|
Post by Trip Rodriguez on Oct 23, 2016 21:19:26 GMT
Accidental post in wrong thread, please delete.
|
|
|
Post by tronicgr on Oct 23, 2016 22:29:08 GMT
I was just reading mhensen thread and found some talk there about the 6DOF board that I think explains why you were saying I will need spare chips! Super sensitive to static. I searched for ULN2981 on ebay and google and did not find anything, can anyone link me where to buy spare chips? Still also wondering what is involved in installing replacement chips, I'm out of town at in-laws currently and can't go look at my 6DOF board. Also tronicgr was saying his wires from AMC to 6DOF were too long and I think mine are maybe even a touch longer. I will try to contact Thanos about this connector he mentions. You can get plenty from here, enough to have spares for long time: m.ebay.com/itm/20pcs-UDN2981AT-UDN2981A-UDN2981-8-Channel-Source-Drivers-DIP-18-/171060432792?txnId=1596768635007If you have the newer 6dof boards with the ic bases, you just pull the old chips and place the new ones (watch the polarity). Using a short connector between the amc1280usb and the 6dof ext board is preferred as it involves SPI timed signals that can be affected by motor noise is cable is used. You can create such connector if you don't have one using two of these: m.ebay.com/itm/20pcs-2-54mm-pitch-1x8Pin-Header-Right-Angle-Female-Single-Row-Socket-Connector-/281396458492?txnId=1579134517018Thanks Thanos
|
|
|
Post by Trip Rodriguez on Oct 23, 2016 23:28:54 GMT
Thanks Thanos! I actually meant to post that question on my own thread but posted it in the wrong tab! Thanks for the links! My boards are 2 years old, hopefully new enough for plug on replacement? Parts ordered, I hope they come fast =).
@ga-Dawg : Once you went to just the single computer setup was the speed no longer an issue on Oculus Rift? I went to 130mm torque arms to try to help with this and now wondering if I should have stuck with 160mm. I can switch the arm length in minutes but my gearboxes are closer together than others so my triangle would no longer be equilateral.
|
|
|
Post by clyevo on Oct 31, 2016 7:18:56 GMT
Regarding the 1uf caps in between digital signal of FWD/REV at 6dof board terminal. Purpose is to filter out high frequency noise right? Does this work for both source and sink input signal? I mean do you put the caps near source or input signal? And just curious what is the frequency that is filtered/attenuated?
|
|
|
Post by paulg100 on Oct 31, 2016 10:30:25 GMT
yes it helps smooth the signal.i put the caps at both ends. I tested at one end with oscope and then at both ends, both ends is better (on oscope and real world performance)
never tried with sink signal but if it is analog signal then i guess it will do the same?
the Wj200 VFDs also have an analog filter built in which you can set with a parameter, maybe your model does also?
|
|
|
Post by clyevo on Nov 1, 2016 6:20:12 GMT
yes it helps smooth the signal.i put the caps at both ends. I tested at one end with oscope and then at both ends, both ends is better (on oscope and real world performance) never tried with sink signal but if it is analog signal then i guess it will do the same? the Wj200 VFDs also have an analog filter built in which you can set with a parameter, maybe your model does also? i am using uln2803 for sinking signal at the 6dof thanos board. So i am not sure the same applies because most of you guys are using udn chip for source signal. In particular should the caps be applied at the inverter/VFD terminal instead. If anyone can confirm this? Would love to improve my simulator because i am still having oscillation issue that i fail to troubleshoot. for the digital terminal input at the inverter it has function of filtering time settable from 0.000 - 1.000s default at 0.010s. But i am still having oscillation, so i ll try increase this value. I did not notice this function. Thanks for bringing it up. But will the caps still has benefit even though the inverter has this function for example if the controller board does not send any signal but because of the noise it is translated as a signal?
|
|
|
Post by paulg100 on Nov 1, 2016 20:28:18 GMT
the caps will work on any analog signal from any device.
i have one cap at the VFD terminal and one at amc end.
i also have one cap at pot end and one at amc end.
in both situations they help with the jitter i have but not enough to increase the sensitivity of my platform where it needs to be for VR.
as i understand it using a 0.1uf cap is a broad spectrum solution. if you can identify the exact problem frequencies using an oscope then you can calculate the best cap to get the best filtering. i havnt had time to look into that more yet.
in real world performance the caps are more effective than the analog filter on the VFD for me.
|
|
|
Post by clyevo on Nov 2, 2016 1:42:22 GMT
the caps will work on any analog signal from any device. i have one cap at the VFD terminal and one at amc end. i also have one cap at pot end and one at amc end. in both situations they help with the jitter i have but not enough to increase the sensitivity of my platform where it needs to be for VR. as i understand it using a 0.1uf cap is a broad spectrum solution. if you can identify the exact problem frequencies using an oscope then you can calculate the best cap to get the best filtering. i havnt had time to look into that more yet. in real world performance the caps are more effective than the analog filter on the VFD for me. I think we need a thread discussing noise and emi. Yeah increasing the filtering time will cause more delay/lag. Which terminal did you put the caps at the potentiometer end and at amc end. Ground, V-in, V-out?
|
|
|
Post by paulg100 on Nov 2, 2016 6:46:40 GMT
|
|
|
Post by clyevo on Nov 9, 2016 3:08:22 GMT
|
|
|
Post by clyevo on Nov 11, 2016 9:08:26 GMT
i noticed how you connect the shielded cable via the pigtail method. I am not really sure why but i read it acts sort of like an antenna reducing the shield effectiveness considerably. full article hereI just found out about this and was doing the pigtail method like you to connect the shield braids. I am planning to fabricate an L bracket with hole in it to connect to the shield braids. but i am not sure how to apply that to the analog potentiometer cable as the shield is suppose to be terminated at GND of inverter and GND terminal of AMC1280. In this case there is no way to connect without twisting the braided shield and connect to the terminal? any thoughts paulg100?
|
|
|
Post by paulg100 on Nov 11, 2016 16:43:54 GMT
yes ideally you want 3 core cable so the pot to amc has a dedicated ground, then ground the shield at the terminal end (AMC) only. using the cables shield as signal ground basically creates an antenna for emi.
a terminal block that has the 6 shields connected to it that then feeds to your main ground point is technically the correct way to do it, so a star configuration to ground.
Mine is a little different as my pot cables feed into an emi enclosure that uses emi glands. The cable shield goes to the gland, the gland is connected to the box and the box is connected to ground.
Grounding at both ends creates a ground loop. (Its ok for the motor mains cable but all signal cable shields should be ground at terminal end only)
Even after alot of searching online, one thing i still havnt found yet is how you check the quality of your main ground? if that is poor then all the grounds will be poor. I looked into putting a dedicated ground rod into garden and running this to the control cab but id like to know a way to check the existing ground quality from the mains before i go to this trouble.
Any ideas how to check the main ground?
|
|
|
Post by riton39 on Nov 11, 2016 17:21:53 GMT
|
|
|
Post by riton39 on Nov 11, 2016 17:22:54 GMT
|
|
|
Post by clyevo on Nov 12, 2016 1:40:59 GMT
Any ideas how to check the main ground? Full articleMine is a little different as my pot cables feed into an emi enclosure that uses emi glands. The cable shield goes to the gland, the gland is connected to the box and the box is connected to ground. Grounding at both ends creates a ground loop. (Its ok for the motor mains cable but all signal cable shields should be ground at terminal end only) So you are using the 3 core shielded cable (as oppose to me using audio shielded cable with 2 core and using the shield for return path or DC ground). And you are terminating the shield at one end only to chasis at AMC side which is connected to main earth ground. The potentiometer is floating. This is to prevent ground loop right? In my case i have abundant amount of 2 core shielded audio cable so i have no choice but to use the shield as return path for ground so that means single ended shield. But I read somwhere single ended shield does not protect against high frequency magnetic field interference.... But your way is convenient in a way to prevent using pigtails (using the gland). Everyone has different configuration eg: like you has the inverter chasis a good many meter away from the motor. Thus difficult to make chasis of potentiometer to have equal potential with amc chasis to prevent ground loop at the shielded cable if connected at both ends. Although, i am still not sure about this ground loop problem. Some uses capacitor at the other end of the shield connected in series to ground just for better shielding against high frequency magnetic interference. Sorry another dumb question, is connecting the shield directly to the main earth ground is better than the DC signal ground path? I would assume the main earth ground is better but any explanation?
|
|