|
|
Post by svart on Jan 2, 2018 10:34:31 GMT -6
So I always have like 5 projects going at any one time.. I've had this one on the back burner for a while, and finally started building my first prototype.. It actually works quite well. Now I need to decide if I want to go beyond making a single one, making a few batches for some lucky folks.. Or what. I could make them out of metal (expensive, a little boring looking, but bulletproof), or design some parts for 3D printing (cheap, can design neat looks, but ultimately could be easy to break)..  I can't seem to get the video player to work on here, so here's a link to the thing working... www.facebook.com/699740678/videos/10155335384920679/ |
|
|
|
Post by EmRR on Jan 2, 2018 11:58:20 GMT -6
Cool. So, I'm sure you've looked at all the versions that went before, just to BE SURE you are aware of them. I've no idea what any of their strengths or weaknesses or price points are/were.
|
|
|
|
Post by svart on Jan 2, 2018 12:10:13 GMT -6
Cool. So, I'm sure you've looked at all the versions that went before, just to BE SURE you are aware of them. I've no idea what any of their strengths or weaknesses or price points are/were. Actually, while I've seen them over the years, I haven't taken any recent looks at anything. I'm just doing what I think is right, and what would serve the greatest purpose for myself first, and then if folks like it, I can have more made. |
|
|
|
Post by Johnkenn on Jan 2, 2018 14:27:39 GMT -6
I totally didn't hear any difference in the sound...  |
|
|
|
Post by svart on Jan 2, 2018 15:20:40 GMT -6
I totally didn't hear any difference in the sound... WUT? |
|
|
|
Post by svart on Jan 2, 2018 15:22:50 GMT -6
The plan is Bluetooth, either through an app, or through a small remote control. Not sure which just yet. The other choice would be whether the actuator would be battery or wall powered..
|
|
|
|
Post by kilroyrock on Jan 2, 2018 16:02:35 GMT -6
I worked on one a while back, never got it off the ground. Bluetooth works, but since you have the mic cable, maybe a cat5 run and then you can skip the pairing troubles Bluetooth is plagued with when it goes to sleep? At 17 cents a foot at home Depot, it seems doable.
Is your price going to be less than the one that's already on the market? I think they had a similar design. I always wanted a two elbow design, with a piston for distance.
Distance sensor in the front, feeding info back to a speaker picture that you could point at a location and the microphone would move to that location.
If you open sourced the software and sold a kit you would make some cool friends...
|
|
|
|
Post by Johnkenn on Jan 2, 2018 16:03:01 GMT -6
I was joking. Like I didn’t hear a difference when you moved the mic...but I guess bad joke.
|
|
|
|
Post by svart on Jan 2, 2018 16:13:44 GMT -6
I worked on one a while back, never got it off the ground. Bluetooth works, but since you have the mic cable, maybe a cat5 run and then you can skip the pairing troubles Bluetooth is plagued with when it goes to sleep? At 17 cents a foot at home Depot, it seems doable. Is your price going to be less than the one that's already on the market? I think they had a similar design. I always wanted a two elbow design, with a piston for distance. Distance sensor in the front, feeding info back to a speaker picture that you could point at a location and the microphone would move to that location. If you open sourced the software and sold a kit you would make some cool friends... I was going to do a simple design first, single axis, with the receiver set up as a peripheral so the pairing would be super easy, as you can set some modules up to auto query for PIN and baud. I guess I could have a backup socket for control over mic cables too. Which one is already on the market? I saw a crowdfunded one a few years ago with a price of 300$-ish, which this one would be less than half of that if I can do it right. |
|
|
|
Post by kilroyrock on Jan 2, 2018 16:26:46 GMT -6
More like 700-1000! dynamount.comYou have my attention with an affordable price. I have seen some great amp positioning achieved by being able to control from your listening point in the control room |
|
|
|
Post by svart on Jan 2, 2018 16:44:53 GMT -6
More like 700-1000! dynamount.comYou have my attention with an affordable price. I have seen some great amp positioning achieved by being able to control from your listening point in the control room Well, I see a few reasons why it's so expensive.. They're using overpowered steppers and generic tracks with belts and all kinds of stuff that isn't needed nor warranted. We're moving a mic around, not sending this thing to mars! So far, I'd have about 40 in CNC'd bearing blocks, another 30 in threaded rod, straight rod and bearings, another 15 for motor, etc. I suspect the driver board will be around 20$ stuffed, and the remote maybe another 20$ depending on if I can find suitable generic enclosures and do little to no machining, or better yet, generic remotes. Then I have to account for overhead, so another 30% on top, then some profit to pay myself for the labor of building and testing, and we'd have something in the 180-225$ range most likely. That's completely ballpark on COGS (cost of goods sold) and reality is that I'd likely be barely breaking even on it at that pricepoint. In any case, my design centers around using CNC'd blocks that both hold the bearings and then use the bearings and smooth rod as their own stiffeners, so that a large skeleton of supporting structure isn't needed. In my prototype, I used a belt due to the small footprint, but I'd entertain the idea of using gears as well, which could work if I can find the right ones for cheaper than the belt. Anyway, my "sewer robot" moniker isn't by chance or jest, I actually spent 5 years designing robots that went into sewers. JohnK thought it was hilarious, so he put it as my description.. So I've done my share of mechatronic design, usually to some ridiculously low price point. |
|
|
|
Post by kilroyrock on Jan 2, 2018 16:52:44 GMT -6
More like 700-1000! dynamount.comYou have my attention with an affordable price. I have seen some great amp positioning achieved by being able to control from your listening point in the control room Well, I see a few reasons why it's so expensive.. They're using overpowered steppers and generic tracks with belts and all kinds of stuff that isn't needed nor warranted. We're moving a mic around, not sending this thing to mars! So far, I'd have about 40 in CNC'd bearing blocks, another 30 in threaded rod, straight rod and bearings, another 15 for motor, etc. I suspect the driver board will be around 20$ stuffed, and the remote maybe another 20$ depending on if I can find suitable generic enclosures and do little to no machining, or better yet, generic remotes. Then I have to account for overhead, so another 30% on top, then some profit to pay myself for the labor of building and testing, and we'd have something in the 180-225$ range most likely. That's completely ballpark on COGS (cost of goods sold) and reality is that I'd likely be barely breaking even on it at that pricepoint. In any case, my design centers around using CNC'd blocks that both hold the bearings and then use the bearings and smooth rod as their own stiffeners, so that a large skeleton of supporting structure isn't needed. In my prototype, I used a belt due to the small footprint, but I'd entertain the idea of using gears as well, which could work if I can find the right ones for cheaper than the belt. Anyway, my "sewer robot" moniker isn't by chance or jest, I actually spent 5 years designing robots that went into sewers. JohnK thought it was hilarious, so he put it as my description.. So I've done my share of mechatronic design, usually to some ridiculously low price point. You continue to fascinate me! |
|
|
|
Post by matt@IAA on Jan 2, 2018 19:46:23 GMT -6
Svart there are some home CNC kits that may be an affordable way of getting controlled encoder / motion. You may be able to find a pre built arduino interface or something on the cheap. Like this: www.inventables.com/categories/machine-components |
|
|
|
Post by svart on Jan 2, 2018 21:11:42 GMT -6
Svart there are some home CNC kits that may be an affordable way of getting controlled encoder / motion. You may be able to find a pre built arduino interface or something on the cheap. Like this: www.inventables.com/categories/machine-componentsThanks for the heads up on the site. I do have a full arduino setup from a number of past projects. I'm doing this half for the fun of it, half to maybe take it commercial, so pre-built won't really cut it.. We'll see. |
|
|
|
Post by kilroyrock on Jan 2, 2018 21:48:29 GMT -6
The pre built controller/brain was where I struggled to see a low cost and stopped persuing.i don't have the background you have to go for it from a scratch. I think you are onto something even the current market hasn't been able to realize
|
|
|
|
Post by svart on Jan 3, 2018 9:23:45 GMT -6
The pre built controller/brain was where I struggled to see a low cost and stopped persuing.i don't have the background you have to go for it from a scratch. I think you are onto something even the current market hasn't been able to realize An arduino (Atmel) can do some cool things. I'd use a DC geared motor, there's no need for a stepper since we're not looking for ridiculous accuracy. We'd just use FET switches for direction control since we don't really need speed control. On/off is fine for this since it moves so slowly and we know the mechanical gear reduction value and the screw-lead/threaded rod pitch. I'd like to try to use current sense to find the ends of the travel during a one-time (or once per use) calibration. Since we know that the motor moves 98RPM and the belt drive is 1:1, then the threaded shaft is turning 1.64 RPS. The threaded rod travel is 8mm per turn, so we know that it's 1.64x8mm = 13.12mm or 0.516 inches per second. Since the atmel has A/D inputs, you can simply use this to measure the voltage across a low value resistor in series with the motor and derive the current draw that way. You set a desired cut-off value, and if it reaches that point, you've either stalled from obstruction, or are at the end of travel. You record the timing to one end, and then to the other, and because you know the distance per second, you know your position. Lets say you have 6.5" of travel, and you've recorded your time-to-end points of 6.5" x 0.516 in/sec = 12.6 seconds end-to-end travel. You've recorded that you've traveled for 2 seconds from an end point, which means 2s x 0.516" = 1.032" traveled from that endpoint. Anytime you stop, that distance/time is put in nonvol memory and recalled next time the unit is moved/powered. The unit could then do a self-calibration of finding the endpoints, and return back to the original position after re-calculating the calibration. Seems like a lot of work, but it's really only a few lines of code for the benefit of having a lot less wiring for sensors and switches. On the remote side, I'd either use two switches (Up and Down), or maybe even capacitive touch sensors in a plastic case (no drilling needed!) while receiving some kind of travel indication. Maybe i'd use an LED bar that shows relative location on the travel. |
|
|
|
Post by svart on Jan 3, 2018 9:44:37 GMT -6
One other thing I noticed is that I'm using 8mm threaded rod as both positional and structural functions. The bearings are tight enough that the rod doesn't wiggle at all in the bearings. The 6mm rod is more of the anti-tilt mechanism as well for torsional stability. Honestly there is no need for the threaded rod to be 8mm in diameter, but there aren't really many 6mm threaded rods available.
Another thing I've found is that the 6mm straight rod actually costs a little more than a similar 8mm straight rod, so I might end up using 8mm straight rods here too. It's not really warranted, but would be a little cheaper overall since they seem a lot more prolific, as well as being a bit stronger.
|
|
|
|
Post by kilroyrock on Jan 3, 2018 10:13:44 GMT -6
What about a distance sensor? So you don't bash the cone?
|
|
|
|
Post by matt@IAA on Jan 3, 2018 10:37:19 GMT -6
Smart. I’m used to industrial applications with limit switches and all. Simple current draw is brilliant.
|
|
|
|
Post by svart on Jan 3, 2018 10:40:30 GMT -6
What about a distance sensor? So you don't bash the cone? This first proto would be single axis, essentially sweeping the cone from edge to dust cover. A 2 axis device could certainly use a distance sensor on an axis used for depth, likely a simple infrared sensor to stop it in case of obstruction. Distance measurement could be done a few ways, one of which could be the standard ultrasound transceiver pair you can buy for arduino, or it could be something complex like a pulsed laser that measures reflection times (probably impossible to do in arduino, I'll have to see if anyone else has attempted it). EDIT: They make laser distance finders for arduino, but unfortunately nothing for distances less than a meter or so. It'd have to be ultrasonic, or some kind of other optical sensor. |
|
|
|
Post by kilroyrock on Jan 3, 2018 12:06:35 GMT -6
What about a distance sensor? So you don't bash the cone? This first proto would be single axis, essentially sweeping the cone from edge to dust cover. A 2 axis device could certainly use a distance sensor on an axis used for depth, likely a simple infrared sensor to stop it in case of obstruction. Distance measurement could be done a few ways, one of which could be the standard ultrasound transceiver pair you can buy for arduino, or it could be something complex like a pulsed laser that measures reflection times (probably impossible to do in arduino, I'll have to see if anyone else has attempted it). EDIT: They make laser distance finders for arduino, but unfortunately nothing for distances less than a meter or so. It'd have to be ultrasonic, or some kind of other optical sensor. I have a 10"Max sensor I used on my first build, I could send it to you some it's just collecting dust |
|
|
|
Post by svart on Jan 5, 2018 13:30:48 GMT -6
 I have some CAD drawings for CNC'd bearing blocks for the ends. They work out to be about 15$ each in quantities of 24. I ordered some plastic project boxes that I hope to use as the end caps that cover the belt and such. Which ones and how they fit will dictate changes to the bearing blocks. Once I get that stuff figured out, I'll start on the micro hardware to control them. |
|
|
|
Post by svart on Jan 9, 2018 12:44:31 GMT -6
a few updates.
I've ordered a few more pulleys and belts to see if the 2GT series of belts work better than the XL series of belts. Seems the XL is harder to get in smaller diameters and pulleys with fewer teeth.
Also, I've settled on a way to cover the ends of the blocks with plastic boxes so that the belts/pulleys and rod ends aren't exposed.
I think I've finalized the overall prototype mechanicals otherwise and I've debated on ordering the metalwork for a set of 12 units now, or waiting a little longer to see if there is any other issues that might pop up.
I think I'll order some more plate aluminum and make a set of parts to the exact CAD drawings that I have to make sure they all look good.
|
|
