Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 When you think about the wires holding the ESC at either end, if it is mounted on adhesive pads, it really isn't going anywhere! I'm starting to think that tie wraps on ESCs are more trouble than they are worth. Too loose and they vibrate loose and can catch on twigs, too tight and you risk damaging the ESC. If the edges of my frame are well finished and smooth, I've taken a liking to plumber's rubber tape. You just stretch it as you wrap it round and after 24 hours, it bonds to itself and is great for keeping the morning dew out of your ESCs! • Posts: 1,773 Threads: 30 Likes Received: 1,199 in 755 posts Likes Given: 714 Joined: Oct 2016 Reputation: 45 Another r thing to think about is the effect on ESCs that are zip-tied down tight (effectively hard-mounted) The newest discussion in BF is the effect of soft-mounting ESCs. I'm not technical enough to understand what effect this has, but I'll be interested in the results of testing. • Posts: 125 Threads: 12 Likes Received: 86 in 45 posts Likes Given: 52 Joined: Jan 2017 Reputation: 8 (24-Apr-2017, 11:25 PM)unseen Wrote: When you think about the wires holding the ESC at either end, if it is mounted on adhesive pads, it really isn't going anywhere! I'm starting to think that tie wraps on ESCs are more trouble than they are worth. Too loose and they vibrate loose and can catch on twigs, too tight and you risk damaging the ESC. If the edges of my frame are well finished and smooth, I've taken a liking to plumber's rubber tape. You just stretch it as you wrap it round and after 24 hours, it bonds to itself and is great for keeping the morning dew out of your ESCs! I have to admit that you done it as well as it could get but i personally "don't like the look". I will add the adhesive tape, fix the ESC with an additional wrap of electrical tape and a softly tightend ziptie. If in the long run damage accour I will of course switch to your choice. (25-Apr-2017, 12:13 AM)RENOV8R Wrote: Another r thing to think about is the effect on ESCs that are zip-tied down tight (effectively hard-mounted) The newest discussion in BF is the effect of soft-mounting ESCs. I'm not technical enough to understand what effect this has, but I'll be interested in the results of testing. I can't really imagine any positive effects besides the durabilaty of the ESC board itself. Like it's been pointed out before, hard mounting will put stress on the board and the components that might cause significant damage on the long run. Performance wise here should be no difference in my opionion since the ESCs are not gyroscopic sensors so movement shouldn't effect their behaviour. And soft mounting the motors should reduce vibrations coming from the motor into the frame that might disturb the measurements of the sensors. I like turtles Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 I'll admit that the rubber tape isn't really going to win any prizes in a quadcopter beauty contest, but once the flying season starts in earnest and my quads have done their fair share of grass cutting, they don't look that pretty any more! • Posts: 125 Threads: 12 Likes Received: 86 in 45 posts Likes Given: 52 Joined: Jan 2017 Reputation: 8 The filter is now connected and the Videofeed is already way better. Still some strong bars scrolling across the screen when i throttle but it's okay for now. I will set both Videogrounds to the minimOSD board instead of the PDB. That will probably make it better. I also shortened all the cables to the VTX and the Radio RX. The cabling in the back looks way better now and the Antennas are now fed through a 3D printed plate with additional heatshrink tubing to make the more stiff and protect them better. Gonna take some pictures tomorrow and post them I like turtles Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 (25-Apr-2017, 07:21 PM)Apicalis Wrote: I will set both Videogrounds to the minimOSD board instead of the PDB. That will probably make it better. It will make a huge difference! I didn't realise that you were not running the grounds to the circuit that is using the signal. If I had, I would have pointed out the error immediately. You'll probably find you can do without the filter as well once you connect the grounds to the OSD. Posts: 1,773 Threads: 30 Likes Received: 1,199 in 755 posts Likes Given: 714 Joined: Oct 2016 Reputation: 45 I've replaced my vtx and camera cables with USB (data transfer, not charging) cable. Pretty tedious job, but it's shielded 3-wire cable and eliminates all interference. In fact, you don't even have to worry about routing it away from power wires. Posts: 125 Threads: 12 Likes Received: 86 in 45 posts Likes Given: 52 Joined: Jan 2017 Reputation: 8 (25-Apr-2017, 08:32 PM)unseen Wrote: It will make a huge difference! I didn't realise that you were not running the grounds to the circuit that is using the signal. If I had, I would have pointed out the error immediately. You'll probably find you can do without the filter as well once you connect the grounds to the OSD. Yeah thought so myself afterwards. I did do the smart thing and used a star mass connection which is of course a good choice for noise reduction... but all the Video Signal masses should've been together on the MinimOSD. But I will keep it this way for now since the video is acceptable. I will probably change something again in the future and when that happens I can change the connections. (26-Apr-2017, 04:03 AM)RENOV8R Wrote: I've replaced my vtx and camera cables with USB (data transfer, not charging) cable. Pretty tedious job, but it's shielded 3-wire cable and eliminates all interference. In fact, you don't even have to worry about routing it away from power wires. Yeah thought about using a shielded cable myself but I have these tiny JST SH connectors on my VTX. Of course I could remove the socket and direkt solder the wires to the pads, but since they are so damn small they will rip out cause of the mechanical stress Genius me thought "Oh I don't have Audio, so I can gut the Audiosignal wire and the GND connection" ... *slaps his own face* I could've used that GND connection to twist with the VideoIn I guess I will order a new cable and when they time comes that I want to do some changes on the Quad, this will be one of the changes that will be done I like turtles • Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 If your VTX has an audio input but you don't have an audio signal to connect to it, you should connect the audio input to ground at the VTX. • Posts: 125 Threads: 12 Likes Received: 86 in 45 posts Likes Given: 52 Joined: Jan 2017 Reputation: 8 (26-Apr-2017, 08:31 AM)unseen Wrote: If your VTX has an audio input but you don't have an audio signal to connect to it, you should connect the audio input to ground at the VTX. Hmmm I don't think it's problematic, but I also can't see a reason not to connect it to GND so I will bridge the two pins on the backside of the JST-SH Connector. I just had a talk with my former boss who is a very highly skilled radio technician. Basically said, I did everything like it's recommended in the books. Using a star ground as close as possible at the power source (PDB). Also of course a star connection from the 12V Output. Using the OSD GND connectors would lead to a slight ground loop since it's going through the FC board and the OSD board. Twisting a GND wire around the Video Signal cables should in theory do absolutely anothing, since it's not a symmatrical signal ( Like it is in Ethernet Cables [Twisted pair cables]) and the shielding effect is minimal. Shielding itself would be perfect if the Video In/Outputs would be decoupled from common GND. BUT It's absolutely possible that all your recommendations will help just like the way you said since this isn't a big cabinet with proper shielded equipment in it. It's a very small space, with lots of different signal types in it ranging from a few kHz to 5.8 GHz and insanely high current draws. So it's basically a huge noise fest inside a Miniquad frame. Everything that you put around a signal wire has a high chance of absorbing some noise ^^ So basically I wanted to tap myself on the shoulder for doing everything by the books ;D but also wanted to acknowledge that you are most likely correct with all your advice to change the wiring. I like turtles Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 26-Apr-2017, 11:26 AM (This post was last modified: 26-Apr-2017, 11:27 AM by unseen.) The only reason for grounding an unused audio input is so that you don't have to go through removing the audio component from your video before you publish it. If the input is grounded, the audio channel will be silent. If it's floating, then it will contain all the noise that the floating input picks up. Your old boss is absolutely correct. I know that the ESC control signal and the video signal are not differential signals and in theory, the benefits that a twisted pair bring to a balanced transmission line don't really apply. However, they are both referenced to ground, and that's where the problems start. As you note, the electrical environment of your average quadcopter is challenging to say the least. I'd describe it as downright hostile! The switching noise from the ESCs as they switch tens of amps through the inductive load of the motors, coupled with the back EMF generated during active braking means that ground is anything but the nice zero volt reference that theory treats it as. The very nature of the interference means that at any one instant, different parts of the quad will experience differing parts of the noise waveform. It's not the same at all parts of the power circuitry at the same moment as just like any other signal, it takes a finite time to propagate. Because of this, if the ground that the video signal is referenced to isn't the same at both ends of the wire, noise will be induced into the signal by the differing potentials of the video camera's or video transmitter's ground and the OSD's ground. Maybe twisting the ground and signal together does nothing at all. Having a separate signal ground and having it grounded to the destination of the signal is important though, regardless of what anyone wants to say about ground loops. Posts: 2,416 Threads: 51 Likes Received: 1,861 in 1,175 posts Likes Given: 3,315 Joined: Mar 2016 Reputation: 74 (26-Apr-2017, 11:26 AM)unseen Wrote: Maybe twisting the ground and signal together does nothing at all. Having a separate signal ground and having it grounded to the destination of the signal is important though, regardless of what anyone wants to say about ground loops. I am just about to solder some RacerStar ESCs to a PikoBLX. The racerstars provide signal and signal ground, but the piko only has pads for signal (and ESC power/ground). In this case, it wouldn't be important to use the signal ground, would it? There is continuity between the grounds on the ESC and they are going to the same "component" (the PikoBLX). On my builds where I use a separate PDB, I do always run signal grounds to the FC. • Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 (26-Apr-2017, 11:44 AM)sloscotty Wrote: I am just about to solder some RacerStar ESCs to a PikoBLX. The racerstars provide signal and signal ground, but the piko only has pads for signal (and ESC power/ground). In this case, it wouldn't be important to use the signal ground, would it? There is continuity between the grounds on the ESC and they are going to the same "component" (the PikoBLX). On my builds where I use a separate PDB, I do always run signal grounds to the FC. If the twisted pair approach doesn't actually help to reduce external interference on the signal wires, then I'd agree that there's no need to use the signal grounds in this case. If it does, you can solder them to the same place as the ESC power ground. Posts: 2,416 Threads: 51 Likes Received: 1,861 in 1,175 posts Likes Given: 3,315 Joined: Mar 2016 Reputation: 74 (26-Apr-2017, 12:02 PM)unseen Wrote: If the twisted pair approach doesn't actually help to reduce external interference on the signal wires, then I'd agree that there's no need to use the signal grounds in this case. If it does, you can solder them to the same place as the ESC power ground. Hmm...what do you suppose twisting the signal with the main ground would do? Different gauges... • Posts: 2,286 Threads: 38 Likes Received: 1,527 in 995 posts Likes Given: 1,881 Joined: Apr 2016 Reputation: 72 (26-Apr-2017, 12:12 PM)sloscotty Wrote: Hmm...what do you suppose twisting the signal with the main ground would do? Different gauges... Nah, the wires need to be the same size in a twisted pair, otherwise it'll actually induce current into the signal wire. |