Planetary Gearhead Motor Servo Project Update!

Lynxmotion Servo Erector Set
61

Jim,

Over at OpenServo (see Barry’s post above), we have a servo bridge design for your new motors that will deliver 10 Amps all day long - so long as we can get rid of the heat generated. Running a 10A motor will produce approximately 15 to 20 Watts of heat on the board. Your plan for mounting the board:

will not allow us to get rid of that 15 to 20 Watts (heatshrink is a great insulator). As Barry said at OpenServo:

If what the airplane folks are doing and using, would always work for robots, there would be no need to do a new servo design - different problems need different solutions.

The servo board needs to be mounted such that it is protected and at the same time allows air flow around the board components and heat-sink. Ideally, the board mounting mechanism would also couple board (FET) heat to the robot frame - the cooler the board, the better the performance.

Back in March, I posted a first cut on my thoughts relating to this problem here (last post - I built it, but nobody came :frowning:). Since you are throwing more motors into the mix, a more general solution is probably needed, but as a starting point for discussion it would help if you could comment on that post - either here or there.

We now have a workable schematic, but to get to a PCB we need some guidance on a mounting plan that is workable. Without that mounting plan, we will be dead in the water.

[Edit: Well, I screwed that one up. :blush: The power numbers should be 2.73 to 3.38 Watts - see my post below.]

Cliff

62

I’m going to have to discuss this with my engineer buddy Mike Dvorsky. I’m a little confused though. Doing the math it looks as if the RDS on would be 0.1 ohms per FET in order to need to dissapate 20 watts at 10 amps. That sounds a little high. Let me get back to you on this.

63

Here is an example of a workable solution to the heat dissipation problem:

http://www.finedesignrc.com/speedcontrols/bkmarine9920.jpg

Include a heat sink and make a cutout in the wrapping to allow for air flow around the heat sink. I own this ESC, so I know it works.

64

Is that thing water cooled ?? What are the rubber hoses…

65

Yup, that’s the boat version. I have the car version which is identical, minus the hoses.

I just wanted to show an example of a heat sink used with a PCB and heat shrink wrap.

66

Thats an amazing idea!

67

hm, as good as an idea seems, i thounk it wodul be better to have the heat shrink to enclose around bottom of the heatsink rather than have it around the top, unless the heatsink is being held in by the shrink wrap

a hole in the top…i just dont think that allows for good airflow,and it would be better to cut the hole the sink of the heatsink then push the heat shrink down and around the heat sink, this way the heatsink would get a lot more air to it :slight_smile:

68

That would work if the heat sink had a flange around the bottom, which might be hard to find off the shelf. I think the cutout is sufficient. It doesn’t have to be ideal, it just has to work. :wink:

69

Jim,

:blush: Sorry, I guess that will teach me not to pull numbers off the top of my head, when I’m tired. Slipped a digit somewhere.

The real numbers: The FETs we are using (IRLR7807Z) have a Rds(on) of 13 mOhms at 25 degrees C (case temperature). The high-side FET will have additional switching losses, drive losses and output losses. The low-side FET will have additional drive and output losses. Also, we have power losses in the gate drive circuit and in the current sense resistor. Driving 10 Amps at 25 degrees C (case), the total power loss will be 2.73 Watts and at 85 degrees C (case), the total power loss will be 3.38 Watts. An Excel spreadsheet used for the calculations is here (the method is shown in the data sheet linked above).

Fewer watts, but we still need to get rid of them.

Cliff

70

could we get a pic of what uve got so far ,might help if we knew what package looks like,

71

Hi,

We have not started to lay out the PCB as yet, as this depends on a few other pieces of information we are awaiting.

At this point, we have a schematic that lays out the plans for a 10A OpenServo motor controller, and we will be making more progress this week.

You can view the entire process over at:
openservo.com/Forums/viewtopic.php?p=2912

Barry

72

Hi Cliff,

I think I should be able to come up with some kind of method to attach the board to the assembly short of the traditional nuts and bolts method. There isn’t any good way to attach the board to the bracket, and keep the brackets symmetrical. Symmetrical brackets are more versatile.

73

Hello ginge,

This is some ineresting stuff. What part do you do? Design, Eagle Cad, etc… ?

Perhaps I can watch and learn from you.

74

Hi Mike,

I am not sure what you mean by the question, whether you mean right now with this design or generally :wink: I will go for both.

I am currently working the design in Eagle PCB. Once we have the schematic right, I can (with the help of the OpenServo guys) start to lay out the PCB (again in Eagle)

In the OpenServo community as a whole I play the role as an admin and also a contributor. Although I do not have much to do with the OpenServo software core (other than constantly ask for new features and changes – I leave the hard work to (our) Mike :wink:) I do have an off tree patch set with some really quite bizarre features. I also create OpenServo derived hardware which is on my website. I have written a few tools and software that makes working with OpenServo hardware very simple. I also try and keep the website clean and documentation up to date. Call me the e-janitor.

Join OpenServo Forums, you will learn plenty (I know I did)

Barry

75

I had no idea…

So, you are the head honcho over at open servo then. I have been there a few times (wiz-bang) but it’s over my head kind of stuff so I have not been around much.

Its nice to see the community come together and work together like this.

76

We can all learn a LOT from Barry (ginge), Cliff, and the other Open Servo folks. :smiley: Anyone who wants to get into the guts of electronics would do well to peruse the Open Servo forums regularly. I do! :smiley: I don’t pretend to understand even a fraction of what they do over there, but I am catching on little by little.

8-Dale

77

You got me beat then, cause I can’t get it. I keep on thinking but nothing happens… :laughing:

78

No I am not the head honcho… That would be Mike Thompson. The creator and all that is OpenServo. I just do the bits no-one else likes to do :wink:

79

Oh, I see…

So they have you do all the eGrunge work. :laughing: Just Kidding

Yes, I remember now, Mike Thompson… Shows how little I have been over there. :blush:

80

Hi Jim,

Believe me, I understand your concern.

I am pushing you and the folks at OpenServo, because I would like to see this product line live up to the high standard that you have established in your Servo Erector Set mechanicals. You have done an excellent job integrating standard R/C servos into parts for robot design - why would you do less for your own servos?

First order, I thought, since you have that large motor hanging off the bracket, that mounting the PCB under the motor would be the best place, since that general space is needed for the motor anyway, but you are right that doing so impacts symmetry. Having said that, the first thing that pops in to my head is: Ok, but if the asymmetry is adjustable, is it still a problem?

In my experience, inventing a solution to a problem is easy. What is not easy is defining the problem. That is the case here and so far, nobody seems to want to deal with defining the problem - not here or at OpenServo.

I would like to help solve the mounting problem, but I don’t have a good feel for the space requirements dictated by how you see the new servos being used. Beyond that, I have no idea what the dimensions of your brackets are and how they would be mounted in the various applications. In your arm picture above, there are no space restrictions around the motors, but I am sure that there are in using the motor in a robot shoulder or hip. In short, I don’t know what the problem is and since I don’t have a room full robot bits to play with, I can’t define the problem myself.

Still, I would be willing to chew on solutions, if I had the bracket and motor dimensions. I have the datasheets on the 22mm motors, but not the rest. Can you post or e-mail me drawings? (Models in .step or .igs would be great, but 2D .dxf would be fine) Also, any guidance you can give me on space restrictions you know about, in planned uses for the new motor, would help. I can post some more ideas and you (and yours) can throw rocks - enough of that and maybe, at least, we can define the problem.

Cliff