Does an AC to DC power brick need pure sine wave or will it be fine on a modified sine wave cheap inverter?

[Deleted member 23531]

Fair enough, Factory400. I agree the cost difference doesn't seem that significant, at least when I bought my inverter it wasn't too much more expensive to get a pure-sine one and not have to worry about it.

By the way, if you distinguish between noise and interference you'll sound [to circuit designers, anyway] more like you know what you're talking about. Particularly to analog and RFIC designers, noise means thermal, shot, flicker, popcorn noise -- basically any random process which is based on physics / statistical mechanics. Interference means an outside or undesirable signal which in theory is deterministic but is out-of-your-control. So inverter "noise" is actually interference; and noise is sort of the layperson term in common usage. I've met people who screen EEs applying to jobs based on this differentiation.

 

[Factory400]

By the way, if you distinguish between noise and interference you'll sound [to circuit designers, anyway] more like you know what you're talking about.

I don't squabble over semantics and all of my successful commercial designs are still successful.

As the business owner and circuit designer - I have to communicate in a way that everyone can pickup on what I am putting down. The engineers around me (that are much better and much more specialized than me) know what I mean and know what we need. I don't do RF or design RF silicon so I have no need to carry on like I do. If I tell pretty much any engineer that 'This section of the system is generating noise that is interfering with the operation of this other section' they know exactly what I am referring to. If they don't I would probably fire them on the spot and move on.

If any of my employees squabbled over sematics with me or others - I would also probably let them find a new job.

Also, I don't apply for EE jobs and most likely never will. My job is to design and manufacture products and that seems to be working out rather well so far.

 

[Deleted member 23531]

That's funny, I thought you worked alone. Don't take it so seriously I'm just joking around.

 

[Deleted member 23531]

To the extent magnetics are involved, PSRR probably isn't linear.
A choke can only choke so much before it can't choke any more.
Most dramatic with common-mode chokes, which are probably limited mostly by resistance heating. A (tiny) 5A common-mode choke carries differential current with little or no magnetic field generated. Small amounts of common-mode can be blocked. But it only takes about 50 mA of common-mode current to saturate the choke and it doesn't do anything any more.

Back to the actual topic...this is an interesting point, Hedges that common mode signal could be higher on a modified sine inverter compared to a sine inverter.

But, where would 50mA of common mode current be going? Power supplies are transformer coupled for safety so there could not be a DC common mode current unless there is a ground loop. There could only be AC common mode signal on the primary side, that I can understand, and that is what a common-mode choke is trying to suppress.

any single-ended chokes trying to block them might be over driven.

Sure, there are more harmonics to filter out; not sure at what level they would crap out. Do high-side rectifiers like you highlighted previously typically have a single-ended choke? I would have to dig up the latest schematic teardown of an Apple power adapter to look.

 

[Deleted member 23531]

Here's the best teardown I've seen. This is now several years old though.

http://static.righto.com/files/charger-schematic.pdf

Features a common-mode (ie, not single ended) choke followed by a high-side rectifier / flyback design like I mentioned earlier.

 

[Hedges]

Maybe not more common mode, just much higher harmonics to be blocked by individual chokes.
But maybe common - there is capacitive coupling from primary to secondary. Lots of these supplies don't have a ground, and are fed a hot and a neutral. The higher frequency harmonics will couple through that better, making stronger common mode.

Any switcher is going to make harmonics on its secondary, which would happily go through common mode chokes. So I would expect single-ended chokes to block it.

I haven't done much design with those things, but dealt with switching noise on or coming from someone else's board. Like 0.9 Vpp on the 1.8V feeding me, which turned my LDO on and off. And noise getting into all the other supplies, several volts p-p. Hash from 75 MHz to 150 MHz, all related to how the brick was or wasn't grounded to the box it powered.

Also found that a Meanwell caged switcher made a -8V power-on transient relative to ground, blowing out boards on the test fixture it was part of (so I substituted bench supplies.)

I like linear supplies. I've got 4 old HP boxes in my rack for 15 programmable voltages.
But for power systems, obviously these are switchers to deliver 95% or 99% conversion efficiency.

 

[Deleted member 23531]

Here's the best teardown I've seen. This is now several years old though.

Next time I'll just link a real schematic like this instead of the dumbed-down one that got everyone at my throat ?

 

[Deleted member 23531]

I wouldn't bet my house on it, but actually it looks like you could power that Apple charger with whatever input signal you wanted: square wave, triangle wave, DC voltage, sine wave, modified sine wave. The L6565 has a 20V zener diode to generate its VCC supply so anything above that would turn it on. At least based on the schematic. There could be some sensing circuitry that I'm not seeing though which wouldn't allow that.

The question is...would the spouse get mad if I blow up our iPad charger.

EDIT: the VFF pin senses the rectified line voltage (divided down) so you'd have to use a high DC voltage or read through the datasheet to know what it would do if VFF is lower. I think it just uses a higher duty cycle since less power is available from the source at lower line voltage. Datasheet says "In applications where this function is not wanted, e.g. because of a narrow input voltage range, the VFF pin can be simply grounded, thus saving the resistor divider".

 

[Hedges]

Rectify, filter, and feed it DC?

 

[Factory400]

That's funny, I thought you worked alone. Don't take it so seriously I'm just joking around.

View attachment 40281

Lol - Factory400 is not my job. If it was.....I would be crazy hungry.
Factory400 is my pet project where I experiment with ideas on how to cram a full suite of design and manufacturing into my garage space at home.
It is exiting, and I do sell things sometimes. But it is just a fun thing to do. I freely admit, that I have not cracked the code to make any consistent money at it.

My wife, however, requires that I maintain a more conventional effort to keep things going. That is not part of my social footprint, sadly. There is not much to show or talk about without causing problems for myself. Honestly, I would enjoy just doing a YouTube channel full time though. It would be fun even if it didn't make much money.

 

[Deleted member 23531]

Ok, great, I don't want to create any enemies. I meant the "sound like you know what you're talking about" comment more as a tongue in cheek comment like this:

How to Sound Like You Know What You're Talking About (Even When You Don't)

We can't all be human encyclopedias, and once in a while, you're bound to stumble upon a conversation on a topic you know absolutely nothing about. To

lifehacker.com

 

[Deleted member 23531]

whatever input signal you wanted: square wave, triangle wave, DC voltage, sine wave, modified sine wave

Well I tried one of those 1 inch Apple "cube" chargers and with 60VDC and it didn't turn on. I might see if I have an older one as that website was from 2012.

 

[Hedges]

Well I tried one of those 1 inch Apple "cube" chargers and with 60VDC and it didn't turn on. I might see if I have an older one as that website was from 2012.

100V to 240V

About Apple USB power adapters

This overview will help you identify a USB power adapter from Apple. Apple USB power adapters and cables are included with many products and are available separately from Apple and authorized Apple resellers.

support.apple.com

 

[Pappion]

Having performed PSRR tests, I was exposed to some crazy expectations from management. PSRR At 6MHZ???
The linear Error Amp which controls the output voltage has a bandwidth, could be 10khz, 50khz. Maybe higher but that invites oscillations. It cannot reject noise above it's bandwidth. The effective source resistance and output capacitance becomes important above the bandwidth, but the single pole RC effect is limited. So more HF trash in more HF trash out.
Capacitors can reduce, but not eliminate. Plate charge can only change when voltage changes. A caps Effective Series Resistance and Inductance also matter. Electrolytic caps are a double spiral wound coil of aluminum, so they are horrible for HF trash.
Coil filters are not much better. They have resistance and capacitance. Being large and expensive they are sized for the switching frequency which is well above the 60hz harmonics.

 

[Deleted member 23531]

100V to 240V

Darn, my power supply only goes up to 60V. I think if I could get the case open, maybe reducing the two 510k resistors would get more current into VCC to kickstart it. The datasheet shows two 75k resistors there.

 

[Deleted member 23531]

I was exposed to some crazy expectations from management. PSRR At 6MHZ???

Yeah that's nuts. This is the best part I've seen in my limited searching:

https://www.analog.com/media/en/technical-documentation/data-sheets/LT3045.pdf

There's a dead zone where it's out of the loop bandwidth but the cap hasn't squashed it yet. Ex: that crossover is likely at about 1.9MHz for the LT3045, where that "spike" in the PSRR occurs.

 

[Pappion]

Yeah that's nuts. This is the best part I've seen in my limited searching:

https://www.analog.com/media/en/technical-documentation/data-sheets/LT3045.pdf

There's a dead zone where it's out of the loop bandwidth but the cap hasn't squashed it yet. Ex: that crossover is likely at about 1.9MHz for the LT3045, where that "spike" in the PSRR occurs.

At least the input capacitor is optional. Imagine the power you need to swing a 0.1uf cap 1vpp at 1-10MHZ. Watch your ground connection placements as you will be bouncing ground too. Test may fail without a part in the socket.

 

[Bud Martin]

Darn, my power supply only goes up to 60V. I think if I could get the case open, maybe reducing the two 510k resistors would get more current into VCC to kickstart it. The datasheet shows two 75k resistors there.

100VAC RMS feeding the rectifier and the the filter cap will be charged up to peak of 100VAC = about 140VDC to run that power supply. You can easily verify that by measuring the DCV on the main filter cap, kust be real careful when doing that.

 

[Roswell Bob]

The cost difference between a modified and a true sine wave is slight. I only ran electronics off a modified sine wave generator because that was what Amazon had in stock for 2 day delivery approximately 3 days before the hurricane. If you have the luxury of time and the choice, get sine wave. If not, generally not a big problem.

Beyond being in a time jam (as you were Just John) I can't understand why anyone would buy a MSW inverter and why anyone would even produce one. The cost to produce differential between the two is about zero. Slightly more heatsink for the PSW, but I would bet the output filter on the MSW would probably gobble up that savings. Most MSW inverters probably don't meet any agency specs on line distortion or RFI so just another reason to run away from those designs.

 

[Roswell Bob]

At least the input capacitor is optional. Imagine the power you need to swing a 0.1uf cap 1vpp at 1-10MHZ. Watch your ground connection placements as you will be bouncing ground too. Test may fail without a part in the socket.

The net power is about zero.