Hello,
I am trying to build a simple 10v DC power supply (for my Arduino Uno), with input from 110v AC wall outlet. I am well aware of wall warts, but for reasons beyond the scope of this discussion I am building this one from scratch.
The Arduino specification recommends an input voltage of 7 to 12 VDC, with max limits of 6 to 20 VDC.
I am using an MCI 04-02-3010 transformer to supply a stepped down voltage. The primary windings are connected in parallel to 110 VAC from the wall. The secondary windings are connected in series so as to provide about 10 VAC. I am using the power supply schematic in the LM317 data sheet. A schematic of my complete power supply design is attached.
With this design, the "free running" output voltage from the power supply (when not connected to a load) is 10.17 VDC. However, when I connect this supply to my Arduino (that is drawing about 30 mA), the power supply output voltage (input voltage to Arduino) drops to about 5.5 VDC - well below the recommended 7 VDC.
Can someone suggest what I must change to hold the power supply operating voltage (output voltage when connected to an Arduino load) at about 10 VDC? Please do not suggest using a wall wart because that is not a reasonable solution in this case.
You need a bigger filter capacitor in parallel with the 0.1uF capacitor to filter the rectified AC out of the full-wave rectifier. Without a capacitor, you have rectified AC, not DC. 1000uF is about right.
The top waveform [u]here[/u] is AC, and the bottom waveform is rectified AC from full-wave rectification.
You don't have a smoothing capacitor, so its not outputing DC at all.
The 0.1uF at the input of the regulator is enough for decoupling, but many orders of magnitude
too small to smooth mains-supplied rectified AC.
For a 1A supply at these voltages a smoothing cap might be somewhere on the order of 4,700uF
The purpose of the smoothing cap is to provide all the power between the peaks of the rectified waveform,
ie for about 8ms or so it needs to provide the 1A without dropping more than a volt or two.
For more current scale up the value, for a higher voltage supply scale down the value.
Wawa:
You do know a 5volt phone/tablet charger with USB lead is cheaper/easier/safer.
Just plug into UNOs USB socket.
Leo..
Not true for a dodgy eBay sourced supply(!).... With a proper mains transformer you get complete
isolation (it needs to be UL approved or equivalent of course) and you can earth the transformer
as well for even greater protection. But you do need to know what you are doing with mains
circuitry of course.
I don't know why the output voltage with the 330 uF was so low, maybe because it is running close to it's max voltage. The one I used is the only 330 uF that I have around. Since all of the values from 1000 down to 100 (with the exception of the 330) provided the same output voltage, I decided to go with the 100.
One of the comments above was that the original cap (0.1uF) was way too low and the output would not be DC. I checked the free running voltage with a scope and found that it was DC and with very little ripple. However, I did not check it when it was under load, and I suspect that it was heavily rippled. It was bad enough that the Arduino failed to successfully power up and run. That's what led to my original post.
I would go for the biggest cap. 100uF is still just a little bit when you load it. Something like a servo is enough load to have a massive ripple again.
cdj15:
I don't know why the output voltage with the 330 uF was so low, maybe because it is running close to it's max voltage.
So close??!! It's well over spec! The 10V AC has peaks of 14V. And that's when the output is really 10V AC. Most transformers can easily be 20% higher when unloaded. That makes peaks of 17V... That's almost double the rating. Pore little capacitor...
cdj15:
The one I used is the only 330 uF that I have around. Since all of the values from 1000 down to 100 (with the exception of the 330) provided the same output voltage, I decided to go with the 100.
But again, that really really really depends on the load....
cdj15:
One of the comments above was that the original cap (0.1uF) was way too low and the output would not be DC. I checked the free running voltage with a scope and found that it was DC and with very little ripple. However, I did not check it when it was under load, and I suspect that it was heavily rippled.
Indeed, the regulator alone is like no load. So yeay, when you don't discharge a cap it will have no rippel at all... The 100nF cap should he there though, but preferably a ceramic type.
And why not make a 5V output? Or at least another output (another regulator to the same transformer/rectifier) that outputs 5V. The onboard regulator of the Arduino is terrible at delivering power. So spare it and make a 5V out as well and use that to power things like servo's, display's etc
cdj15:
I checked the free running voltage with a scope and found that it was DC and with very little ripple. However, I did not check it when it was under load, and I suspect that it was heavily rippled.
Ripple depends on load.
2200uF/Amp is normally used, but for such a low voltage I would go for 4700uF/Amp.
Since your transformer is rated for 175mA (DC) max, a 1,000uF cap is ok. 100uF is not.
When ripple on the smooting cap dips below "output voltage - dropout voltage", dips will appear on the regulated output. Dropout of the LM317 is >1.5volt. 10.17volt out (diagram) is too close.
Drop the output voltage to 7.5volt to prevent that.
I would ditch this whole setup.
Use the transformer center tapped to ground.
Two diodes and a 2200uF cap to make a 6.5volt unregulated supply.
Feed that into Vin, and use Arduino's lower dropout onboard regulator.
Twice the current, so you can also use some LEDs and sensors.
Leo..
I am trying to build a simple 10v DC power supply (for my Arduino Uno),
Your off to a bad start. You should be building a 7V (7.5V max) power supply. The arduino onboard regulator is a Low Drop Out (LDO) regulator and doesn't need more than 7 Vdc. Anything more than that is just going to be dissipated as heat and serve no purpose (wasted energy)
DC with a lot of ripple is still DC - because the current doesn't alternate (the A in "AC" is for "alternating", alternating current.)
Rectified AC is DC: whether half-wave, full-wave, filtered well or not, lots of ripple, very little - no matter. The current resulting doesn't alternate, it doesn't reverse, it only goes one way.
Runaway Pancake is right and it supports this statement:
You need a bigger filter capacitor in parallel with the 0.1uF capacitor to filter the rectified AC out of the full-wave rectifier. Without a capacitor, you have rectified AC, not DC. 1000uF is about right.
The input cap you chose is from the datasheet which assumes a filtered unregulated dc. I've used the LM317 many times with decent unregulated supplies with no issues. I still think you should set the output voltage to 7.2V dc, not 10V.
a MCI 04-02-3010 transformer that only can source 25mA
0.25 A is 250mA; it should be fine for powering an Arduino.
so as to provide about 10 VAC
10VAC should provide a bit more than 14V when rectified and filtered, maybe more when lightly loaded, and that should be fine for the lm317 circuit and an Arduino.
Are you sure that you have the 317 connected as per the schematic? Your pin-numbering is weird (132?) but your picture/schematic match the datasheets I have.
Wawa:
I would ditch this whole setup.
Use the transformer center tapped to ground.
You can parallel the secondary windings into the full-wave rectifier, transformer runs a bit cooler.
Assuming they are exactly the same turn-count (which they should be).
MarkT:
You can parallel the secondary windings into the full-wave rectifier, transformer runs a bit cooler.
Assuming they are exactly the same turn-count (which they should be).
5volt AC would be too low for a bridge rectifier.
You theoretically have only 5.6volt DC left (not counting transformer regulation).
Then a 6.3volt transformer would have been better.
Leo..
