Failed resistors

[Diver300]

I've been repairing some of these:- https://www.redlion.net/product/225-6-digit-led-counter and they have the quite neat feature that the supply voltage is 22 - 240 V, AC or DC.

They use these:- https://www.power.com/sites/default/files/product-docs/top242.250.pdf for the flyback controller. In order to work over such a wide voltage range, the built-in shunt regulator in the flyback controller isn't used, and there is an external shunt regulator with a buffer transistor instead.

The current limiter is made of three 0805 resistors, each of 392 kOhm. I assume that three resistors are used to keep the voltage across each one within the rating for the resisotors. The power is very small at around 30 mW when on 240 V ac. The resistors appeared to be precision ones, and I can't think of any reason for that.

On all of the units that had failed, one of the three resistors had failed open circuit. There was no visible fault. The photo was taken after I had replaced the resistor.

Has anyone got any idea what caused the resistors to fail?

 

[DrG]

What else, apart from heat, would be expected? [asking, not telling]
edit: or is is assumed that it is heat and then it is, why did they get so hot? Again, just asking.

 

[Diver300]

I don't think that it can be heat, because there are three 392 kOhm resistors in series, and the maximum voltage across them is rectified 240 V. I think that comes out at about 33 mW.

It could be that the resistors cracked, but I can't see why those resistors and not other on the board were affected.

The voltage is quite large, at around 110 V max for each resistor. The design uses three resistors in series, and uses 0805 resistors rather than 0603, and looking at this data sheet:- https://www.vishay.com/docs/28705/mcx0x0xpro.pdf that agrees with the specification. It's not clear why they used three 0603 resistors, when two 1206 resistors or five 0603 resistors , but the voltage seems to be what defines how many resistors are needed. The operating voltage is quite close to the rated voltage. It is only way that I can see that the resistor is being taken anywhere near it's maximum rating

 

[Nigel Goodwin]

The resistors are is series for three reasons:

1) To increase their wattage.
2) To increase their voltage rating.
3) Cost - it's cheaper to use three small resistors rather than one larger one.

They reason they fail is quite simple - they VERY commonly do when used in this way - it's a VERY, VERY common design flaw, and the designers (as they don't do repairs) and manufacturers (who obviously ignore the warranty claims and all the many reports from their service agents), seem oblivious to the fact.

A very common service technique is to look for resistors in series or parallel, in high stress areas - you'll often easily find the fault in that way.

Don't over think it, it happens, it's INCREDIBLY common (and has been for decades) - if you can replace the entire string with a single higher wattage, higher voltage resistor - as should have been done in the first place, it will never fail again.

 

[Diver300]

In this application the three resistors seem to be to increase the voltage rating.

I agree that a single resistor would be a neater solution, but it would probably need a through-hole one.

Can you explain what the failure mechanism is? I know that the resistors are run near their voltage rating. In this case 110 V when they are rated at 150 V, but does having three in series make it any more likely that any one will fail than if each one was supplied at 110 V.

Or is this just a case of components being less reliable when near their voltage rating? When the designer choses how many to put in series, each extra one has a cost, so they end up near their voltage rating. Anywhere else in the circuit, the maximum is 24 V or so, and the resistors are far below their rating.

 

[Nigel Goodwin]

In this application the three resistors seem to be to increase the voltage rating.

I agree that a single resistor would be a neater solution, but it would probably need a through-hole one.

Can you explain what the failure mechanism is? I know that the resistors are run near their voltage rating. In this case 110 V when they are rated at 150 V, but does having three in series make it any more likely that any one will fail than if each one was supplied at 110 V.

No idea - but as I said it's a VERY bad idea to use multiple resistors, and they WILL fail.

 

[kubeek]

They reason they fail is quite simple - they VERY commonly do when used in this way - it's a VERY, VERY common design flaw, and the designers (as they don't do repairs) and manufacturers (who obviously ignore the warranty claims and all the many reports from their service agents), seem oblivious to the fact.

Why is that a problem?

 

[Diver300]

Why is that a problem?

Well the three modules that I repaired yesterday, by replacing three 0805 resistors on each one, would cost over £1000 to replace

 

[Nigel Goodwin]

Why is that a problem?

Because equipment would be vastly more reliable if designed properly.

However, to be fair, the designer 'may' have specified a single larger resistor, and it 'could' have been altered during production to save money. But I think it's more likely the designer is responsible, as they tend to be theoretical with little or no practical experience.

 

[kubeek]

I meant why three identical resistors in series to achieve higher wattage/lower temperature or higher voltage rating is a bad design?

 

[Nigel Goodwin]

I meant why three identical resistors in series to achieve higher wattage/lower temperature or higher voltage rating is a bad design?

Because, as in this case, they fail - it's a seriously common design flaw that causes failure in a vast range of equipment.

When I used to go on courses for the release of a new TV chassis, I'd always glance through the circuit diagram and circle the parts that you just knew were going to fail - I was rarely wrong

As for the failure 'mode', it might just be as simple as that the values commonly used are in an 'unreliable' range - historically 47K resistors in particular were a really common one to fail. As for the series/parallel ones the failure range was probably mostly between 10K and 500K. Grundig used a 12K and 18K in series (to give 30K), and just as expected they used to fail - replacing them with 1W resistors ensured they never failed again.

 

[gophert]

The smaller components get, the more closely they track the instantaneous input power rather than the RMS power.

 

[Diver300]

The smaller components get, the more closely they track the instantaneous input power rather than the RMS power.

That is true. However in this case the voltage is near DC. The resistors are fed from the rectified supply, and the instantaneous power is well below the maximum for the resistors.

 

[DrG]

I have been following this thread and want to raise a point as a possible reason for failure.

As has been stated, I don't doubt that this is a "known" problem in certain cases. I am curious as to understanding why it occurs.

I read a couple of items that got me thinking about the puzzle. The data sheet (as posted), this stackexchange and this **broken link removed**.

To make a long story short, the data sheet includes these two notations:

When the resistor dissipates power, a temperature rise above the ambient temperature occurs, dependent on the thermal resistance of the assembled resistor together with the printed circuit board. The rated dissipation applies only if the permitted film temperature is not exceeded.

These resistors do not feature a limited lifetime when operated within the permissible limits. However, resistance value drift increasing over operating time may result in exceeding a limit acceptable to the specific application, thereby establishing a functional lifetime.

For the first point, the film temperature is the upper operating temperature, so that is understandable.

But the second point is much less specific. If drift, were measured as a percentage of initial resistance (it may not be), the effect would not be the same comparing one R vs 3X R/3. Let's say, for the sake of argument only, a 10% drift. The case of a single R drops to 10584, where each of the R/3 drop to 353.

My thinking (which is admittedly not thorough) is that a "limit" could be exceeded in the case of one of the R/3, before the R. The result is consistent with what is observed - 3 X R/3 fails before 1 X R.

 

[shortbus=]

This is a UK product? Maybe the designer was someone who comes to online forums to get ideas?

 

[Nigel Goodwin]

This is a UK product? Maybe the designer was someone who comes to online forums to get ideas?

As I said above, it's a VERY common design error, and almost always causes problems with resistor failures.

 

[Diver300]

I have been following this thread and want to raise a point as a possible reason for failure.

As has been stated, I don't doubt that this is a "known" problem in certain cases. I am curious as to understanding why it occurs.

I read a couple of items that got me thinking about the puzzle. The data sheet (as posted), this stackexchange and this **broken link removed**.

To make a long story short, the data sheet includes these two notations:

When the resistor dissipates power, a temperature rise above the ambient temperature occurs, dependent on the thermal resistance of the assembled resistor together with the printed circuit board. The rated dissipation applies only if the permitted film temperature is not exceeded.

These resistors do not feature a limited lifetime when operated within the permissible limits. However, resistance value drift increasing over operating time may result in exceeding a limit acceptable to the specific application, thereby establishing a functional lifetime.

For the first point, the film temperature is the upper operating temperature, so that is understandable.

But the second point is much less specific. If drift, were measured as a percentage of initial resistance (it may not be), the effect would not be the same comparing one R vs 3X R/3. Let's say, for the sake of argument only, a 10% drift. The case of a single R drops to 10584, where each of the R/3 drop to 353.

My thinking (which is admittedly not thorough) is that a "limit" could be exceeded in the case of one of the R/3, before the R. The result is consistent with what is observed - 3 X R/3 fails before 1 X R.

The resistors are being run far below their power rating. They are also being run well below their voltage rating, and they appear to be precision resistors. The voltage across the resistors will depend on their values as they form a potential divider, with nothing connected to the mid points. There would have to be a huge variation in the resistances before they exceeded their voltage rating.

 

[Diver300]

No idea - but as I said it's a VERY bad idea to use multiple resistors, and they WILL fail.

Could it be a transient effect? If there is any capacitance from one of the mid points to switching parts of the circuit, the voltages would be uneven.

Would adding small capacitors in parallel with each resistor solve the problem? It is sort of like the way that resistors are put in parallel with capacitors that are in series to equalise the voltages.

It is all making it look like one larger resistor rated for the job is much more sensible. However, I curious as to what the root cause is.

 

[shortbus=]

As I said above, it's a VERY common design error, and almost always causes problems with resistor failures.

You didn't get my meaning.

 

[JimB]

You didn't get my meaning.

I guess you mean questions of the form:

"Do you believe that three series resistors are better than one for dropping high voltages?"

JimB