Rail-splitting a wall wart

It's 4 ohms in unipolar mode, which is better but doesn't do reverse protection the obvious way.

I want a higher-power version.

Consider, maybe, a biggish SSR and a small ssr or optocoupler to measure the ON voltage drop for shutdown.

Reply to
John Larkin
Loading thread data ...

Not the kind of mission-critical product I think I'd want to personally take responsibility for selling as some kind of black box-ish module at this time given, as you say, the details need to be right.

I think such a design (even just the "code" part) would need more than just my eyes on it to ensure that, and more time testing than I'm willing to commit ATM.

But if you have ideas of what an implementation might look like on the analog side I can post a "code" sketch this month, and people could make closed-source derivatives if they like. Honestly the Greenpak Designer is very straightforward to get a handle on, and as its mixed-signal designs simulate more like SPICE than an FPGA sim.

There are a bunch of overvoltage/reverse polarity protection ICs availabe; most seem geared to automotive applications where the hazards are mainly load dumps and someone connecting the 12 volt battery backwards, relays obviously aren't helpful to defend against the type of over/under voltage conditions caused by load dump transients.

Reply to
bitrex

Nah, I'll stick with 70 cents worth of small FETs and a resistor pack for the grounded-wart fault condition, and let the polyfuse look after overcurrent. The TCA0372 will be pretty much loafing at current levels I care about for this use.

The FETs would be cheaper still except that I need >24 V_GSmax--for PFETs, I wound up with DMP3056es for all three, and for NFETs, DMG3406L pass devices and a MCC 2N7002A driver.

Cheers

Phil Hobbs

Reply to
Phil Hobbs

This really isn't bad.

formatting link
A variation would be to make the diode a TVS and trigger the SCR after the over-voltage was there for a while, before the TVS fries.

Reply to
John Larkin

Or dump the big diode and use a triac.

Reply to
John Larkin

Fun.

"Go ahead, wallwart--make my day." ;)

Cheers

Phil Hobbs

Reply to
Phil Hobbs

On Wednesday, February 8, 2023 at 7:33:14 PM UTC-8, John Larkin wrote: [about power inlet protection]

Me, I'd stick with two zeners and a resistor. For polarity reversal, and overvoltage, make it a fusible resistor.

Reply to
whit3rd

Then, when a user plugs in the wrong wart, the unit dies hard. That's exactly what we're trying to avoid.

A polyfuse and a TVS is pretty good but is not safe in all cases.

Seems to me that all resistors are fusible resistors.

Reply to
John Larkin

The obvious response is USB-C.

Reply to
John Larkin

Your followup in turn leads to the intuitively obvious question: does USB-C supply 3.25A @ 19.5VDC?

Hint: the assortment above pertains to laptop power adapters.

Danke,

Reply to
Don

Yep, potentially USB-C can negotiate up to 5A at 20V. However it could be as little as 5V 0.5A if for example an original USB-A supply is used with an adapter cable.

piglet

Reply to
piglet

USB-C has the potential to end the mockery of "Universal".

It could be the only connector that most boxes need.

Reply to
John Larkin

Everyone loves simple solutions. It's a shame such "one size fits all" constructs continually constrain until things become unworkable. Contemplate USB-C's 5A (possibly more) separated by a few hundred micrometers from GND - what can possible go wrong? In the wild you see USB-A males jammed into RJ45 ports. You experience USB-A males excessively forced into USB-A females until the internal structure crumbles. Subsequent shorts sometimes take out power supplies.

Danke,

Reply to
Don

That would have been USB-B, the device side of the cable fitted perfectly into the 1 GBit RJ45 port of an expensive motherboard.

Arie

Reply to
Arie de Muijnck

Both variations happen in practice, because the more tech savvy people realize both ends of a USB cable must be connected for it to work. The people who cram USB-B males into RJ45 ports may believe only one end needs to be connected, as is the case with antenna, for instance. Bless them for they keep me busy.

It occurs to me how ?hundreds of millions? ?billions? of people in the wild correctly connect USB-C chargers to their phones. There's tentative hope.

Danke,

Reply to
Don

All the AC outlets in my house are compatible! And all the cable TV connectors! Something ought to be done about that.

None of mine have gone wrong. Snark aside, it seems to be an excellent connector.

What's the required insulation distance for 5 amps?

Sure, we need a single USB connector standard: C.

I doubt that many people confuse USB-C with RJ45.

Eventually usbC should replace RJ45. One cable per device, power and data and everything, would be ideal.

Reply to
John Larkin

B is the only unambiguous USB connector and the only one (until C came along) that is easy to mate first try. We use a B connector on our instruments whenever we have the room.

formatting link
Customers just use a common printer cable. This box enumerates as a serial port so no drivers have to be installed.

Reply to
John Larkin

Don't know about F threaded coax connectors, but as for AC outlets, there's a known incompatibility issue:

formatting link

... and I can see several examples from this chair.

Reply to
whit3rd

Distinct 20A AC outlets circumvent the 15A constraint. My allergy to cable TV makes me not comment on it.

Danke,

Reply to
Don

No, but it can do 5A at 20V, that might be close enough.

Reply to
Jasen Betts

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.