Close

Testing - Round 2

A project log for Goliath - A Gas Powered Quadcopter

A BIG Gas Powered Quadcopter

peter-mccloudPeter McCloud 10/07/2014 at 12:241 Comment

10/15 UPDATE: Good 3rd Gas Engine test! See bottom for details...

Ok, so things are finally in a good state to start doing the second round of testing. As I did before, I'll be updating this project log as I go, instead of making multiple logs. I'm trying to be cautious so that I don't have a repeat of breaking propellers so I'm breaking down the testing into smaller increments to make sure everything is good for the final hover test. I've already done two tests previously, so the next tests will be:

So it took a while to get here because of the redesign of the belt tensioners, redoing the prop mounts and supports, etc. Last week I was in the process of making the final adjustments and alignments when I realized there was going to be an issue with the vehicle due to some of the changes I had made. One change I forgot to document was that I had flipped the propeller pulleys over. This is important because of the bushing attached to the pulley. By flipping the pulley over, it added a half inch of additional clearance between the propeller and the belts.

What I didn't take into account is that this reduces the clearances above the prop. The prop would fit, but the prop tips will deflect upwards and I didn't want to risk destroying another propeller. There was lots of clearance beneath the pulley, but it was as low as it could go do to the center beams and engine shaft length. You can see this a little in the next photo.

The solution I came up consisted of several parts. The first part was to replace the lower center beams which stood 2 3/4" tall with 1 1/2" tall beams. These aren't carrying the load of the engine like the ones above, I had just chosen to match these beams to the ones above them. This gave an additional 1 1/4" to move the belts downwards, but the pulley was already as low as it could go on the shaft. The next part was to lower the upper center beams with the engine attached. This was done by replacing the bolts one at a time with all-thread and then turning the nuts to slowly lower the assembly. The all threads and nuts act as stand offs. The engine was lowered 1" and the final result is shown below

This modification also gets the cg lower which will hopefully make control that much easier.

Test 3 - Single Sided Belt only with Dummy Props, Starter Only

This test was done on 10/6 and it went well. To get a better view of the belts, a second camera was used, attached to the frame. The belt didn't appear to oscillate.

Below is where the second camera was mounted. I simply put the camera on the frame and screwed the mini tripod onto it through one of the slots.

It shouldn't take too long to complete the belt guards and run the next test.

Test 4 - Single Sided Belts only with Dummy Props and Belt Guards, Starter Only

The last of the upgrades to address the issues from the previous round of testing are the belt guards. These are intended to keep the belt from being able to interfere with the propeller. They were constructed from a five inch wide strip of 24 gauge galvanized steel bent to shape. Last night (10/7) I completed making the first 2 and bolted them to the frame.

On the upper surface, 1 one inch wide strip was folded over so that if the belt does hit the belt guard, there isn't a sharp edge that could cut the belt. The picture below shows the clearances. I had originally intended to have the belt guard in between the propeller and the belt around the tip, but with the lack of clearance, it seemed having it just beyond the propeller was more practical.

Things are all set for the next test. When I get home tonight I'll set up the test and make sure these are going to work okay.

10/8 8:30 UPDATE

So the test went well. Even with the tight clearance shown there was no signs of the belt impacting. I moved the second camera to beneath one of the active rotors to get a view of the belt guards. It's a bit shaky, but looks cool.


Next steps are getting the throttle servo connected and the new gas tank attached and I'll run the gas motor with belt hardware as is.

Test 5 - Single Sided Belts only with Dummy Props and Belt Guards, Gas Engine

10/10 9:30 PM UPDATE

The throttle linkage is now hooked up. It took a couple of iterations to get it right. The linkage is made from 4-40 all-thread rod with Ball and Socket Rod ends. The servo is a standard RC servo and the servo mount is made from a foam block coated with epoxy to stiffen it up.

The fuel tank is a 50 oz (1500 cc) tank for large RC planes. I've mounted it inside the frame, on the opposite side from the pulleys.

All that's left to do the next gas engine test is to straighten up some wires and fill the gas tank. So I plan to conduct the test tomorrow (10/11).

10/11 2:15 PM UPDATE

Today's test was successful! After having some initial difficulty getting the engine started, it came to life. The problem seemed to be that it needed the choke to get started up. There's no servo on the choke, so if it continues to need the choke to get started, then I'll need to think about adding a servo.

The test started at about half throttle and after a few seconds, brought it down to 25% throttle. It was run for a good 20 seconds and the belts and the dummy props seemed to hold steady. A visual inspection afterwards didn't turned up anything of note. The audio really doesn't do it justice, it's loud and intimidating. (For those safety conscious people, we are ducked down below a chain link fence during the test).

With this test out of the way, the next step is to hook the second belt back up and make sure the system as a whole behaves before added on the real propellers.

Test 6 - Both Belts with Dummy Props and Belt Guards, Gas Engine

10/13 3:30 PM UPDATE

Well if you haven't seen today's announcement for The Hackaday Prize, [spoiler alert] Goliath wasn't selected. The judges have picked some great projects and the best of luck to the remaining contenders.

Meanwhile, progress with Goliath is continued today with Test #6. This was the first test with both of the belts connected. During the test the gas engine was run for about 30 seconds. Nothing seemed amiss during the test, but reviewing the footage it seems that the double sided belt is still moving around quite a bit. It's also interesting to see some of the harmonics present, particularly with the engine, moving about a bit.

I'm probably going to adjust the belt tensioners and repeat this test to see if the belt oscillations can be further reduced. There are thunderstorms going around today and I was lucky to get this test in before it started raining, so we'll see when I'm able to run the next test. I'm also starting to set up the test rig for the hover test.

10/13 10:00 PM UPDATE

I did get to test a second time today, repeating the conditions for Test #6. I had tightened up the belt tensioner for the double sided belt to try to reduce the oscillations. There wasn't much improvement, but I think I may have found the culprit. The prop pulleys were a bit higher than the main drive pulley so that they weren't alinged as well as they should be. I think the oscillations could be triggered by the belt riding up the pulley, and then bouncing against the other belt.

The total test time was one minute. With the longer testing, the paint on the exhaust is starting to off-gas a bit. The other thing that was apparent was that the battery mount is inadequate. It's currently slung underneath a wide piece of plate metal and it was bouncing up and down visibly from where I was viewing the test, (but not seen well from the 2 other cameras). I'll need to add a support underneath it so that It's mounted more securely. I was lucky I stopped the test when I did because a bolt was rubbing against the battery box and could have worn through it.

The prop pulleys have already been adjusted so once the battery mount is fixed, it'll be time to test some more. Even if the belt problems are fixed, it may be wise to increase the test times to make sure all the bugs are worked out before adding the propellers.

10/15 1:00 PM UPDATE

This morning I did another repeat of the Test 6 conditions. New battery plate held together nicely and the test was for 1 1/2 minutes, bringing the total run time on the engine up to 3 1/2 minutes. Everything held together, but the one belt still seems to be bouncing around. I could probably move forward with some risk, but I'd like to get the belt system rock solid before I move on to the hover testing. I've got some additional ideas, but it'll mean getting some additional parts and standing down from testing for a week or so. I'll talk about them and a Round 2 testing conclusions in my next project log.

Discussions

zakqwy wrote 10/09/2014 at 02:27 point
Looking good! Excited to see tests with gas power. Should be terrifying and awesome.

  Are you sure? yes | no