PCB Where did I go wrong?

[goldfinger]

Hi All,

It's only my 2nd design for a PCB so be kind did a proto pcb of the attached board and unfortunately it didn't work.

When I hooked up the transformer got hot (as did the secondary wires) so I am assuming its some sort of short somewhere but can't work out what I've done.

I'm not sure where I have gone wrong so any feedback would be appreciated. I'm sure it's something simple but I can't seem to see where! 

The attached is a two layer (was meant to be a regulated power supply using an 1085) The 1085 is attached to the board via two screws and heat sink. The screw holes on  no net and have a decent keep out around the screw holes from the pour.

The top layer is for power signal and the bottom is a ground plane. No power signal is routed in the bottom layer only ground connections. The traces are 1.6mm I was planning to run a max of around 1A through the board, but realistically it wouldn't see more than a few hundred ma. R5 is to fine tune the output voltage.


If anyone has any ideas what could be wrong I would appreciate the feedback, I suspect I've done something simple and stupid, but well I can't seem to find where.

Thanks heaps guys really appreciate any feedback great bunch of members here who have always been helpful. Thanks all

[eugenenine]

What is C1?

[goldfinger]

What is C1?

Hey Eugenenine
It's meant to be  small snubber mate

[Tac Eht Xilef]

How's the temperature on D1-4? R1? U1?

That'll give you a clue as to where the fault is...

[Ammar]

What is the value of C1? Also, what is the AC voltage supplied at the inputs?

[goldfinger]

What is the value of C1? Also, what is the AC voltage supplied at the inputs?

Hey Ammar
20vac, c1 is 0.1uf rated 100v


Sent from my 486 IBM WIN 3.1x Desktop...

[goldfinger]

How's the temperature on D1-4? R1? U1?

That'll give you a clue as to where the fault is...

Hi tac,

I'll have to check as when I noticed that there is no D.C. Output at all and the transfomer is getting hot (inc secondaries) I've turned off straight away

No connected the transformer is fine measuring and no heat at all.

As soon as I connect it heats up, with no D.C. On output.

Diodes are over spec for circuit, elec caps are rated 63v 4700uf each.

Wirewound resistor is 5w 1ohm and has around 4mm clearance under it

I doubt (although could be wrong ) that it's soldering etc as I've checked board and cleaned it and doesn't seem to be any residues or spill over to gnd plane. I've placed pretty big through hole to plane spaces as you could probably seee.


That's why I'm assuming I've done something Silly with board and shorted something, highly probable as new to electronics and even less experience with pcb design (which probably shows)

[digsys]

Well, the circuit is fine and straight-forward. Time to disconnect the Regulator and use diode test on your DMM. Also reistance checks to GND.
Regulator incorrectly wired / faulty?

[sleemanj]

Traces on C1 from AC1/2 are very (needlessly) close together, check for (manufacturing defect) short (also, check for short of C1).

Net C6_1 runs very (needlessly) close to ground on DC1, again check for short there.

In fact you have quite a few needlessly close clearances between nets and pads not on those nets,  check them all.

[goldfinger]

Well, the circuit is fine and straight-forward. Time to disconnect the Regulator and use diode test on your DMM. Also reistance checks to GND.
Regulator incorrectly wired / faulty?

Hi mate,

Thanks the 1085 is already soldered to board and bent back and screwed to board.

Any advice on measurements I should take would be great, fairly new so any further info on measurements I should take would be awesome


Thanks heaps!

[sleemanj]

Thanks the 1085 is already soldered to board and bent back and screwed to board.

Note that the tab is output positive.  Make sure you have not shorted it to your ground plane via a screw.

[goldfinger]

Thanks the 1085 is already soldered to board and bent back and screwed to board.

Note that the tab is output positive.  Make sure you have not shorted it to your ground plane via a screw.

Hey sleeman thanks bud, when you say tab are you referring to the regulator package, e.g.back area that is connected to heatsink/board?



Sent from my 486 IBM WIN 3.1x Desktop...

[sleemanj]

Hey sleeman thanks bud, when you say tab are you referring to the regulator package, e.g.back area that is connected to heatsink/board?

Yes that is the tab.

[goldfinger]

Hey sleeman thanks bud, when you say tab are you referring to the regulator package, e.g.back area that is connected to heatsink/board?

Yes that is the tab.

Ok thanks heaps, sorry thought so just wanted to check

[goldfinger]

Thanks for everyone's feedback thus far appreciated all


Sent from my 486 IBM WIN 3.1x Desktop...

[Cubdriver]

This is obvious and you've doubtless checked, but just to be certain, the diodes are all installed in their correct orientation, right?  Can you take a photo of the assembled board and post that?

And to eliminate the possibility of the regulator tab being grounded, a quick check would be to remove the screw and lift it slightly off the board, then try applying power again.

-Pat

[eugenenine]

I can't say I've ever seen a snubber cap in the middle of a bridge like that.

You could de-solder r1 to isolate there.  If you get good results left of R1 with it removed then that half is ok.

[M4trix]

Speaking of which, regarding the schematic document. Why do you have unnecessary junctions at the end of some component pins ? That usually happens when you rotate components, drag or flip them while they are connected with wire.   

[goldfinger]

Traces on C1 from AC1/2 are very (needlessly) close together, check for (manufacturing defect) short (also, check for short of C1).

Net C6_1 runs very (needlessly) close to ground on DC1, again check for short there.

In fact you have quite a few needlessly close clearances between nets and pads not on those nets,  check them all.

Hey there,

Thanks is the best way to check is just use continuity function on the DMM? E.g., check gnd against c6_1 etc


Sent from my 486 IBM WIN 3.1x Desktop...

[sleemanj]

Thanks is the best way to check is just use continuity function on the DMM? E.g., check gnd against c6_1 etc

Or resistance.  But if you are doing this on a populated PCB you will have to think about how components could affect the measurements. 

For starters though, break out your magnifying glass and a good light source and take a close look.

[goldfinger]

This is obvious and you've doubtless checked, but just to be certain, the diodes are all installed in their correct orientation, right?  Can you take a photo of the assembled board and post that?

And to eliminate the possibility of the regulator tab being grounded, a quick check would be to remove the screw and lift it slightly off the board, then try applying power again.

-Pat

Yup I'll remove soon and check with the regulator off board. I'll post a pic too

Cheers



Sent from my 486 IBM WIN 3.1x Desktop...

[goldfinger]

Speaking of which, regarding the schematic document. Why do you have unnecessary junctions at the end of some component pins ? That usually happens when you rotate components, drag or flip them while they are connected with wire.

I would say rotation is probably the reason as mentioned pretty new!!

[goldfinger]

Thanks is the best way to check is just use continuity function on the DMM? E.g., check gnd against c6_1 etc

Or resistance.  But if you are doing this on a populated PCB you will have to think about how components could affect the measurements. 

For starters though, break out your magnifying glass and a good light source and take a close look.

I've actually got two unpopulated boards I could look at. If I'm using a bare board and I'm looking for power net referenced to ground with very little or no resistance?

[Cubdriver]

Thanks is the best way to check is just use continuity function on the DMM? E.g., check gnd against c6_1 etc

Or resistance.  But if you are doing this on a populated PCB you will have to think about how components could affect the measurements. 

For starters though, break out your magnifying glass and a good light source and take a close look.

I've actually got two unpopulated boards I could look at. If I'm using a bare board and I'm looking for power net referenced to ground with very little or no resistance?

If you have bare boards, I'd do resistance checks between ALL the traces and ground.  NOT finding a short doesn't rule out a manufacturing defect that happens to be only on the board you populated, but if you DO find one it would definitely be something specific to look for on the populated board.

-Pat

[goldfinger]

Well, I've done a few things and not much luck. Removed the regulator from heat sink no change.

Got a board and couldn't find any links between gnd and nets.

Diodes for rectification all seemed ok measuring forward v only.

I cut the link to R1, tried desoldering and it was a prick so just cut one side of it.

Even after I cut the link to it and tried measuring just after the rectification I was only getting up and down mv.

The secondary wires were also getting hot as if it was drawing too much current!

So with R1 removed it has to be around the ac entry/rectification area.

As even cutting off the rest of the circuit it still got issues.

At loss what to try next. Should I build a new board with just the entry and 4 diodes and then measure to see if it's still the same issue?


Sent from my 486 IBM WIN 3.1x Desktop...