Problems when using external power

I was having problems with my arduino/LCD/RTC setup freezing and the RTC saying it is 33-o-clock. I thought it was a software issue, so I ran the setup attached to my computer (usb power) for a couple weeks with, you guessed it, no problems. Back to the external power supply and the problems returned.

The power supply seemed to function correctly on other devices, but I tried a different approach. I used a wall-to-usb power supply (ipod charger) and the problems returned.

How are you guys powering your ardunios?

Well, since you're having such strange problems, I think you should consider the possibility that you've slipped into an alternate dimension where it really is 33 o'clock :slight_smile:

All of my Arduinos are "serial", not USB, and most of them have on-board regulators, so I feed them with wall warts that put out voltages ranging from 7.5 to 12V. Which probably isn't much help to you.

When I use one of the ones that doesn't have a built-in regulator, it's either on a breadboard (with a solid bench supply) or built into some gadget that has a regulator for the whole system.

It's just a wild guess, but your ipod "charger" thingie may only be good for charging: it could have poor regulation, or an occasional dropout, that you don't notice when you're charging batteries, but do when it's running something that needs "all the Volts, all the time".

The other power supply might be giving you problems because it's overtaxed: does your LCD have a backlight on it that's drawing lots of mA? How does the power consumption of your Arduino+LCD compare to the other things that appear to work okay with it?

Ran

My power supply says 25ma... That may be a problem! I have ordered a 1a 9v supply. Hopefully that fixes the problem.

My power supply says 25ma

Are you sure? That is the lowest current rating I have ever seen on a mains unit. :o

It is from an OLD TI calculator. I saw the 9v and thought "free power supply". Here it is:

It also says 8W which doesn't tie up with 9V at 25mA. If it is an 8W supply at 9V then it would be about 900mA.
8W can't be the input power either as it would be fantastically inefficient burning 8W to produce 0.225W.

What ever the "25 MA" means I don't think it is the current rating.

Oops that's "25 mega Amps" - too much Marston's Pedigree.
Mmmm, Marston's Pedigree.

What ever the "25 MA" means I don't think it is the current rating.

Well I think it does, but without testing the transformer we are both just using our gut feelings. The input power rating could be something that has to satisfy the UL classification for input power requirements, not a simple input power equals output power minus transformer efficiency.

A useful test might be to load the output with a variable resistor load and measure the resulting DC output voltage. Determining the secondary impedance value would give one a better understanding of the true secondary current capacity.

Lefty

PS: Decades ago I have a similar low MA current rated wall wart transformer that came with a TI calculator that would only trickle charge the rechargeable Ni-cad batteries.

that would only trickle charge the rechargeable Ni-cad batteries.

Yes I could go with that theory.

I have had dealings with UL and they are not the most sensible of organizations, in fact they are down right perverse and their main purpose seems to be protecting the U.S. markets from imports.

But as you say measuring the Voltage / current graph would show it.

If it is 25mA then for arduino work it is in the chocolate teapot class of devices.

chocolate teapot class

I like english expressions like that, we would just say useless. However I do recall some US jokes about screen doors on submarines. :wink:

Lefty

Well, if you would like it to test your theory, it's yours for the shipping cost. This transformer is probably 30+ years old if that changes anything.

Well if that offer was to me then thanks, but I have more wall warts then I know what to do with. I pick them up at thrift stores for a buck or two and lately some switching transformers are starting to show up also.

Lefty

I have more wall warts then I know what to do with.

The hardware hacker's equivalent of coat hangers...

Thanks but I don't have 120V mains. :wink:

OK, I got a new power supply and I have the same problem. What I found is that when plugged into a circuit with my aquarium (that's what this project is for) the RTC goes crazy and spits out bogus times. When I plug into another circuit in the house, everything is great. Any suggestions other than the obvious, plug it in far from the thing I am trying to control?

That's pretty weird, but you probably already figured that out.

What does your arduino hardware look like? Are there any long wires involved? (e.g., a 3' cable on a temp sensor?) We can probably discount power leads.

What other electrical devices are near the aquarium? pump, heater, refrigerator or elevator on the other side of the wall, etc.

When you tested on other circuits did you use an extension cord and get the arduino close to the aquarium again? Or, alternatively, run the arduino off a battery while it is physically where you plugged it in and it flaked out?

The problem could be one of:

  • there is something generating EMI (radiating through free space) that is confusing your arduino
  • there is something generating EMI (coming through the mains connection) that is confusing your arduino
  • there is something wrong with your arduino circuit in some intermittent way, and it's coincidence that it only shows up near the aquarium
  • something else :slight_smile:
    -j

I think there is emi? from a metal halide light/ballast. When it is not on, the whole thing works great. I am guessing this is coming through the mains. Would a ferrite filter do any good on the power cable?

A ferrite certainly would not hurt. Put it near the arduino.

Are there any other long wires besides the power connection (e.g. to peripherals, like a temp sensor)? A wire can act as an antenna, pick up noise, and feed it into your digital system.

-j

There are a couple ling wires: temp sensor and LCD, but both are shielded.

Is the shield connected to ground?

If the temp sensor is analog (e.g. LM35 style, or thermistor) throw a ferrite on it as well (again, near the arduino). You can try one on the LCD or a digital temp sensor, but it may do unpleasant things to the digital signals.

Hmm, could be that the only thing hosing up is the LCD, if that is the only place you see an error.

-j