my core makes a good heater, at a loss for why?

[bdu]

I was trying to avoid asking for help, as I like to puzzle these things out for myself where possible, but I'm at a loss for ideas.

Here's the scoop:

A few months ago I built a core board (r3 from smashTV), did something really stupid after assembly (put 12v dc on j2 instead of j1, whoops), replaced the sensitive components (all ICs, anything that smelled bad), tried to power up (correctly this time), and all I got was very hot rectifier, 7805, and pic. No midi, nothing, and I'd have to power down inside of a couple of minutes due to the excess heating.

OK, time to cut my losses and start over, as I'd made a real stupid move with that first board, and I didn't feel like shotgunning it.

So a few weeks ago I bought another core board (this time r4 from smashTV, much cleaner, BTW), assembled it, and I get the same issue.

Here's the details for anyone wanting to think about the issue a bit to help me out:

1) I've checked all the 5v pins and all the gnd pins, they're all good whether the pic is in place or not.

2) I was very careful to get all polarized components in with long leg positive. That is the standard, right?

3) All ICs are in with pin 1 in the correct place.

4) I'm supplying 9v AC 1A at j1, the 7805 is putting out a good solid 5v. I've also tried DC supplies at lower voltages... they all do the same thing.

5) If I pull the pic from the board, the 7805 doesn't heat up like this, this seems the best clue?

6) I went through and did a reflow on any and all joints that looked dodgy.

7) If I power up with my LCD in, I get a bunch of block chars on the bottom row.

8) I don't believe there's any midi comm going on, at the very least, no requests for bootload.

9) I definitely can't leave it up very long with the PIC in place, as it all heats up dangerously hot.

Now, the PIC is one that I'd had before (the second one from the first core board), I've ordered a new one, but might a busted pic be the source of all my trouble here? Is this a known failure state for them or?

I'm not a total n00b, I've successfully built other things (x0xb0x, midisense, midi2SDSII, etc) without any major trouble, and given that my voltage test points all seem good, I doubt it's my soldering. Any ideas, guys?

Much thanks in advance for any help you can offer.

[stryd_one]

Sounds like you fried the sid and then tried to use it again ... I'd see how you go with a new PIC... but you might want to check out the other components first, and certainly test them before you insert the PIC

[bdu]

Sounds like you fried the sid and then tried to use it again

You mean pic, right? I never got far enough to even think about a sid! ;0)

Yeah, I'll double check my voltages and maybe replace my 7805 before putting the new one in (just in case the extra heat did it any damage).

Thanks.

[dj_statikfire]

Mr. BDU! Nice to speak with you again!

What can you do without power? How about doing a DMM check on the Vcc rail to the Ground rail and see if anything is leaking thru?

Remember, troubleshooting is all about isolation.

Also, it couldn't hurt to doublecheck your power source.

Thx, and best of luck! Waiting for Mr. Smashtv to send my kit, ordered last week!

THX,

Jason D

[stryd_one]

You mean pic, right? I never got far enough to even think about a sid! ;0)

Yeah, I'll double check my voltages and maybe replace my 7805 before putting the new one in (just in case the extra heat did it any damage).

Thanks.

Yes sorry PIC :) I was a bit tired when I wrote that! You might want to check thecaps and stuff too....

[MRE]

Judging by the fact that you said everything was tested with the PIC out, I would say that is your culprit.

By putting 12 volts on j2, you put 12 right into the pic and the optoisolator (IC2). These are the only components that are not 12 volt tollerant.

Also, you mentioned blocks on your lcd.. so.. you DID manage to get the pic to run far enough to initiallize the lcd. so. it sort of runs. random appearing data tells you that the pic is not running RELIABLY. Normally with such indications you should check that the OSC is getting a good clock signal, and that the pic is not 'browning out.' Brown out would be caused by sag in the 5 volt regulator.. which is usually only an issue with batteries. Brown out is very apparant on an LCD as.. you guessed it: Random charactor data.

It sounds like something is seriously damaged inside the PIC, and it is just managing to limp along.

You didnt mention anything other than the LCD connected, so I am assuming that is the case. If not, then go over your I/O boards just to make sure they are in good shape. While 74xx series chips are somewhat tollerant of 12 volts, they dont exactly like it (and certainly wont operate right at 12 volts). They could be damaged as well.

Finally.. you did the swaptronics method of troubleshooting. The only thing you carried over from one board revision to the other is the pic. So.. at this point.. it is about 98% obvious that the problem is the chip.  ;)

[bdu]

Judging by the fact that you said everything was tested with the PIC out, I would say that is your culprit.

By putting 12 volts on j2, you put 12 right into the pic and the optoisolator (IC2). These are the only components that are not 12 volt tollerant.

You didnt mention anything other than the LCD connected, so I am assuming that is the case. If not, then go over your I/O boards just to make sure they are in good shape. While 74xx series chips are somewhat tollerant of 12 volts, they dont exactly like it (and certainly wont operate right at 12 volts). They could be damaged as well.

Finally.. you did the swaptronics method of troubleshooting. The only thing you carried over from one board revision to the other is the pic. So.. at this point.. it is about 98% obvious that the problem is the chip.   ;)

Note that while the PIC is the only thing carried over from the previous board, that's not the same pic I gave 12v, that one was very obviously dead, this was the second pic on that board, one I didn't put in until after I thought I'd ironed other stuff out.

Nothing else built but the LCD connected, and even that not all the time.

I agree that it's the likely culprit, given that nothing heats up like that with it out, but then again, nothing is drawing current if the pic isn't in... I was just wondering if anyone had seen a dead pic heat up like that, to confirm my inclination, since it all tests fine with the pic *IN* too, just the heat issue, you know?

[MRE]

Note that while the PIC is the only thing carried over from the previous board, that's not the same pic I gave 12v, that one was very obviously dead, this was the second pic on that board, one I didn't put in until after I thought I'd ironed other stuff out.

Nothing else built but the LCD connected, and even that not all the time.

I agree that it's the likely culprit, given that nothing heats up like that with it out, but then again, nothing is drawing current if the pic isn't in... I was just wondering if anyone had seen a dead pic heat up like that, to confirm my inclination, since it all tests fine with the pic *IN* too, just the heat issue, you know?

Wowo, I am at a loss then. Yes I can confirm that I have smoked a PIC and it cooked eggs... but can also confirm that I smoked a PIC and it was cool to the touch while 'dead but working' (well.. partially dead. some ports worked, some didnt, and code was unstable).

Under working circumstances, it is to be expected that the 7805 will get hot. Hot enough to hurt without a heat sink mounted when runing 9 volts 1 amp in. It is NOT customary for the PIC to get hot!

The 7805 is a very 'lossy' VR.. it burns off anything over 5 volts (plus some for its own circuitry) as heat. So, the closer you get your input supply to ... say.. 6 volts (I dont know the exact lower threshold off the top of my head) the better. I.e. less heat. But that in no way relates to the PIC. Dont forget 1.4 volts drop from the Diodes.. so you need to be about 7.5 volts for ideal supply without a lot of excess heat.

The only reason for the pic to be drawing enough current to heat up noticably is when it is over driving its outputs. With no other boards attached to the core, and you ARE using a printed circuit board, and you DID double check your solder work.. There is very little in the core module circuit design that could be causing a problem like this.

The only other possibility is constant resets.. You mentined polarized parts..

In fact, as for polarized parts, only the supply circuit has any. The Diodes forming the bridge and the filter caps. If your diodes were wrong, you would get no Vout, or a very unstable Vout. If your caps were wrong you would let the smoke out of them soon enough. In fact both diodes and caps are hard to get wrong anyway as they have markings to keep you straight.

There is litterally nothing else on the core design that would cause this. I might suggest you take a good hard look at your power circuit. If you are supplying 9 Volts DC to the core, and you get 5 volts out, then your bride diodes are fine. Thats not to say that one of them isnt reversed, since in DC mode, a bridge only has two diodes 'on' at a time. So, its possible one is backwards, but not being used anyway, so it doesnt matter. If you get this far with DC supply, I have no other ideas. :(

However, if you are using 9 volts AC to supply the core, AND one diode is backwards,  you are only getting half of the power cycles. This results in the caps and voltage regulator working double hard to keep a 5 volt stable output. Likely it wont be able to accomplish that, and thus would supply a lot of ripple to the PIC. This in turn would cause the PIC to brown out a lot. Constant repetative reset cycles would cause crazy stuff on the LCD, as well as a lot of heat in PIC.

That is the ONLY logical explanation I can come up with without actually looking at your board.

I CAN confirm that constant resets will heat up a PIC fast. I was working on a project with a PIC that drove several high current motors (H bridge) all of a single large battery. The PIC was also supplied from the same battery. When more than two motors were on, the 7805 started browning out, causing the PIC to brown out and reset. It was controlled by the serial port and we'd get crazy data. It all spiraled out of control and the only way to recover was to disconnect all the motors, let the PIC come back up, and issue the command to turn them all off. Eventually this was solved by a seperate source for the PIC.. but yeah it got pretty damn hot.

[henrygr]

I ha heating problems. It eneded up being minute parts of bad soldering. In my case it was C4 and T1. Also, check C6 if you have power problems.

[bdu]

In the end, swapping in a new, programmed pic did the trick. Now to wire up all that AIN...

[MRE]

huh.. well.. shucks.  ;)

Yeah.. I have cooked one or two pics on projects before and they 'appeared to still be ok' (serial port was talking) but then I/O would not work and boom it resets or overheats or whatever..  strange lil buggers sometimes.