OpenServo.com

[ginge]

Hi Bren,

Take 2.

http://www.headfuzz.co.uk/files/openservo/openservointerface-test2.zip

You will need the qt3-mt.dll file from the previous posted zip. This one does a save after the I2C address is set. It tries to force the issue so you may see some failed writes in there.

You will need to power cycle the OpenServo before trying this.

Fingers crossed!

P.S still not addressed the voltage issue

Barry
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

[bren]

Will try in the morning as I've just got home from the lab.

Bren

[robotjay]

I just wanted to say that I got the Pendragon Robotics site working, and you can purchase from there if you wish. I've also created a new thread with more information here. Talk to you soon.

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

[ginge]

Bren,

I have figured out where the problem lies. There is infact no problem, and all of the application patches above are pretty much useless, although the last upload I did will work bus is less than elegant.

Here is the procedure to change the i2C address.

1) Change the address in the box.
2) Write the address using write.
3) Click Save
4) reset the servo using reset or preferably by power cycling.
5) rescan the bus, and presto it should change.

The programmed address is only initialised at startup, and therefore you need to write and save before it works as it should.
I have verified this procedure and it works for my hardware.

I hope this works for you

As this is not a V3 hardware issue as such, I recommend we start a new thread if issues are still occurring.

Barry
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

[bren]

Hi Barry

Thank you so much that last bit of advise works like a charm. Now to install 18 of these boards .

Bren

[bren]

Hi All

I'm in the process of assembling all the servos. I have a question, I've read that people put a capacitor between the motor connectors, is this recommended and what capacitance. It's a standard ish servo that I'm going to running at about 8V.

Thanks Bren

[ginge]

Hi Bren,

I tend to go with a 0.1uF ceramic cap (marked 104J). It seems to work well. if you find that you are getting servo resets you might want to bump that up a little.

Barry
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

[kbb]

Over the next week or so I plan on swapping all the 20K R6s (battery volts divider) on the V3 boards I have here for 10K resistors.

Although they are even more fiddly to do, I also intend changing R17 and R18 (BEMF dividers) to 10K at the same time. Unless someone can think of a reason not too?

Kevin.

[ginge]

I know it is a bit late in the game to be requesting hardware changes, but hopefully 3.1 can accommodate this?

These requests are pretty simple, and come from prolonged use of the v3 hardware.

1) PD2 (EN_A) and PD3 (EN_B) are also the AVR interrupt lines. If we want to support encoders without too much messing around with firmware, would it be possible to move these to PD1 and PD0 (the serial lines) or preferably PD8 and PD9. That way we have the external interrupt lines available as well as serial.

1a) Would it be possible to bridge PB4 and PD2/3 so we can have an external interrupt source? I have had a couple of scenarios where something like an emergency stop would have been extremely handy. I realise the new broadcast features in firmware can do this, but a hardware pin would be nice.
I also have a basic OpenServo running from software USB instead of I2C, the hardware limits this from working natively without bridges all over the PCB.

1b) Given the extra space we have, can we have some through hole/SMT connections to the above interrupt pins? This is not essential, but would be extremely nice to have. They don't even have to be near each other.

2) It is a pain in the backside to solder the pot wires into those tiny holes. Now we are using the smaller MCU, can we make these bigger?

3) The motor mounting holes are tiny. Can we make this a copper pour so I can solder onto the tabs instead of going through hole? Maybe we can just have bigger holes?

Again, sorry to drop this on you all so late on

Barry
_________________
http://www.headfuzz.co.uk/
http://www.robotfuzz.co.uk/

[bren]

Hi there

[ginge]

[kbb]

I’ve just added one (click here and here) version of Cliff’s heatsink solutions to a V3 board: cutting complete heatsink shapes out of copper (click here) is a lot of work (especially given the size) and I am not sure if the extra dissipation that they would provide is necessary.

Talking of heat dissipation this has led to two mounting scenarios for the board and I haven’t been able to decide which is “best”. So before I do the rest of the servos, does anyone have any thoughts on which of the following is the best position?

I have just noticed that with option 2 means there may be a clash with the mounting of the servos in the Servo Erector Set brackets, in some cases.

1. Heat-sink inwards (towards the pot/gear assembly).

One side effect of this is that the motor gets really hot and the heat-sink is, of course, right next to the motor. This may reduce slightly (or more) the effectiveness of the heat-sink solution.

In fact with continuous unloaded running between each extreme position, whilst I can’t really detect any heat from the MOSFETs on an unmodified V3 with my fingers, the motor gets too hot to touch.

A possible advantage of this solution is that the temperature sensor is facing the back and is so moderately insulated from most of the heat from the motor by the V3 board itself.

2. Heat-sink facing the back (away from the pot/gear assembly).

One thing about this is that you need to cut a hole in the back plate of the servo case for the heat-sink to emerge from: this isn’t as hard as it seems it took me around 45 minutes to cut a reasonably neat looking one with a sharp knife (cutting from the inside makes for a neater looking hole when viewed from the outside).

However, the problem with this position may be that the temperature sensor is now inside the main enclosed space of the servo with that nice hot motor, and it is also, of course right next to the motor. Possibly reducing the usefulness of the temperatures it reports?

3. Images of the conversion.

The following images are on "ImageShack": I will look for somewhere else to put them, if there is an issue with the adverts (sometimes Norton complains about them).





Kevin.

[robotjay]

Kevin,

Glad to see that you are pioneering the heatsink mod that Cliff suggested. With regard to the placement of the heatsink, your guess is as good as any. Either has its benefits. IMHO, protecting the FETs is more important than accurate temperature readings, and placing the sink outside the case is going to have the most effective results.

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

[jharvey]

Seems like most of the design info is in this thread, so I thought I'd post this link

http://www.ladyada.net/library/pcb/costcalc.html

The above is a nice tool for predicting costs of a pcb run, before you place an order. I put in the V3 board, with a min qty of 10, some prices came back with a price for each board of less than $.50 and total cost of $100 or less. Granted these are double sided not 4 layer, and likely are 1oz Cu or less. But still a nice prediction tool. I hear it's fairly accurate.

[edited to fix link]

[guru]

This is a group purchase? I would like to share in the group purchase of the pcb blanks. Count me in for 25 or so.

Colin