Voltage divider failure damage

Hi,

I was just looking a solder job I did on a voltage divider that is pretty craptacular (supposed to be through hole, but I put the resistor in on the wrong side of the board and couldn't get the resistor leg out of the hole, so it turned from through hole to smd) , and I was wondering: if the solder joint fails, won't that damage the ic, since the circuit is possibly exposed to higher voltage?

Do engineers plan for this possibility by adding by protection of some sort or since the problem would be in board, just let the ic fail?

"I was wondering: if the solder joint fails, won't that damage the ic, since the circuit is possibly exposed to higher voltage? " Yes, especially if the lower ground joint fails.

"Do engineers plan for this possibility " Joints are inspected via visual means (human eye or camera), or electrical when not visible, to make sure they are good before power up. Parts may be added if it doesn't interfere with signal performance.

But it's not only for voltage dividers... You have the same problem if a Vcc or GND pin of a micro fails. If the Vcc fails and you phantom power it via a pin that can damage the pin.

Or voltage regulators, cut the GND lead and you probably end up with the input voltage on it's output...

“You have the same problem if a Vcc or GND pin of a micro fails.”
That only happens if you have a bad part - or if the design overstresses the pin, like all these HC595 designs that drive LEDs with insufficient current limiting and put way more than 70mA thru the power or Gnd pin. I am really surprised more parts haven’t failed.

Qdeathstar: Hi,

I was just looking a solder job I did on a voltage divider that is pretty craptacular (supposed to be through hole, but I put the resistor in on the wrong side of the board and couldn't get the resistor leg out of the hole, so it turned from through hole to smd) , and I was wondering: if the solder joint fails, won't that damage the ic, since the circuit is possibly exposed to higher voltage?

Do engineers plan for this possibility by adding by protection of some sort or since the problem would be in board, just let the ic fail?

Not sure what you mean by solder joint failing. You probably need to explain what you mean by it.... like.... if the solder joint fails..... then what will destroy the IC? Eg a loose resistor leg getting shorted to nearby high voltage source etc?

If so.... then the problem could maybe be rectified if the original mistake is removed. That is, get the resistor mounted where it should be mounted - on the correct side of the board, which means getting the resistor leg out (which couldn't be removed before).

If what you did isn't life threatening, or won't cause any significant impact if the solder joint fails, then just go with it and leave it as it is.

It makes my eye twitch to think about it, but: FMEA

@southpark

|500x190

If R2 comes loose, the ic will be destroyed, unless there is additional protection. Could you replace a voltage divider with a mosfet or opto-isolator?

Sure, but what if that fails?
And it won’t work if you want to read the analog value.

Qdeathstar: @southpark

|500x190

If R2 comes loose, the ic will be destroyed, unless there is additional protection. Could you replace a voltage divider with a mosfet or opto-isolator?

Hi there QDS, maybe some kind of over-voltage protection. Just depends what sort of voltages or currents are involved. Like.... zener diodes across the input.... and if a current flows through the zener(s) due to over-voltage, then a current sensing resistor (in series with the zeners) could turn on a transistor, and then shut down the supply etc. Kind of crowbar protection.

Boards are typically designed with the assumption that the board is correctly assembled, without bad solder joints.

A bad solder joint on the Vcc or Gnd pin of a micro exposed to external voltages could easily trash the micro; a floating sense pin could cause a switching regulator to output a destructive voltage, and so on. Adjacent pins shorted could cause all manner of havoc. And there are always single points of failure.

It is common to design for EXTERNAL faults - issues on other connected boards, external devices, etc. But not internal ones, unless that condition is expected to be common.

Most boards are not even designed with the assumption that repair will be attempted if the board is damaged in operation - the assumption is that they will be discarded, because anyone skilled enough to perform diagnosis and repair on PCBs would need to be paid so much that repair would not be economical. Normally the whole board is replaced.

Qdeathstar: @southpark

|500x190

If R2 comes loose, the ic will be destroyed, unless there is additional protection. Could you replace a voltage divider with a mosfet or opto-isolator?

I don't believe you need any additional protection. If R2 comes loose, then R1 will conduct. The Vin side will be at 12V (or whatever) but the current is limited by the resistor. Since there's a protection diode between the Arduino pin and the 5V rail, the Vout will be 5.7V. What current will flow to make that voltage drop? It depends on the resistor.

With appropriately high values for R1 and R2, say in the 5K-10K range, the fault current will not be high enough to overheat the protection diode. The pin can be in this state forever with no damage. You need to do the calculations for your own voltage and select your own resistors. Use 2mA as the fault current for your calculation.

Indeed what MorganS says, if you pick proper resistors the current is limited. But a solder joint doesn't randomly go bad. So just solder it right, check it, recheck it and you're fine.

@CrossRoads, don't need a bad part or a design problem. For example, if the GND pin of a micro is lose and you have a button to GND connected to a IO pin, if you press the button you phantom power the micro. On it's own it will probably survive but if the micro now starts to sink 50mA of current on other IO pins (combined) you are for out of spec... And I wouldn't be surprised if it will break then.

Do engineers plan for this possibility by adding by protection of some sort or since the problem would be in board, just let the ic fail?

Where I work, the production test calls for bringing-up the voltage slowly while measuring.* (But the truth is, we usually only do that for the 1st board of a new design, and sometimes for the 1st board in a production run.)

But, one thing we do consistently is keep the supply voltage "reasonable". For example, we have a few boards with an on-board regulator and a jumper to bypass the regulator. These are 5V boards, and we test them at 6-7V. That allows us to confirm that the regulator is functioning, and if someone screws-up and the jumper is in the wrong position or if the voltage regulator is wrong/bad, 7V won't kill any of the chips on the board.

Of course, if the regulator fails in the field (I've never seen that happen) the board can be fried. (I have seen boards that the customer has over-voltaged.)

We do small production runs (100 boards or less) and we do repair (or attempt to repair) the failed boards. Usually a failure is caused by a soldering problem (open or short) and it's rare that a soldering error blows a chip (or other part). But, we can change bad/blown parts. Components are super-reliable, and the bare boards are tested before assembly so about the only problem we can have is an assembly error.

  • We don't have any designs where a bad voltage divider would kill a chip, but we have lots of voltage regulators.

That voltage divider.... it's not for powering an IC I gather, but is for conditioning an input signal to be of a suitable voltage level presented to an input of the IC for sampling or something?

no, it was to see if i have 12v in or if i'm running off of battery. I have to 12v divided to 3v. I was mainly asking just to know i have a lot of good info now :)

AFAIK you can't power ic's from voltage dividers.

Qdeathstar: no, it was to see if i have 12v in or if i'm running off of battery. I have to 12v divided to 3v. I was mainly asking just to know i have a lot of good info now :)

AFAIK you can't power ic's from voltage dividers.

I see! Thanks!

It is possible. It just depends on how much power can be delivered to the IC without going over power handling rating levels for any component. It should be ok. Just need to do the usual and put a energy reservoir - like a suitable sized and type capacitor next to the IC power pin.