Arduino Micro Regulator Defect?

[adafruit2]

YES! if you have some 1-2 ohm resistors do that, like this:

5V ----/\/\/\/----|(------GND

see if that helps?

[kengineer]

YES! if you have some 1-2 ohm resistors do that, like this:

5V ----/\/\/\/----|(------GND

see if that helps?

Thank you for all your help and enlightenment :) I guess I have some reading to do with regards to this ESR of output capacitors topic.

[liudr]

I have arduino mega 2560 using the W34 7CE50Z. It shows ripples after some thermal stress of using 12V external power. I don't know if it is this regulator or selector circuit but I'm not using USB, just 12V. I used a 12V 2A ac adapter to power it. With ethernet shield and sd card, the arduino mega start showing weird behaviors after some time. It recovers after I disconnect it from power and allow it to cool down. If I reset it by depressing reset, the ripple on the 5V is gone. If I let go of reset, ripples come back after a few minutes. If I don't have the ethernet shield (I have 3 such shields from 2 suppliers they all show same thing), then there is less current consumption (60mA compared with 250mA with ethernet shield and sd card). Here are two screen shots (I don't have an adapter to get the screen image and the screen has some damage):

Ripples shown with DC coupling and bottom of screen is 0V:


No ripples immediately after reset:


My other arduino mega 2560 has C||3rC 17-50 marking (the C is backwards and the vertical bars are two sections). I have not tested this one yet but it's for a different project with somewhat less power consumption. Other mega that I'm using successfully are clones with various regulators that most of them work just fine.

Adafruit,

What's the status of the talk with Arduino LLC? I believe my client purchases arduino MEGA from you and sparkfun so this defective one is likely yours. This one doesn't have white silk screen printed on the side of the headers.

[adafruit2]

i think your issue is different, your regulator isnt unstable, its just noisy. The fact that it goes away 'after reset' just points to it being an overcurrent operating problem
Honestly, I would not power a Mega+ethernetshield+whatever with 12VDC, I'd use 9VDC as there's no switching regulator, and those 3V drop has to be burned off.

if your customer purchased it from us, they have to contact us with their order details

[liudr]

i think your issue is different, your regulator isnt unstable, its just noisy. The fact that it goes away 'after reset' just points to it being an overcurrent operating problem
Honestly, I would not power a Mega+ethernetshield+whatever with 12VDC, I'd use 9VDC as there's no switching regulator, and those 3V drop has to be burned off.

if your customer purchased it from us, they have to contact us with their order details

I didn't attach an oscillatory screen shot I got and I didn't zoom in to this one. How do you tell it's unstable but not noisy or vice versa? According to most spec sheets, voltage stays within 10mV. Yes, it is a different issue but maybe Arduino LLC should specify more than range of voltage for the DC barrel or just go with switching regulators. Two other mega boards I monitored had not problem with your "overcurrent" statement. One of them has a larger regulator so I assume it dissipates heat better. The other one has SOT223 and works just fine. I have car batteries charged with solar panels. They are 12V and sometimes a bit more. I need 12V to turn on pumps. I can't just go for a 9V adapter. I don't have the spec sheet for this particular regulator but I can easily calculate power dissipation from a typical TI 1117-5.0 regulator with the copper trace area it should handle the 250mA current over 7V it needs to dissipate. This regulator is not. I wish I had spec sheet for it.

I'm just going to bypass the regulator with a switching 5V supply and end the trouble. If you have the power to make changes on your shop pages, add some comments with arduino boards like this one below, say recommend 9V supply and watch for overheating if your load is to big, give example such as Ethernet shield with sd card and 12V supply cause many people do projects with car batteries.

https://www.adafruit.com/product/191

[adafruit2]

i think your issue is different, your regulator isnt unstable, its just noisy. The fact that it goes away 'after reset' just points to it being an overcurrent operating problem
Honestly, I would not power a Mega+ethernetshield+whatever with 12VDC, I'd use 9VDC as there's no switching regulator, and those 3V drop has to be burned off.

if your customer purchased it from us, they have to contact us with their order details

I didn't attach an oscillatory screen shot I got and I didn't zoom in to this one. How do you tell it's unstable but not noisy or vice versa? According to most spec sheets, voltage stays within 10mV.

unstable power supplies oscillate wildly, like 1-2VPP, with low frequencies, and at zero-load. it's usually due to mis-spec of in/out capacitors.

Yes, it is a different issue but maybe Arduino LLC should specify more than range of voltage for the DC barrel or just go with switching regulators. Two other mega boards I monitored had not problem with your "overcurrent" statement. One of them has a larger regulator so I assume it dissipates heat better. The other one has SOT223 and works just fine. I have car batteries charged with solar panels. They are 12V and sometimes a bit more. I need 12V to turn on pumps. I can't just go for a 9V adapter. I don't have the spec sheet for this particular regulator but I can easily calculate power dissipation from a typical TI 1117-5.0 regulator with the copper trace area it should handle the 250mA current over 7V it needs to dissipate. This regulator is not. I wish I had spec sheet for it.

Picking the right power supply setup is *really hard* and depends completely on your setup. You may be operating at a 'margin' of functionality, some Arduinos happen to work, some dont! its hard to debug :) Post up a new thread (this one is for the Micro regulator questions ken & i were working on) and we'll take a look at your schematic/setup and maybe have some suggestions.

[cgram]

I can confirm the LDO related problem on the arduino micro. I recently ordered a batch of 15 to update an older project and all 15 have the same 17-50 regulator. I haven't tried all of them but the ones I have used have similar intermittent buzzing issues depending on how they are hooked up. I have an older model with an ON regulator that does not have the problem. One additional bit of information is that there was also a board revision that seems to have happened at the same time as the component switch. The SCK silkscreen location moved on the board. I haven't had a chance to chase down specific differences between the boards otherwise but at a minimum they made a silkscreen change also. So the issue may be tied to other changes that happened at the same time as the regulator. I also removed the two power switching mosfets (T1/T2) and jumpered the source and drain pads where T2 was as I did not need the power switching functionality and I thought (before finding this thread) that they could be the source of the noise. The buzzing problem still happened after this modification on the boards with the 17-50 regulators.

[liudr]

It's been a month. I'm working on a different project and this time is the new arduino wireless sd card shield. The shield has an on board 3.3V regulator with backwards C ||3mF 17-33 marking. Having an xbee module and sd card on the shield causes the regulator to buzz or ring audibly and more than 50% time it either outputs nothing or has a large swing with +- 250mV range. My xbee radio will not work.

Adafruit, has arduino team got back to you regarding this regulator issue?

[adafruit]

the arduino team received our emails and let the factory maker of arduinos know they did not give us any further follow ups(s) or information.

==========
Thanks for the heads up!

Katia can you contact SmartProjects and report this issue?

Katia will reply with SP's answer, unfortunately this is a
manufacturing issue and we have no direct control over it
==========

[adafruit2]

Since we have no reply from Arduino, we can suggest either
using an external regulator such as
https://www.adafruit.com/products/2166
or
https://www.adafruit.com/products/2165

or

try adding 100uF capacitance on the output of the 3.3V regulator, such as
https://www.adafruit.com/product/2193

[kengineer]

Since we have no reply from Arduino, we can suggest either
using an external regulator such as
https://www.adafruit.com/products/2166
or
https://www.adafruit.com/products/2165

or

try adding 100uF capacitance on the output of the 3.3V regulator, such as
https://www.adafruit.com/product/2193

47µF put me back in to the stability region but now I am using 100µF just to have a margin of safety.

No problems since then.

[liudr]

I was thinking of replacing the regulator with a working one. Anyway, is 100uF or even 47uF really necessary? I'll give it a try. I only have 10uF caps at hand. I'll get a big one tomorrow. With 10uF caps the usb port is already complaining with the transient current when I connect my boards. Should I add a serial resistor to deal with this? Thanks.

[kengineer]

I was thinking of replacing the regulator with a working one. Anyway, is 100uF or even 47uF really necessary? I'll give it a try. I only have 10uF caps at hand. I'll get a big one tomorrow. With 10uF caps the usb port is already complaining with the transient current when I connect my boards. Should I add a serial resistor to deal with this? Thanks.

Adding a series resistor will do nothing (perhaps even make it worse) because you will be connecting it in parallel with two other ceramics. You will only decrease the ESR even more unless you remove the on board ceramics. It is better to connect a high capacitance to get in to the stable region of the ESR Vs Capacitance curve.

Check the attached graph. Reference: http://www.ti.com/lit/an/slva115/slva115.pdf

[adafruit]

here is the response we received from arduino today 12/17/2014

it seems that the issue has already been solved and was related to a specific batch of production. The customers can get in touch with the distributor whe they bough them and they can have the substitution of al the boards without any additional charges.

[liudr]

I ended up replacing the regulator with a 1117 regulator. Took about 15 minutes cause I was trying to save the bad one for more studies but couldn't. The arduino team should really move the regulator away from xbee socket. It's too close. Guess they don't replace bad regulators themselves.