0
\$\begingroup\$

Are there any real negative drawbacks - outside of the double drop of efficiency - of DC-DC boosting a power line that was already DC-DC bucked?

I'm building a modular system, serially linked through contact points. Most modules need 5V and low amps, but a couple need 12V and 1A. I'm intending to use a 12V power supply as the main source, but would like a battery option.

The most straightforward option is to have dual power rails, one 5V and one 12V, and let each module just draw from whichever rail works for it. This means I would only have to implement the boost/buck converter once, in the controller. But it would mean having to use six contact points (12+, 12-, 5+, 5-, and two data lines) on the sides of each module, which cost significantly more than the four-contact strips I'm looking at. (I also just want to reduce complexity with the connection)

I could also just run one 12V line down to each module, but this would mean implementing a buck converter in each 5V module, so again not very economical. In the case of a user opting for the battery pack, this would also require a boost converter in the battery pack to provide the 12V rail.

So I could also just run a 5V line down the modules. I buck the 12V power supply once in the controller, or just let the 3 or 4 AA battery pack run straight through. This would mean including a boost converter in the modules that need the extra voltage. This works out economically because it's a relatively small number of modules that need this. But does it work out electrically?

There are two possible issues I can see -- first is my title question. This would mean bucking 12V to 5V, then later boosting 5V back to 12V. I'm not worried about the loss of efficiency. But are there any other concerns?

Second is, would the batteries be able to handle boosting from 3.6 (3 * 1.2 rechargeable) - 6.0 (4 * 1.5 alkaline) volts to 12V with a 1A draw? The vast majority of the time my 12V module will draw 1A in very small pulses. 1-5 pulses lasting 10-50ms each, with 3-15 second pause between them. Fairly low overall power. But every now and then the user will need to trigger a routine that will pull the 12V/1A for 30s at a go, perhaps multiple times. Assuming an ideal boost converter, 12V/1A would pull 6V at 2A, so I've got to assume it could be even larger with real world components. Can AA batteries handle that? (another option: this routine is not the primary function, so I could possibly just tell the user they can't do that with a battery pack)

Thanks for any advice. I'm new to this field but having fun learning. :)

\$\endgroup\$
  • 1
    \$\begingroup\$ Your question is very long, i haven't read it all. In general, as engineer you do what it takes. Just treat carefully the layout, and everything will be fine \$\endgroup\$ – Gregory Kornblum May 13 '16 at 4:54
  • 1
    \$\begingroup\$ The most obvious issue is the potential EMC nightmare. Boost converters are noisy. You have to ensure that both converters have a good layout, and also ensure that the boost converter doesn't leak noise back to the 12V plane. Doing these kind of layouts isn't trivial. \$\endgroup\$ – Lundin May 13 '16 at 9:14
  • \$\begingroup\$ @Lundin are buck converters also noisy? Do Cuk converters have noise issues? \$\endgroup\$ – David Woods May 13 '16 at 14:54
2
\$\begingroup\$

Apart from the lost of efficiency that you have already identified, the other drawback is that the current that will flow through the 5V line will be higher than if you'd provide 12V. So it means that you may have to run thicker cables that can carry more amps (more expensive, less safe).

If you need to provide 12V at 1A to two modules, the 5V line will have to carry at least 4.8A. This becomes to be quite large. Check your wire gauge and connector specs. Moreover, this could cause a significant voltage drop if the wiring length is long.

You said that "Most modules need 5V and low amps". So if the 5V modules actually need relatively low current (in the range of a few tens of mA), it may actually be acceptable to deliver 12V to the modules, and use a simple linear regulator for the 5V, and no buck. Overall, it may be more efficient than using boost for the 12V modules, you'll be immune to voltage drops for the 5V modules, and you'll carry less current through you whole installation. It really depends on your application, but choose wisely.

Regarding the batteries ability to provide the required current, it seems to be OK (check the discharge curves, you can find some easily on the internet). But the lifetime will be very short. After drawing 20/30 times the power to produce 12V@1A during 30s, I think you'll need to change the batteries.

Note: if you draw 2A from the alkaline batteries, their voltage is not 1.5V anymore. It will be more like 1V. Check this, as well, because 4 batteries will therefore not be enough to provide 5V. You'll need a boost converter, even for the 5V.

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.