I want to design a shield which would take power from a 24V supply and use a switching regulator to downconvert it to 5VDC or 3.3VDC to limit power losses and heat dissipation. However, I need 3.3VDC on the board for other components.
I believe that I cannot simply bring externally produced 3.3VDC into the Arduino board through the 6-pin POWER connector without making it unstable because that pin is an output from the FTDI chip as opposed to an arbitrary input for power. So, I was thinking instead about bringing in 5VDC on the 6-pin POWER connector, and then trying to extract 3.3VDC from the arduino instead of putting yet another regulator on my shield.
Does the FTDI chip produce the 3.3VDC output if the board is powered from a 5V supply provided by direct connection to the 6-pin POWER connector?
From everything I can see in the schematic, the components on the board will be perfectly happy running from a 5VDC supply attached in this way, and my understanding of the switchover circuit at the top of the schematic is that if 5VDC is present in the board, it is automatically connected directly to the VCC5I and VCCIO and RESET lines on the FTDI chip, with USBVCC only connected to +5V if VIN/2 is less than 3.3VDC/2 (which in my case it will be).
So I think that this gives me three states then, but I'm not sure:
-
+5VDC is supplied directly to the +5V pin. No USB cable is plugged in. Because +5VDC/2 is not larger than 3.3VDC, the USBVCC will be connected to +5V, but as there is no connection no current will flow. Offboard power supply wins, and powers the FTDI chip, producing 3.3VDC on the 6-pin POWER connector.
-
+5VDC is supplied directly to the +5V pin. A USB cable is plugged in. Because +5VDC/2 is not larger than 3.3VDC, the USBVCC will be connected to +5V. In this case there are now two supplies on +5V, which is bad as they could have different voltages and result in large current draws due to the low impedance connection and high sourcing capacity. In principle, 3.3VDC is provided by FTDI chip, but the connected supplies are a serious problem.
-
+5VDC is not powering the +5V pin. A USB cable is plugged in. The USBVCC will be connected to +5V and the board will also output 3.3V on the 6-pin POWER connector.
Is this a real problem? Is the solution to remove the 5V supply (MC33269D-5.0) from the arduino board in a custom version and replace it with my switching supply capable of handling 300mA and 24->5V conversion without overheating, and then just supply +24V to the Vin pin of the arduino? Or combine my board with the arduino schematic in a giant monolithic solution? Is there a better way?
Thanks for your reply.
I am talking about the FTDI USB->USART chip contained on the main Arduino board ("Duemilanove") and am referring to the schematics for that board alone. If the other Arduino boards behave differently, perhaps in a more desirable way, I would be happy to look at different versions, as I am not wedded to that design.
I am talking about the FTDI USB->USART chip contained on the main Arduino board ("Duemilanove") and am referring to the schematics for that board alone. If the other Arduino boards behave differently, perhaps in a more desirable way, I would be happy to look at different versions, as I am not wedded to that design.
Well as you may already know the main limitation of the standard Arduino 3.3vdc volatge source is it's 50ma maximum current limit. Also wiring a external regulated +5vdc to the Arduino's +5vdc power pin is something many do, but is not considered good engineering practice in that if the Arduino is plugged into the PC for uploading or communicating, you then have two +5vdc sources hardwired together.
I recently posted some information on how the Seeeduino clone boards differ in their power management and more robust 3.3vdc source that might might relate to your questions: http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1284225531/11#11
Lefty
Hi Lefty,
Thanks for the reply.
It sounds like the 3.3V supplied by the FTDI chip will be unsuitable, thanks for pointing out the current limitations, I believe this chip requires a few hundred mA.
Given that limitation, would you agree with me that the best solution for using a standard Arduino board (I imagine I would also be redoing a lot of the power circuitry if I did an integrated derivative board) would be to have my switching regulator supply 7VDC instead of 5VDC, and connect that to Vin?
I think that this will cause the Arduino to be running "as expected" with power connected to Vin, and that it should then provide me with a reasonably high current 5V source, right? Can the Arduino onboard regulator take another 0.6W of heat generation?
With this solution I would not have two supplies on any one net, I think it adds a fairly minimal amount of power dissipation requirement to the Arduino, I still use my switcher to drop the majority of the voltage, and I still get the +5V and +3.3V supplies I need for my shield.
Thanks again,
Brian
By the way, in case you're interested, the shield I am designing will control a very high power LED light controlled over WiFi. It's very fun to play with. I was previously using a nasty perf board for connections, so I wanted to make my own shield to handle power supply, rerouting 24V and digital control signals to the light, and handle wireless communications.
http://led-artwork.com/applications.html
I imagine I'll make a real announcement once I get all the quirks of the controller design worked out so that people can just plug in their own arduino, upload their own code, and control it however they like.
Given that limitation, would you agree with me that the best solution for using a standard Arduino board (I imagine I would also be redoing a lot of the power circuitry if I did an integrated derivative board) would be to have my switching regulator supply 7VDC instead of 5VDC, and connect that to Vin?
Yes that would be better in my opinion. You might consider running the Vin at 7.5-8vdc as there is a critical minimum voltage needed by the voltage switching op-amps before the switching circuit operates correctly. If going through the external power connector there is also a diode voltage drop to consider also.
I think that this will cause the Arduino to be running "as expected" with power connected to Vin, and that it should then provide me with a reasonably high current 5V source, right? Can the Arduino on-board regulator take another 0.6W of heat generation?
I really can't answer that for you. It's the total current draw X the voltage drop that determines the total heat dissipation of the on-board +5vdc voltage regulator. Thing is the voltage regulator will self-protect by shutdown down automatically if the dissipation is too high, so it will be a go/no go type of situation.
With this solution I would not have two supplies on any one net, I think it adds a fairly minimal amount of power dissipation requirement to the Arduino, I still use my switcher to drop the majority of the voltage, and I still get the +5V and +3.3V supplies I need for my shield.
Just be sure to consider any external components +5vdc requirements in addition to the basic board's requirement of around 50ma or so.
Lefty
Can the Arduino onboard regulator take another 0.6W of heat generation?
This is discussed as an example here:-
http://www.thebox.myzen.co.uk/Tutorial/Power_Examples.html
with the background here:-
http://www.thebox.myzen.co.uk/Tutorial/Power.html
Thanks! That's just the sort of information I needed. I can never remember what the heat transfer coefficient to standing air is... I wish there was an easy way to just tie the regulator to my aluminum case. But in any case, 2.5V drop at (154+50)mA shouldn't be a serious issue and that's all the wifi chip plus the arduino should be looking for. All the high power stuff is powered directly on my 24V bus!
Thanks again,