NCP1117 Board Load Test (Part 2)

In my last post, I tested my NCP1117 5v voltage regulator board with a couple of power resistors. In those tests, I used a wall plug rated at 9v 1A. Since then, I received a new variable power supply which I used to see how different voltages and currents would affect the regulator board.
IMG_20150704_181821For these tests, I stuck with the 5-ohm resistor. I set up the power supply to 9v and “unlimited” current so that the board would draw whatever it wanted. It drew roughly 890mA, read at both the input and output of the board. (The multimeter is measuring the output current.) Unlike the previous experiments, the regulator was able to stay steady for the three or so minutes I left it going. Even though it was able to stay on, the board becomes too hot to handle with bare hands. The current seemed to hit a ceiling at 890mA. Going any higher than 9.5v would cause the regulator to quickly hit its thermal limit and the current would start dropping rapidly. I suspect the wall power supply is slightly higher than 9v printed on it which is why it did the same in my tests with that.


 

Calculations

To calculate the power dissipation that the regulator is dealing with:

Power (P) = Voltage (V) * Current (I)

P = (9v – 5v) * 0.890A

P = 3.560W

From the datasheet, the thermal resistors junction-to-ambient, RθJA, and junction-to-case, RθJC, is 67°C/W and 6°C/W, respectively. Together, it’s 73°C/W, which can tell us how hot the regulator should get:

73°C/W * 3.560W = 259.88°C

Yeah… It needs a heatsink, though the current design doesn’t really allow for a proper one that screws into the circuit board.


I’ve used a similar board that uses the AMS1117 regulator on many projects that were running 24/7 for months. I noticed that the regulator did get very warm but I wouldn’t really call it alarmingly hot as this board was during these tests. I didn’t have the bench power supply by then and I didn’t do any current measurements (doh!), but I can estimate that none of those projects ever pulled more than 200mA from the regulator board. They’ll still be good for those types of projects where I can’t get an already regulated wall power supply. However, I was also using them for prototyping but, now that I have the variable bench power supply, I won’t be using them for that anymore.

I hope these experiments were interesting to you. Thanks for reading!

Advertisements

Atmega328p Breakout PRO

The Atmega328p Breakout Board has been out of stock for a while. I have half of the new components in hand, and the PCBs have been shipped and are on their way. The rest of the components will be ordered in the next couple of days and, if it’s like the last time I ordered, it will arrive the next day. Friday is Good Friday so I have to keep that in mind, and hope that everything comes in by then. I’ve made the decision that I will no longer sell unassembled kits so I hope I can use this long weekend to put them together so I have a decent stock ready for the next week.

Anyways, while I’ve been waiting for the new batch, I’ve been working on something else…

draft1To the right is the first draft of the Atmega328p Breakout Pro. The current base version will likely drop the ISP header. The new Pro version will sport the ISP header, in addition to a voltage regulator circuit. As I talked about in the last post about voltage regulators, I’ve been using my AMS1117 voltage regulator boards with my Atmega328p Breakouts in my own projects. With the built in regulator on this Pro version, I won’t need that extra board. The regulator I’ll be using this time is the NCP1117.

I believe the NCP1117 is the same regulator as the one on the Uno, or at least similar. It can output a fixed 5v at 800mA, even though at 800mA it’s best to just use another supply if you’re driving things that need that much current. When I was learning about PLCs, I was taught to separate the power supplies between the control unit (the PLC) and the actual peripherals (sensors, indicator lamps, etc). I don’t see why that lesson can’t be applied to Arduinos, even if they draw less current than a PLC.

I’ve added in the reverse input voltage protection diode so the total drop out is roughly 2v, just like on the Arduino. You’ll need to supply it at least 7v to get the 5v, and I’d limit the input to 12v tops just because of heat dissipation.

I’ve been trying to find out if it’s acceptable to have an external voltage applied on the output of the regulator. I want to be able to apply an external supply to Vcc (the output of the regulator, and the input power supply of the microcontroller). That way, you could use a battery or supply that’s already in the operating voltage range of the Atmega328p (1.8-5.5v). I’ve been looking at some Arduino and Adafruit schematics and it looks like you can, if you consider USB Vcc as an external 5v supply. They have it connected to the same 5v net as the output of the regulator (although the Adafruit schematics use a different regulator). I’ll keep researching and probably put together a prototype demo circuit and see what happens.

That’s it for now! Thanks for reading and I hope you’ll stop by my Tindie store and pick up some of my Atmega328p Breakout Boards when they’re back in stock!

74HC595 Shift Register Board Revision B Preview

I said when the year began that I wanted to get back into Eagle and revisit some of my PCB designs, as well as start some new ones. My first project for the year in Eagle is giving some attention to my 74HC595 Shift Register Board. The original version, even with its flaws, was used in the latest Light Show to help control 10 RGB LEDs. It worked out quite well but, again, it had its flaws. I’d love to have a revised board in my hand by the time I’m ready to go back to the Light Show.

sch1

This is the new schematic for the board. It includes a breakout of the Output Enable pin which gives you some PWM control. Using this pin will PWM all outputs over the two shift registers (16 outputs), so my idea is to chain three together so I can dedicate each board to a color (red, green, blue). Of course, it’s not as flexible as something like the TLC5940 which has 16 channels you can PWM individually, but it’s cheaper…

Anyways, if the application doesn’t need PWM, I added another ground pin next to the OE pin so you can just connect them together. When you do that, the outputs have no PWM control. The first version of the board had the OE pin already connected to ground in the schematic.

brd

Admittedly, my first PCBs were auto-routed as I was just getting acquainted with Eagle. I’m still learning but I managed to route this one manually. There seems to be a lot of technicalities on what you should and should not do when laying out a board but I don’t think my design is too much off what the auto router would have done. I tried the auto-router earlier and it did some really odd looking turns and loops around pads… Anyways, it’s just like solving a puzzle, though it takes me a few tries to get it right.

If you’re interested in buying these boards, let me know. I’m trying to decide on how many to get made.

Thanks for reading!

Attiny85 Programmer/Breakout Rev B3 Preview

Let’s go back to basics, shall we?preview1_frontI decided to throw out the whole breakaway section idea. It complicated things and there’s always the chance of a bad break. I put everything back on one board, which is now down to a 3x3cm footprint, 38% smaller than Rev A. The space savings come from less silkscreen text (removed the word “Pin” for the pin numbers) and using SMD components for the power LED indicator and resistor. I would have gone with an SMD capacitor but it costs more than the usual through-hole electrolytic capacitor, at least from where I get my components for these boards. With the price appearing to go up to get these boards manufactured, I’m looking for savings.

preview1_backAnother thing that changed with this sub-revision is that all of the pads are circles and are bigger than they have been on previous boards. They were kind of a pain to solder because any circular pads were really thin. I’m still working out the right size but I’m happy that I know  out how to address that issue now.

To date, the Attiny85 board was my most “Watched” item on eBay so I take that as an indicator of interest for this kind of thing so I’m set on getting these made. Being 3x3cm, I can’t fit two on one 5x5cm board which is what I have to work with to get manufactured. I’ll probably do something simple with the rest of the space but I want to nail down this one first. Thanks for reading!

Chinese Power Adapter Teardown

IMG_0198I got this power bank yesterday from China. Obviously it’s not a Samsung product but the power bank does seem to work fine. I wanted to make sure that it wasn’t connecting to data so I sliced the cable open. It only connects the power wires, thankfully. I bought a cheap USB voltage and current measurement tool on eBay but it won’t be here for a long while, especially with their national holiday. I’d like to measure the current coming out of the outputs.IMG_0187Anyways, it came with this really sketchy power adapter. The USB slot is like a half USB slot where your plug doesn’t fit in all the way or latch in properly. You can also feel the assembly was loose inside of it by just shaking it. I was definitely not planning on using the plug so I decided to open it up and see what was inside.IMG_0188Three sides of the front plate were already loose. I just had to slide a knife down the fourth side and a quick tug on the plug popped the assembly out.IMG_0189There’s a transistor between the transformer and a capacitor but there’s barely enough space for it so they put it on an angle. For whatever reason, the other transistor nearby is also at an angle and they’re both pushing over another capacitor. I thought it looked funny.
IMG_0190Dave from the EEVBlog explains in a video what’s crap about it, more than I ever could, but it was interesting to look at one myself.
IMG_0199Here’s a closer look at one of the boards.

So yeah… I’m not expert enough to rip on the board, but anyone can point out it’s a piece of junk just by holding the thing.

Attiny85 Programmer/Breakout RevB2 Preview

A little while ago, I posted about some updates to my Attiny85 Programmer & Breakout PCB that showed off a new design with a breakaway Programming section of the board. I threw that out and started over again. There’s still some work to be done but I wanted to show where I’m going with this.

render_frontThis is a render of what the board looks like. The program I use to render places the design on a rectangle so keep in mind that the purple outline would be the edges of the board.

The Power In and Attiny85 have swapped places. I’m going with an SMT LED and resistor for the power indicator to open up some more room on the board (labels are in the design, won’t be printed on the board).

The board is designed so that the Program section would be broken off once you’re ready to use the the board as a breakout. What if you realize later on that there’s something wrong in your code? You could connect up your Uno using the Breakout section, but you’d still need another capacitor for the reset pin on the Uno, and this kind of defeats the purpose of the board. I’m working on an experimental idea to make it easier. The pads next to the Power In area is a temporary programming area where you’d connect the Program section. The problem is making the connections. You can’t really connect the Program section back directly using headers because then you have no way of making another set of connections to the Uno. The best way would be to make the connections on a breadboard. It’s not ideal, but it’s still easier than looking up which pins of the Attiny85 go where on the Uno.

render_backI’ll put a URL to an updated manual on the board again. I still want to get some instructions on the board with the space I have but the small size of the board makes it difficult to put anything really specific. I’m on the fence on whether the text on the Program side will stay (the lines it’s referring to are on the top side of the board, I need to add lines on the bottom side too).

I’ve been making an effort to getting the sizing of text (and traces) right since I made the text on my AMS1117 board a bit too small.  I use mm to design my board, but a lot of help resources and parts of Eagle use mils. I made a table of conversions between mils and mm. I put it up for download as a Word document and pdf in case you’re interested. I got the table from here but formatted it to print.schematicI redid the schematic, chopping it up into sections and adding labels. Everything was directly connected to each other in the first schematic. It was messy, but not having things directly connected to each other does worry me in case something isn’t connected properly. I’m constantly checking connections and working on the other details of the board I talked about above so I’m taking my time. There were people watching the Attiny85 Programmer and Breakout Rev A board on eBay so hopefully that means there is some legitimate interest in these boards. I would like to get them manufactured.

Just a note for people who are Googling around for the warning, “Segment of net [name] has no visual connection”, I deleted the label of that net and placed a new one to clear the error. I couldn’t find much help when I was searching around so hopefully people land here if they run into the same warning.

Anyways, thanks for reading!