Flakey power-ups?

[Jidis]

Hi,

I've got a core here which doesn't always boot correctly on the first attempt. Sometimes it comes on to the top row of black bars, and if I switch power off and back on quickly it inits properly. I thought I remembered someone mentioning a similar problem here, but I'm not sure.

There's been a couple different AC/DC "wall-warts" involved, so I don't think that's the cause. The current one is putting out 9VAC at 850mA. The previous one was higher amperage with DC out. My 7805 heatsink is hot as crap, but I guess that's normal(?). I'm also swapping between two different 18f452's, when I screw one up, so I don't think it's the PIC, unless the new JDM is doing the same thing to both of them. I'm using the regular MBHP Core v2 board.

Anything in particular to check?

-Thanks! 

[pay_c]

If you can, you should check how much current your stuff is drawing. A core+LCD without backlight  not more than 50 mA. Core+LCD with BL not more than 300 mA. Core+LCD with BL+ DINs+AINs not more than 400 mA (max), the same + DOUTs (depends on how much LEDs are connected) up to 600 mA or even more.

If it´s to high there´s a short somewhere (ICs in right direction? Especially the Optocoupler?)

For the heating of the 7805: Either go to a lower voltage (min. about 7,5 volts) or attach a heatsink. If it gets to hot or there´s to much current running through there it will shut down automatically.

Are the caps in there the right way? They should filter the voltage to a stable value. If your stuff is drawing to much power (anyhow without shorts) try higher capacity values (e.g. 4700 µF as the first big one in parallel with a 100 nF one).

[smashtv]

I built a core recently with the same exact symptoms, and I noticed I could push down on the pic and it would boot every time, so obviously I'm thinking bad soldering on the socket, but it all looked great under the magnifying glass.

After re-flowing the solder it works fine, and now gets a lot of abuse on my test rig.  :)

Also I have noticed that the MB64 app is hesitant to boot sometimes when the inputs are floating (nothing else connected) but what wouldnt act strange when seeing thousands of events a minute while it's power supply is ramping up..... ;)

</edit>removed everything about power supply issues since pay_c beat me to it, now this post is kinda bare and worthless!  He did a much better job explaining than my usual ramblings......

Best!

Smash

[pay_c]

LOL  *FIRST!*  ;D ;)

Anyhow *I did* forget that resoldering stuff you mentioned!  ;)

[Jidis]

Thanks for the great tips guys!

I did bring my multimeter home with me. I'll check the current. I'm pretty sure it is pulling too much. I was running a Greg board with the exact same core parts and I don't remember the startup problems, or all this heat. This regulator actually burned me last night :-X

I'm pretty good about checking for shorts and continuity after etching, but as Smash mentioned, there is good possibility that there's a bad solder joint or something from the assembly.

Smash, Speaking of test rigs, you'd like this one ;) -- 32- 5mm LED's with 4- 595's in banks of 8, on a board about 2.5"w x4.5" tall, 20- tactile buttons and a 12 key numeric keypad on a board about the same size horizontally with 165's on the back, 16- miniature 2" faders on a board about 2x5", a core board, a 2x40 display, a 1.5x5" 8 digit clock display with 3- 595's and a board with about 20 different single and dual pin headers and sockets, linked together in pairs with the two MIDI connectors and power jack on it for custom patching. All the i/o boards have 5 pin headers on them, except the faders, which have a ribbon (I still need to make a dual 4051 board for that one). All six boards and the LCD are mounted with rubber standoffs to a plywood panel (about 6" tall x 20" wide) at 45 degrees, in between my keyboard and monitor at home with my online computer. The standoffs give me enough room to run the cables under the boards if I want. I'm looking forward to using it for MIOS programming. The jumble of breadboard junk was causing me more problems than it was worth, and now I've got plenty of buttons and lights for testing. I'd like to get the JDM on there with an in-circuit programming hookup. I've had to pull these PIC's too many times now.

                                                         - Take Care

                   George

PS-- Those shift registers are still: pins 6,5,4,3,11,12,13,14 for 1-8 aren't they?  :-[ 

[Jidis]

FWIW-

I never did get this straightened out. I think I've checked just about every connection and replaced about 80% of the parts. There's a new 2200uF cap, a new regulator, a new 3904 and 4001(?) diodes in place of the rectifier. All the 5v pins seem to be getting a little over that (5.1 maybe). I think the opto has also been switched and there's two different PIC's being tried (they both verify and program well on the JDM). I also re-heated and checked all the PIC solder joints as Smash recommended, plus a few others.

I couldn't find anything between 9v and 6v here, so I had to stick with a 9, but I figure there are a bunch of others using that here. The problem has gotten worse, but I've got a modified core board I'm going to finish at the studio tonight, so maybe that one will work OK.

I remember a thread about crystals which don't start reliably. Is the faulty, black square LCD/no output messages behavior something you'd get from that?  I haven't swapped the crystal yet.

                     -Take Care

George

[pay_c]

- W e I r D -  ???

Well, ok. What crystal do you use? 10 MHz, right? Sometimes swapping the pins can help, but that´s more or less some seldom thing. Did you install the small ceramic bypass caps as mentioned on UCApps.de? They could help, too. Did you already attach a heatsink to the 7805? And do check the current.

Still hoping...

[Jidis]

pay c-

Thanks! I did eventually get the second core working, so I'll check the other later when I've got time. Between the two, I think I've made every careless mistake there was to make, and desoldered/resoldered and replaced components before finding whatever stupid thing it was that I did. For the grand finale on that first core, I replaced the rectifier with four power diodes, and while checking some voltages, I managed to short two of the wires from the diode jumble, and got a nice "burnt stink" smell. The PIC survived, but I think I zapped one of my 40x2's. :'( I now get a screen of two rows of solid black squares, and if I wait a few seconds, a blank hole starts to burn through it like the intro to Bonanza. It then spreads throughout the screen and wipes all the squares. I should've used a 2x16 until I got everything straight.

The new one is definitely working now, and I've got the "LED digits" app spitting out an 8 segment LED display :) My meter, for some reason, wouldn't give me an amperage reading in series with the core power. I now have nice big heatsinks on both cores and will probably add one to my Greg, but I think with the boards and components all working right, the amperage and heat will be in the right range. I don't have many parts left to swap on that bad core, so I should hit the problem spot with a bit of tinkering.

PS- The base of that transistor going through 1k to the wiper and to one side of the 10k pot on the schematic hits the third leg of the pot instead, on the core layout. I found  a forum mention of a mismatch between the board/schematic in the luminance area. This is actually correct right?

Also- Has anyone seen that "burning black squares" thing on an LCD? Is it most likely a lost cause now?

                        -George

[pay_c]

I can´t help you with the schematic thing right now, but at least your second core works! I´m sorry for the 2x40´s! That´s hard! If the new core works, just try them again! And: Nope, never seen a fading like you described on a LCD...

For the current measuring: Some meters do have: First a fuse (inside) which is burnt really often, second another In only for amperage stuff. Did you try? In series means that you used your meter as some sort of power switch, right? One pin (core) goes to power, the other first into the meter, from the meter to the power, right?

Don´t give the displays up already! Normally they´re pretty indestructible (with what do I know secure stuff). *thumbspushing*

[Jidis]

pay_c

Yeah, I had it in-line with the supply. I think this one does 200milli down to 200micro amps, if that sounds likely.(?) Maybe I was overrange. I still get some decent heat on this one, after it's been on for a while. I may try to check the amperage again if that's not normal, or bump down to a smaller input supply.

I brought my other 40x2's back from the studio tonight. I think this one is indeed "toast". :-[ Funny thing, I ran it all through my initial temporary setups with hot irons and crossed wires and junk all over the table, and then bring it home and mount it to a nice clean board and I end up cooking it. ;D I shouldn't feel bad. I got it with two other brand new ones on eBay a while back for 9.95 USD. I didn't even realize that was a good deal at the time.

-George 

[pay_c]

For the amperage: Yepp, if you got a LCD with BL you´re pretty sure above the 200 mA limit. *If* you still want to measure it, just detach the LCD. Should be well below 200 mA then. But if the 7805 does get *decent* warm like you described everything should be just very fine. Those (sometimes indestructible -  ;) ) 78xx work up to a temperature of appr. 80°C. So you only want to be concerned if you are likely to burn your fingers.  ;D

Hoping for another 9,95-deal for you!  ;D

Greetz!