Fake… resistors

This note is a small laboratory work on measuring the “fine” parameters of the most common blue metal film resistors with a spread of 1%, type MF-25. The peculiarity of these resistors is that they are fakes from China. Unfortunately, now this fake “metal film” is sold everywhere, so if not in the nearest radio store, then on marketplaces you will most likely buy it. Surely this is no longer a secret to many, but many are still blissfully unaware, thinking that “all the parts are made in China and bottled from the same barrel.”

Problem

Now, in the era of the triumphant surface mounting, output components are gradually becoming exotic, and this is especially true for low-power resistors and ceramic capacitors. Even radio amateurs now use them more often in order to play with a simple circuit, sticking parts into breadboard or soldered on a breadboard. After all, the use of chip components sharply reduces the number of holes in the boards, which is important when you make the board yourself, and their parasitic parameters – capacitance and inductance – are noticeably better than those of output ones. But there are cases when it is necessary to use them: composite thick-film chip resistors suffer from prohibitively high levels of excess noise. Using them in the first stages of sensitive amplifiers often means seriously ruining their characteristics. Thin-film SMD resistors, as well as metal-film resistors in a leadless MELF package, which are free of this drawback, are practically not available in retail sales. At the same time, the most common metal film outputs have not gone away from sale yet; it is not difficult to buy them in any radio parts store.

In one of my recent designs, I was faced with the fact that the output signal noticeably “floated”, reacting unexpectedly strongly to warming up. My search led me to a voltage divider, which, for the reasons stated above (and also due to the high voltage), was mounted from lead metal film resistances. The instability was many times higher than calculated, based on the reasonable TCR values ​​of these resistors, especially since it had to be compensated, leaving the resistance ratio unchanged. Then I remembered that I took one of the resistors in this divider from old supplies, and the other from a recently purchased set – 2600 pieces from 1 Ohm to 10 MOhm. Aha – I thought, this is it.

I didn't like them from the very beginning. There was something wrong, something bad about them. And the conclusions are somehow not the same – thinner and more rigid, and the markings are somehow crooked… I rummaged through old stocks and finally found the required denomination there, having soldered it out of “donor”and voila! – everything immediately became good.

So, we have another Chinese fake in our hands. And now let’s explain this Chinese “owl”.

Accuracy

The first thing that comes to mind is to measure several resistances of various values. To do this, we will use the GDS-8246 desktop multimeter, connecting the resistors directly to the clamps inserted into its sockets in order to avoid affecting the accuracy of the probe resistance measurements. Its measurement accuracy at 5, 50 and 500 kOhm is 0.1% plus 2 least significant units. The accuracy is tested on standard resistance wires of class 0.05 and corresponds to the class of the device. At the limits of 500 Ohms and 5 MOhms the error is higher, 0.2%, I had nothing to check it on, so I chose the ratings of the components being measured in order to avoid these modes. I randomly selected 5 resistors of different values. The results are in the table.

Table 1. Set resistor measurements. Bold obvious “departures” for admission are shown, italics – borderline cases.

In general, the resistors “almost” fit within the error, but there are also obvious deviations from the tolerance, sometimes reaching 3%, except for one of the kilo-ohm resistors, with an error of as much as 20% (obviously, this is just a 1.2 kOhm resistor, mistakenly labeled , like 1K). Of course, it doesn't have to be that way. Fail.

For comparison, here are the resistances of ten such resistors on one “machine-gun strip”, forgotten for many years in a characteristically bent corner and sealed with a stapler chip-and-dip plastic bag: 10023, 10028, 10031, 10021, 10029, 10014, 10024, 10030, 10018, 10022 Ohm. PI found this racket among the inheritance that I inherited from the previous owner of my desk at work, and judging by the price of 5 rubles indicated on the faded sticker, they were bought back in the early 2000s.

Thermal stability

Metal film resistors have a positive TCR (temperature coefficient of resistance), their resistance increases with temperature, like any metal. Its typical value for MF-25 and MFR-25 with an accuracy of 1% is no more than 100ppm, that is, when the temperature changes from room temperature to 100°C, the resistance can increase by 0.75%.

Let's set up the experiment as follows. Using thermal paste, we will stick the resistor to the plate of the “bottom heater” for soldering heat-intensive circuit boards, equipped with a thermostat, we will place the sensor of an electronic thermometer nearby, and we will cover the whole thing with mineral wool for thermal insulation. Let's turn on the heater and begin to gradually raise the temperature, measuring the resistance of the resistor as we go. The result of the experiment is on the graph.

As you can see, here the resistance decreases with temperature. TCR near room temperature exceeds 500ppm, and in the temperature range 25-104°C the resistance decreased by 3.4%. So it's anything but metal film resistors. Perhaps metal oxide – but in general, such a TCR value is typical for the most ordinary carbon resistors of low quality.

Crash test

What should happen if 16 V is applied to a 1 kOhm resistor with a rated power of 0.25 W? In theory, nothing. It must withstand operation at the maximum permissible power for at least 1000 hours (more precisely, you need to look at the manufacturer’s specifications), without getting burned and not changing the resistance more than specified in the approval.

What happened in reality? It darkened noticeably to the eye after just a couple of hours of execution. I left it for a day and measured the resistance after it cooled down. Up to: 993 Ohms. After: 1013 Ohms. So, the resistance increased by 2% and left the tolerance zone after 24 hours of operation at rated power! And it doesn’t look good, like it’s burnt…

A similar real metal film resistor passed the test without changing either its appearance or resistance by more than one in the fourth digit during the same time.

Verdict: here too there is a complete discrepancy with the stated parameters.

Conclusion

It would also be possible to conduct experiments to measure the noise level, conduct a chemical analysis of the resistive film, etc. But it’s already clear: these resistors have nothing to do with what is stated. It is strongly not recommended for purchase and use.

And here are the characteristic features of such details. The first is sloppy markings. The trait is not constant; on some specimens the color rings are quite even and clear, but there are curves, of different widths, displaced, unpainted or peeling. The second is conclusions. They are thinner, and at the same time rigid, with slight elasticity. The presence of rust on some of them indicates that they are made of tinned iron wire. Magnetic, but magnetism is never an indicator of low-quality radio components, including resistors (otherwise there is such a common myth). For completely branded electronic components, the terminals can be made of kovar and also have magnetic properties.

All these signs cannot always be seen. You may be lucky with the quality of the markings; the terminals may not be rusty, but the terminals of high-quality resistors may also become magnetic. A randomly selected specimen will most likely fall within the resistance tolerance. But there is a simple, reliable and express test. As we remember, the fake is carbon resistance with negative TCR. And the original is a metal film with a positive one. Let's connect the resistor to a digital multimeter in ohmmeter mode and heat it up – with a soldering iron, a hot air gun, or even a lighter. The counterfeit resistance is immediately noticeable will fall. But a real metal film resistor is only barely noticeable will increase resistance. This check can be done even right in the store or pickup point.