Had a problem with power, I had the 5v line tied to a PSU. Inside the case there were a couple of fans tied to 12v + and 5v for the ground rail to slow the fan speed down.
I suspect that this blew the arduino. I've since gone back and changed my power to 12v and removed the 5v line that caused my issues.
Here is whats going on (I'm not asking for project troubleshooting this is just relevant info), when I power on my project, all the fans come on, the water pump is spinning, life is good. If I attempt to plug my wiring harness in from the Arduino the PSU immediately goes into protection mode and shuts off.
I have a thermistor on a voltage divider tied to a analog pin, it is not reading at all. My digital temp sensor is still working (this is when I leave the PSU off and plug the harness in and power Arduino from USB).
I have printed the raw analog output and I get 1023 on all 16 channels, regardless of the input. Input is 1.6v (tested with meter).
Another symptom, when plugged into USB I am back feeding power into Vin (4v to be precise). The 5v rail is at 4.7 and 3.3 is dead on.
I have already replaced the voltage regulator as it was d.e.d.
Is there another simple component that has gone out that I am missing? Or does it matter, is the chip itself toasted?
Maybe half of the components are damaged. Would you risk that there is a damaged component that you missed and it will cause a lot of trouble in a few weeks ? It is best to replace the board.
It is possible that the ATmega328P chip (used in Arduino Uno) has a damaged analog section and the rest is still working. But with negative voltages, probably more is damaged.
When powered with USB only, a voltage of 4.7V is normal. For myself I say 4.5V is the minimum.
Some voltage at VIN is normal, since current is going backward through the voltage regulator.
Inside a computer, it is allowed to connect a fan to 12V and 5V. As long as more current is used from the 5V than the fan pushes into it, the computer is not damaged (maybe).
For a project, you should never mess around with the ground. The ground is what everything is based on. A fan from 12V to 5V could easily lift the 5V voltage. The power supply might also be damaged.
Going over my steps from the last couple of days, I suspect that the board was blown while bench testing with a bench top PSU, not when installed in the project. There was never any negative voltage applied. I believe that my bench top was sitting around 5.4v during my testing. I messed up and didn't read the datasheet prior or I would have kept it at or below 5.0, as there is a possibility that it spiked up to or over 6v as a load was removed.
I wasn't aware that putting a fan between the 12 and 5 could potentially raise the level of my 5v rail! I understand how important ground is I never did anything to mess with it, unless it was inadvertent and I am not seeing what I did?
I already have a new board on the way as I want to get the project finished up. I was just hoping for some insight into the operation and components that have gone bad, I have not had a chance to sit down with the schematics yet, it was late when I found the issue and my time during the day is used up by the paying job
I am going to repair the board and hopefully use on another project. Even if every component on the board had to be replaced, it would still be cheaper than buying an entirely new board!
I didn't keep a scope lead or a meter on the 5v leg of my bench top. I really really hope it didn't spike above 8! If so I'm going to blow up some awfully expensive boards when troubleshooting.... I'd better test the unit!
The reason I think the arduino blew up prior to installation - Before I ever attempted to power up my entire project, I applied power to the arduino with USB to ensure the correct program was installed. Once the program was loaded I did not have the analog readings like I was expecting.
I'm even considering using a voltage regulator to smooth the voltage to get around an 8v input from the 12v I am going to use now, just in case the fans on the 5v are causing any issues, rather be safe than sorry!
I'm not sure how you have connected things and how it is powered. Perhaps you could make a drawing or schematic of your wiring; tell what Arduino board you have (official or fake clone ?); give links to the PSU and other things you use; make a photo of it.
There are so many things, that I can think of. Here are a few:
Some power supplies require a minimum load.
Applying 12V to an Arduino board is only possible if the Arduino doesn't have to supply any power to leds or relays.
Connecting 5V to the Arduino 5V pin to power it can be done, but a protection diode is needed, and current flowing into the computer can damage things.
Sorry if it's a bit cramped, I threw this together real fast from my other drawings. There are quite possibly mistakes in the drawing that I missed.
I don't have the model of the PSU on hand but will post it tonight. If I remember right 3.3 required load is 0.5amp, 5v is 2amp and 12v was 1amp. The PSU is a IstarUSA 1000w (TC-1000 something).
The voltage regulator on the Arduino is new. When all the issues happened it was 5v hooked the 5v on the PSU
Four MAX4420 mosfet drivers is all the load for the 5V. When they are idle, the current for a MAX4420 is 0.045mA. But since the MAX4420 will work up to 18V, you won't notice that the 5V is not stable.
The MAX4420 is all about fast switching and supplying a peak current for the gate of a mosfet. I hope you calculated if they can be used to drive fans as well.
Using a 8V voltage regulator is good and safe.
When I power an Arduino with 5.0V to its 5V pin, then I use a power supply that can do perhaps 1A, and I add a protection diode from 5V to VIN.
You should not apply that massive 55A 5V to the Arduino 5V pin. It might blow the voltage regulator (reverse current) and it might damage the Arduino and the computer as soon as the usb cable is plugged in.
Is the 15uH okay for the Arduino default PWM frequency, or are you going to change the PWM frequency ?
If that circuit is never powered by batteries, then I would turn the peltier on/off once per minute. Then the PWM is not needed and the circuit with the coil and capacitor is not needed.
Is there a datasheet for those mosfets ? I can't find it.
I apologize for not putting it in the schematics (I knew I had missed something....) There are two load resistors, one for the 3.3v, 10watt 6Ω, and one for the 5v, 15watt 2.5Ω. There was a third, 50watt 8Ω on the 12v line, but the exhaust fans and pump never turn off while the unit is running and should provide more than adequate load for the 12v so I removed it.
I bet you are on the right track as far as the voltage. I'm betting that I had a 5v issue that cooked the board, I still think that it was probably the benchtop test supply I was using, no clue how or why, but I'm sure that it's completely my fault. Hence the reason I am moving over to Vin off of a voltage regulator. I figure that the voltage regulator will help to smooth any spike my hard switching on the 12v line will cause, and if any translate through, they should still be within acceptable limits and the Arduino will smooth them farther with it's regulator. I can learn from my mistakes
I didn't check to see if the 4420's would be ok driving fans. I've looked at the scope and they send a nice clean signal out, but I'm not sure how to calculate if they are safe for PWM control on fans. Where would I go to find those calculations or data?
Does it have sense wires for the 5V ? or are there only sense wires for the 3.3V.
The input capacitance of those mosfets is only 5nF. That is a low value, that's good. The MAX4420 will probably not get hot.
MAX4420: about 500mW (depends on package), typical 1.5 ohm, peak 6A.
The output pushed current into a gate and also pulls the gate to gnd.
If I'm right then: I = sqrt ( P / R ) : I = 0.58 amps
Because of the high output impedance of 1.5 ohm, the continuous current can be only 0.58 A.
I'm ignoring what happens on the rising and falling slopes, that's too complicated for me.
MAX4420 for mosfet: The MAX4420 is powered with 5V, so the ouput is 5V or 0V. That is okay for the mosfet.
MAX4420 for the fans: That is also powered with 5V, but the fans are 12V. That is not allowed. Power the MAX4420 with 12V, and the output is maximum 0.58A.
Only fans that are made to work with PWM can be controlled with PWM. Some fans have a capacitor inside and a PWM to the power wires is not okay, others have a seperate pwm wire (a CPU fan in a computer).
Ah, I see where you are coming from now. The fans on the driver are PWM fans, I should have specified, they are designed to accept a 5v signal (the same that your computer puts out for the control line).
I used to have some IRFZ44N MOSFETS which were not logic level, but the ringing from the inductor blew the driver. I added a snubber circuit, but the ringing is still high enough that I just went ahead and moved the entire switching circuit to 5v rather than complicate the circuit anymore or spend a ton of time on the snubber.
Not sure about the sense wires. Aren't the control loops are usually built into the PSU? They read the voltage right at the output on the circuit board and adjust from there? There is one 20ga 12v line and one 20ga 5v line. Edit: and a 3.3 20ga. If those are the sense lines, do I need to hook those up to anything in particular?
Again, thanks a ton man, I can't tell you how helpful it is to have someone looking over this for me.
Edit: Ayup, doing some reading and digging. Apparently that PSU needs each sense line hooked up, AND I will need to balance the loads. Crap. I'm going to need more resistors or a new Haswell ready PSU....
Thank you for saving me from my own idiocy! So going back through my wiring I saw that I had the 'little' 20ga 'power' wires hooked up as... well power. Nope. Not gonna happen. Gonna blow stuff up every time. They ARE sense wires... you know what happens when you have a sense wire hooked up as a load? The PSU tries to turn itself up until it hits over voltage protection and shuts off..... I could have sworn that I had tested those with a meter and saw voltage on them.....
Crap, I was SURE that the sense was internal to the board. There I go, assuming things. No clue why that didn't trigger a warning when I saw those itty bitty wires on a 1000w PSU.
Not 100% sure that was the cause of my arduino popping though. The 5v sense line was what I had tried to use as power in. le sigh Man do I ever feel stupid right now.
So, hooked sense lines up as.. well, sense lines. Tied the Arduino to an actual input voltage, and lo and behold, it no longer shuts the PSU off! Still no analog signals, definitely fried something there, but definitive progress.
I should have a new board here tomorrow and I can try, try again. Thank you sir! +karma added!
I have a computer power supply for high current 5V and 12V. I tested a few, and it turned out that a very old one (without sense wires) had an output of 5V even without load. I only had to connect the green "turn on" wire. Perhaps older is better
Is the PWM a seperate wire to the fan ? Then the output from the Arduino can be used. Perhaps with a protection resistor of 150 ohm to 1k. An extra buffer with the MAX4420 at 5V is okay of course.
The newer Haswell ready PSU's are designed that way. The 3.3v and 5v are regulated from the 12v, so the only regulation is on the 12v.
I fired up my older PSU with around 20 amps last night after I hooked the sense lines back up correctly. The 12v was rock steady even with the high load on 12v and low load on the 5v. Hopefully it'll be alright and I won't need to balance the 5v to match the 12v. I'm expecting around 50 amps 12v. I guess only one way to find out
The fan is a four wire, - + RMP PWM. The reason I put a driver in there is I ran out of room in my harness, I didn't even run the RMP lines back. So I am running three fans off of one PWM. Granted the intel spec says less than 5mA per fan, but I didn't want to fight the switching noise getting back into the power on the Arduino. This completely isolates them. I can always just put a decoupling cap between pin 9 and ground to see if that will smooth it instead of the driver.
I never did anything to mess with it, unless it was inadvertent and I am not seeing what I did?