SMALL - espritmodel.com SMALL - Telemetry SMALL - Radio
Closed Thread
Thread Tools
Old Aug 17, 2013, 11:00 PM
Registered User
United States, GA, Atlanta
Joined Dec 2010
2,061 Posts
Quote:
Originally Posted by Samdamang View Post
Just so everyone is clear, Evvaldis' post said he is pulling 1.5A on a 3 cell with 5010 motors (from RCTimer http://www.rctimer.com/product_870.html). Huge difference between those and the 5206/5208 motors. I use 5010's for yaw on a gopro rig for example whereas I would be expecting to handle DSLR + lens with the others.






I highly recommend keeping it under 24v or so with even the 1.3 board. I posted (many pages ago now) that I got a micro puff of magic smoke at 24.75v before instantly unplugging the battery, but it seemed like it might be ok at 23.75v. I didn't keep it plugged in for more than 5 or 10 seconds though at 23.75v so not sure how hot things would have continued to get. Been waiting forever for a 5 cell to show up to move forward with testing.

Overtuner is on the right track with what he is working on I believe, though I am curious enough to see if the experimental code makes any difference so I will probably give it a shot tmrw (after hopefully reading your results later tonight tmarone).
Just a fast pass, but it looks like Toshiba makes a similar 3v3 regular that is good up to 29V, and .6A... I'd have to do more research, but it might be a drop in replacement for the 25V version...


However, lots of the 1.2/1.3 parts are only good to 25V
aadamson is offline Find More Posts by aadamson
Sign up now
to remove ads between posts
Old Aug 18, 2013, 12:39 AM
Qick'nDirtySolutionExpert
Germany, BY, Bobingen
Joined Mar 2013
73 Posts
Quote:
Originally Posted by Samdamang View Post
I highly recommend keeping it under 24v or so with even the 1.3 board. I posted (many pages ago now) that I got a micro puff of magic smoke at 24.75v before instantly unplugging the battery, but it seemed like it might be ok at 23.75v. I didn't keep it plugged in for more than 5 or 10 seconds though at 23.75v so not sure how hot things would have continued to get. Been waiting forever for a 5 cell to show up to move forward with testing.
right. forget the 1.2 and 1.3 for voltages over 3S - maybe with some code tuning its able to handle 4S. I believe no.

Quote:
Originally Posted by Tom Frisch View Post
Do we have a list of changes that we need for this? It may be too early, but I figure its worth asking.
STM f4 chip?
Different fets?
More pins out from the STM for various uses?
yes. all of them

Quote:
Originally Posted by Samdamang View Post
Does it make any sense (if possible) to separate the logic and the power in the v2 setup? Sparate driver board (or boards) for the motors to receive maximum power without hurting the low voltage control components? An extra 100g 1 or 2 cell on my rig to power the controller wouldn't cramp my style at all...
thats what I'm thinking about too. I was looking at some small STM boards with the real neccesary things on board (3v3, crystal, SWD, usb) but found only with 2.54mm pitch pin headers. it would be very sexy, to have a full equipped STM 1/2/4 board in small size with 1.27 headers. then we could design a "mother" board with different power logic.

Quote:
Originally Posted by OSHStencils View Post
Who's the actual core designer of the board for this project? Seems like a handful of contributors, but who's controlling the revisions between 1.2, 1.3, 2.0, from the hardware level?
good question.

Quote:
Originally Posted by aadamson View Post
Just a fast pass, but it looks like Toshiba makes a similar 3v3 regular that is good up to 29V, and .6A... I'd have to do more research, but it might be a drop in replacement for the 25V version...
However, lots of the 1.2/1.3 parts are only good to 25V
we need a 5v0 rail - at least for the r/c connectors. so we use a bigger 5v regulator - (recom or traco) and the small 3v3 behind them.

my next steps:
1. design and tests with TC4452 in SOIC-8 has finished. 6 channels with enough power for 5010 motors on 4S. there's a bad limit - the 18volts max voltage, but I haven't found a replacement with higher voltage, thats why:
2. design of a high-voltage (up to 6S minimum) driver. at the moment I plan to use some stuff like IRS2301 or, better: IR2104 and (ripped off some cheap china-esc for stealing ideas ) some fat dual FET like fairchilds FDS8949 for the small version and single FET for the big toy - behind them.
parts are on the way, mouser.
3. some ideas for the next board:
- more r/c inputs. with smaller connectors - i believe the most micro-servos use Molex Pico - so why not. and - only the first with 3-pin, the others only with signal pin.
- 5v rail with switched regs
- maybe a (unstabilized) 12v rail too? btw. I need 10-20v for the fet drivers..
- seen at martinez: a voltage divider to sense the lipo voltage for future adaption of PID for different battery state.
- jumpers to the edge, leds too - maybe some fat alloy will be in top of the board
- i2c power should have a (solder) jumper for 5v0 or 3v3
- is there a new imu coming? - what about a SPI connector?
- more tantal cap
- as a tribute to aleksej -> a button onboard
- I got some test parts - only 4 small smd-pads on the back side! or - much more coolness-factor: an xbee-socket
4. I'm learning EAGLE since the last 2 days...

ok guys, lets go on.

Lutz
Overtuner is offline Find More Posts by Overtuner
Old Aug 18, 2013, 06:48 AM
Registered User
United States, AL, Birmingham
Joined Jul 2013
47 Posts
Hi Alan,

I have some news regarding the experimental branch of firmware.
I have only tested the straight compile, not the debug.
Everything seems to work normally, current consumption is about the same as the gold standard firmware. No FETs were harmed in the test

I found that if I decreased the motor power, things worked better, and I have a theory why that is. My frame picks up vibration from the motors and transfers it into the gyro chip. I have tried to figure out an isolation for this problem, but haven't come up with a good solution yet.

This is the frame I have:
http://cgi.ebay.com/ws/eBayISAPI.dll...m=300938964697

Lowering motor power helps greatly, but does not fully fix the problem.

Alan, could you compile a version with a higher switching frequency, and maybe the vibration will go away?

Is there a way to programmatically filter out the motor vibrations?

I had no problem loading the experimental firmware on either board.
Values of P and D do not seem as jumpy as before.
Motor current from .3 to .55 amps each. Running 35 to 45% power on motors.
Input voltage 12.3 VDC total current was 1.7A max.
FETs heat up evenly and are warm to slightly hot in use, no runaway temps.
When the vibration is at minimum, the scope waveforms are almost textbook perfect!

This vibration is just like what happens when an open microphone in a PA system starts to ring and feedback. It starts low and quickly builds into a vibrating loop. If you folks think I just have bad P or D settings, I will try to tune more, but I have methodically been through over 200 iterations on a single axis.

If there is anything specific that anyone wants to know, just ask.
Thanks again for letting me test this, I appreciate it greatly.

Tony
tmarone is offline Find More Posts by tmarone
Old Aug 18, 2013, 07:51 AM
Registered User
Joined Jun 2013
136 Posts
Quote:
Originally Posted by tmarone View Post
Hi Alan,

I have some news regarding the experimental branch of firmware.
I have only tested the straight compile, not the debug.
Everything seems to work normally, current consumption is about the same as the gold standard firmware. No FETs were harmed in the test

I found that if I decreased the motor power, things worked better, and I have a theory why that is. My frame picks up vibration from the motors and transfers it into the gyro chip. I have tried to figure out an isolation for this problem, but haven't come up with a good solution yet.

This is the frame I have:
http://cgi.ebay.com/ws/eBayISAPI.dll...m=300938964697

Lowering motor power helps greatly, but does not fully fix the problem.

Alan, could you compile a version with a higher switching frequency, and maybe the vibration will go away?

Is there a way to programmatically filter out the motor vibrations?

I had no problem loading the experimental firmware on either board.
Values of P and D do not seem as jumpy as before.
Motor current from .3 to .55 amps each. Running 35 to 45% power on motors.
Input voltage 12.3 VDC total current was 1.7A max.
FETs heat up evenly and are warm to slightly hot in use, no runaway temps.
When the vibration is at minimum, the scope waveforms are almost textbook perfect!

This vibration is just like what happens when an open microphone in a PA system starts to ring and feedback. It starts low and quickly builds into a vibrating loop. If you folks think I just have bad P or D settings, I will try to tune more, but I have methodically been through over 200 iterations on a single axis.

If there is anything specific that anyone wants to know, just ask.
Thanks again for letting me test this, I appreciate it greatly.

Tony
Hi Tony,

That is good news. Perhaps I will go back to that branch. I get vibration when there is a balancing issue or parts of the gimbal creates friction due to misalignment. If not bringing up the D value usually solves it. Generally I am able to run on 100% power and even managed to get a 5D mkIII (1.5kg) working on motors as small as the 4008.

I made a sort of combination of 0.3 and the experimental branch which uses the timer settings from 0.3 incase. Weirdly enough I noticed as soon as I plugged in the motors I got intermittent I2c errors. I know I2C is prone to noise but I was wandering if anybody else has experienced this?

I would like to be able to cancel out the possibility of it being a software problem as it only happens when the actual motor is connected. I am not all that familiar with the I2C protocol.

Other than that I am getting good results. Just received four 5208's so I will give them a try next week on low power settings and watch the current with the supply.
PaulPhill is offline Find More Posts by PaulPhill
Old Aug 18, 2013, 08:59 AM
Registered User
United States, AL, Birmingham
Joined Jul 2013
47 Posts
Hi Paul,

Are you telling me that you get no noise as the motor changes positions? As the pulses change to move the motor, I hear a buzz and then quiet. In certain positions , the buzz is more pronounced and conducts into the frame and back to the gyro. This results in the oscillation I have described. I suppose my frame could just be resonating at the frequency being produced. Let me know what you think.

By the way, I had no I2C problems.

Tony
tmarone is offline Find More Posts by tmarone
Old Aug 18, 2013, 09:19 AM
Registered User
United States, GA, Atlanta
Joined Dec 2010
2,061 Posts
Quote:
Originally Posted by tmarone View Post
Hi Alan,

I have some news regarding the experimental branch of firmware.
I have only tested the straight compile, not the debug.
Everything seems to work normally, current consumption is about the same as the gold standard firmware. No FETs were harmed in the test

I found that if I decreased the motor power, things worked better, and I have a theory why that is. My frame picks up vibration from the motors and transfers it into the gyro chip. I have tried to figure out an isolation for this problem, but haven't come up with a good solution yet.

This is the frame I have:
http://cgi.ebay.com/ws/eBayISAPI.dll...m=300938964697

Lowering motor power helps greatly, but does not fully fix the problem.

Alan, could you compile a version with a higher switching frequency, and maybe the vibration will go away?

Is there a way to programmatically filter out the motor vibrations?

I had no problem loading the experimental firmware on either board.
Values of P and D do not seem as jumpy as before.
Motor current from .3 to .55 amps each. Running 35 to 45% power on motors.
Input voltage 12.3 VDC total current was 1.7A max.
FETs heat up evenly and are warm to slightly hot in use, no runaway temps.
When the vibration is at minimum, the scope waveforms are almost textbook perfect!

This vibration is just like what happens when an open microphone in a PA system starts to ring and feedback. It starts low and quickly builds into a vibrating loop. If you folks think I just have bad P or D settings, I will try to tune more, but I have methodically been through over 200 iterations on a single axis.

If there is anything specific that anyone wants to know, just ask.
Thanks again for letting me test this, I appreciate it greatly.

Tony
Tony,

Great to hear... The PWM frequency is already turned up to 18k in this version. Evaldis turned that up in 0.3g. The vibration, do you think it's vibration or oscillation? Try experimenting with a lower P and see if it's oscillation or vibration.

I'll look and see what the next step in the pwm might be and how to turn them up.

Thanks for posting, keep an eye on things and let me know if anything changes.... I got my stencil yesterday so I'm hoping to build boards today... I spent the evening laying everything out and sorting parts...

Alan
aadamson is offline Find More Posts by aadamson
Old Aug 18, 2013, 09:22 AM
Registered User
Joined Jun 2013
136 Posts
Quote:
Originally Posted by tmarone View Post
Hi Paul,

Are you telling me that you get no noise as the motor changes positions? As the pulses change to move the motor, I hear a buzz and then quiet. In certain positions , the buzz is more pronounced and conducts into the frame and back to the gyro. This results in the oscillation I have described. I suppose my frame could just be resonating at the frequency being produced. Let me know what you think.

By the way, I had no I2C problems.

Tony
Under normal operation there is no noise or vibration.

if it always happens in the same position this could either be that it is not perfectly balanced or the frame is causing friction. To test whether it is the frame rotate the IMU so that it is no longer level and test whether the vibration still occurs at the same physical angle on the gimbal. Try and distort the frame with you hands and see if the vibration goes away.

I found that vibration generally occurs when the motors are under strain. This could be contributed to control loop issue but one cannot tune for multiple scenarios.

Ok something us up with the I2C.

Hope that helps.
PaulPhill is offline Find More Posts by PaulPhill
Old Aug 18, 2013, 09:29 AM
Registered User
Joined Aug 2013
15 Posts
Quote:
Originally Posted by Overtuner View Post
Dion,
I have the 5206 here for testing - the little brother of 5208. imo you'll have no fun with them on the current EvvGC boards - there's too much power required than the boards can deliver right now (they can, but with a lot of heat at the FET's).

I'm on the way to get a new pcb layout with other drivers, the work fine and could bring the 52-class motors to warm up your coffee
but: 1.- i don't have an Altium license and 2.- I'm still to dumb to layout a completely new one with Eagle.

Lutz
Thanks a lot for the tip Overturner, its nice to get a reply from you.
DionPaz is offline Find More Posts by DionPaz
Old Aug 18, 2013, 09:30 AM
Registered User
United States, GA, Atlanta
Joined Dec 2010
2,061 Posts
Quote:
Originally Posted by PaulPhill View Post
Under normal operation there is no noise or vibration.

if it always happens in the same position this could either be that it is not perfectly balanced or the frame is causing friction. To test whether it is the frame rotate the IMU so that it is no longer level and test whether the vibration still occurs at the same physical angle on the gimbal. Try and distort the frame with you hands and see if the vibration goes away.

I found that vibration generally occurs when the motors are under strain. This could be contributed to control loop issue but one cannot tune for multiple scenarios.

Ok something us up with the I2C.

Hope that helps.
I agree here with Paul.

It sounds like either a balance issue or a slight misalignment issue... When the motors are holding, because they are acting like stepper motors, there should be no noise. If there is any friction that is causing the motor to get *quite* get to where it needs to be, there will be an offset on the IMU and the processor will try *hard* to get the motor to that perfect position. If there is a balance issue, you have the same issue... The IMU will be at it's desired position, but the holding force will need to be greater and greater so that the movement can't creap towards the out of balance.

Balance on these things and I know I sound like a broken record is...

With no power, you should be able to put the camera is *ANY* position and should *stay* there, if it moves at all, it's not in balance. And you'll have to balance for 3 axis, which the yaw axis could be fun to do that for ...

Alan
aadamson is offline Find More Posts by aadamson
Old Aug 18, 2013, 09:46 AM
Chad
FL
Joined Nov 2006
1 Posts
Do we need to worry about the crystal being shorted out?

Quote:
Originally Posted by acc007 View Post
Hi guys,

while trying to check USB hardware on v1.2 board I've discovered one major fault on my board. And it's all because I was gathering components to be ready for v1.3ce2...

It's all about the crystal. 1.3ce2 BOM suggests digikey #887-1452-1-ND. Someone in the thread suggested digikey #535-9720-1-ND as an available replacement. I also got similar crystal for my boards. But... the thing is that these crystals are 4P - have 4 little squares to solder.

PCB v1.3ce2 is ready to accept them. PCB v1.2 is NOT - it's prepared for 2P crystal, having only 2 larger areas to solder. The whole thing is about the internal connection diagram for 4P crystal, have a look:

So... pins 2 and 4 are connected. PCB v1.2 has only 2 pads to solder, so in fact pin 1 connects to 4, pin 2 connects to 3, therefore everything is short-circuited. And - that makes crystal not working, there is always a short-circuit between crystal legs.

As a result - cpu fails to start up external crystal and runs internal 8Mhz oscillator only. Therefore - motor's PWM are not rated at 18kHz, but 2kHz only ! But they work "somehow" :-) And everything else works correctly as well (GUI, leds) Arrggh... that was really hard to find out, I wouldn't have done that without my logical state analyer, I was really surprised having 2kHz on my motors :/

So - you guys having assembled v1.2 board on your own and using the crystal with 4pads - better fix it. The easiest way to fix is to replace the crystal with the one having 2 pads. But - it's not required, it can be fixed. All you need is hot air gun (to desolder the crystal) and a bit of kapton tape.

First I tried to connect external crystal with cables:


and... it worked like a charm. So... all I need was putting some kapton tape on pads 2 and 4, exactly the same as in the second picture. Then I soldered my crystal back to v1.2 PCB and it is working now as it should

And at the end, one sweet thing to all of you. I've checked VCP (virtual com port via USB) with some loopback demo and it works on v1.2 hardware! So in the near future we should have USB support builtin

Ahh - and now I can really enjoy the motors silence at 18kHz, compared to 2kHz PWM
Didn't see anyone else comment on this.
I have a 1.2 board and it would appear the crystal is shorted.
Thanks
user1 is offline Find More Posts by user1
Old Aug 18, 2013, 10:01 AM
Registered User
United States, GA, Atlanta
Joined Dec 2010
2,061 Posts
Quote:
Originally Posted by PaulPhill View Post
Hi Tony,

That is good news. Perhaps I will go back to that branch. I get vibration when there is a balancing issue or parts of the gimbal creates friction due to misalignment. If not bringing up the D value usually solves it. Generally I am able to run on 100% power and even managed to get a 5D mkIII (1.5kg) working on motors as small as the 4008.

I made a sort of combination of 0.3 and the experimental branch which uses the timer settings from 0.3 incase. Weirdly enough I noticed as soon as I plugged in the motors I got intermittent I2c errors. I know I2C is prone to noise but I was wandering if anybody else has experienced this?

I would like to be able to cancel out the possibility of it being a software problem as it only happens when the actual motor is connected. I am not all that familiar with the I2C protocol.

Other than that I am getting good results. Just received four 5208's so I will give them a try next week on low power settings and watch the current with the supply.
Paul,

The I2C lines are very susceptible RFI/EMI from all the components. This is a well known issue and it usually shows itself when the leads are long. One thing to try. If they aren't twist or braid the leads to create some inductance..... You can try either ferite beads on the leads or a toroid with the leads wrapped around it. Be careful however as you can lock up the I2C buss (or one of the devices can). You'll know if that happens as everything still just stop working.

You could also try some shielded cable and only ground the shield at one end.

Anyone else got the experimental branch working? I'm looking for someone who is willing to test what they know works with 0.3g and then move to the experimental branch. What I'm really trying to verify is that the SI unit change that we made to the experimental branch, is functional.

Alan
aadamson is offline Find More Posts by aadamson
Old Aug 18, 2013, 10:03 AM
Registered User
United States, GA, Atlanta
Joined Dec 2010
2,061 Posts
Quote:
Originally Posted by user1 View Post
Didn't see anyone else comment on this.
I have a 1.2 board and it would appear the crystal is shorted.
Thanks
The 1.2 board uses only 2 pads for the crystal, if you use the wrong crystal (the 4 pad version) as acc007 did, you'll need to do what he shows. If you use the correct crystal, you just need to make sure there isn't a solder bridge between them.

Note the 1.3 board uses a 4 pad crystal hence why some are using the 1.3 crystal when building 1.2 boards.

Alan
aadamson is offline Find More Posts by aadamson
Old Aug 18, 2013, 10:21 AM
Registered User
Joined Jun 2013
136 Posts
Quote:
Originally Posted by aadamson View Post
Paul,



Anyone else got the experimental branch working? I'm looking for someone who is willing to test what they know works with 0.3g and then move to the experimental branch. What I'm really trying to verify is that the SI unit change that we made to the experimental branch, is functional.

Alan
I would gladly do it but I only have one board left and takes ages to get another one down here in SA.
PaulPhill is offline Find More Posts by PaulPhill
Old Aug 18, 2013, 10:35 AM
Qick'nDirtySolutionExpert
Germany, BY, Bobingen
Joined Mar 2013
73 Posts
wanna look at some excercises from my "learnig phase" ?
Overtuner is offline Find More Posts by Overtuner
Old Aug 18, 2013, 10:37 AM
Registered User
United States, GA, Atlanta
Joined Dec 2010
2,061 Posts
Quote:
Originally Posted by PaulPhill View Post
I would gladly do it but I only have one board left and takes ages to get another one down here in SA.
Yeah I hear ya Paul... NO worries... I'll have a board up this afternoon and I've got a gimbal all ready to go as well.

Thanks,
Alan
aadamson is offline Find More Posts by aadamson
Closed Thread


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Discussion Open Source Brushless Gimbal edge0xc Multirotor Electronics 3052 Yesterday 07:13 PM
Idea Open Source Gopro Brushless Gimbal icais Multirotor Talk 16 Mar 28, 2014 12:07 PM
Discussion SkyLight SBG - GoPro 2 Version Released (+ GP3 Beta + 3 axis Alpha) !! ed_jza80 Multirotor Talk 494 Jan 03, 2014 10:09 AM
Sold open source brushless gimbal controller jy0933 Aircraft - Electric - Multirotor (FS/W) 57 Mar 26, 2013 07:02 PM
Sold [closed] open source brushless gimbal controller jy0933 Aircraft - Electric - Multirotor (FS/W) 16 Feb 26, 2013 09:19 AM