Fried a regulator? « LeafLabs Garden

LeafLabs Garden » Support » Maple Hardware Support

Fried a regulator?

(10 posts) (7 voices)

Started 3 years ago by peax
Latest reply from cmartin

peax
Member
After selecting EXT power and connecting my Maple R5 to a 12v wallwart (power socket on board says 7-18v so I thought this would be safe). It lit up, flashed a bit, and then went dark. Now it will no longer light up at all, even if I switch back to USB power and only connect the USB cable.

With my very limited knowledge of electronics I have spent some time searching both google and this forum for an answer to what could have happened. What I am currently suspecting is that one of the voltage regulators have fried.

I measured both regulators (HU95) next to the power socket with the board connected and powered by USB:
- Vin is 5v on both regulators
- Vout on the regulator closest to the power socket gave 3.3v
- Vout on the regulator furthest away from the power socket was shorted to GND
Would I be safe to assume that I've managed to fry the second regulator, or is there somewhere else I should be measuring as well before I go look for new components?

I believe these two regulators are of the type MCP1703T-3302E/CB (3.3v SOT-23A-3)

Any idea why this happened? Was 12v too much for this board to handle despite the 7-18v markings?
Posted 3 years ago #
lucky13
Member

Hello peax. I was facing the same problem before some time.
Have a look at my post..http://forums.leaflabs.com/topic.php?id=1378
"Was 12v too much for this board to handle despite the 7-18v markings?"
Simple answer. YES.
The MCP1703 is a crap and is highly important to be changed in the next revisions of the maple..
Did you have any sensors-leds hooked to the maple when you powered it with 12V?
From the MCP1703 datasheet if you give 12V to the maple input leaves only 10mA max! (in real life is smaller from a personal bad experience..)
So this means that putting an external regulator for powering your components (such as sensors ,leds etc) is a one way.
Another method to avoid this to be happened again is to use something like a LM7805 regulator in order to be as close to 3.3V..That will give you about 200mA to use..
The one regulator that you measure 3.3V at the output is the one for VREF thats why is working properly, the other is for powering the uC and gives the 3.3V VCC pin.

Posted 3 years ago #
siy
Member

Well, MCP1703 is good enough for the purpose of sourcing MCU and on-board peripherals from low voltage sources such as USB. Taking into account required current, 9V probably is the maximum safe input voltage even without external loads such as shields. This limitation is the main reason why I included dedicated regulator for the external loads in Mini64.

Posted 3 years ago #
mikep
Member

Yep, peax, sounds like you found the culprit. It's also somewhat documented here: http://leaflabs.com/docs/hardware/maple.html#power-regulation-on-the-maple

The totally underpowered voltage regulators of the Maple are one of it's biggest design flaws in my opinion. The way they are dimensioned, you can't really put any shields on the Maple without having to have a separate power supply for them. So you have to power the Maple and then power the shield too. That kind of defeats the purpose of having the barrel jack connector on the Maple. Hopefully the new board designs will address that.

Posted 3 years ago #
peax
Member

Thank you guys,
Both quick and very helpful :)

lucky13: It was actually your linked post about regulators that made me measure them in the first place, so my hat's off to you. I had nothing but the bare maple connected to power when the lights went out, not even USB.

One of the most annoying parts to this is that while a replacement regulator of the same type is super-cheap, the shipping of one is like 15 times the price of that single component so I'll order a couple of them just in case, and some other random stuff that could be useful.

Thanks for your help!

Posted 3 years ago #
siy
Member

This is usual situation with many components. I've recently purchased 100 pieces of tiny KMR211 switches just because in smaller quantities final price (price + shipping) of one switch looked like each switch has an AI inside and is smart enough to fetch newspaper and bring can of beer.

Posted 3 years ago #
poslathian
LeafLabs

Yea, the power situation with those linear regulators was bit of a mess, as you guys are all up to date on that story as I see in the links. The good news is were going the switcher route on all the new boards in dev, with a separate LDO for the ADC power plane. That should clean a lot of stuff up. For now, all the new boards are coming with the silkscreen changed to indicate just a 5V input.

Posted 3 years ago #
gbulmer
Moderator

poslathian - maybe LeafLabs could make a small 'switcher' power supply board so existing users of Maple could upgrade to something which will handle 7-18V?

Posted 3 years ago #
peax
Member

Got my new regulator today, however as it turns out the regulator was not the problem, or at least it wasn't the only culprit.

So far, looking at the schematics found at github I have removed IC3 (the VDDA regulator), the neighbouring C13 capacitor and also the tiny C5 up by the processor P$12/P$13 and I still have a short from VDDA to GND, or actually it is a 3.0 ohm resistance.

Does anyone have any tips on what else I can remove / check? I'm currently out of ideas, and I haven't really got a clue as to what components usually fries from too much voltage and pretty much short the circuit.

Posted 3 years ago #
cmartin
Member

It looks like the only things left connected to VDDA is the CPU and J4.

Did you try the new regulator? The CPU may show such a resistance reading, especially if it's reverse polarity. If your meter has it try the diode test setting and put the positive lead on VDDA and negative on GND. It should show open circuit.

You might want to check what voltage it's showing from VDDA to Gnd. If it's a diode drop then you are probably looking for a diode or damaged IC.

Also when you say 3.0 ohms I take it that means as measured with an ohmmeter across the terminals. The meter leads can easily have an ohm or two. If your meter has a zero function then short the leads, zero the meter and then read the resistance. Otherwise short the leads, take the reading, put the leads on the terminals, take a reading and subtract the first reading.

When I've had a very low resistance short on a board I've used the following technique:
Connect a current source (or a voltage source with some resistance) to the circuit. You want enough current flowing to give a good voltage reading but not high enough to damage something.
Using a voltmeter, start at the terminals where the power comes in and note the voltage.
Now leaving one meter lead in place start probing around with the other lead. You are looking for a lower voltage. Once you find the lowest reading start moving the other probe around the same way (leaving the first probe connected to the lowest voltage point).
When you are done hopefully the probes will both be on the same component and that will be the source of your problem.

The one problem with this method is you may need to pump too much current through the circuit. You need enough to develop some voltage drop in the traces.

It's "possible" things got too hot and some of the PCB get carbonized. But it's not common.

Posted 3 years ago #

RSS feed for this topic

Reply

You must log in to post.