Hacker News Reader

Comment by cestith | Hacker News Reader

A 10:1 transformer would work wonders. Lots of lawn equipment is 40v or 80v. That's less efficiency losses than charging a smaller battery and running off of that.

View on HN

Transformers only work with alternating current. You would be better off using a direct buck-boost conversion scheme, rather than DC->inverter->xformer->rectifier->DC.

I was assuming you'd already have an inverter for 120v AC. Wouldn't it be less loss to convert at 400v, step that down to both other voltages, and convert back to DC at 40v than to have a buck convertor for 40v and an inverter for 120v separately? Of course then you'd draw fewer amps to just go ahead and plug 120v tools in and not convert back.

If you're just skipping the NEMA wall outlet and going to 40v DC that's different. You might want too go with a non-isolated switched mode there for efficiency. Still, an inverter above 20kHz and transformer at high frequency in an RCC or resonant forward configuration shouldn't have terrible losses and the transformers at high frequencies can be small and multiple for different output voltages if necessary.

I think the design goals regarding space, weight, cost, flexibility, and reliability really come into play here.

About that buck convertor... I'm just a hobbyist when it comes to power supplies but since a buck convertor tends to use semiconductors I'm thinking a 400v input might be an unusual design. The highest input voltage rating I find for a "high voltage" DC buck convertor is around 140 volts. If you tap the battery a quarter of the way through you could do that, but then you're loading that portion of the cells more than the rest. I can find 700 volt "AC bucks", but there again you're talking AC power. Everything I can find that bills itself as a 400v capable down-stepping DC to DC buck convertor has an isolated portion labelled, which means it in fact has at least a 1:1 transformer somewhere connected to the buck. If you're talking about flipping mechanical relays quickly instead of transistors and diodes, you're going to have a lot of noise, slop, and reliability issues aren't you?

Transformers don't work on DC. You'd have to use a switching converter (which itself may use a transformer internally).

A 50-60Hz transformer would be less efficient than a buck regulator. And it's 400V DC, not AC, so a transformer won't work for the purpose anyway.

You could convert, but you'd have the converter loses too as bob1029 pointed out. A switched DC going into a transformer can act as a convertor, then a step down transformer, then... yeah. A buck regulator would be better.

If you have something to run that needs constant motion besides the current, a motor-generator might be useful. That can convert pretty directly. I'm thinking like the outboard drive shaft on older midsize tractors.

Switching DC through a transformer is just what you do in a flyback converter, so the two windings of the transformer act like separate inductors to the circuit and the turn ratio sets the duty cycle instead of the output voltage. Tends to be more lossy than a buck or SEPIC converter (though less lossy than a low-frequency AC transformer), but provides isolation easily.

And a motor-generator is far lossier than a flyback converter, they're mostly good for high-power applications where you need to convert from single phase to 3-phase and need isolation at low cost. Even then a VFD (variable frequency flyback converter or inverter) will be more efficient, just more expensive. Though days VFDs are getting quite cheap for their power.

Have you found a buck convertor that takes 400v DC input that isn't actually paired to a transformer for isoolation?

They're almost certainly all paired with a transformer for isolation (unless maybe the whole thing is in a sealed module, but that would be rare). But it won't be a low-frequency high-loss AC transformer, it will be driven at a much higher switching frequency to minimize losses. That will also mean the core material will be different, though as I'm not a power supply designer I can't tell you the most appropriate formulation for such applications. Probably not silicon steel used for 50/60Hz AC transformers though, likely some sort of ferrite.

Basically not all transformers are equal. You need to get the properties right to match the application. In the case of their use as a component in an isolating DC-DC converter (or inverter, or other similar system) they're typically not mentioned separately, since the converter is usually a module.

Well, yeah, but a high-frequency transformer can still itself have a coil differential for step-down and is still a transformer. They make them that operate at tens of kilohertz just fine, which is about where the inverter is going to operate.

Though transformers only work with AC.

Reply on news.ycombinator.com