Feeding a SMPS with DC instead of mains

[nickb]

There are plenty of off the shelf solar -> battery -> AC inverters on the market for off-grid applications. That's the way I would go. Once you've got a standard distribution voltage and frequency it's easy to hook to other things.

[eschertron]

With the special case argument presented, I wonder what 'special case' you are solving for by using a SMPS?

If your solar-power friend is looking to get 110 or 220 VAC line power to run his home/camper/mars rover then maybe an "off the shelf" variable frequency drive would work. Most modern units have AC line inputs to the rectifier, but also DC bus connections specifically designed for load sharing. You still need to size it for the expected max load and minimum expected DC voltage, but that's minor compared to the minutia discussed above. Try looking at VFDs made by Yaskawa for starters. Generally low-cost and available.

[R.G.]

Solar power is generated in about half-volt units. The photovoltaic junctions are stacked in series to get you to battery charging voltages, generally 20V or so open circuit to charge 12V batteries, and higher to charge 24 or 48V batteries. Your friend is probably stacking a bunch of solar cell arrays to get up to 220V, or floating many of the arrays to charge each battery from a floating solar cell, but provide the power out as high voltage DC.

The question keeps coming back - why? If the charging scheme is built on the idea of running switching power supplies from high voltage DC, you get all of the quirks mentioned above. Some SMPS use doublers, some use active power factor correction (which isn't generally a problem, but there are different variants to take into account) and others will happily run on DC fed in. Pretty much, you're going to have to know the schematic and operation of every SMPS you connect to this setup. Each one will have a potentially different set of quirks.

The solar power industry is built on the idea of putting solar-collected energy into the cheapest mass electrical storage available, then converting that to usable standard AC as efficiently as possible. That has amounted to arrays of lead-acid batteries for availability and cost reasons, although advancing lithium battery tech is getting there. This approach uses power inverters to make standard AC power from variable DC power. It trades off some inefficiency in converting twice, once to AC standard power, and again to the thing being powered, for ease and low cost of storage.

If you're going with stacked, high voltage battery storage, you get the same problems unless you can eliminate a conversion step, and you take on the additional problem of now having to tinker and select for things that accidentally can be powered by high voltage DC. It gets a lot more special-cased.

[nickb]

Thanks for your input. nickb mentioned this too. But this supply has a voltage-controlled oscillator, it'll work with any voltage from 90 to 240 VAC.

That does not mean there is not a doubler that is switched in as needed. Measure twice, cut once...

[jmaf]

You should be fine with DC input if this SMPS is less than 65W.......
Over 65W then you have Power Factor Correction on the front end of the switcher....depending on circuit used you should be OK..
Best to get schematic..
I recently designed a 5kW SMPS for a MIL application....
One requirement was that it needed to accept both AC or DC input...
The AC input frequency range had to be huge...DC to 100kHz ...
The 90V to 240V range is standard for commercial SMPS... For Dell I designed them from 80 to 300V to account for India's wild grid voltage swings...
The frequency usually is fixed on most commercial SMPS....It's the Duty cycle range that changes...
It becomes tricky at very low duty cycle , less than 10%, to control the inner current loop, so instead the Off-Time ramp is used to predict the On-Time slope ....
Best to get a schematic....
Can you vary this 220V DC you have planned to use ??

That's awesome info, thank you.

I think he spot-soldered a ton of batteries together, I haven't seen it in person. I guess he could derive taps along the way if several DC levels are needed, assuming he didn't pot them in gook. Would lower voltage taps be useful somehow?

[glebert]

I think the 220VDC is a battery bank.

[cerrem]

You should be fine with DC input if this SMPS is less than 65W.......
Over 65W then you have Power Factor Correction on the front end of the switcher....depending on circuit used you should be OK..
Best to get schematic..
I recently designed a 5kW SMPS for a MIL application....
One requirement was that it needed to accept both AC or DC input...
The AC input frequency range had to be huge...DC to 100kHz ...
The 90V to 240V range is standard for commercial SMPS... For Dell I designed them from 80 to 300V to account for India's wild grid voltage swings...
The frequency usually is fixed on most commercial SMPS....It's the Duty cycle range that changes...
It becomes tricky at very low duty cycle , less than 10%, to control the inner current loop, so instead the Off-Time ramp is used to predict the On-Time slope ....
Best to get a schematic....
Can you vary this 220V DC you have planned to use ??

[jmaf]

220V DC and 220V AC are completely different.
108V different to be precise.
Now that I gave you the clue, YOU tell me why
Proof you actually read and understood it

Thanks for your input. nickb mentioned this too. But this supply has a voltage-controlled oscillator, it'll work with any voltage from 90 to 240 VAC.

famous last words

LOL

But yeah I'm concerned. I told him so many times.

[J M Fahey]

220V DC and 220V AC are completely different.
111V different to be precise.
Now that I gave you the clue, YOU tell me why
Proof you actually read and understood it

I suggest you read a little on how Offline SMPS work, to save you reinventing the wheel time and time again.

(I've warned him about safety and 220 VDC with no circuit breaker at the terminals, he says he'll be careful.)

famous last words

[jmaf]

Ah, we may be talking past one another.

Yes, that buck takes a DC voltage and works with it. But where did that DC come from? In our amps, the DC supply is switched off and on through the primary of a transformer. But that DC is made by rectifying the mains AC.

Exactly. I'm thinking of injecting straight DC where the AC would be....where mains would be... Instead of pushing in 220 VAC I'd give it 220 VDC, the rectifier would be redundant...

Sorry for double post, Enzo replied while I typed.

[jmaf]

Yes, an SMPS is DC after the rectifier, but every PC I worked with plugged into the AC mains. Lap tops run on low DC.

Thanks Enzo. Yeah this is definitely unconventional. I've never done it, though should be simple to test at the bench, problem is I'm not near the shop now. Can't test. So I wanted to bounce the idea off you guys before I mushroom-clouded a SMPS.

This is classic SMPS DC-DC conversion: https://en.wikipedia.org/wiki/Buck_converter

This definitely a good analogy. What I'm asking about is just a buck converter after all, duh!

I just gotta look for what nickb said about it having some AC-only circuit at the input, doubler and so on. Other than that should be cookbook buck converter.

I got off the phone with some more feedback, he said it's a 90 to 240 VAC automatic SMPS, so the other problem nickb said about it working on lower voltage won't be an issue I guess, probably has a voltage controlled oscillator that'll increase PWM frequency for higher voltages or something like that.

[Enzo]

Ah, we may be talking past one another.

Yes, that buck takes a DC voltage and works with it. But where did that DC come from? In our amps, the DC supply is switched off and on through the primary of a transformer. But that DC is made by rectifying the mains AC.

[glebert]

This is classic SMPS DC-DC conversion: https://en.wikipedia.org/wiki/Buck_converter

[Enzo]

Yes, an SMPS is DC after the rectifier, but every PC I worked with plugged into the AC mains. Lap tops run on low DC.

[glebert]

I actually think of SMPS as working MOSTLY on DC, but then again I was working in the PC industry where we pretty much assumed everthing to be DC. Don't know about at 220VDC though.