decade power supply

On a sunny day (Thu, 13 Oct 2022 19:31:00 -0700) it happened John Larkin snipped-for-privacy@highlandSNIPMEtechnology.com wrote in snipped-for-privacy@4ax.com:

Replace the 5 V switcher wih the clean switching LM we discussed, add a series diode in the input against reverse voltage. Leave the rest as is?

PS the series diode is really needed, there exist 'universal' wall-warts, I have 2:

Bought if from Reichelt. Always check with voltmeter before use!!! Nice thing really.

or more

very nice work, thanks John. I assume your company includes a "correct" AC power adapter with each product. And you are just being cautious, if say someone grabs a similar looking adapter in a rush. You could control the problem by using special DC connector, but maybe raising the cost?

Of course in consumer space, its the

What would happen if a 5-VDC was inputted?

soon USB PD switchable between 5-9-12-20V is probably going to the be more common

nothing, the max809 only enables the switcher above ~8.8V

I need a lot of +5 so may as well use that as the intermediate. It makes reliable sequencing easier too.

We've considered a zillion options on this supply. Efficiency (ie box cooling) and parts size are just a couple of dominant issues. The old

This supply forced me to use four new parts, all from TI, which is a good thing in the long run.

I like TI. They make good stuff and are not exploiting the parts storage. We can get parts direct from TI at sensible prices, which I suspect keeps the distributors and brokers in line.

We do, but about half of our customers have their own DC power and buy without the wart.

But on my bench, I sometimes have several warts with their leads tangled and it's easy to grab the wrong one. I paint the connectors to color code the voltages, but it seems that is not commom practice. We sell maybe 700 a year of this box and maybe 5 come back with the 12v transzorb fried.

This design won't start up until about 8.8 volts, so 5v wouldn't harm it. I do want a quiet and stable +5 supply, so would prefer to make that internally. Our biggest user buys it board-only and puts it inside a camera, and supplies +12.

I guess we could use a buck-boost Sepic to let it run from most anything. But I'm too tired of this to start over, and we've bought a lot of parts because we could.

I might change the inductors, after we do some breadboarding.

The left side of sheet 1 is responsible for supplying +5 with guaranteed brownout behavior to the downstream stuff. Specifically, a short dip in the +5 is impossible; that lets the critical low voltage supplies discharge and makes the powerup sequencing reliable.

That MAX809 guarantees a minimum low time of 140 ms. The MAX809 is a great part, especially if you don't buy Maxim.

On a sunny day (Fri, 14 Oct 2022 08:16:38 -0700) it happened John Larkin snipped-for-privacy@highlandSNIPMEtechnology.com wrote in snipped-for-privacy@4ax.com:

Did not somebody show the 'perfect diode' with a MOSFET and 2 transistors here recently?

It can be done with a single Pfet.

fredag den 14. oktober 2022 kl. 18.46.55 UTC+2 skrev John Larkin:

as long as you don't need it to stop back feeding the source

Just what we need: Still more options and configurations. NOT!

Jeroen Belleman

Agreed, Lasse - the topic of "ideal diode" is worthy of its own thread. Many circuits Ive seen failed to meet all the requirements.

We used some TI Efuse parts. Mostly blew up.

The polyfuse and TVS can handle reverse supply voltage. The TVS is just a big diode in its forward direction.

lørdag den 15. oktober 2022 kl. 00.29.32 UTC+2 skrev John Larkin:

what did you do to them?

Used them at the input of a pockels cell driver that ran from a 48 volt supply. TPS26600, rated for 60 volts.

With a power-pad TSSOP package, it takes about 200x as much time to replace as a real fuse.