Power supply not switching to battery backup

Hmmmm sounds like a power supply is dipping which can of course also mean the earth return wire is giving a glitch into the micro to reset it...

Without a circuit diagram its difficult to comment on the circuit, but from your problem I'd say it was a supply spike occuring when switching, if you've tried decoupling the supply with caps what about having a seperate local regulator for the microprocessor circuit, as well as having a seperate earth return wire to the power source...

You don't say whether the battery is a long way - in wiring terms, from the ac power supply but it could make the difference enough to cause the earth wire to supply the spike which means capacitive decoupling wouldn't work in itself...

Try a seperate supply for the micro as its obviously resetting itself at every swithing point.

John.

I am using DesignWorks, which can't Save As anything except its own format. It can Print to .pdf. It can Export to .dxf, .wmf, or to formats matching everything from EAGLE to Zuken. It can't do .jpg, .bmp or anything like those.

Be that as it may, the long lead idea had value, because one thing I didn't mention last time was that I actually have four boards that all get power from the same source. I daisy chained the power and ground to cut down on the clutter. The board I am testing is physically next to the power board, so I thought I had short wires. But when I tracked them down, it turned out that power was actually fed to the other end, and then back through every other board, meaning the power leads to this board were about three feet long.

So, I added some six inch (24 ga) wires between the power board, and the board I am testing. I left the existing wires. There was no change in operation at all.

Next, I will change the switcher to 8 or 9V, and add a three terminal 5V regulator in series on the power board (too dangerous anyplace else). There is very little room available to do this, so the answer will be delayed until tomorrow. I added this message so you wouldn't think I was blowing you off.

Assuming you have a circuit diagram, save or export it as a JPEG image. Then when you are posting your message, click on the camera icon and navigate to the image - it will be inserted wherever your cursor is at the moment.

Without that circuit, it is very difficult to follow your problem...

As the others have said, a schematic would be nice. For the meantime...

It sounds like your battery is not being switched in fast enough or the supply is dipping too much before your battery comes in to take the load. I'm saying this because you said that when you tied Q2 to 5V, (effectively, this connects the battery to the load continuously, I gather) the circuit worked fine. There's practically no delay in switching in the battery.

However, I didn't get it why doing this disables the switching off of the charger circuit. These should be separate circuits and work independently. I may be misunderstanding your description though.

Just on a hunch, try increasing the power supply's capacitor to a really big value, say 2000uF. That should keep the DC voltage stable long enough (milliseconds?) for the battery to kick in and take the load.

Have you ever heard of Winsnap? or any other utility for taking screen shots?

I use Winsnap all the time for saving Eagle output that is postable in several forums that I am a member of.

I forgot where I got it or even if it's a freebee, but it is worth checking out.

I got it fixed. Actually, I ended up with two solutions.

Raising the switcher output and adding a linear post regulator worked quite well.

I finally thought to measure the load current instead of merely calculating it (you'd think I'd have learned THAT lesson after all these years...) It turned out to be over twice what I had calculated. So, after removing the linear regulator, and resetting the switcher, I tripled the output cap, and managed to reduce the load a bit by removing some of the "features". This approach also seems to be working just fine.

So, thanks for all your help.

Thanks for letting us know... Glad its all sorted.

Its a shame that most others never respond to their question on here and so nobody knows what has happened...

John.