Replies to This Discussion

DeveloperPermalink Reply by R_Lefebvre on January 14, 2012 at 7:56pm

I already used some soft rubber (servo bushings), but it's not enough in my case.

So here's what I came up with today. 

The square aluminum is exactly 1x1x1 inches, and I had to do this because of the odd shape of the sonar board, with the square thing sticking out the back, plus the RC filter resistors I glued on the backside, etc. The blue stuff is "Moon Gel" from a music store, it's the softest material you'll find anywhere.  I've made a sandwich with the middle layer being the frame mounted bracket.  There is a single mounting bolt, and the hole for the bolt in the frame bracket is quite large, so the screw doesn't touch at all.  The top piece of aluminum is a clamping plate so I can sandwich that frame bracket.  The other metal squares on the outsides are lead plates I added to try and get the natural frequency up a bit.  Kind of a double edged sword.  It helps get the NF down away from the rotor headspeed, but it makes low speed shaking worse.

The mount works, the only problem is that it's a bit too easy to swing it around, and it will move a lot if the heli is shaking such as on spoolup.

Anyway, here's a comparison of the old sonar data compared to the new one.  Big improvement.   There's one bad spike of noise in the middle of the new graph, I'm not quite sure what that was, I may have touched the ground or something and the frame shook a bit.

I'm not sure if this is as good as it gets, but it's obviously a huge improvement.

• ▶ Reply

Permalink Reply by Lloyd DeForrest on May 18, 2012 at 10:49am

OK, I officially hate this sonar! (MB1200)

R_Lefebvre,  Do you have an update to this sonar mount? 

I have encountered the same problem with vibration.  Everything worked fine with 850KV motors using the 3DR sonar mount.  I added a camera mount and went to 880KV motors and now the sonar reports 30 to 40 cm whenever the motors are turning regardless of height.  When the motors are not turning the sonar reports the correct altitude.  After un-mounting the sonar, letting it hang by its wires and re-testing, all is well again. 

APC props are balanced.

Looking for a better (easier to construct) sonar mount that reduces vibration significantly.

 

• ▶ Reply

DeveloperPermalink Reply by Randy on May 18, 2012 at 8:17pm

@Lloyd,

      You know, it could also be interference from the motors.  Maybe when you're letting it hang, it's also further away from the motors.  How about leaving it mounted but removing the props.  Or how about holding it in your hand (so it can't be the vibration from the frame) and moving it close to the motors to see how it reacts?

• ▶ Reply

Permalink Reply by Lloyd DeForrest on May 19, 2012 at 11:10am

Hi Randy,

I considered EMI effects. The test with the sonar hanging was in the same location (slightly lower) as the bad test in the mounted position and the differences were dramatic. Also, I added a shielded cable with no effect.

So I built an anti-vibration mount yesterday whose design is based on the one from R_Lefebvre.  It uses some anti-static foam that single ICs are shipped on (it was the only foam of this density I had laying around). I may upgrade to a moon gel equivalent later.  It needs more tests, but early results are encouraging. See results in plots below.

I also have sonar unit that I will try to duplicate the results in first the mounted and then on the anti-vibration mount.  Perhaps this maxsonar is just super sensitive to vibration.  I don’t see a lot of complaints about vibration from 880kv motors from others.  BTW, I can’t feel a noticeable vibration increase with the 880kv motors over the 850kv motors. 

 

 

Attachments:

• Mount Front View.jpg, 541 KB
• Mount Top View.jpg, 507 KB

• ▶ Reply

Permalink Reply by Lloyd DeForrest on May 19, 2012 at 11:26am

Randy - here are the before/after plots

They were made with the following sequence starting with the copter on the floor:

Attach battery => arm motors => raise copter to above head => raise throttle to ~60% => hold for a few seconds => lower throttle to zero => lower copter to floor => disarm motors => remove battery

Attachments:

• Bad Sonar plot.jpg, 30 KB
• Sonar Plot After Vib Reduction.jpg, 31 KB

• ▶ Reply

DeveloperPermalink Reply by Randy on May 29, 2012 at 7:45am

Lloyd,

     I missed this reply for a bit.  The second graph looks like quite a lot better.

     One thing is that the height looks too low.  Still, maybe that's because it's bouncing off your legs or something.

     Anyway, looks like the vibration dampening has worked!

-Randy

• ▶ Reply

Permalink Reply by Grips on June 15, 2012 at 5:33pm

Does anyone have online links to the resistors and capacitor mentioned here.

• ▶ Reply

Permalink Reply by Dave on June 16, 2012 at 12:41am

It's on the Maxbotix site in the Tutorial/FAQ area.

The signal filtering is described HERE

The power supply filtering as shown above is described HERE

I made both of these mods and used shielded cable (shield only earthed at APM end), and my sonar seems to work fine.

• ▶ Reply

Permalink Reply by Greg on June 16, 2012 at 2:53am

Grips the components can be sourced from from a variety of good electronics suppliers.  I used 0.5W 10 Ohm Metal Filament resistors with 1% tolerance from Soanar.  For the Electrolytic Capacitor you can use 100uF 6.3V or 16V types   I am from Australia so I tend to get these parts from Jaycar or RadioSpares.  However in the US there is Mouser, Digikey, Radioshack Sparkfun etc

Capacitors 100uF 6.3-16V Electrolytic

http://www.digikey.com/product-detail/en/ECE-A0JKA101/P803-ND/6910?...

http://www.mouser.com/ProductDetail/Panasonic-Electronic-Components...

Resistors 10 Ohm Metal Film ).5W, 1% Tolerance

http://www.electron.com/mfr-0207-10-ohm.html   Also available from Mouser and Digikey

Hope this is what you are after

• ▶ Reply

Permalink Reply by Chris Huitema on June 16, 2012 at 4:27am

The sonar thing is a problem for me too... i thought i had it all sorted, see my post here http://diydrones.com/forum/topics/sonar-noise

Here is the before

and here is the after

and all was looking sweet.

but now its gone back to worse than before!

the quad has changed alot since fixing the sonar, its been changed to an octoquad with coax motors and ive changed from mystery ESC's to JDrones esc's. but the main system is unchanged. the same Ubec is powering the AMP, all the mounting is the same, with the AMP, Sonar, and all the other accessories.

My problem is even without the motors running, the readings are bad... real bad! this is raising and lowering the quad from 0.5m to 1m

I need to check for a bad connection, but i have soldered all the connections so i dont think thats the problem. vibration cant be an issue because the motors arent running, have even tried with the ESC's removed (powering the APM, RC, and sonar with ubec from a battery and leaving everything else out)

Ahh that reminds me... ive changed Recievers.. to one with telemetry! and also changed the XBee to 38400 baud.. i wonder if one of these things will have changed anything???

Changing the Xbee to 38400 was to try and fix a connection issue with the ardustation but didnt change anything, so i will change that back to 56k anyway. and i can try without the receiver and arm using a joystick thru the Xbee. this will rule out both of these or show for sure if there is a problem with one of them!

Shame im offshore again for another month or so :(

• ▶ Reply

Permalink Reply by Greg on June 16, 2012 at 5:21am

I think your on the right track by looking to what has change and isolating one thing at a time.  I use the Jdrones ESC's {30A} but I plan to switch these out for PlatinumOpto isolated versions.  I do not use the ESC bec's Instead use a Castle UBEC to power the APM and Receiver.  One of the extremes I went to for checking interference was to hook up the sonar using the Tx RX UART port and monitor the sonar data realtime using a laptop This helps to see actual source of interference and find an optimal spot to place the sonar.  The details on hook the sonar up this way are on the Maxbotix web site.

• ▶ Reply

Permalink Reply by Redemptioner on January 1, 2013 at 3:38am

Hey All,

 

I have run into the same issue with the same sonar mentioned here (no surprise there). At first I  was not getting any readings while in flight (stayed at 20cm) even though it work on the ground. I played around with a few things without much luck until my father was back in the country where he could give some help with his wealth of knowledge seeing he is an electrical and computer engineer and works in robotics R&D.

 

A few things that have been pointer out,

-        firstly the power filter, he strongly recommended not using either 10 or 100 ohm resistor on the +ve line, as it was said earlier in the post you are asking for trouble with voltage drops and the start-up calibration of the sonar. He has recommended a choke with 50 or more turns on it, this will not affect the power and will do a far better job of removing the noise the 10/100 ohm resister is meant to be handling, it may however turn it into an antenna but this will be corrected by the 100uF capacitor. He assumes the reason for the 10/100 ohm resistor is that it very cheap and very small, but noted a 100 turn choke for the small current is not much bigger than the capacitor and only costs about $2 more. The capacitor and choke should both be directly soldered onto the sonar and then a shielded cable attached.

-        the shielded cable being used needs to be braided copper around each of the internal wires to be effective, using aluminium as a shielding source is not that good, so don’t bother with the cables using it such as USB cables or ipod cables (an to be honest apple equipment is cheap nasty rubbish anyways so their cables are not much better [current iphone 5 cost less than $90 to make] not too many people around that don’t have an apple device with a cable that is coming apart….but I digress). A decent shielded cable will also have the cores twisted as well.

-        There is no way to put a filter on the signal wire without undue effect on the handling….he said don’t even try this

-        Using the I2C mod will give the best results as PWM is the only way to go, keep in mind that if the power is unstable you may still get issues.

-        Solder everything, don’t use plugs or connectors when it comes to passing a signal through, the power is fine but solder the signal cable

-        Shield the power and signal separately

 

Originally I would not get any sonar reading as soon as the motors spin up, this was using an unshielded cable that ran under the ESC’s and parallel to the motor cables. I changed to a crappy shielded cable without any better response, so I move the cable away from the ESC and yet still not any better. It was not until my father looked at the setup and pointed out that the motor cables and ESC were both mount on carbon fibre and so was the sonar mount and that carbon fibre is notorious with carrying EMI, moving the sonar away from the carbon rod fixed the issue immediately, well at least it seemed that way.

 

Above about a meter high the sonar starts to spike all over the shop, making Alt hold no good above a meter, the funny thing is it does not seem to do this with the props off (so it could be acoustic/vibrations or still be EMI as the motors have no load on them without the props. I have not yet filtered the power, but the power for the sonar is coming from the APM but its power is supplied from a decent low noises filtered 5amp SBEC, the power for the OSD/RX/FPV are all on separate SBEC’s (yes that’s 4 SBECS in total), the only other things powered by the APM is the GPS/optical flow/telemetry and the SBEC’s are power off 2x 8,400mah high C lipo batteries in parallel so there is plenty of juice and a reasonably stable supply and there are no servos connected. I do however feed the sonar cable up through the middle of my power distribution board and then under the APM before it is connected (although shielded all the way).

 

Next step tomorrow is to go down to jaycar and grab a 100+ Turn choke, a capacitor and a high quality braided shielded cables with 2 cores and one with a single core. Dad has recommended running the signal wire on its own and have this go directly to the APM and not go near anything else with this cable, he does not think the power is less likely to be the issue in my setup but worth spending the $6 to make sure.

 

The only other thing is to make something that is vibration and acoustic proof to mount the sensor in, or at least reduce both. I noted that in most of the vibration dampened mounts I see have a bolt running through them, does this not defeat the purpose of the dampening material in the first place. Anything with a high density (ie. the bolt) will conduct the vibrations through the dampening material, sure the dampening would help somewhat but most of the vibrations would just traverse the bolt. Would you not be better to simply glue the material to each surface, it’s not like the sonar weighs anything? I am thinking to use some acoustic absorption foam and put the sonar in-between 2 pieces of the foam with around 3 -4cms of excess foam around the edge and the foam on the downward facing part to be level with the sensor array, this would “help” absorb any stray noise coming in as well as work as the vibration dampener, then I could just glue the top side to something with some rubber cement. If it works I will run the same shielding and filter on the Optiflow sensor to be sure it is noise free also.

 

If this does not work then I will be ordering my arduino mini tomorrow afternoon, I assume it would still have enough headroom to run the 3dr OSD mod as well?

Update to come in the next day or so when I have tested the mods.

• ▶ Reply

• ‹ Previous
• 1
• …
• 5
• 6
• 7
• 8
• 9
• 10
• Next ›
• Page