm EDF but this time with the 2 control surfaces to control roll. Looks like adding the control surfaces and having them in full deflection reduces the full-throttle thrust 0.16 Kg. Really not that bad, and the rocket will not be rolling excessively.

I've had a suspicion that the 80 amp Eflite ESC might not be providing the full power to the EDF, so I used a Avian 80 amp ESC and 2200 mAh 4 cell battery and found out that the 70 mm EDF is capable of 1.64 Kg thrust, not the 1.46 Kg recorded with the Eflite. So, again reminded that the full-throttle thrust depends on the right combination of EDF, ESC, and battery.

At this point, I'm going to wait for the 3S/4S 64 mm EDF I ordered. I hope that this EDF along with the 60 amp Avian ESC will provide over 1.3 Kg of thrust for the rocket and reduce its total mass.


80 amp Eflite ESC (left) and 60 amp Avian ESC (right).


80 Amp Avian ESC along with 2200 mAh battery on the test stand. 70 mm EDF. With the control surfaces on to control roll, the full-throttle thrust was 1.46 Kg.