MovieMaker
First off , I want to say that you will NOT find a Better system ANYWHERE that is close to the Quality, support, and Performance that the EZ-Robot System Provides. Or, better people to do business with. D.J. and his staff are Fast and Accurate, and GOOD People to do business with.
Having said that, I want to mention a few other things. I have been into building robots for the last 41 years. In all that time, I have NEVER really been satisfied with the robots built. It seems they where always too slow. Now with the new technologies, you have a chance to improve the speed and reliance.
When , I got the EZB, I decided to put together a robot that would have many features and I started with the neck. I copied the design from this site, a site which always has so many resources. Next, I wanted to use the Roomba platform. I was in luck because support for it was already installed.
I wanted it to be cheap and simple. So, I used PVC pipe. I built a robot called Marty. I found that if I could not figure out how to do some thing, all I had to do was wait. Many new updates were happening continuously and much to my amazement most were for the things on my personal want list.
Now, I was all set! I made TWO BIG mistakes, and i KNEW better. But, I did it anyway. I ordered four PING units from china. They appeared to be ok at first. Next, I also ordered a HUNDRED servo cables for about Ten Bucks. When they came in, I say to myself "Good Grief!" they were so small and lightweight. But, I decided to use them anyway because my wife was gonna kill me if I spent one more red cent.
From the beginning, I had problems with the Sharp IR units and the Ping unit. This robot did EVERYTHING I wanted it to do but Navigate. I could give it verbal commands and it would do JUST fine. But, let it roam on its on and it ran into me, the closet, the wall, the chairs, the tables, and Everything. It would also backup into furniture.
I changed out the Ping units, I changed out the IR units, I changed out the wiring with more crappy wiring. My Ping would give me 255 if nothing was there and would register about 12-15 units up close upto 255 far away.
One day, I just set it there with the PING showing what it registered. I noticed that I had three types of problems. One was the PING was 255 all of the time. Or the PING was Zero all of the time. Finally, I got it where it was measuring. I had at first swapped the white wires for the trigger and echo. Anyway, I wound up with four very BAD PING units and ONE Ping that registered very very flaky. I changed out the wiring again and ran it straight to the EZB from the PING. Once it was working sort of properly I put everything back where it went. But, it still gives me a flaky output. So, it does EVERYTHING but Navigate properly. If I would've bought heavy duty servo cables, I would have saved myself a lot of headaches. It may be a year or more until I can save up to get more Pings and cables.
Also, I learned from Hazbot not to put the servo directly onto the end of the arm. He lets the arm balance on the servo so it is easier to move. i have to do that when i rebuild Marty.
Also, there are some servo savers out there to prevent the servo from burning up on you like three of them did on me.
I hope this advice helps someone.
Good Luck with your projects!
That's some good info, I was looking into getting a bunch of servo cables, guess I'll be careful about where they come from and what gauge the wires are.
@MovieMaker To clarify, you are saying the quality/gauge or the connecting wires from the ping to the EZ-B is/was your issue?
I haven't been able to properly test collision avoidance due to limited test space, so I am unsure if my results are proper or not, but this is my setup...
I have the single PING that came with the kit, I have it wired to the EZ-B with about 3" of servo wires (The PING is currently in the nose of my transforming car). I have it angled up by perhaps a degree or two and sitting stable with a wall 4 1/2" away. Based on the radar (set to scan min50 and max50 - no servo - and 100ms interval) I get a constant reading of 85-87... but what I notice is that it flickers between the detected distance (with blip on radar) and 255 (no blip) every few seconds... sometimes a quick 255, sometimes lasting a second, occasionally more. Almost like the sensor data just dies.
I guess my question is... is this normal? Or am I getting signal degradation on cables due to proximity to the EZ-B or Camera (right beside PING) or just expecting too much from a PING sensor?
Mine was bad. I MAY have heated it up too much with the hot Melt glue. My cables are absolutely horrible. I think that them being short and flimsy had the most effect on it. The ultrasonic goes from 0-255.It should stay on ONE reading until something shows up in front of it. Now, another thing is, I have them LONG to reach from the EZB to the front of the Roomba. That may ALSO be an issue. I may need some sort of filtering. As far as Normal. It seems if the way is Free, it should tell you. Like 255. IF the way is blocked by an Obstacle, it should tell you. If it is Right in your face, it should be a low number like 12-15. I have used PING units that worked reliable. But, I have also paid $29-$39 a piece for them instead of $4 Each. I also thought since I was using 4 or 5 sensors that it was causing problems. So, I made it down to only ONE sensor. Still had the same problems. I think the gauge and distance matters.
OK, I did a little research and noted that voltage levels seem critical for proper operation of the HC-SR04 PING sensor... as in it wants 5v, not less... some have reported giving it 9v with good results.
So, I wanted to know what voltage it might be getting... my initial though was asking in the forums if there is some form of internal voltage readout on the EZ-B... but then I had my own little eureka moment and jumpered the 5v from an unused digital port to the sensor pin on an ADC (probably could use any 5v pin on the EZ-B.
I am running off of the basic 9V power source that came with the car (6x AA 1.5v batterys). I currently have 1 standard servo, 1 mini servo, 1 H-Bridge and the PING sensor running.
As I ran my car on the bench I noted the the EZ-B voltage ran normally at 4.98v with no action, down to 4.3-4.5v average with jumps to 5v with one standard servo in holding position but dipped as low as 3.9 depending on h-bridge actions. Voltage from source, as I tested, dropped as low as 6v (these batterys also power the H-Bridge & motors). The Ping sensor seemed to give proper results, when it did, but was still erratically jumping between something and nothing.
I think the solution might be independent regulated 5v power to the PING? I might try to rig up something to test that.
ADDENDUM: I input 12v for the power source (a little 12v Lead Acid battery) and ran the tests again... this time the EZ-B voltage stuck around 4.88-4.98 regardless of actions. However, a few minutes into testing, I experienced repeated brownouts with EZ-B voltage seeming to drop to under 1v... , even though I was only running servos and no H-Bridge. I confirmed solid connection to battery...confused
But If anything, the PING sensor was more erratic... more often then not giving the 255 reading and only occasionally alternating between the incorrect and the correct signal? Close range seems the most stable... anything further then 17 and I get strange results. Aside from that, clean, sufficient power seems part of the key, go fig :)
2nd (and last ;)) ADDENDUM: Somehow, the 12v battery power showed the more stable EZ-B power reading of 4.88-4.98v, but caused the PING to behave irraticly and cause multiple brownouts when using the standard servo. Going back to my 9v batterys cleaned up the PING (but still not where it should be) and stopped the brownouts but show the EZ-B as getting fluctuating power again. Test for one issue, discover others... ahh but what fun I am having getting into robotics tech :D
Thanks for the info. I am running a battery at 7.2 and 3800mha. Sometimes it will go down to 5 or 6 volts.