100v diodes blow @ 14Vac...normal ?

[petemoore]

  I used 1n4003 rectifier diodes for a full wave bridge rectifier, the input voltage being 14vdc [which might be somehow 22.5 vac...either I nearly halved the transformer output symmetrically...or the DMM is a POS...]
  Should these diodes be failing?
  The amp is 'kwerking' as in working...putting out loud AC [sounds about like 60 cycle hum]
  GGG LM3886tf amp.
  Now I'm thinking I should have paid more attention to the current rating, 2.5a

[Sir H C]

Diodes usually fail for overcurrent, could be a short somewhere.

[Paul Perry (Frostwave)]

Yeah, something is wired wrong, I expect.
How big is the amplifier? Those diodes are good for more than an amp of current, at least.

[JimRayden]

Hmm, check the orientation of the diodes. I know at least one setup that you can fry 'em with.

---------
Jimbo

[Nasse]

I put 35 amp bridge on mine, and I know one commercially made 3886 kit that uses four 1N5402 diodes, I think they are 3 amp diodes. I dunno if heat sink is necessary for diode bridge but I think I have seen it used.

SOmewhere I learned a "design rule" that they don´t tell in every school or book, it makes sense to use a lot overrated bridge rectifier in this kind of amplifier because

a) they are cheap
b) if your rectifier diode gives out blue smoke, and you don´t have a big series cap in the speaker output (old single supply amps usually do have) you can have DC voltage on your speakers, then the blue smoke comes out of them... Counting what good big speakers cost you see what I mean...

[alextheian-alex]

Just for kicks, I threw some shottkys in as a bridge recto in my last LM3886 build, and fried them in a couple seconds.  I figured that it was a turn-on surge that popped them, so maybe you are having that problem.  You may have a short somewhere too as Sir H C mentioned... double check for sloppy solder joints and stray wire hairs.  You could try an anti surge MOV or something, or a simple light bulb that is rated a bit above the max current you need.

Also as a side note, make sure to strap a snubber across the diodes... those chips seem to pick up the diode switching noise pretty badly.

[petemoore]

  Ok so...normal when stuff ain't right.
  So I pull that board, test some diodes twice at least the diameter of the 1n4003's...these seem to hold up better.
  Doesn't really mean the 1n4003's are not suitable or were wired right.
  I figured out how to check easily the diode bridge, the two with the anodes meeting ------>l---l<------ just bring the --- to V+, the two with the cathodes facing --l<---->l-- bring the ---- to V_.
  The transformer secondary winding goes across to the other two corners.

[Sir H C]

Also the amount of line filtering you use can smoke the diodes.  The less ripple, the less time the diodes are on charging back up the filter capacitors.  This means that the peak current the diodes are seeing can often be several orders of magnitude greater than the average current and again smoke the diodes.

[paulruby]

The start-up current could be to blame. Just some quick math, if you are trying to make a 2A supply, then you would have about 2000uf of filtering. If you are using a 14VAC supply, then the output voltage is 20V. The filter caps will pretty much fully charge in one-half AC cycle. In the best case scenario, the power turns on right at the zero crossing of the AC so that the filter cap can be charged gradually during that time: 8ms.

Here's the math for this case: 2000uf, charging to 20V in 8ms. That's a total of 0.04 coulombs transfered in 8ms, which is about 5A of average current for that time. This is not the problem since the spec says 30A can supported for "one cycle" in the spec.

Turning the power switch on is not very likely to be right at the 0-crossing of the AC supply. If you come on anywhere else, then you get a huge spike of current limited only by the effective resistance of the transformer and diodes. The IV curves in the specs sheet don't go out to 30A, but it looks like the forward drop will be ~1.5V or so to hit the 30A peak. The transformer voltage is likely to drop 5% at full spec. If it is 2A, the let's say it's effective resistance is 0.35 ohm. At 30A, it is dropping 10.5V. So, 10.5V drop across the transformer, but as much as 20V from the transformer means there is still plenty of capability to supply a peak current well beyond 30A. If my assumptions are in the ball park, then the peak current could be up around 60A during start up for a simple 2A supply.

If I had to make a rule of thumb for this, presuming no other forms of current limiting in the circuit, then use a diode with a peak current rating of 100X the average current you want to supply.

If you can tolerate some voltage drop in the primary supply, then adding a series resistor between the diode and first filter cap can protect the diode. A simple 1 ohm resistor, for example, will ensure the peak current doesn't go beyond 20A during start up. The cost is a 2V loss when supplying the 2A rating. But, a bigger bridge recto that can support high peak current is probably cheaper than a 5W 1ohm resistor and the wasted power and lost voltage headroom.

- Paul

[Minion]

I was trying to use some 1N4007 (1000v) diodes in a Bridge config in my LM3886 PSU useing a 100Va Toriod and every time I started the amp the Diodes would Fry because they couldn"t handle the Inrush current....

I ended up switching the 1n4007 diodes for a RS606 Bridge rectifier (rated at 600v rectifying 6A) and the Problem went away......

a Soft Start curcuit would also be an option......

Make sure you have enough Filtering in your PSU or you might have a Lack of Low end definition ...I personally use 10,000uF per Rail per LM3886 Chip Pluss at least 1000uF close to the Supply pins of the LM3886.....


Cheers