steve.neal
Australia
Asked
— Edited
I seem to be having trouble with one of my Sabertooth/kangaroo motor controllers. The red error light is on and the the motor will not work. Up until now it has been working fine. I suspect what may have happened, (not sure if this is a thing or not), I had the motor/encoder unplugged from the sabertooth and had run the initiation script while testing other things on the robot. I have tried re-auto tuning but this does not help. How do I reset the red error light? Everything seems to be connected. I cant see an obvious reason for it not to work.
Steve
I'm thinking the Sabertooth has met its maker
Also make sure your motor is working and not damaged, recheck all your wiring and connections. Without a complete and working circuit you may also get a red led. Did you accidentally swop the input power input wires? That will kill your Tooth dead.
Hope you find the issue. If we lived closer I'd send you a new 2x12 Tooth. I've got 6 new one's sitting here that I never used because I upgraded to the 2x32. I wanted to get away from having to use the dump battery for the regeneration power off of the tooth.
I have checked all the wiring and there is no shorts. I have taken the radar off and run the motor with a battery, that works fine. The motor is rated at 5 amps stall current and 0.3 amps free run when supplied with 12 volts. I tested it with a 6 volt battery and it drew about 0.3 to 0.4 amps while turning the radar for load. The Sabertooth is rated for 12 amps and 25 amps peak so that motor load is well within its capabilities.
Whats strangest is the radar motor has been working fine for a while, I got a good tune first shot, set the ramping parameters with the Dimension Engineering software, and it was working great. I haven't touched the connections to the motor or the sabertooth, It simply just stopped working.
I don't know if this could have caused the problem, but I have just tested the battery I use for dumping the re-gen power into. Its a sealed lead 12 volt battery. I just measured it and it's only putting out 5.5 volts, so along with the Sabertooth, it seems this battery is also buggered... (an Australian term for not working, or stuffed,....another Australian term with the same meaning). Could this buggered battery have caused the death of my Sabertooth?
This "buggered" battery is there for dumping re-gen power from the radar motor and the bubble lifter linear actuator. I have another battery the same in the leg section for the sabertooth that runs the waist motor. I have three Saberteeth in total (I think that's the plural of Sabertooth ;)) The other battery is measured at about 11.5 volts at the moment. If the dump battery IS in fact the cause of the problem, I'm thinking I might have to upgrade all three Saberteeth to the 2x32 and use a resistor for the re-gen power to avoid future issues.
One question, If I do upgrade the Saberteeth, and re-use the existing Kangaroos, will they all need another auto tune?
Steve
I should also point out that I have connected the battery's and the power supply using Dave's MOSFET and Diode circuit so the Sabertooth is not back fed by the battery when the robot is switched off thus draining the battery and still allows a path back to the battery for re-gen voltage if the motor is rotated while off.
If I upgrade to the 2x32 Sabertooth I can forget about the battery's.
Steve
I can say this, I've had this same setup in the leg section in my B9 with a 12v SLA battery for the past few years in my leg section and have never charged it. It's charge level stays at about 11v and I have no issues. Before I switched over to the 2x32 sabertooth in the torso section I also was running this setup for a long while with no issues. I never charged them and they stayed about 11v by just receiving the regen power from the Sabertooth.
With that said, RR is right about SLA'a needing to be charged. Lead acid batteries must always be stored in a fully charged state and be periodically recharged even when not in use. Leaving them discharged is a sure way to ruin them. Continuous over charging or under charging is the worst thing for a SLA battery. It's possible in this setup that you overcharged this battery. However like like rr said there is really very little Regen power going back into the battery. Again, how much is to much?
Now, we'er not talking bout discharged batteries here. You "are" charging them and if your other circuit and my setup is any comparison to go by the level should stay about 11v on a 12v battery. This has been a constant in my system for years now and I've never had a battery fail. I'm not sure that just one volt low or high will really make a big difference in the battery life (but I'm not 100% sure). I would say there is more risk of overcharging in this system.
Also, something else to keep in mind when thinking this through. There's a lot of talk now everywhere and on this forum about LIPO batteries. Don't confuse the two. SLA's don't have the same issues as LIPO' s when you discharge them. With LIPO's they certainly die when you drop below a certain level. This is not an issue if you kill a SLA once on a while. I cant tell you how many times I've had to jump start a completely dead car battery that's been drained from the cold or a short somewhere in the car and have it last for years after that.
I really think you have something else going on here that has nothing to do with discharged batteries. If this was a undercharging issue I'd think the battery would fail over a longer period of time. The flag for me is that you have a very low voltage reading on your SLA. The other SLA in your other system is preforming nicely and has been. Here's my thoughts on possible causes:
*You battery just failed on it's own. This happens. Do you have a battery store near you? We have them here (Batteries Plus) and they gladly test suspected failed for you and give advise. If you end up getting a new battery make sure it's got a "minimum" capacity of at least 1.3Ah.
*If the battery has failed (or even if it can be saved) it could have been caused by a short. I know you said you checked for this but what about the cutoff circuit you built? Could something in that be shorted or have a failed part?
*The Sabertooth failed on it's own. I've been reading posts from people that have recently had one fail. Unless you accidentally shorted out the power input I don't think there is anything your little setup could do to kill this tank of a controller. It protects it's self pretty well.
*The Kangaroo failed. Same as above. Just a note and question here; Do you have 5v feeding "into" the 5v port on the control side of the Roo? This is a 5v "out" port and you'll kill the Roo if you feed power "into" this port.
*There is something shorted in your circuit or power supply. Battery, homemade shutoff circuit, wiring, power supply, Sabertooth, Kangaroo. Dont forget to check the wiring between the Roo and the EZB.
Try removing or replacing parts of your setup to see if you can get the red led to go off. I'd start by making sure everything in the circuit is attached and ready then :
*Removing the homemade shutoff.
*Replace or remove the battery.
*Move the Roo then Sabertooth one at a time to the working circuit. When you move the Roo and attach it to the other Sabertooth did it's LED turn red? When you move the Sabertooth in question to the new circuit does it's red led go off? When you attach a known working Roo to the Sabertooth in question does the red LED got off? Be advised that when swapping around devices that may be bad you may cause damage to something in the working circuit. However because of the nature of this circuit and the protection built into the Roo & Tooth I think this is unlikely. Still if the red led comes on when it shouldn't remove power as soon as possible.
*Try running the Sabertooth without the Roo attached. You may need to place the dip switches in the proper setting.
In the long run I would advise replacing all your Saberteethes 2x12 with the 2x32's. I loved being able to remove the SAL's and not have to worry about things like what your going through, a failed battery down the road or ever an acid leak. Running a power resistor across the power input/output of the 2x32 as a voltage clamp was so much cleaner and less hassle. The hardest part was finding the proper value to use. That was more or less trial an error.
EDIT: One more thing to answer your question; I've found when ever you change anything including adding a different Sabertooth you need to do a re-tune of the roo. This may not be the case if you swap out like for like between 2x12's. My experience was when I upgraded to the 2x32. I'd advise to see if your motor is acting as it should after the swap and before you re-tune. You might get lucky. I never did though.
Good luck ad have fun!
OK, let's say at 11v we can assume the battery is empty. If it's just empty the red LED should't be on because Steve's still feeding the circuit with a 12v power supply. I guess it's possible if the battery is hosed and shorted on the inside or if the bad battery is sucking all the available power from the Sabertooth and creating a brown out.
This is why I suggested first removing the battery from the circuit to see if the red led goes out. If this is the cause I still think the Sabertooth will be OK and not damaged. If the Sabertooth is bad I don't think a low voltage on the battery is what caused it.
I still cant get past the fact that I've been running this setup for years with three Sabertooth's and with their own dump batteries and have never had this issue. Maybe it's because I have better quality batteries then Steve. Maybe it's something else. I don't know, that's why troubleshoot methods need to be preformed. I've also measured my dump battery voltage levels many times that are in parallel with the power supplies after the robot has been powered up and after the Sabertooth was doing it's thing. After power down my battery level is always 12v. I'm not saying this is the right way to charge a battery. I'm just stating what I personally found in real life.
I still feel Steve has something else going on here. It could be something as simple as a bad battery that just puked on it's own. It could be something more sinister. He should check out the other parts of the circuit just to be sure. In the end I still feel (if people can afford it) that it's best to use the 2X32 Sabertooth set up with a voltage choke when you want to supply power with a AC to DC power supply and be done with all this.
Thanks Guys for all you're input with this problem.
Dave, I have tried some of the things you suggested already, but not transferring known working parts to see if the problem follows or disappears, I guess for fear of damaging other working components. With the Roo, I don't have any connection to the 5V terminal, only Ground, S1 & S2 to the EZB. I didn't think however of the home made battery switch possibly being the cause. I have tried disconnecting everything including the Roo from the Tooth leaving only the 12 volt supply and the red error light remained. In saying that though, I didn't think, as you suggested, as it is still fed via the home made switch there could be a bad component on that. I have tried with the battery disconnected from the switch with no success but not with the switch disconnected as well.
I must head out now, but I will look into this further tonight. I am probably going to replace all the Teeth with 2x32s anyway, but I would like to know what caused this in the first place.
Thanks again and I will keep you posted
I have tried to figure out if the home built switch is to blame but my results are inconclusive... mainly because I am not very knowledgeable with trouble shooting electronics. I think I will just buy the 2x32 Teeth, remove the dump batteries, remove the home made MOSFET switches, rewire everything and be done with it. This robot owes me well over 10K to date, and he's not finished yet, so whats another $600,.....right? No sense taking short cuts..... right? I just need to order them when my wife is not looking ;).... Actually, she's pretty cool with the my whole B9 Robot obsession...... I might have to buy her another pair of shoes though
Steve
It looks like your dip settings are 1-4 Off (down), and 5 & 6 On (up). According to the manual If switch 6 is in the UP position, then the Sabertooth is in standard R/C mode. It enables a Timeout Failsafe, which will shut down the motors if the Sabertooth stops receiving correct signals from the receiver.
It looks like you have this set in RC mode right now. If dip is up and you don't have a signal from a RC receiver the failsafe is shutting down the Sabertooth. This could be why your getting the Red LED as it sits now. I'm not saying this is your problem but your not getting reliable information from the Tooth's LED. I'd suggest switching dip 6 to Up and seeing what happens. If you still get the red LED then add the Roo and set the dips on the Sabertooth to 1 & 2 Down (off) and 3-6 Up (on). If you still get the red LED i'd say your Tooth has failed.
Why did it fail? We may never know. It may just have given up the ghost. However, maybe there's a short in the battery or somewhere along the power train.
*Remove the battery, the sabertooth and the power supply and do a continuity check between the power supply and the ground side of the circuit for a short. *Make sure your battery is good. Use a battery charger to see if it will bring it up to full power and hold a charge (or take it out and have it checked). Some chargers know if the battery is bad when you hook it up.
*If all this checks out you may have accidentally shorted out the circuit somehow. Most likely it's a manufacture defect and the Tooth just failed.
When things like this happen we normally ask yourself "What did I do?". Sometimes the answer is "Nothing, it was something else".
The dip switches are set 1 & 2 Down (off) and 3-6 Up (on). Its a bit hard to see in the last photo. Its easier to see in the closeup on the previous page. The dip switches are mounted upside down in relation to the text printed on the PCB. Looks like it is in fact dead. I have ordered the new 2x32 Teeth to replace the three 2x12s I have now. I may have to pick your brain if that's OK, when it comes time to work out what size resistors I need for them
Steve
Let me know how that works for you.
Does anybody have any suggestions?
Steve
Can you remove the limit switches and use one of the other tuning options and have it do a successful tune?
I am hoping to have some time tonight to swap the roo's around to see what happens. I'm not sure if the roo was the cause of the last issue with the Sabertooth because the Sabertooth had the red error light on with nothing else connected to it..... unless..... the roo IS faulty and it damaged the tooth.... doesn't seem likely though.
I'm planing to move a known working roo to the non working tooth and see if that fixes it. I'll let you know how that goes.
Steve
I have confirmed the limit switch circuit continuity already. I checked the plug connections for tightness, broken terminals, the limit switches themselves for correct operation and the whole circuit by checking across both sets of limit switch terminals on the rear of the Roo with everything plugged in.
I still haven't had a chance to investigate any further and it looks like I will have to wait until tomorrow for another go at it. Too busy at the moment with other stuff and my poor robot is having to take a back seat *tired*
Thanks again for your thoughts Dave.
Steve
If this is a old roo that you have already did an autotune on (or tried to) you may want to try to get into the software and do a reset. You may be picking up the old limit switch limit settings.
Once in the software try to tune it from within there. You'll find it easier to see what's going on and make adjustments.
I already tried resetting the roo back to "Default" with the software, I could have sworn that was going to work....... but no dice.
I also was thinking of trying the "Teach" tune, but I have my doubts that will work because I still suspect the roo is faulty. I will confirm that by swapping them around on the weekend.
Steve
Now that I have a spare motor, I tried connecting it directly to the sabertooth and plugging the encoder directly into the roo and simply bridging out the limit switch terminals to simulate N/C contacts. This worked and I was able to start a tune and indicates there might be a problem in the permanent wiring in the robot. I then took out the bridge on the limit switch terminals and plugged in the limit switches I have been using in the robot and I was still able to start a tune. This proves the limit switches and wiring are OK as well.
The encoder wiring has been extended from the motor to a plug that's mounted on the top of the CSS and then continues to the roo mounted in the CSS. I have checked continuity of the encoder wiring via the plug with a multi-meter because I was thinking there may be too much resistance due to cable length and the plug, but I only got a reading of about 0.15 ohms on all cores. I'm not sure if this is the cause of my problems or not, but I figure the only way forward is to rewire everything and also try a better quality plug
This problem has been really frustrating because up until the original 2x12 sabertooth died, everything was working perfectly
Steve *tired*
Have you tried completely removing the limit switch circuit and trying to do a Teach Tune?
In the end you may be right. Most issues I've come across in troubleshooting all kinds of devices end up to be connector or wiring related.
Yeah, I've tried the Teach tune several times. It still comes up with another error code
(2 flashes - "System Range. The system can’t tune in the range provided or reached one of the limits. Make sure you started the tune in the center of motion"),
even though the limits are not even connected.
If the motor works when directly connected to the sabertooth but not when connected via the normal wiring installed in the robot, then there must be a fault in the wiring or plug. Its so frustrating because it was all working with the wiring the way it is with the previous sabertooth
Anyhow, I'm off to work now, Catch ya soon mate
Steve.
https://www.robotgear.com.au/Product.aspx/Details/724-131-1-Metal-Gearmotor-37Dx52L-mm-with-64-CPR-Encoder
If this encoder ends up needing a pull up resistor due to cable length, what value and how would it be connected?
If I remember correctly, everything works outside of the robot (and with shorter wiring )? When using the robots wiring this is when you get the errors?
On the other hand, the roo has different error codes for feedback and limits. You could give DE a call and get their advice.
Just a thought
Also, as Dave said it might be worth contacting DE
The error I get when I try to auto tune is six flashes on the kangaroo which is according to the kangaroo manual is a limit switch is fault but I have proven the limit switch circuit is fine by disconnecting the new motor and connecting a second identical motor directly to the sabertooth ( this is the original motor I was using but replaced with the new one in the hope that would fix this problem ) With the motor connected directly to the sabertooth/kangaroo, and the limit switches in the head section still connected to the roo and the auto tune will start. While the motor was connected directly to the sabertooth and the auto tune works, I extended the encoder wiring with 1 metre servo extension leads thinking the issue may be encoder signal voltage drop, but it still worked.
I have proven the new motor itself in the head section works by disconnecting the wires from the sabertooth motor terminals and connecting across a battery and the motor runs.
I am clutching at straws here, but the wires from the VGA plug are connected to the motor wires by in-line twisting and soldering with heat shrink over the join. This doesn't sound possible to me but, is it possible the soldered connections are causing the issue with feedback from the encoder?
There has been mention of the possible need for a pull up resistor, Does anyone know if this will help and how and where to connect it?
I hope someone can help me as this thing is doing my head in. *stress*
Steve
This may be just a stab in the dark but I see you have a big speaker mounted right next to where your wiring from the encoder passes. If your using a Hall effect encoder maybe that big magnet from the speaker is messing up the signal. Try removing the speaker and any other magnetic devices and see if that helps.
I'm sure you will get this fixed. You have a very solid and well built system there. You just need to figure what is keeping the dam auto tune to work.
Do you have your old working tune saved? Maybe you can load it into the new roo and manually adjust some of the settings if needed in the DeScribe Software. That way you don't even need to do an auto tune.
The delay in re-posting is mainly due to being really busy with home stuff and I haven't had much time to work on the robot. That and I am fast becoming really frustrated with this issue and I had already exhausted all known avenues over and over, (well, avenues that I, as an electrician can understand. Electronics are a bit of a mystery to me). I found myself having to walk away from it to collect my thoughts, calm down and try and think of something I hadn't already tried 20 times already, .......such as removing the speaker to see if magnetism is causing the issues. Thanks Dave, I'll give that a try. Why didn't I think of that? I've had issues in the past with EMF while working on the waist motor and the 240 volt fluorescent task light I was using to see.
Unfortunately I didn't save the original tune from when I was using the 2x12 Sabertooth, good Idea though.
I going to try removing the speaker right now and I will let you know how it goes
Thanks Dave, for all the help you've given me
Steve
Sorry to say Dave, but removing the speaker didn't work.
Back to the drawing board I go.......
*tired*
Steve
If you are able to start a tune with the motor out of the robot there's a reason why you can't start one when mounted in the robot. Something is keeping the feedback signal from getting back to the Kangaroo or causing it to be confused. The error message that the kangaroo shows is usually spot on on my experience. Trust me, I've caused most of them to trigger. Lol.
You said you tried to get a tune to start when you disconnected the limit switches and tried a Teach Tune? Did you get the same 6 flashes?
Alan
Steve
There is no pull up or down resistor in the circuit. If required, where and how would that be connected?
- keep the sensor connected to the roo
- use a multi meter and set the voltage for DC
- probe the signal wire against the sensor's GND and tell me the voltage
- probe the signal wire against the sensor's PWR and tell me the voltage
The Voltage measurements with everything plugged in are as follows.
Supply voltage to Sabertooth is 11.99 volts
Voltage measurements at feedback pins of Kangaroo.
(As labeled on the Kangaroo)
5V to GND = 5.00 volts
GND to A = 15.00 volts
GND to B = 15.42 volts
A to B = 0.42 volts
I hope this is the information you are after
The motor/encoder I am using is this one
https://www.robotgear.com.au/Product.aspx/Details/724-131-1-Metal-Gearmotor-37Dx52L-mm-with-64-CPR-Encoder
Steve
I'm not certain of the kangaroo can support more than +5 on the encoder logic input pins. It may be damaged.
1) The motor wires of the motor/encoder should be connected to the HBridge pins of the sabertooth
2) The encoder power and signal wires of the motor/encoder should be connected to the encoder input of the sabertooth
Also I have omitted the index wire.
Is this incorrect for my application?
GND to B = 15.42 volts
I think DJ has something here. The above voltages to A&B of the feedback from the encoder seems very high. I don’t understand why if you are feeding 5v to the encoder (as your reading show, 5v to grd) your getting 15v back on pins A and B.
I need to think about this, go over some specs and test my roo pins to see what I have. My first impulse is there's a short or miss wiring. Dont know how you can get 15v though. The other possibility would be that you can't get a good and understandable read from these pins because of diodes or caps in circuit.
Edit: it's hard to see by your picture but how do you have the dip switches sat on the kangaroo? They should all be in the on position for your application. It kinda looks like they're in the off position. Please confirm this.
Edit again : according to the specs below your encoder input voltage can be as high as 20v so I doubt you damaged it by overvoltage. Double check your wiring to make sure your wiring color map to the proper pins as shown below :
Channel Hall effect encoder is used to sense the rotation of a magnetic disk on a rear protrusion of the motor shaft. The quadrature encoder provides a resolution of 64 counts per revolution of the motor shaft. To compute the counts per revolution of the gearbox output, multiply the gear ratio by 64. The motor/encoder has six color-coded, 11" (28 cm) leads:
Color Function
Black motor power
Red motor power
Blue Hall sensor Vcc (3.5 20 V)
Green Hall sensor GND
Yellow Hall sensor A output
White Hall sensor B output
Just confirming the dip switch positions on the roo, I have to head off to work soon. This is a zoom of one of the photos of the roo from an earlier post, sorry its a bit fuzzy as I dont have time till later to take a better one. It shows all dips are on.
I am going to try changing motors again and re-measuring the voltages at the roo tonight to see what that shows
Steve
I just do some testing on my roo's encoder pins.
After power up I get:
5V to GND = 5.00 volts
GND to A = 0 volts
GND to B = 0 volts
After command is sent for motor to turn one way:
5V to GND = 5.00 volts
GND to A = 2.5 volts
GND to B = 0 volts (this could be wrong. Could be 2.5 also)
Values change slowly as the motor turns the encoder
After command is sent for motor to turn the other way:
5V to GND = 5.00 volts
GND to A = 0 volts (this could be wrong. May be 2,5 also)
GND to B = 2.5 volts
Values change slowly as the motor turns the encoder
These are ruff voltage readings. I was twisting around trying to stick the probes in tight places. It could be that A&B both jumped to 2.5 volts but I was having trouble getting a good reading. The thing that sticks out is that when A&B pins tested to ground show zero till a command is sent then the voltage reading jumps to life. Looks like each pin gets 1/2 of the 5 volts supplied to the encoder.
I wonder if it makes a difference if I have no tune saved in the roo.
All I did to take the measurements was to turn the power on and measure the pins. I also cant understand how I get 15 volts out of the signal pins when the supply to the encoder is only 5 volts. The power supply feeding the whole robot is only 12 volts *confused*
Looking forward to doing some more investigation after work
Thanks for going so far above and beyond the call of duty Dave
Steve
Good news is that is the problem - there is a short or wire connected incorrectly. This could be inside the motor encoder, or simply a wire of incorrect color to the wrong place. So you do have some direction to take.
Bad news is it's pretty difficult for us to help diagnose the wiring, it's up to you
What you can do is disconnect the motor/encoder entirely from the kangaroo. Then probe the wires at the kangaroo without the encoder hooked up. You should get 5 volts and 0 volts, never 15.
Also probe the motor encoder wires unhooked from the kangaroo. If you get 15 volts, then the issue is at the encoder.
Oh, that reminds me, Double check your crimps at the molex connectors that attach to the encoder and limit switch pins on the roo. I found one or two that I thought were good but actually crap squeezes by me and caused problems. I've made a thousand of these connections over the years but once in a while I squeeze a bad one.
OK, I also asked myself why you're getting 15 volts at pin A&B. Wish I knew. Only things think I can think of:
*You're testing wrong (sorry)
*Bad tester (low battery?)
*One of your circuits is shorted to your 12v circuit somewhere and back feeding through the encoder to the roo.
*God only knows
Too bad you didn't save your last working tune. You could jump start your roo with it. Don't fee bad I made the same error at first. Cost me a lot of time to get back to where I was. I've attached a zip file with my working tune of my radar. You can install into your roo and it may help get you started. Before you start it up your going to have to load it into the DeScribe software and change the feedback from Analog (Pot) to Digital (Encoder). Your using a different motor then me also so the other settings wont work well and the end points will be wrong. However this will be a way to see if your set up at least starts up. Let me know what happens.
Radar2.zip
EDIT: I see DJ answered while I was typing and getting the ZIP file ready. He as usual is giving good advice and in a more elegant way then I could. Good luck.
Another Edit: If you are in fact back feeding 15v into the roo's A&B pins you may have damaged the roo. *tired*
There is no battery in the robot at all, just a power supply that puts out exactly 12 volts. I'm certain as I can be the wiring is connected correctly, but stranger things have happened and I will check it all again for the umpteenth time ;).
I am intrigued to measure the voltages of and find out what the second motor will return, (which is actually the original motor I had in the robot and swapped to try and fix this problem). Maybe the brand new motor/encoder, which is in the robot at the moment has a faulty encoder.... hard to imagine though as the same problem was there with the original motor too, but the original one is the one that works if I connect it directly to the sabertooth...... ARRRRH, my brain is melting. I will do some more testing when I get home tonight and keep you guys posted. Thanks for all the advise.
Steve
GND to A = 12.3 volts
GND to B = 4.98 volts
Keeping in mind that this motor works with the tooth/roo, I'm still very confused. No zero volts in sight.
To further my confusion, I lengthened the encoder wires with 1 metre servo extension leads. The motor still will start an auto tune but with the longer leads, the voltage measurements are slightly different.
GND to A = 13.47 volts
GND to B = 4.97 volts
Not sure why the voltage goes up on the GND to A pin. I would have thought if anything it would have gone down with voltage drop caused by the thin gauge longer wire.
I have also tried unplugging the encoder on both motors and with the multimeter set to ohms, I measured from the VCC pin to both the A and the B pins and the readings were the same for both motors
VCC to A = 5.08 ohms
VCC to B = 5.10 ohms
A to B = 10.19 ohms
GND to all the other pins = 0 ohms
I'm going to have to resort to dismantling the radar section, yet again and cut all my hard work to remove the new motor.., again, and first see if it will work connected directly to the sabertooth as the other motor does. If not I will have to put the original motor back in the radar and try to get it working by temporarily connecting it and then trying the permanent wiring a section at a time.
*tired* Steve
Did you try DJ'S suggestion of pulling off the connection at the roo and voltage testing both encoder side and roo side while energized to see what your voltages are? That will tell you what direction the problem is at least. However I'd be surprised if you got any voltage from the encoder side because it should get its power from the 5v pin of the roo. But that is what testing is all about.
Did you see my previous post where I attached a zip file with a working copy of my Radar roo tune? Try adjusting it like I suggested and uploading it to the roo to see if you get any changes. Don't forget to power cycle the roo after the upload. If you can get this tune to move the motor at least you can have a little control hopefully.
One last thing for now. Take another look at the limit switches. Physically discount them from the roo, clip on a continuity tester and physical open and close them. Watch and listen. Make sure they are operating properly with clean open and closes. I've had more than a small share of new micro switches malfunction. Old switches will fail like this more often after a few thousand operations. I'm always replacing micro switches on my pinball machines because of this.
I was only taking measurements with the motor not running.
I did try DJ's suggestion to remove the encoder signal wires, I removed them from the plug that plugs onto the roo while leaving the +5 volts and ground wires in the plug so I would have supply to the encoder and I could read what was coming back from the encoder on the signal wires. Unfortunately I didn't wright down the results at the time and now I cant remember for the life of me. It definitely wasn't zero, it was something high again. Wish I could remember but this fault is really beginning to get to me and my brain is beginning to shut-down, possibly as some sort of self defense mechanism.
Thanks for the Zip file of your radar tune. I haven't tried it yet, if all else fails I will have to try and work out how to load it into the DIScribe software and adjust it for my needs and see if that will work.
As for the limit switches, I have proven their functionality with a continuity tester multiple times and I also have by-passed them by putting a direct short on the limit pins on the roo by using a modified 4 pin plug with shorted wires to simulate NC contacts
I'm just about to start ripping the new motor out of the radar and I want to test it out of the robot to see what happens. I'm starting to wonder if this is all just because kangaroos just don't like Hall Effect encoders.
Steve
I truly understand the flusteration you're going through and feel for you. Been there and done that. I'm impressed by how you're sticking with it. Understandable why you shut down sometimes. Keep up the good fight. Your not only helping yourself but others down the road.
I'm not sure but I don't think you should be getting a higher voltage back from the encoder. You may be right that the roo doesn't like that type of device. Have you tried calling DE yet? They may have some answers.
I'm quietly..... (very quietly), confident that I may have found the gremlin in the system..... I think..... I hope. I removed the new motor out of the radar, the one that refused to work. I connected it directly to the tooth/roo and it worked. So with that in mind I wanted to see if EMF is the culprit. I connected the encoder and limit switches using the VGA plug and the power to the motor run completely separately and it worked. If I reconnected the motor power through the VGA plug, it stopped working. It seems it was EMF after all.
I'm now going to redesign the disconnect-ability of the head section ... again and plan to use separate shielded cables and plugs for the encoder and power. This time I'm going to do some bench testing to confirm that shielded cables will work as they will still need to reasonably close to each other inside the collar so I don't have tangling problems with the other cabling in there for the ears and brain lighting which needs space to move when the radar turns and the bubble goes up and down.
I'm afraid I'm not going to have much time to work on the new design till next week but at least I think there is a light at the end of the tunnel with this problem. If it works I will finally be able to start moving forward again. YAYYYYY.
I will keep you posted on how it works out. Thanks for keeping me from loosing my mind
Steve
Before you do anything try, one just more thing. Do a continuity test from pin to pin on that big connector. Maybe you splashed a little solder between them and caused a short.
I'm really looking forward to seeing your final post where everything starts working for you.
Once you get a good and successful tune make sure you save that thing. *eek* Do the same for each roo.
The 15 pin VGA plug and lead is a molded factory made pre-terminated extension lead I bought from the local electronics shop, I simply cut it in half and used the cut ends to connect to all the components. You're still right though, factory made doesn't mean that it cant be faulty. I will buzz between all those cores out and see if there is a problem in the plug itself.
Thanks Dave
I can finally call this issue "resolved"
The problem was definitely corrupted data due to EMF from the motor power wires. I completely rewired the radar section, installed metal shielded plugs and used Category 6A shielded cable and earthed the shield and plugs. Once installed I tested the system, and you wouldn't believe it, I got the same error. at that point I had to just walk away or I was going to explode with frustration.
A few days later when I was in a better frame of mind, I did a bunch of testing and fiddling It seems the only way I could get it to work was to use the shielded cable for the encoder and NOT use the shielded cable for the power cables. If I just used normal unshielded cable on the motor power it worked.... Cant explain it.... it just did. So I rewired the power wires permanently, selected the expropriate Auto Tune...... and off it went. But... the first tune failed half way through so I separated the cables some more and I got a successful tune....Finally
Next problem....., I had issues getting both the waist motor and radar motor to run with my existing ARC initiation script. The bubble lifter worked fine though. The bubble lifter uses a potentiometer and the radar and waist use encoders and need to "Home" when initiated. The problem was that I needed to fiddle with the homing speed and also with the initiation script a little and I finally got it all working. YAAAAAYYYYY
Hopefully this is the end of this issue but It's taken me so long to fix, and tested my patience so often, I'm not 100% convinced its over. I still keep expecting to see that FU@*!/# error light. Fingers Crossed
Steve
Your solution is fascinating and curious. So just to be clear; you had to use unshielded cable for the power and shielded for the signal. When you had the power cables running in shielded cable were they running side by side and close to the signal cable? When you say you separated them to get them to work what were you separating them from? I wish I knew what the actual issue was with the EMF and where it was occurring. If you had to separate the + & - of the power cable to get them to work that should not have mattered. If you had to separate and move the unsheilded +& - power cables away from the signal cables then you were still getting emf interference. Either way it was still bleeding over somewhere. Maybe at a connection point? I hate to keep beating a dead horse and put this out into the universe but all that moving cables could have reconnected a bad wire or connection somewhere. I hope I'm just full of crap on that last statement. It's just nice to know what the actual cause of these things are for the next time something like this happens to us.
When you recover from this traumatic event (LOL) would you be willing to take a few pictures of your wiring setup from Roo to motors to power supply? That is if you don't have him all buttoned up yet.
So I'll keep a positive attitude and be sure this issue is behind you. Thanks for including me in your process. It was great to watch you work through this to success.
I will add some pictures tonight with explanations. It's such a relief to be able to move forward. I felt I couldn't move forward at all by working on another part of the robot knowing this issue was still there.
I must head off to work now ....at NASA
Steve
Three, Two, One..... Blast Off!
Steve
Here's a few photos of my setup with regards to the radar motor and ear motors
I have a couple off topic questions if your willing to answer;
Have you had the torso in place with the CSS and the EZB sealed inside? If so have you experienced any connection problems? I see your EZB is mounted on the second shelf down. That would place inside the torso and depending on your WIFI signal direction you may have lots of metal and devices between it and your router. I was afraid of this same issue so I have my two EZB's mounted on the top shelf to get the most unobstructed signal possible. One EZB is for all the torso and head robot functions along with sound streaming. The other EZB just runs both arms. I was also worried with two EZB's being so close to each other that the WIFI radio signals would interfere with each other. I have each mounted about 5 inches apart and there seems to be no issues with interference. However I was having trouble at first with one EZB connecting and not the other. I set a static IP address for each EZB in my router and also have and Engenius Access Point that they attach to. That gives me very strong signal strength close to the robot. Once I did these two things I have had no issues with connections.
The second question; When your radar rotates back and forth (along with the bubble lifter pipe that runs down through the CSS) do you exuberance the collar twisting with the movement and unlocking it's self? I see you have a hook latching system for your collar to lock it down the CSS. I assume you set the collar down on your CSS and twist it in place to secure it. My Collar tend to move a little when the radar rotates and after a while I need to re-lock it in place. After viewing Chris P's HD pictures on his DVD I see the orignal B9 had a clip and screw in the very rear where the collar joined to the torso. I bet it's there to keep this from happening when Bob May used his head to move the radar back and forth and bubble up and down. I plan to add this clip and screw to keep my collar in place.
Third question. What are you using in the center of your CSS to guide the Bubble Lifter pipe. It looks like a PVC fitting. I'm also using a PVC fitting but it's much more bulky and bigger. I had problems with binding and had to do a lot of sanding and aligning from level to level. It was a pain but is heavy duty and solid now. I'm not sure if it was really necessary to go that HD and was a pain in the but. LOL.
Anyway, Nice work and please keep in touch. I'm very interested in your path.
Have fun!
I haven't forgotten about you.
I should be able to spend more time to upload the pics and answer your questions tonight.
Steve
With regards to the EZB mounted inside the torso, yes I have had the torso on with seemingly few problems with connection. Sometimes It does disconnect from the router but the router is installed way up the other end of the house and the robot is down the other end. I do want to upgrade my router and move it to a more central location in the house as I have had some problems with the laptops in the house disconnecting or having a poor signal also from the router. This sort of leads me to question of my own. Every time I run the INIT script, It wont connect. I then have to scan in the Connection box in ARC to find the EZBs and the addresses always have changed since the last time I connected them (From one day to another). I think you may have answered this question in your above post when you said you set a static IP address in the router for each EZB. Not sure if this is the answer or even how to do it.
My radar doesn't seem to loosen the collar from the CSS with radar movement. Just luck I suppose. (More unicorn than class). My collar has three locating blocks which fit into slots cut into the top shelf of the CSS and they fit snuggly between the torso top ring and the CSS. Also the metal hooks are snug as well and when I locate the collar onto the hooks I need to give the collar a quick twist to lock in the hooks. Seems to be firm enough not to twist off.
EDIT: (more unicorn than class).......? Wait.. What? That's not what I typed... rhymes with class though... Laughed my unicorn off!
The center guide in the CSS for the neck tube was again, more luck than good planning
Anyhow, I think I have waffeled on long enough
Steve
Thanks for sharing.
I am proud of how he's coming along so far but there are still plenty of things I would like to improve on. I've lost count how many times I have re-made, re-mounted, replaced or re-wired something because I wasn't happy with it. I still now go to add something to the CSS only to find out that it will impact or clash with the placement of something else not previously thought of and again, the snowball effect happens all over. I try so hard to think of everything that will be in the CSS and how best to locate everything for functionality, space saving and aesthetics, but despite all that thinking, I still run into lots of hurdles... but in saying that, this B9 is a labor of love and I get a tremendous amount of satisfaction from it.
Your CSS on the other hand, in fact the whole robot is world class and I can only imagine how much time and effort has gone into it. In my opinion, your robot would have to be the best autonomous B9 in the world. Truly a work of art and an inspiration to the rest of us B9ers, especially me. Whenever I need inspiration, I just watch some of your videos. Just incredible
I will try to make a video of the radar and bubble lifter in action today and then try even harder to work out how to upload it
Thanks again for the kind words Dave
Steve
I don't think i've ever seen a robot build with so much care in the wiring - that's an entire project just in itself. Amazing
I did a video of the bubble lifter (linear actuator) in action. I thought Dave might be interested in it. The camera work and editing leave a little to be desired though *eek*
https://www.youtube.com/watch?v=3oFhUPcIjbg&feature=youtu.be
Seriously though, The lifter action is so smooth. That actuator is a lot faster then I thought it would be. Seems like a very natural and human speed. Are you able to stop it at different positions and control the speed? Are you using the Sabertooth and roo to drive this? Sorry if you have already shared this info with me and I forgot. Heck somtimes I cant remember what I had for breakfast. LOL.
If I had to rebuild my lifter I'd go your route. I'm not entirely happy with my servo driven fork design.
I really did enjoy that vid. Thanks for sharing.
Thanks for showing off.
I cant take credit for the neck tube roller ball castor idea. I saw in in a video by Greg Logue showing his new and improved Bubble Lifter design.
https://www.youtube.com/watch?v=BOICtuLYBoA
You can see the ball castor at the 2:22 minute mark. I was thinking to do the same as Greg and use a motor with an off-center wheel but decided to go with a high speed linear actuator instead.
The roller ball castor, I bought it off ebay. Here's a link to one the size I used but there's heaps of choices
http://www.ebay.com.au/itm/2-4-10x-32mm-Wide-Screw-Mounted-Furniture-Trolley-Rack-Ball-Swivel-Caster-Castor-/401044714205?var=&hash=item5d6020b2dd:m:msE1VXcFDlouHeIhgD8eZUg
I was planning to pull the neck tube apart tonight to show you how I made it but yet again I have run out of time *tired* I will endeavor to get it done for you in a day or two
Steve
Thanks for the links they really help and are a pleasure following.
Too Late
I too was worried about the linear pot not working well with the sabertooth but it seems to work fine. Something I do need to test though is the linear actuator with some weight on top of it. I need to add about 5Kg to it, (my head section weighs about 5Kg), and test it with that. The actuator is apparently rated to 10Kg according to the specs, so I'm hoping it will do it with ease. Time will tell.
Steve