6
\$\begingroup\$

The resistor in question is R10 (523K 1% 1/4W) in the circuit below:

Schematic

The resistor is reported to fail as open circuit and there is no visible damage. We have shipped some boards and got some damaged ones back with R10 damaged. So this is not a single instance.

The AC input is 220V. The resistor specification from the manufacturer is:

  • 1/4 W
  • Max working voltage: 300V
  • Max overload voltage: 500V
  • Max dielectric withstanding voltage: 500V

Since R10 withstands an AC signal, the Vrms it sees is less than 300V. The wattage it withstands is also less than 1/4 W. But R10 on a few boards just failed and they are now open circuit/huge resistance and there is no burn mark on the outside.

We did a few tests, trying to amplify the problem. We overload one board with 253V AC (230V*1.1) and let a blower heat up the board. We also put a more-than-usual load on the board. The board did not fail after several hours of continuous test.

I have no clue on how R10 would die. Any pointers are greatly appreciated. Is there any way to examine the dead bodies of R10 to determine the cause?

The excerpt of the resistor (Metal Film type) is shown below (I do not have the full datasheet): enter image description here

The measurement on R10 and GND (the voltage drop across R12 is less than 3.3V) is shown below (the peak is about 320V):

enter image description here

\$\endgroup\$
9
  • \$\begingroup\$ What is the resistor datasheet? What is the critical resistance value for that type? \$\endgroup\$
    – markrages
    May 17, 2014 at 2:31
  • \$\begingroup\$ If you can figure out how to do it safely, what does a scope show for voltage differential on the resistor? Also some pictures of this might help. How are you checking continuity on bad resistors? In or out of circuit, etc? \$\endgroup\$ May 17, 2014 at 2:36
  • \$\begingroup\$ @markrages I have edited the post to add the "partial" specification of the resistor. I do not have the complete datasheet either... \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 2:50
  • \$\begingroup\$ @ErikFriesen I have added the waveform at R10 in the post. The bad resistor was taken out of the circuit to measure its resistance. \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 2:57
  • \$\begingroup\$ When I was working on projects involving measuring mains, the standard approach was to chain a number of smaller value resistors. use 4 or 5 smaller resistors in series. As a bonus, you get better creepage, which is critical if you're doing this with SMT components. \$\endgroup\$ May 17, 2014 at 8:03

5 Answers 5

8
\$\begingroup\$

It is my experience (failure analysis) that 1% metal film resistors are not reliable long term with high voltage DC on them.

I've also seen huge changes in value without them complete going open.

You might try to fit a MOF (metal oxide film- also known as metal glaze) type if you have room. I recommend Vishay (BC Components nee Philips) VR25 resistors for this service if you care about reliability. They're tested at 100VDC, but have a "rating" of 1600VDC.

enter image description here

There's a hint here that moisture (probably plus elements of the lacquer) play a role, so if you want to recreate the failures I'd suggest putting them in your environmental chamber at elevated temperature and 95% RH and see what happens. Obviously, you want to avoid condensing humidity when power is applied or bad things (tm) could happen.

\$\endgroup\$
7
  • \$\begingroup\$ The Vrms seen by R10 should be well within its max working voltage. Or should I just look at the peak voltage? The peak voltage in the usual worst case would be 230V*1.1*1.414=358V. I am totally confused. \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 3:01
  • \$\begingroup\$ Yes, all this is within the rated working voltage. It's DC that does it- maybe plating off of the film due to contaminants. A common failure mode of SMPS (startup resistor). \$\endgroup\$ May 17, 2014 at 3:13
  • \$\begingroup\$ @SpheroPefhany You mean this is a quality issue as the quality of the resistors batch may not be consistent? \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 3:22
  • \$\begingroup\$ Working voltage is defined in terms of DC, but I suspect that consistent application of this voltage for a few milliseconds, at a 100 or 120 Hz rate for prolonged periods of time, may well count as exceeding the working voltage. Since the polarity is always the same, it makes sense to me that in the long term it would behave "almost as good as" DC. \$\endgroup\$ May 17, 2014 at 3:56
  • 5
    \$\begingroup\$ @SpehroPefhany After quite a few experiments, the resistor in question indeed has some quality issues. High humidity plus 220VAC input will increase its resistance slowly. Probably the coating is bad. We applied the same test to a resistor from another manufacturer and it survived. We tested a batch of resistors from both manufacturers and the results are consistent. Thank you for the directions. \$\endgroup\$
    – Brian Wang
    Jun 11, 2014 at 6:50
7
\$\begingroup\$

R10 is seeing almost all of the 220V input full wave rectified (there is a voltage divider with R12 but R10 is more than 100 times larger). The peak voltage across R10 is thus given by 220*1.414 which is 311 volts. This is larger than the specified working voltage of 300 volts. It is quite possible over time that this overvoltage is causing the resistor to fail. Since it is not due to overtemperature but overvoltage, the failure will not be visible on the outside. Most likely, there is internal failure of the insulation. In any event it would be advisable to get a resistor with a higher voltage rating.

\$\endgroup\$
6
  • \$\begingroup\$ Does it matter if this peak voltage (> 300V) is a DC or AC signal? I thought the max working voltage is specified as Vrms/Vdc. \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 3:14
  • 2
    \$\begingroup\$ I would interpret the maximum working voltage as the maximum voltage allowed across the resistor independent of time. Thus AC or DC is irrelevant especially since we are talking about peak voltage. RMS is an average so that a waveform could have a low RMS value but a very high peak value (for example, a signal composed of narrow pulses). The RMS voltage is more appropriate for checking the power being dissipated in the resistor. The peak voltage is appropriate for checking against a maximum voltage specification. \$\endgroup\$
    – Barry
    May 17, 2014 at 3:39
  • 1
    \$\begingroup\$ The old trick is to use several resistors in series, instead of single small high-impedance resistors. 1/4W resistors are not that expensive. Also, I don't think 1% tolerance is necessary in this application. \$\endgroup\$
    – markrages
    May 17, 2014 at 3:45
  • \$\begingroup\$ @Barry Should we look at the max overload voltage instead of the max working voltage for the peak voltage? The max overload voltage is 500V, which is well above the peak voltage seen here. \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 3:49
  • 1
    \$\begingroup\$ Adding to this answer: mains power can sometimes exhibit nasty transient pulses, e.g. from switching off a heavy inductive load (just a vacuum cleaner may do the trick). These might exceed 500V, violating the overload specs for the resistor as well. \$\endgroup\$
    – marcelm
    Sep 4, 2017 at 10:32
2
\$\begingroup\$

The resistor in question is not protected against lightning strikes to live and neutral relative to earth. This could be the problem I.e. both L and N get raised to some absurdly high transient voltage whilst earth remains pretty much steady at ground potential.

\$\endgroup\$
3
  • \$\begingroup\$ Do you have any suggestions on how to protect the board as a whole from lightning strikes? Is the MOV, R5, at the mains input too slow for that? \$\endgroup\$
    – Brian Wang
    May 17, 2014 at 14:28
  • \$\begingroup\$ Given that the diode bridge is also subject to the same potential voltage spike and if that has no known failures I'd consider upping the spec of the resistor to exceed the voltage ratings of 600Vpk, 420Vrms. Other than that I'd consider putting X or Y rated capacitors on hot and neutral to earth after the common mode choke, maybe 47nF or greater. Another idea is to derive your sync pulse with a capacitor dropper from either live or neutral to earth. Another idea is to dispense with the earth connection completely - does it need to connect to earth? \$\endgroup\$
    – Andy aka
    May 17, 2014 at 14:59
  • \$\begingroup\$ Thank you for the advise. I will talk to our hardware engineer to see if they can be done. The high voltage side of the board is not connected to the earth, I think. \$\endgroup\$
    – Brian Wang
    May 18, 2014 at 10:35
0
\$\begingroup\$

I sell thousands of small dusk to dawn sensors which have a 100R 1/2W metal film resistor in the AC circuit. I used to get returns with exactly the same symptoms, with the resistors appearing normal upon inspection.

The way I overcame this problem was to use a 1/2W carbon film resistor. Problem fixed!

\$\endgroup\$
0
\$\begingroup\$

I have found 1/4 W & 1/2 W metal film and carbon film resistors going open circuit for no apparent reason in circuits where the voltage drop is nowhere near its design voltage. Example a high value 220K bias resistor in Hartley oscillator circuits in Conn electronic organs. To begin with the resistor goes high in value till eventually in will go open circuit over time. While this is happening the oscillator will drift in frequency, actually going higher as the bias current becomes less due to the resistor drifting high in value. It seems that current flow in the resistor causes this over a great amount of time. It appears to happen mostly in high value resistors > 47K. Microscope examination with the paint removed would have to be done to ascertain the actual cause but it would have to be a gradual decrease of conductivity somewhere in the spiral carbon film if not all of it. If the resistor is subjected to a fairly high voltage drop as in this case and Switched Mode Power Supply startup resistors then it appears to happen sooner in the life of said supply. Maybe the answer is to design the power supply with a much greater specification resistor than the circuit actually requires, eg say a 1W high voltage resistor and use a cracked carbon resistor instead of metal or carbon film

\$\endgroup\$
1
  • \$\begingroup\$ I have a few high value resistors, metal & carbon film in my possession removed from power supply's and other circuits. Maybe it's time to study them under a microscope to see what the cause actually is to make them go open circuit without showing any sign of damage on the outside of the resistor. One thing about the curious phenomenon it is well known to experienced service men of a certain age and is an east fix. \$\endgroup\$
    – user251283
    Apr 30, 2022 at 12:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.