HobbyKing.com New Products Flash Sale
Reply
Thread Tools
Old Oct 04, 2012, 06:01 PM
13brv3's Avatar
Navarre, FL
Joined Mar 2002
3,839 Posts
Quote:
Originally Posted by MauiNate View Post
aircraft spruce has great 1/2 aluminum square tubes...
I just got these yesterday, and as I feared, they're quite rounded. They'd be fine if you're going to drill holes in the center, but they won't work for the shallow groove clamps I'm planning to make.

Of course you could weld them, key word being "you"

Cheers,
Rusty (please don't make me weld...)
13brv3 is offline Find More Posts by 13brv3
Site Sponsor
Reply With Quote
Sign up now
to remove ads between posts
Old Oct 04, 2012, 07:08 PM
Registered User
United States, NC, Tryon
Joined Sep 2005
211 Posts
Quote:
I just got these yesterday, and as I feared, they're quite rounded. They'd be fine if you're going to drill holes in the center, but they won't work for the shallow groove clamps I'm planning to make.
So why not use "plates" for the joints instead of "shallow groove clamps"? Wouldn't that be easier to produce for a kit? I was thinking of using plates that are riveted instead of through-bolted so there wouldn't be any possibility of flex in the joint. .060" aluminum would be plenty stiff and light (also cheap!), and it would work with the rounded corner tube from aircraft spruce. Seismics frame looks like it would work with that tube as his holes are drilled more towards the center of the arms. Then again, I'm just the peanut gallery and you're the one doing the work...

Whatever the decision, I'm in for a kit or two when the details get worked out...
soarsimon is offline Find More Posts by soarsimon
Reply With Quote
Old Oct 04, 2012, 08:46 PM
13brv3's Avatar
Navarre, FL
Joined Mar 2002
3,839 Posts
I told Hansen that one of the main problems is that there are too many valid ways to do this

My plan is to not make users drill any holes. The plates will work as clamps, with a shallow groove that holds square cornered tubing well. They will also have holes that line up with the center of the tube, so you can use those a drill guides if you prefer to drill and bolt through the tube.

I've just assembled the first test yoke using 10mm tubing, and I'm pretty darn impressed with how stiff it is in short pieces like this. From what I've seen so far, 10mm is almost certainly going to be the final size tubing for the kit.

I've got most everything I need to finalize this, but it will take a bit more time.

Rusty
13brv3 is offline Find More Posts by 13brv3
Site Sponsor
Reply With Quote
Old Oct 04, 2012, 09:26 PM
Registered User
Hawaii
Joined Feb 2003
5,952 Posts
Quote:
Originally Posted by SDSURFnGLIDE View Post
@seismic...

where are you sourcing the pulleys and belts from? and also what are you using for the bushings/bearings?

just trying to get an idea of what i might need once i get around to putting one of these together...

thank you
Belts and pulleys from here:

http://www.sdp-si.com/

Andrey and Rusty also have them.

For bearings I am using 1/4" X 3/8" flanged bearings that I get from eBay on the arms. For the roll cage I am using 1/4" X 1/2" again from eBay.
SeismicCWave is offline Find More Posts by SeismicCWave
Reply With Quote
Old Oct 04, 2012, 09:27 PM
Registered User
Hawaii
Joined Feb 2003
5,952 Posts
Quote:
Originally Posted by MauiNate View Post
hmm. There was another thread in the multi discussion where someone was using welded square tubes to make a multi frame. Might be worth looking into...

Thanks for the posts!
If know how to weld aluminum it may be interesting to weld the tubes together instead of using plates.
SeismicCWave is offline Find More Posts by SeismicCWave
Reply With Quote
Old Oct 04, 2012, 09:28 PM
Registered User
Hawaii
Joined Feb 2003
5,952 Posts
Quote:
Originally Posted by 13brv3 View Post
I just got these yesterday, and as I feared, they're quite rounded. They'd be fine if you're going to drill holes in the center, but they won't work for the shallow groove clamps I'm planning to make.

Of course you could weld them, key word being "you"

Cheers,
Rusty (please don't make me weld...)
I just got mine today and you are right. They have very rounded corners.
SeismicCWave is offline Find More Posts by SeismicCWave
Reply With Quote
Old Oct 05, 2012, 09:47 AM
Destroyer of G-10
askman's Avatar
United States, OR
Joined Jul 2004
10,085 Posts
actually, they are brazing aluminum, which is ok. welding thin aluminum is very tough without correct gear.
askman is offline Find More Posts by askman
RCG Plus Member
Old Oct 05, 2012, 12:09 PM
Registered User
GGoodrum's Avatar
South Orange County, CA
Joined Mar 2003
8,087 Posts
Just getting caught up in this thread as well...

One comment about servos, I have tried a bunch of "high-end" servos, including the Hitec 7940TH. The problem is that while most have what looks like "perfect" specs, it is what they don't publish that is important. I used to think it was the resolution of the electronics in the servo that is important (i.e. -- 12 bit, vs 11 bit or 10 bit...), but then I found 12-bit Savox servos that didn't work as well as others we've tried. This is because it turns out their deadband rating is higher. It is around .002-.004ms, depending on the model. The Hitecs have an adjustable deadband setting, but at the lowest setting, it is still around .002ms. The MKS servos we've been using have a deadband rating of .0008ms. This is why with these servos, small movements/corrections are buttery smooth.

The second most important factor, in my opinion, is a servos's centering ability. What this means is how well it stays put, at the neutral point. Again, after trying a ton of servos, what we've found is that models with brushless motors do a much better job of centering. This factor you can tell just by listening. If the servos "buzz", they are having a hard time centering and when that happens, you will see small vibrations/movements in the video. Brushless servos make so little noise, it sometimes is hard to tell that everything is powered up.

One of the reasons the HoverFly Gimbal board works so well is because of its 800Hz output rate. The MKS servos we use have internal rates of 333Hz, but to really get the smoothest performance out of them you need to use a gimbal controller output rate that is at least double that. The original Naza/WKM gimbal outputs were designed for analog servos, used in cheap direct-drive gimbals, so the output rate was only 50Hz. Recently, DJI puit in an adjustable output rate in the firmware, but the highest setting is 400Hz. There is a clearly noticeable difference that you can see with your eyes, between 400Hz and 800Hz, if you are using servos like the MKS BLS950s we use, that have really low deadband ratings.

Finally, one other big plus for the HFG board is the ability to adjust the acceleration rate for each channel. This lets you dial in the performance to each setup. I find I need to use different accelerations, for different weight cameras, for instance.

Anyway, once you get through all the mechanical/flex issues, to really get "zen-like" performance, you will need good servos and a good dedicated gimbal controller that is mounted on the isolated side of the gimbal.

-- Gary
GGoodrum is offline Find More Posts by GGoodrum
Reply With Quote
Old Oct 05, 2012, 09:35 PM
The Engineer
prodriverex's Avatar
Singapore
Joined Dec 2009
621 Posts
Wow 800 Hz output rate! That's high end stuff! I better chk mine and see if I have over paid!
prodriverex is offline Find More Posts by prodriverex
RCG Plus Member
Latest blog entry: Gimball Drone
Reply With Quote
Old Oct 05, 2012, 10:04 PM
Creative director
Canada, QC, Montreal
Joined Dec 2011
25 Posts
I've read about that 800Hz around here a few times, and simply can't wrap my head around it... The signal sent to an rc servo being a frequency of pulses with a varying length between 1-2ms, it means that technically the fastest signal sent cant exceed 500hz (otherwise it would be impossible to send 2ms pulses).

So isn't it the internal processed loop that is at 800Hz? And the signal sent would be more around 400...? A very short deadband is indeed very important, and if the pot/encoder can back it, it ultimately defines the resolution of the servo.
alevesque is offline Find More Posts by alevesque
Reply With Quote
Old Oct 05, 2012, 10:24 PM
Destroyer of G-10
askman's Avatar
United States, OR
Joined Jul 2004
10,085 Posts
best servo can take 333hz. due to timing offsets, it is best to have it about 2x that. so 800hz is ok, it will just ignore the rest.
askman is offline Find More Posts by askman
RCG Plus Member
Old Oct 05, 2012, 10:30 PM
The Engineer
prodriverex's Avatar
Singapore
Joined Dec 2009
621 Posts
Heres the parts used in my Triax. Hardly used, or not used yet so sun ask me how's the performance..

This is the Roll servo, Spectrum.. no model printed..


This is the Pan/Yaw servo, Hitec 5645MG.


And the Tilt servo, Hitec too but 5625MG.


Hope the info is useful in your selection..
prodriverex is offline Find More Posts by prodriverex
RCG Plus Member
Latest blog entry: Gimball Drone
Reply With Quote
Old Oct 05, 2012, 10:41 PM
Creative director
Canada, QC, Montreal
Joined Dec 2011
25 Posts
The servo can't ignore it, as 800Hz won't allow for a 2ms pause in the signal, so if you were actually sending an 800Hz signal to an rc servo, and granted the servo electronics would actually take it, you could only send from 1ms (or less, some servo will take as low as 800us and extend their travel) up to 1.25ms - which would significantly limit your travel... It's simple maths, you simply can't send a 2 milliseconds pause more than 500 times per second (and actually, less, since 500Hz would result in a 0% duty cycle)

The 800Hz can only really be the internal loop of the processor/stabilizer, and that's good cause it allows cycle time for signal smoothing, etc, but there's no way 800Hz PWM will work with a regular servo...
alevesque is offline Find More Posts by alevesque
Reply With Quote
Old Oct 05, 2012, 11:23 PM
Registered User
United States, CA, San Diego
Joined Apr 2012
752 Posts
Quote:
Originally Posted by askman View Post
best servo can take 333hz. due to timing offsets, it is best to have it about 2x that. so 800hz is ok, it will just ignore the rest.

Yeah, you want to sample at a minimum of 2x due to the Nyquist criteria to prevent any sort of clipping or aliasing of the signal. My chief engineer gave me a little lecture on that the other week
jcmonty is offline Find More Posts by jcmonty
Reply With Quote
Old Oct 06, 2012, 04:06 AM
Begin with end in mind...
power's Avatar
Eagle Lake, Minnesota
Joined Apr 2003
5,600 Posts
Looks like extruded 1/2 in square tube here with sharp corners. Looks like the 1/2 is available in 6063-T6 I could not find the 6061-T6 (they say it's sn option). The 6063 is a bit softer material, not that it will matter much with this project. Less than $1.50 a ft .

http://www.metalsdepot.com/products/...cc=%20&aident=

Mike
power is offline Find More Posts by power
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Question Can you make Embedded Pictures show up with Show Attachments in this Thread? Richard N Site Suggestions / Complaints 4 Sep 04, 2012 01:03 PM
Help! Can you tell me what make and model this car is? KiTCyclic Electric Power Cars 25 Aug 14, 2012 11:41 AM
Careful! Always make sure you clarify functional condition prior to agreeing to buy!!!!!! -= GYRO =- Trader Talk 84 Jun 15, 2012 01:06 PM
Help! Is there any gimbal stabilisation unit out there that can do camera lock? hexa Multirotor Talk 9 Apr 17, 2012 06:46 PM
The new momie hopperz f/d channel rc you can make a full functional micro plane saab2fast Scratchbuilt Indoor and Micro Models 10 Dec 16, 2004 07:43 PM