Possible Shield Requirements Analysis

[jmccartney]

This thread will be used to discuss basic requirements for a possible shellmo shield.

What do you want your shellmo to do?
--------
I think some big questions that need to be answered is:

Logic voltage?

Do we support the bio wire? Requires MOSFET integration.

Do we support motor drivers?

How much of the Arduino do you want to utilize? How much do you want to leave for future expansion, shields.

Price point per shield? A rough guestimate I did today is the PCB for the shield including lead free hasl will cost about 3.50 US a board. This is JUST the PCB. So going from their not including build labor (we assume this is free since this is our hobby how much do you want to spend on a single board. Just so we can keep things sane lets just base things on 10 test units to start.

----

Help me come up with more questions at first, then we can start answering them and find the best solutions for each requirement. Then we can design.

[steamboat]

Voltage

3.3/3.7 Volt with Lipo or
4.8 Volt with rechargeable batteries

What do we get:
3.7 Volt Lipo: leight weight, energy saving processor mode, high sophisticated, modern

4.8 Volt battries: easy to surce, you find it in every household, you can run most shields and sensors with it, all servos will work with this.

Disadvantages
3.7 Volt - will all servos work - will all shields and sensors work. Special charging process/unit necessary

4.8 Volt - heavy for it's capacity.

Because with our motors to run the shellmo is not an energy efficient vehicle which could work days in the field. Going to 3.3 Volt could cause some trouble for people to source or to use some stuff. My BT has an converter 5 - 3.3 volt on board...


Shield or not a shield

If we are going to make a PCB I would vote to produce it with such an Atmega 32u4. This gives us the possibility to deliver ready to run pre programed boards to people who just want to play. With the chip 7$ the USB Connector 1,5$ and the Crystal we will raise the price - but who cares. (maybe we even should not talk to much about costs here) In reverse we don't have to solder the shield pins and - the biggest advantage - we can design it as we want.

To go crazy we could use two atmega 32u4 on our board. This would be something special and even if one of it would be enough for most applications it would make things much easier to program for people like me. One chip for the walking action including bumpers - one for the eyes and other sensors - speech/noise?


Users/customers behavior/request- what to do with Shellmo?

This is a wide field of question we should ask ourself - but the key to become popular
... No time for this in the second...

[jmccartney]

2 chips is crazy, but one thing came to mind when you mention was some sort of intelligent "sleep mode" where the whole "system" isn't powered up some sensors are always on and if the right circumstances arise can interrupt the chip that is asleep to wake and control movement etc etc.

This kinda solves one of my issues with these types of creatures where they seem very superficial, they are either off or on. Having that blur a bit will make them feel more "alive".

----
If we did not go the shield route the PCB would probably be a little cheaper, you could go off a 10x5cm base instead of 10x10cm that the shield would require (Arduino shields are a little wider than 5cm thus requiring the move to 10cm width) Obviously then you need to design an entire power system....which will add to dev time. My one good power system is all Lipo based and I don't think its going to work here.

[jmccartney]

Another thought would be stay with the "shield" but build in a 32u4. Then you could stack the shield on top of a uno or leonardo etc etc if you wanted some more intelligent sleep mode etc etc. Lots of options there.

[jmccartney]

I have a teensy 3.0 which is really fun. Not sure if anyones interested in going 32bit. It runs arduino and sketches exactly like your uno, but you have a lot more horsepower and it has a RTC.. They are 19us.

https://www.pjrc.com/teensy/index.html

[steamboat]

Sleep mode and power management are two things we have to deal with our electronic calipers at work stead rest

This electronic system measures the quill travel in the steady rest and displays it on an LCD. In the mill/lathe we can not power it by wire. So a rechargeable battery has to do it. There is no on/off switch because the user will forget it anyway. So we switch it off completely after some minutes to avoid a battery damage. When pressing (the only) button it comes to live again.

First off all you need a circuit which can switch on and off all main consumers. This could also be used when you are working only with 500 mA USB power to switch off all the hungry muscles.

Additional we should watch the battery Voltage and switch off if to deep.

What else of power management do we need?

---

(05-12-2014 10:55 PM)jmccartney Wrote:  ... going 32bit...horsepower and it has a RTC...

If we make thinks common on one side maybe more people will/can follow. If you write "ARDUINO" many flies will "detect" it ;-)

My two MC where only a brain storming idea - and didn't came out of an power request/need

[steamboat]

My tests with Sho's BT App revealed the necessity of an circuit - or at least a switch - to power down the servos manually.

First it would be nice to save battery energy.

Second it could prevent the servos from damage.

Third and important: while you are experimenting and programming Shellmo.

Also an LED for each servo could be nice to see if the program tries to move the servo without having the servos attached. One hour of servo movement is a torture for you and yours!

[jmccartney]

(05-13-2014 05:34 PM)steamboat Wrote:  Also an LED for each servo could be nice to see if the program tries to move the servo without having the servos attached. One hour of servo movement is a torture for you and yours!

I am trying to wrap my head around what this circuit would look like..It would have to rely on the PWM input for indication of servo activity right? But how would you lay that out? Analog is my weakness.

[steamboat]

No analog stuff please. Why not an LED attached to the port and if you soft-switch to test mode the power to all servos will be cut off and you only see the led.

Maybe all this are special wishes. To sum it up: I would like to have a digital component which can switch off all servo power an withstand about 1 or 1.5 Amp max. without to big losses. If we are generous we could take two of this to stop moving and eyes seperately - maybe not necessary but would save energy. As an extra we could install a third thing like that to switch on and of other "power" consumers like a LED, a Camera and so on.

Maybe energy saving is not a major point in this application but it is more fun to work with such a toy more than 5-10 minutes before recharging 2 hours or more. Walking will not be the major field of this creature - I guess. So it is worth to switch off that stuff. On other hands I never measured the Amps needed by the servos - will do that tonight. Maybe all this thoughts are an invest in nothing...

[jmccartney]

Here is a logical diagram, and a rough circuit schematic I worked on yesterday. I am really stuck on the idea of having "3 modes" a semiconscious mode operated by what I am calling for now the "brain stem" is in sleep mode (woken by interrupt of a sensor), a conscious mode operated by the brain stem fully active, and a active mode where a Arduino Leonardo etc etc can take over "higher level" functions.

The brainstem would monitor power and one or two basic sensors, maybe vibration and sound or light. This can then "wake up" the rest of the "brain" the Leonardo for instance using a combination of mosfets and interrupts on the Leonardo.

Theoretically you could keep a shellmo powered on for days and only a few moments where user interaction really "wakes" it up will power be used in great quantities.

I just think that if your going to make it look like its kinda "alive" it needs to act that way....