|
|
GY521
Jun 12, 2016 6:16:29 GMT -7
Post by mayujsw on Jun 12, 2016 6:16:29 GMT -7
hi max, i'm a newbie to DOBOT and also trying it with ramps1.4. i'm just so excited when i found this open-source place for study, thanks very much ! i read the RAMPS wiki at git hub in these two days, and have a question about GY521's position and the direction, is it working as below picture shows (on the top of existing accelermeter):  drive.google.com/open?id=0B4eAidlHaiD1QVpVYlZ0c3hBMEE |
|
|
|
GY521
Jun 12, 2016 7:08:08 GMT -7
Post by Max on Jun 12, 2016 7:08:08 GMT -7
Correct. Just make sure there is no electrical contact between gy-521 and the original accelerometers or the arm.
When mounting try to make it as parallel to the original accelerometers as possible, but you'll be able to find any offset with calibration-tool.py and set it with the instructions I'll make available in the next week along with some other changes.
Just a heads up - I found there is a problem reading accelerometers when original stepper drivers (A4988) are enabled - they produce quite a lot electromagnetic noise that is picked up by all other components and is strong enough to disable I2C. There is only one fix for it right now - enable motors only after accelerometers have been read, which very inconvenient as it means the arm will fall down every time a program is executed. Much better solution would be to replace the drivers with DRV8825, with are much better drivers themselves, more quiet and provide 1/32 microstepping as opposed to 1/16 on A4988 (twice the resolution).
Stay tuned, some fixes and improvements are coming.
|
|
|
|
GY521
Jun 12, 2016 23:03:03 GMT -7
Post by mayujsw on Jun 12, 2016 23:03:03 GMT -7
Thanks max for so quick reply, looking forward for your improvements ! and sorry that seems my picture cannot be displayed  so the 8 connector holes of GY521 is upward, point to joint 2 |
|
|
|
GY521
Jun 13, 2016 0:17:09 GMT -7
Post by mayujsw on Jun 13, 2016 0:17:09 GMT -7
after i mount the 2 units on front and rear arm, and connect the AD0 of front GY521 to 3.3v, but both the units' LED is red, is there anything i missed ..
|
|
|
|
GY521
Jun 13, 2016 0:18:58 GMT -7
Post by Max on Jun 13, 2016 0:18:58 GMT -7
The link to the picture works ok The pin holes of both modules are expected to look at the back of the robot (same direction where the wires would go). However, as this is an open project you are free to redefine it any way you like. The angles are acquired by _initializeAccelerometers function in SDK and is easy to adjust to any either way you module is mounted. In case you mount one of the modules with pin holes looking the opposite direction all you need to do is to change the sign of the angle returned from accel3DXToRadians. I've spent some time today drying out DRV8825 and other ways to make gy521 work when A4988 are enabled and haven't had much luck. DRV8825, apart from the high-pitch noise, which is very annoying, don't behave well on either 1/16 nor 1/32 microstepping - not smooth at all. There is some fine tuning with soldering possible for those, but people report them as not that good at all. Even if I manage to make them work nicely it won't be a straightforward solution, hence not available to everyone. So for now I'm sticking with A4988 and trying to make gy-521 work when the motors are enabled (I can't stand seeing the arm fall down every time an example is executed). I've ordered couple of those www.adafruit.com/products/757 , but I suspect they are not really "logic level" shifters, but rather just "signal scalers", hence they just "scale" the noise, which is the problem in the first place. There are two other options: 1. A server based on SDK that would connect to the arm once and initialize accelerometers while the motors are still disabled. The server is planned anyway, but the idea of working with disabled motors doesn't seem nice to me. 2. Get rid of the accelerometers altogether and use limit switches/photointerrupters to calibrate the arm. This approach is ok for the base, but for the links it is not that good because of the lack of accuracy (there are just that many places to mount them) and complexity to calibrate them. I don't see any simple way to calibrate them on the links. Every way I can think of requires something special (an extra sensor, a handmade sensor, a 3d-printed part, etc.) Any ideas are welcome. |
|
|
|
GY521
Jun 13, 2016 0:22:26 GMT -7
Post by Max on Jun 13, 2016 0:22:26 GMT -7
|
|
|
|
GY521
Jun 13, 2016 19:50:35 GMT -7
Post by mayujsw on Jun 13, 2016 19:50:35 GMT -7
Thanks Max, I tried the steps, and the angles reported seem right as below, however, both LEDs are red  Rear arm: 74.67 | Forearm: 48.27 | End effector height: 98.79 | Raw rear arm: 16017 -148 -4387 | Raw forearm: -12310 -549 10965 |
|
|
|
GY521
Jun 13, 2016 19:56:44 GMT -7
Post by Max on Jun 13, 2016 19:56:44 GMT -7
Thanks Max, I tried the steps, and the angles reported seem right as below, however, both LEDs are red Rear arm: 74.67 | Forearm: 48.27 | End effector height: 98.79 | Raw rear arm: 16017 -148 -4387 | Raw forearm: -12310 -549 10965 Great! Thanks, I'll make a note on wiki that the LED's are not necessarily of different colors. I'll push a simple fix soon to make the firmware read accelerometers before enabling motors to make that at least reliable while I'm looking for a better solution. I'll probably make it a release as it fixes accuracy too and simplifies SDK a little bit. |
|
|
|
GY521
Jun 18, 2016 18:14:58 GMT -7
Post by Max on Jun 18, 2016 18:14:58 GMT -7
There is interference problem that prevents reading accelerometers when motors are enabled. See wiki for a fix. Also, with open-dobot 1.3 just released ( github.com/maxosprojects/open-dobot/releases/tag/1.3.0 ) there is no special procedure required on RAMPS to read accelerometers online. |
|
nad
New Member
Posts: 13 
|
GY521
Aug 15, 2016 4:20:21 GMT -7
via mobile
Post by nad on Aug 15, 2016 4:20:21 GMT -7
Hi, i got gy521s but when i try to calibration, says "Cannot get board version. Giving up" message. I connected vcc,gnd,sda,scl and wired 3.3v to one gy521 from arduino uno's 3.3v out. What can i do? Sorry for bad language
|
|
|
|
GY521
Aug 16, 2016 13:31:37 GMT -7
Post by Max on Aug 16, 2016 13:31:37 GMT -7
|
|
nad
New Member
Posts: 13
|
GY521
Aug 17, 2016 2:47:06 GMT -7
Post by nad on Aug 17, 2016 2:47:06 GMT -7
i think this can't cause. Becasue i didn't connect 12v to ramps so stepper motors can't make electromagnetic noises
|
|
nad
New Member
Posts: 13
|
GY521
Aug 17, 2016 5:08:32 GMT -7
Post by nad on Aug 17, 2016 5:08:32 GMT -7
|
|
|
|
GY521
Aug 17, 2016 20:59:31 GMT -7
Post by Max on Aug 17, 2016 20:59:31 GMT -7
1. Is one of the mpu6050 modules lying on a motor without any dielectric material between?
2. It was hard to trace connections to one of the mpu6050 modules, but the other one is not shown at all
3. Which version of firmware have you uploaded? FPGA, RAMPS? Version number?
4. Which version of the calibration script are you trying? Current from master?
Try increasing wait time on line 70 in DobotDriver.py to 10 seconds. Also try resetting Arduino - on Windows 10 with the latest arduino driver it doesn't seem to get reset when you just open serial port.
|
|
nad
New Member
Posts: 13
|
GY521
Aug 20, 2016 11:07:22 GMT -7
Post by nad on Aug 20, 2016 11:07:22 GMT -7
It does not contact the motor. it was for photo
the problem is understood.I couldn't flash firmware to arduino because I don't fully understand the instructions about flash firmware. I've never done before this.
|
|
so the 8 connector holes of GY521 is upward, point to joint 2
