OpenServo.com

[Cliff]

poorobot,

Throwing sticks now! And here I thought you were a rock throwing kind of guy.

Well in any case prepare to duck - they are coming back your way.

There are two problems with using the TPS76550 [or TPS76150 see below] on this board:

a) The TPS76550 recommended Max. input voltage is 10V and the absolute Max. is 13.5V, where the battery voltage for this board can be as high as 16.8V.

b) If the board is powered by a 5V supply through the programming connector and the battery is not connected, the TPS76550 protection diode will conduct current (that is not limited) until it can charge the capacitance connected to the battery supply to 5V. On this board, battery supply capacitance will be around 1000uF, which will take a lot of current to charge (initially it will look like ground) and since neither max. current or max. power are specified for the diode, it is hard to tell if it would survive. The diode was only intended to discharge 5uF or so of output capacitance into an input source that had been shut off and was sinking due to connected loads. Here is the statement from data sheet about the protection diode:

[poor-robot]

Haha, well, whatever's handy, Keeping with the vehicular motion theme

Shoot, I missed the 16.8V voltage input requirement. How about adding a ~470 resistor on the input side. That could work, yeah? It would prevent the diode from reverse-conducting beyond its limit, and given a maximum current draw of 15-20 mA out the 5V regulator your LDO should still be functional. With a minimum current draw of 4 mA you might actually stay within spec

Probably not the best solution though. Definitely not the cleanest. For a 12V input I think it would work just fine because you wouldn't be counting on current draw to keep your regulator within spec.

[Jim Frye]

Hello everyone,

First off I want to thank all involved in getting this project off the ground. I have actually caught up a bit on things, enough that I could read this thread. I'm trying to cover a lot of stuff while I have the chance. I don't know how long my caught up status will last. I'm trying to remember any questions or required feedback here. I hope I don't forget too much...

The issue of mounting the controller. I originally wanted to make it like the RC aircraft "speed controllers" as they are very powerful, and completely encapsulated in heat shrink plastic. The fact that I see lots of them available in high current packages made me think it was a viable option. Mounting with double sided tape seemed to be reasonable too. Note: I dunno how they pack so much current capacity into such small packages. Now I realize they are only one direction, no reverse, so maybe that's why they can make them so small...


I am now thinking it would be better to go ahead and make it attach directly to the motor via a small aluminum "L" bracket. One end sandwhiches in between the motor and the mechanical servo bracket, and the other end holds the controller. The actual mounting hole placement is not critical. It would be nice if the controller board / bracket was limited to 1" to 1.25" on one dimension, and 2.125" on the other if possible. This will allow the board to take about the same area as the smaller motor we are wanting to use. Please use 0.125" mounting holes. The bracket may also be used as a heat sink...

I want to keep the project as inexpensive as reliable operation will allow.

I would like to try and incorporate I2C, RC, TTL serial, incorporated into a single row of pins if possible.

Our customers immediately wanted to know where the encoder connections were. I think it would be an excellent addition if there was room for it. We are getting 37mm motors with 12 CPR hall effect encoders built onto the rear of the motor. They are 200 rpm, 30:1, spur gear motors, not good as a servo, but excellent for rovers. We are using the following pin format. Standard .025 square posts spaced 0.1" apart.

1 = GND
2 = Index (not used on our motor)
3 = A Channel
4 = + 5vdc
5 = B Channel

I can't think of anything else. I have a few motors available for testing. I'm doing my best to keep things going... Thanks again...
_________________
Jim Frye, the Robot Guy
Lynxmotion, Inc.
home of the Servo Erector Set
http://www.lynxmotion.com

[Cliff]

Hi Jim,

Glad to see that you have found a few spare cycles for this project.

It would really help if you could send me specifications on the motors you plan to use with the new servo board, also dimensions on the motor brackets. You could PM the information to me here or if you prefer, you could sent the information to Mike Thompson and he will get it to me.

In the mean time, here is a straw-man design to throw rocks at - I'm still guessing on dimensions, so look at this model as a concept only:



A few notes: The large screw terminals and capacitors are needed to handle 10 Amps. The strain reliefs could be eliminated to shorten the bracket by 1/2 inch, but cable stain relief needs to happen somewhere. The connector on the side is a 20 pin 2mm connector (for size), a 0.1" connector could be used but it would need more head room and board space. The motor shown is the small motor (pghm04 / pghm05). The 0.125" x 0.25" x1" aluminium strip shown (between the PCB and mounting bracket) couples the FET heat to the mounting bracket and will require an electrically insulating pad between the PCB and the coupling strip. Also, when thinking about applications, recognize that the servo board bracket can be rotated in 90 degree steps around the motor bracket.




The following dimensions are only so you know what you are looking at - as I said, I'm still guessing.




Comments will be appreciated.

Thanks,

Cliff

[Cliff]

Hi All,

Wow, really heated discussion on the Lynxmotion Servo Bracket. Hello, testing 1, 2, 3 . . . is this thing on?

Oh well, I guess I'll have to pretend it is working and continue with some thoughts on connectors for this board. The last pass, Barry posted, on the board schematic had the following connectors and signals:

AVRISP
1- MISO / INTR
2- VISP
3- SCK
4- MOSI
5- RESET
6- GND
7- NC
8- NC
9- PD5
10- PD6

I2C / SERIAL
1- SCL / RXD
2- +5V
3- GND
4- SDA / TXD

RC Pulse
1- GND
2- RC_PWM
3- +5V

POT
1- +5V
2- WIPER
3- GND

Additionally, Jim Frye requested the following:

ENCODER
1- GND
2- INDEX
3- Channel A
4- +5V
5- Channel B

What I would like to propose, instead of the above, is the following:

Pot / Encoder
1- INDEX (SCK)
2- Channel A (PD5)
3- Channel B (PD6)
4- +5V
5- WIPER
6- GND

I2C / Serial / RC Pulse / AVRISP
1- SCL / RXD
2- GND
3- RC_PWM
4- GND
5- SDA / TXD / MOSI
6- MISO / INTR
7- VISP
8- SCK
9- RESET
10- GND

This connector arrangement would mean that however the board is being used, there would always be a six pin connector used for motion feedback and a ten pin connector used for servo communication.

Six pin connector examples:
Pot feedback would use pins 4,5 and 6.
Encoder feedback would use pins 1, 2, 3, 4 and 6.

Ten pin connector examples:
RC Pulse control would use pins 3 and 4. (Jumper pins 1 and 2)
RC Pulse control w/ serial feedback would use pins 3, 4 and 5. (Jumper pins 1 and 2)
I2C control would use pins 1, 2, 4, 5, 6 and 10. (no Jumper)
Serial control would use pins 1, 2 and 5. (Jumper pins 8 and 10)
AVRISP would use pins 5, 6, 7, 8, 9 and 10. (no Jumper)

[edit : I made the following comment in an earlier post in this thread, that I think is worth repeating here:

[Cliff]

Hi - I'm back,

In my initial look at the servo bracket, I used a 20 pin 2mm (2x10) connector just to get an idea of available space. Since I have proposed using a six pin and a ten pin connector, let's start by replacing the 20 pin connector and adding the mating connector bodies, so we can see what the space issues are.


Fits in about the same space as the 20 pin connector and we can see below that the mating connector body overhangs about 1/8 inch.


Since, it seems, everybody else is big on using 0.1" center, 25mil post connectors, let's give those a try.


Well, they will fit - but the overhang with the mating connector body is getting a bit less friendly.


The problem is that the connector can not go inboard due to interference with the motor.


The only help for it, is to raise the motor by 0.2" and move the connector inboard. The problem is when the connector moves inboard, it interferes with the forward PCB mount - which is solved by moving the connector to the far-side of the PCB, moving the mount point outboard and changing the mounting screw from a pan head to a flat head screw. The connector in its new location w/o the mating connector body:


With the mating connector body:


Here is a look at the dimensions with the motor raised:


The big trade-off, using the 0.1" connectors in the inboard location, is that the top side of the PCB can no longer be used for mounting components. This is going to make the circuit layout very tight and may require growing the length of the board. The question is: Is it worth making the bracket and PCB bigger, just to be able to use the 0.1" connectors?

The 2mm parts I've shown on the above drawings are Molex, but I think the Hirose 2mm (DF11) parts would make it a no brainer for users, as DigiKey stocks pre-crimped wire (2"-12", 24 - 28 awg) for the Hirose connectors. The difference in my mind is waiting a day or two for an order from DigiKey (2mm parts) v.s. instant gratification from Radio Shack (0.1" parts).

So, let me know what you think about the above.

Thanks,

Cliff

[Jose.Torres]

Hi Mike, Jim, Cliff and everyone else,
I'm unsure of the status of this project but the initial design with the 6 and 10 pin connector @ 2mm is very clean and convenient. The an AVR-ISP .1" to 2mm adapter can easily be made.

Has the motor been decided yet? I noticed the GHM-09 initially used in the high-torque servo design by Jim has been phased out in the lynxmotion store.

I have found an alternative motor from Shayye named RA-37GM, 11 & 12 Type running at 41RPM @ 12V.

[edit] As I had to create a new account I can not post urls yet so the datasheet has been redacted.[/edit]

Love to see this project come to fruition. I would like to build a humanoid robot based on this Open Servo design as a seed project for the Open Humanoid Initiative, a non-profit organization for the development of a open full-scale humanoid robot. I believe such an endeavour would be inline with the originators of this project.

Any news or status updates would be greatly appreciated.

[robotjay]

_________________
"Nothing is fool-proof; For we fools are ingenious and will find a way."