Switchmode regulator for R/C use

[BrianG]

Some of you may be familiar with these, but for hobby-class electric radio-control vehicles, many people use a small switchmode regulator (called a BEC for Battery Eliminator Circuit) to efficiently regulate various battery voltages from 11.1v (3s lipo) up to and over 42v (10s lipo) to provide 5v/6v @ 3A+ to power servos and the radio receiver. Linear regulators would be far too wastful. Here are a few examples of such circuits we use:
- **broken link removed**
- **broken link removed**

Anyway, these switch mode regulators have the unfortunate characteristic where they can fail, and when they do, can output the full battery voltage. This is "not good" for the 6v servos and receivers, both of which can run over $100 each. So, I was thinking of making a small and cheap catastrophic safety circuit that can be added to the output to keep this from happening. Size and weight in these vehicles is a concern, so the device has to be small and light. I've been toying around with a few ideas. Two of my favorites are:

- Use some type of shunting device (like a TVS or Zener diode) rated for around 6.5v. Hook this to the output via a fuse. If the BEC fails and outputs full battery voltage, the TVS would shunt the extra energy to ground, and thereby blow the fuse.

- The other idea uses a comparator that cuts the output when it exceeds a set value, probably 6.5v. This method is less destructive, but a little more complex (and therefore more prone to fail) and I am concerned about the response time.

Yeah, the BEC is still toast in either of these solutions, but at least we've saved the expensive stuff downstream.

Obviously, the best solution would be to have these makers put some type of protection in their products, but until they do, we have to do what we can.

Any ideas, suggestions, comments?

Edit: Sorry, I just realized I might have posted this in the wrong area...

 

[dknguyen]

Both of those will work. The TVS+fuse is probably more appropriate as it is more universal, cheaper, smaller, lighter, and simpler. The only real point with a "non-destructive" method is that it can reactivate. But with a faulty BEC that's not going to happen. The plane has to go through a refitting anyways, and replacing a fuse is not a big deal.

The TVS diode is MUCH faster than anything else. The problem is that a large TVS might be required since you want the fuse to blow, and not the TVS. You need a TVS large enough so it does not burn out while supressing the voltage until the fuse can blow.

 

[Chippie]

Anyway, these switch mode regulators have the unfortunate characteristic where they can fail, and when they do, can output the full battery voltage. This is "not good" for the 6v servos and receivers, both of which can run over $100 each. So, I was thinking of making a small and cheap catastrophic safety circuit that can be added to the output to keep this from happening. Size and weight in these vehicles is a concern, so the device has to be small and light. Yeah, the BEC is still toast in either of these solutions, but at least we've saved the expensive stuff downstream.

Obviously, the best solution would be to have these makers put some type of protection in their products, but until they do, we have to do what we can.

Any ideas, suggestions, comments?

Edit: Sorry, I just realized I might have posted this in the wrong area...

The idea sounds fine in principle for ground based r/c stuff but not for airborne....No control would lead to destruction of the airplane unfortunately....This has been discussed at length over on R/CGroups....

 

[MrAl]

Hi,


The kind of circuit you are after is usually called a 'crow bar'. It simply shunts the
output to ground in the event of an over voltage.

There is a catch however, and that is that none of the circuits are truely
instantaneous, so the load has to be able to handle a very small time length
over voltage, or you have to add a small inductor in series with the output.
The small inductor prevents a very fast rush of current into the sensitive
load, while meantime the crow bar can act fast and clamp the output
(before the inductor) to ground.
Crow bars these days can probably be made to act very fast, so only a
small inductor would be needed. A small cap on the output too would be
a good idea, to keep the inductor from kicking back and blowing out the
load all by itself.

For example, if you can get your crow bar to clamp in 100ns an inductor
of only 1uH would be enough to limit the voltage to the load while the
crow bar has time to take over. A couple turns of wire would do it.
This is all assuming 5v and 1 amp load.

 

[indulis]

You could try building your own:

**broken link removed**

Throw in a comparator (hi speed) to monitor Vout and have it control a low side MOSFET on the battery.

 

[BrianG]

I guess I should have mentioned that this is primarily for ground-based vehicles. I too think a destructive method is fine because the BEC is toast anyway. I'd rather pay ~$30 for a new BEC than $200 for a new servo and receiver.

And for those interested, here is the scenario that happened to me: UBEC Failure - RC-Monster Forums

Yeah, a crowbar protection is basically what I'm leaning towards. It would be the easiest and cheapest solution that would offer some type of protection.

As far as building my own BEC; I don't think that will be a good idea. Most BECs use a circuit and IC similar (if not the same) as the one posted, so anything I would build would be basically a copy with the crowbar circuit added on. And many ESCs are coming with switching BECs nowadays anyway. So, a tiny circuit insterted inline would be the best solution until manufacturers build protection themselves.

I will probably do some testing on the simple crowbar idea, and if I see it doesn't react fast enough, will then look into adding an inductor.

Thanks a lot guys! At least I know I'm on the right track.

 

[Mr RB]

What about a redundant system? If you have a second BEC (or second regulator) and connect them both by diodes either one can go shorted and you will still have a fully operational plane. The primary regulator will be the one with the fractionally higher output voltage.

 

[dknguyen]

I'd just use a receiver battery if you're so concerned and bypass regulation. THat has an additional safety measure. But I'm thinking about planes.

 

[BrianG]

Well, a redundant system still has the same shortcomings. If one fails, yeah, the second one will take over, but only until that one fails too. And, the first one will have to have about 1.5-2v higher output than I really want to compensate for the drop out voltage of the second stage. I did think about using a switching BEC for the first stage to get the output to around 7.5v, and then use a linear regulator for the second stage. Linear regs tend to be darn near bullet proof. The first one will do the majority of the work while having high efficiency. The second linear stage won't have much voltage differential, so it should be fairly efficient at 3-5A, unless the switching one fails. At that point, the linear reg would get quite warm/thermal and definitely notify the user that something is wrong. That's another possibility. Unfortunately, only a few of these switching BECs have adjustable output that can go higher than 6v, and that's the Castle unit (linked above). The rest are usually switchable for 5v or 6v, or set to one or the other. The solution I'm looking for ideally should be used for any unit.

Most of us car guys hate using receiver packs. It's just one more thing to charge up. Much easier and lighter to use an electronic solution that runs off the main batteries.

 

[indulis]

You could also build an isolated DC-DC converter, then a primary side MOSFET failure won't affect the output. However, the efficiency won't be as high as a non-isolated buck converter... probably around 85%. Synchronous rectification would help, but it would also complicate the circuit. The size of the PCB would also be a bit larger... ~ 1" x 1.5" would probably do it with SMT components.

 

[Mr RB]

Well, a redundant system still has the same shortcomings. If one fails, yeah, the second one will take over, but only until that one fails too. And, the first one will have to have about 1.5-2v higher output than I really want to compensate for the drop out voltage of the second stage.
...

No, you connect both supplies in parallel, each has a diode going to your receiver. The supply that has (say) 0.1v high voltage will do all the work unless it fails then the other one takes over. The only down side is the cost of a second BEC and the voltage drop from a diode, if you use a schottky that's only about 200mV drop at maybe 80mA for a receiver.

And the chances of both BEC systems failing in the one flight???

 

[BrianG]

I don't understand how that will help. Usually, these BECs fail closed, where the output will be at the same voltage as the input. And since we usually run anywhere between 11.1v (3s lipo) and 37v (10s lipo), the result can be very "bad". This is why I was thinking that you meant to run these BECs in series. I can see the parallel working if they failed open, but sadly that is not usually the case.

And since this is in a car, it really doesn't matter if the receiver loses power. The car will just coast to a stop. In an aircraft, the results are a bit more dramatic. So, this is why I was shooting for a solution that somehow disconnects the load from the BEC when it fails. Sorry for not being clearer.