Let's Make Robots!

Step-down regulator problem

I've been using an LM2940CT-based regulator to power the logic on a Dagu 4ch motor controller. It's one of OddBot's designs from his walkthrough and it works just fine.

I decided to buy a Pololu D15V35F5S3 to replace it simply because it's much smaller than the one I built.

Before using it, I decided to have a look at the output on my scope and got a reading of 0V. I checked the input terminals and found a good 8V so I unplugged the battery and tried again. This time it was close enough to 5V. After much testing, and occasionally getting 0V output, I noticed that it's quite sensitive to the way you connect the battery and shuts down if I don't plug the bullet connectors together cleanly.

After digging through their website, I found that this seems to be documented where it talks about ensuring that the rise time on the input should be less than a few milliseconds or it will latch into an error state.

I asked about this on the Pololu support forum a few days ago but didn't get an answer so I'm going to stick with the LM2940 to make sure I don't get any magic smoke from applying the motor supply without power to the logic.

Is anybody successfully using this reg or has found the same behaviour and solved it? I'm sure it's a nice reg but just seems too risky for this particular application.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

I forgot to mention...

For the robot, I'm using a 2S 3000mAh 30-40C Lipo through the same good quality 4mm bullet connectors that I've been using for years in RC aircraft. I only have one 2S plane (most are 3S or 6S) and it only draws around 38A at full throttle from the same Lipo packs (but that should be an adequate test for a Rover 5). A 6S 4000 version of the same brand pack in a larger plane supplies 80A continuously through the same bullet connectors and only shows 0.2V drop after 2 minutes at full load. It remains cold to the touch after that test. (China has been making very good quality Lipo cells for the last few years.)

Running from the bullet connectors, I have short lengths of the largest diameter cable I can physically fit into the screw terminals on the Spider, on the 4ch motor controller and on the regulator in question. I think it's around 14awg - after I've tinned the core, it's around 2mm in diameter.


Anyway, I've built a more compact LM2940 based reg than the previous one. It's powering the Dagu board quite happily and doesn't mind if my hands shake a bit when connecting the battery. I was hoping to gain a better understanding of why the other reg gives up so easily (perhaps there's a very good technical reason?) and maybe how I could prevent it doing so. Before I stopped actively building electronic toys and gadgets years ago, all my previous regulators had been based on 78xx devices so I'm still coming up to speed on more modern technologies. (Thanks again OddBot for steering me onto the LM2940 - it's an improvement)

But the cute little reg is now on my test bench and, as I said, when I get a chance I'll capture some traces with the DSO as I connect it up.

Thanks for the suggestions.

 

OddBot's picture

Switching regulators draw pulses of high current. If your batteries or wiring are not up to the task then they shut down. Some people have had a similar problem with the Spider controller. This is where linear regulators are a better choice for beginners.

Hi birdmun,

I read that section in the application notes a couple of times to make sure I hadn't missed anything and it says that EN can be tied to VIN or left disconnected but it can't do any harm since I definitely want the output enabled at all times. So good suggestion, thanks.

Just for the sake of experimentation, I might also look at the input terminals with the scope and capture how much "noise" is required to make it play dead.

 

birdmun's picture

I am curious if you have considered trying to tie the ENABLE to the VIN?