Motors shield for the Panda

@ Geir : here’s what I’ve got with your suggestion.

I’ve added a LM317 and its circuitry to obtain 6.46V and 4 3-pins headers for the servos. They are connected to IO68, IO61 to IO63 of the Panda pins.

The driver for the servos can then be a modified version of the GHI driver, I think. Only pinout to edit. But I will check.

I don’t know if the voltage is good. Should it be exactly 6V ? In this case, I will try to put 2 resistors in parallel to obtain the correct value.

That looks great Bec a Fuel :clap:
In stead of the LM317 I think I would go for a regular 5V regulator. 6V will give you marginal more torque but a straight 5V sufficient.

If you can make this work, I’ll definitely buy one!
Mark me down for a signed copy no. 1
;D

So I can put some kind of L7805 and remove 2 resistors. Is that ok ?

[quote]If you can make this work, I’ll definitely buy one!
Mark me down for a signed copy no. 1[/quote]
If you don’t mind, I will keep #1 for me :wink: But you will get signed #2 :hand:

The only thing I fear, apart from a design flaw, of course, is soldering the DRV8821 ???

There will be also some time after that to have a complete driver or set of drivers.

As fare as I can see the L7805 will work just fine.
I’ll settle for #2 :slight_smile:

No rush, I haven’t any robot plans for the moment but I will definitely build one on this shield.

When I look at the L7805 datasheet, it says that both capacitors are optional.

For one, they say [quote]Although no output capacitor is need for stability, it does
improve transient response.[/quote]

And for the other [quote]Required if regulator is locate an appreciable distance from
power supply filter.[/quote]

What is the “transient response” and does it apply to such a board ?

What is an “appreciable distance from power supply filter” ? 5cm ? 1m ? 10m ?

Not that I don’t want to pay for 2 resistors, but they would give some more room on the board if they can be removed :wink:

I’m only guessing here but I think you can drop the capacitors as they are there for smoothing the voltage. As this will only power motors/servos I don’t think there will be any problems.
A diode to eliminate problems with wrong polarization would be nice.

By the way, there are some good tips here [url]http://www.robotroom.com/PCB-Layout-Tips.html[/url] like the test points, that is always handy.

Thanks Geir !

I have then replaced the regulator by a LM1117-5.0 with a 1N4148 on its output line. Also removed unwanted resistors/capacitors that were attached to it.

It seems fine to me, now. What do you think ?

Edit:

[quote]Welcome to the hardware challenges. This is the biggest problem GHI face,


…[/quote]
I won’t satisfy everyone here, but 4 more motors seems reasonable, doesn’t it ?

Hi Bec a Fuel,

Transient Response, as I understand it, is the time it takes for the regulator to adapt to current changes.

The regulator gets caught by surprise when the current suddenly increases, causing the voltage to drop. The regulator will adapt and bring the voltage back into regulation, but it takes time. The cap helps to minimise that voltage drop(or spike if the current drops suddenly).

Hope this helps.

Thank you Errol.

Then it may not be a good idea to remove at least one of the caps in this case. :think: I can’t assume that nobody will not start the four servos at the same time, hence causing a big voltage drop that the cap would help to deal with.

But what response times are we talking about? Motors and servos are slow reacting devices so if you loose some uS on responstime I wouldn’t think it would matter.

“transient response” I take as meaning not so much a response time, but the reaction to something occurring, specifically an instantaneous high amperage demand causing a decrease (“sag”) in voltage.

For once Wikipedia doesn’t have all the answers, [url]http://en.wikipedia.org/wiki/Transient_response[/url], but I think the generalisation is that the caps help dampen the size of the “transients” when for instance a motor is first powered on. That’s obviously more critical when powering a sensitive device, but a motor may not need this (or may not need as much damping).

Thanks a lot Geir and Brett.

Then I will consider the current board and layout as “definitive”. I also have made bigger traces for motors power, as suggested before or in another thread.

Since I have no DRC failures and assuming (well, not assuming, I’ve checked it) there’s no design flaw because I’ve followed the datasheets, I will now put the different texts at their places to help place components.

Unless someone is seeing something obvious that I could have missed, of course :wink:

I find it very hard to find flaws in your design based on a gif image with all layers turned on. Maybe if you posted some images of only bottom and top copper layers it would be easier?

One small thing: Why the mounting holes? Those will serve no purpose.

One large suggestion :slight_smile:
How about extending the board to the right to give room for a micro SD card? Then the user could solder one on or leave it alone or even cut it of if it wasn’t needed. :smiley:
(link removed)
It’s only 8 wires and I believe there is a SD footprint in the Eagle lib.

Here they are :

If you have Eagle, I can send you .brd and .sch files.

Indeed. But since they do not disturb the overall thing, I’ve left them in place. I can easily remove them, of course :wink:

[quote]One large suggestion
How about extending the board to the right to give room for a micro SD card? Then the user could solder one on or leave it alone or even cut it of if it wasn’t needed.[/quote]
For this board, I didn’t want to go outside of the original dimensions :think:

And I think it will become a “monster” if we add too many things :wink: And the driver will also grow. I have to admit that I’m not really for this extension :-[

:slight_smile:
No problem, it’s your design after all. And by the looks of it it will be great.

Is someone interested by this board ?

If yes, would you please contact me at : #becafuel#@ #lsp-fr#.com (remove # of course :wink: )

Are there 3 (rash) guys here ready to take some risks ? :whistle:

Why 3 ? Because 4 PCB are only 10 euro more than the amount for 2 PCB ::slight_smile: And also because I have 4 DRV8821 and some other components.

It’s 133 euro for 4 boards (33 euro each if someone gets in) and a delay of 8 days at PCBPool. Silkscreen and soldermask on both sides, surface "Ormecon - partial chemical tin for ultra-flat pads ".

This board will (should) :

  • control 2 DC motors via PWM and L293D (Panda, Domino)
  • control 2 bi-polar steppers with Texas Instruments DRV8821DCA (Panda only)
  • control 4 servos with modified GHI driver (Panda only)
  • use EEprom to store various informations via I2C (Panda, Domino)
  • expose 4 more I2C headers, to add other sensors if needed (Panda, Domino)

Used pins are marked “N/A” (not available) on the silkscreen to avoid using them with another shield.

Here’s a zip file that contains top face, bottom face and partlist.

[italic]Edit: [/italic]
@ GHI staff : if this post is not correct regarding forum rules, please do not hesitate to tell me and I will remove it. Sometimes, speaking about money or business between members is not well accepted :think:

As long as you are doing things to help FEZ (not compete with FEZ) and help the FEZ community then you can sell whatever you like on this forum.

It’s not about selling, here. It would be rather a sort of mutualizing fees, nothing else. :hand:

If nobody is interested, then I will take care of the whole amount for me, that’s all. There will be only one batch of 4 boards, not more. At least for me.

Unless everybody wants one, of course, in which case I would proceed differently. :smiley: