next topic »

[on:]

I'm at a bit of a loss here and hope someone might be able to help me.

I have a LM7805 (DPAK) being fed 12VDC from a wall-wart whose line is protected by a 100mA PTC. It's running a board I put together, is cool to the touch, etc. It's output is stable (i.e. 5VDC), there are several caps to help it stay stable, etc.

Cascaded one step further (i.e. being fed from the 5VDC power supply), there is a L4931AB33, which until recently was producing 3.3V quite stably... I am planning on using it for a SD Card as well as a radio like the RFM12. I never saw any smoke, the chip isn't hot or warm to the touch, etc. but the SD card that I used to be able to write to is now dead and the output voltage of the L4931 is now 4.2V. Now I realize that this chip is overkill for a SD card and the -245 converter chip to drop the voltage from 5V to 3.3V but I wonder why its voltage would swing so much out of spec.

How could I have managed to damage this chip to behave this badly? Was the output simply too low? If I were to right-size a linear power supply chip, what are typical power draws for radios like the RFM12 and/or transflash memory while they are in use? Many thanks in advance, Constantin

[Reply #1 on:]

100mA might be a bit undersized for radio transmissions, but I doubt the PTC effects would cause damage to your regulator. It even has internal current limiting.

If the output of the 3.3V regulator is 4.2V and it never got hot, you never applied 5V to its output (or other stresses like that), then it could be electrostatic discharge (ESD). How careful are you when working with electronics? Unfortunately, ESD is one of those things that happen without warning and the end result is exactly as you describe: it just "stops working". Make sure you work on a grounded surface (ESD laminated bench or ESD mat), keep yourself ground using a wrist strap, use properly grounded instruments, etc.

--
The Flexible MIDI Shield: MIDI IN/OUT, stacking headers, your choice of I/O pins

[Reply #2 on:]

Thank you so much for the replies to date. I will try to post a circuit diagram later today. But here is a quick word description of what is on the board:

If memory serves, the output of the 5VDC regulator has three 10uF ceramic caps attached to it (in parallel). Another six 0.1uF ceramic caps are distributed across the board to help the various ICs maintain a stable power supply. The 3.3VDC voltage regulator gets it's power from this 5VDC bus. The 3.3V regulator is adjacent to the 5V output of the LM7805 equivalent.

The output side of the 3.3V regulator has a 10uF ceramic cap, as well as two 0.1uF ceramic caps (one for the voltage translator chip, the other for the microSD card holder). Is that enough in terms of capacitors? I would think so based on the data sheet... Should I perhaps have chosen different capacitors?

On the other hand, the static discharge theory could very well explain the issue, though I don't have carpet here and it isn't the height of winter either. Are there more 'rugged' linear regulators than the one I chose in this regard? Many thanks!

[Reply #3 on:]

Hmmm....I wonder if you have too much capacitance (30uF). The large inrush current when you first turn on power together with stray inductances and whatnot may have caused a voltage spike at the input to the 3.3V regulator. I'm not going to defend this theory too much as it does sound fairly improbable, but I have read app notes on switching regulators that mention this phenomenon (ceramic capacitors are *too* good at their input and electrolytics/tantalum are recommended).

A-ha...here it is:

http://cds.linear.com/docs/Datasheet/3411afb.pdf

Check out page 12, the section titled "Ceramic Input and Output Capacitors". I quote:

However, care must be taken when ceramic capacitors are
used at the input and the output. When a ceramic capacitor
is used at the input and the power is supplied by a wall
adapter through long wires, a load step at the output can
induce ringing at the input, VIN. At best, this ringing can
couple to the output and be mistaken as loop instability. At
worst, a sudden inrush of current through the long wires
can potentially cause a voltage spike at VIN, large enough
to damage the part.

You don't have long wires, but.....who knows.

I'm still betting on ESD. There may well be more rugged linear regulators out there but there are other components on the board that are equally susceptible. I would suggest developing better ESD practices rather than finding more rugged components 

--
The MegaRAM shield: add 128 kilobytes of external RAM to your Arduino Mega/Mega2560

[Reply #4 on:]

Thanks so much!

I replaced the L4931CZ33 with a MCP1825S that I had lying around. Yes, the latter is a bit bigger than the original L4931CZ33 it replaces but for prototype work it's good enough. The output is a stable 3.3V, so I seem to be OK now. However, the past circuit also produced 3.3V at first, so this is not that comforting. The good news re: the MCP chip is that it is explicitly for use with ceramic caps on the output side. I wonder if the other linear regulator was not as tolerant of having a 10uF and 0.1uF ceramic capacitor combination.

So much to learn. I suppose a 25-100uF 16V Aluminum electrolytic on each output stage would have been a better choice?

As far as input spikes go, I wonder if that's possible since the wall-wart is producing 12VDC that has to pass through the other linear regulator first. I suppose the other regulator may not be doing a great job but the output shouldn't be that high either. Additionally, the L4931CZ33 has a much higher input voltage range than the MCP1825S, and for whatever reason the MCP128S is working fine. I suspect your suggestion re: better ESD practices is spot-on.

Thanks again.