I've been putting my Peggy2 together today (great kit btw ) After much faffing around and resoldering some loose bits I had it displaying a single solid row (or is a column? It's on the same axis as the text is printed) of LEDs with the default firmware. At that point I somewhat naively decided to try and throw a different sketch on it (16levels) and now things have got a bit strange

Most of the time it just flashes on off, which I guess means it's trapped in a reboot cycle. Oh and only every other pair of LEDs lights up. i.e. 00110011001100110011001100110. Sometimes it looks like it's running 16levels for a few seconds, during which the right half of the board doesn't do anything.

Uploading new sketches simply results in the usual "not in sync" error. I've tried rebooting and stuff to make sure nothing else is messing with my serial port. Pretty sure it's not that.

When I plug the USB-TTL programmer into J3 everything gets a lot dimmer and it just keeps on rebooting.

I can sort of influence the behaviour a bit by plying with with VR1 and VR2 but since I don't know which LEDs and controlled by which of those, and whether to turn the screw clockwise or anti to make it dimmer I really have no idea what's going on there.

I'm sure it's still entirely possible I have some dodgy solder joints on it. I don't get why things changed so much after I uploaded the sketch though

This sounds like a problem that we've seen in *some* of the Peggy 2.1 kits, for which we found a fix of installing a capacitor in the VCC/GND holes on the left side of the circuit board. (If you don't have a capacitor installed there, we can send you one-- please request it through the Evil Mad Science Shop contact form.)

If you *do* have a cap installed there already, I'll have to think about this a bit more-- it sounds like your computer setup is mostly correct, since the Peggy is getting the reset signal from it.

The brightness controls are fairly straightforward-- turn them clockwise to make it brighter, counter-clockwise to make it dimmer. Turn both to opposite extremes to see which LEDs are affected by each.

A response on a Sunday afternoon :O

I do have the Tantalum capacitor that was included in my kit in that spot. Between VCC an GND on the left side of the ATMega. I'll try playing with the trimpots again and see if there's any correlation between high and low and what actually happens. I did notice before in some configurations the white LEDs stopped lighting up, but there wasn't generally much noticeable variation in brightness when I was playing with them.

Hmm... Make sure that the correct side of the tantalum cap -- The one with the stripe or the "+" marking is towards the VCC side. Definitely re-check the soldering on that, and also the soldering and polarity of the big capacitor.

If only a single row of LEDs is lit up, that usually means that the microcontroller is not up and running-- particularly so if which row is lit seems to wander around.

The problem could be that there's a low-voltage condition-- the voltage is high enough to start up the microcontroller-- it's clearly starting since it can send a dotted line out as part of the 16-level program. But, the voltage seems to fail pretty quickly once it starts sending high current to the row of LEDs. Are you running on batteries or on an external DC power supply? Which ever is the case, you might try the other, to see if that makes a difference. Any information that you can find might help point towards a solution.

You might try dialing down the brightness to full-low on both halves, to see if that affects anything. If possible, try uploading the "bounce" demo as well, which only lights one LED at a time.

It doesn't have a stripe. It just says "+10\n10V" on. I put the left pin with the text facing you on VCC. I'm running on the DC wall wart that came with it.

I'd love to try the bounce demo but uploading sketches doesnt work anymore. The Arduino IDE reports that avrdude throws a "not in sync" error.

I'm trying to dial the brightness down but no matter how much I rotate it anti clockwise it seems to flash equally bright which seems rather strange although the flashes are so brief I imagine it could just be a high peak as the cap discharges a bit.

The screw spins in either direction without ever getting any stiffer or anything, how do I know when I've reached it's limits?

>It doesn't have a stripe. It just says "+10\n10V" on. I put the left pin with the text facing you on VCC.

Whichever side the "+" is closer to should be the one closer to VCC.

>I'm running on the DC wall wart that came with it.

Please do try batteries, at least momentarily to see if that's the issue. You don't even have to solder down the battery box to try it.

>The screw spins in either direction without ever getting any stiffer or anything, how do I know when I've reached it's limits?

It's a 20-turn dial. Once you've gone 20 turns, that's as far as it goes. It will make a faint clicking noise every turn or two once you've passed the end of the travel.

Also: You may be able to get it to program by pressing the reset button at the right moment-- there should be a few seconds at the beginning, after restart, where it's just waiting to be programmed, and not executing the new program yet. If you press reset by hand during this time, that can sometimes help with an Arduino board that's not otherwise responding correctly.

Ah the trimpots make more sense now thanks If I turn VR2 all the way up then the white LEDs stop turning on most of the time. I don't have any D-Cells to hand atm but I'll pick some up tomorrow and try that out.

Hitting the reset button then immediately trying to upload the bounce sketch causes nothing to happen for 2-3 seconds followed by the pairs of LEDs on my only soldered on row to blink quickly three times followed by a sync error from avrdude and the Peggy2 continuing to flash dimly as it resets over and over.

Also check: Make sure that R4 is in place and soldered correctly.

Here's a rather poor picture of the blinking it does when it's rebooting constantly.

http://yfrog.com/4oxcpj

You can see in the reflection on the PCB that only every other pair of white and red LEDs is lit. When it does actually seem to be running the 16levels program the 2 right most pairs lit in that picture don't light up at all and the other 4 pairs do light up and seem to fade out. The fade out they do do though seems a bit broken and no like what I'd expect from the video of that sketch.

I've examined the solder joints on the caps and they look OK. There are some other ones which look like they could use a bit more solder so I'll go and touch those up and see what happens.

I do have a multimeter I have no idea what would be useful to actually measure to figure this out though

Thanks a lot for the help.

Well, it does make it considerably harder to tell what's going on when there's only one row of LEDs.... From your first few messages, it sounded like only one row of LEDs was lighting up-- and that that was for some reason other than there only being one row of LEDs.

This may be (at least partly) a programming issue then. Have you tried the "Set RTS on close" fix yet? Device Manager - Comm Ports - USB Serial Port - Port Settings - Advanced button - Set RTS On Close.

Next: Which processor do you have (ATmega168 or 328?), and what "board" do you have selected in the Arduino menu? It should reflect the correct processor. Then, are you 100% sure that you have the USB-TTL cable end oriented correctly?

I've been trying this on OS X so I'm not sure there is a way to do the Set RTS on close fix. I'll try programming from my Vista PC and see what happens. I've had rather bizarre issues with OS X and serial ports in the past so I wouldnt be entirely surprised if that was partly to blame.

I have a Peggy 2.1 with the ATmega328. Arduino is configured for Deumilanove w/ 328 and I have the proper serial port selected. I absolutely definitely have the green wire of the TTL cable attached to the pin labelled green on the Peggy.

Do I have to have an entire row of LEDs wired up before any of them will light? I've tried stuffing LEDs in some of the other holes but they wont light up at all even when carefully made to touch the pads.

I'll solder on some more LEDs and and tidy up some parts tomorrow and see what happens, It's getting late over in the UK

Sorry if I'm a bit useless at providing relevant details. I'm new to this hardware stuff but I know from being a software developer it can be pretty frustrating dealing with silly people who don't describe things properly.

>I've been trying this on OS X
Ah, nevermind. Half of the potential problems just went away.

>Do I have to have an entire row of LEDs wired up before any of them will light?

Nope. It's sounding more and more like there really is that voltage dropout problem. I've not seen that problem persist once the "extra" cap is installed. But, I would really suggest trying the batteries next.

You might also try other caps to see if anything changes. My board's malfunction was the reason for adding the tantalum cap to the kit and it didn't actually fix my problem. I'm not sure why. I found that a 10uF 35V electrolytic I had laying around did fix my problem though. My board works very well with that. I still have issues programming it but manually hitting reset and retrying a few times will eventually program it so it doesn't really bother me.

I haven't tried batteries yet (local shops didn't have any, going to Maplin tomorrow so will try then. I'll pick up some 10uF caps to try out in place of the one included in the kit) but I have soldered on an extra row of LEDs and tidied some bits up. No change to the patter of LEDs that light up or the constant rebooting. The 2nd row never lights up. However if I hold down the ON/RESET button for a few seconds they do light up, and the same pairs of columns as the first row turn on. The first row doesnt light up while I'm holding the button down.

Another very strange thing I've noticed is that waving my hand around over the right side of the empty battery box causes the LEDs to dim slightly!

Oh, and could you explain why batteries are less likely to suffer from whatever voltage drop-off issue you suspect this to be? I'd have thought the AC converter would have been pretty reliable. Just curious.