How can this resistor be 1/2W ?

[alank2]

I ordered a SFR16S0004700JA500 resistor which says it is a 1/2W, but when it arrived, it was much smaller than expected.  The size of a 1/8W resistor!

What do you guys think?  Would you trust this for 1/2W ?

https://www.digikey.com/en/products/detail/vishay-beyschlag-draloric-bc-components/SFR16S0004700JA500/7351757?s=N4IgTCBcDaIM4DMBOBGAbHADNgLAdmwCsBDAVhAF0BfIA

[tooki]

Half a watt isn’t a ton of heat, and either way most of it is going to escape via the legs, not contact with air. As long as they selected the materials to handle the slightly higher operating temperature, which I’m sure they did, I don’t see why it’d be a problem.

[tooki]

P.S. Also, take this order as a lesson that not all through-hole resistors are the same size! The standard size for a 0.5/0.6W resistor is 0207.

[sleemanj]

Half a watt doesn't mean it can handle half a watt without potentially getting very very hot, just that it's not going to fail and will remain (given specified tolerances and temperature coefficient) the correct value.

The datasheet indicates that at rated power, a temperature rise over ambient of at least 90 degrees C would be seen.  At half a watt in a typical ambient temperature environment, it's going to be well over 100 degrees.

[bobbydazzler]

Datasheet says 155c max temp, I think with thick leads it could dissipate that.  Solder it to some pcb and put 15.3v through it and see if the case scorches or it melts its solder.

[tkamiya]

Yup.  I have one like that from Vishay.  I contacted DigiKey support stating exactly what you said.  The person called back saying she verified with technician that it IS indeed rated 1/2 watts.  It doesn't look like it to me but that's the official statement.

I was going to test it by applying 1/2 watts to see if it will tolerate it but forgot about it.

[Vovk_Z]

The size of a 1/8W resistor!
What do you guys think?  Would you trust this for 1/2W ?

Yes, I have a bunch of 0.6W resistors with a size of about 1/8W or less. Rated power depends on the materials used, how temperature-stable and fire-proof they are.

[tkamiya]

Datasheet says it can withstand 1/2 watts.  I would trust Vishay.  But at 155C max temp though, installation and material near by needs to be examined.  Regular fiber glass PCB material will deteriorate fairly quickly.  Pads on PCB has to be large to dissipate the excess heat.  Put 10 in closely packed formation and it will dissipate 5 watts.  That's pretty significant power.  Considering some installation might also involve in envelope that doesn't allow for air circulation, I'd probably de-rate it depending on particulars of usage.

On ceramic substrate or aluminum based PCB, probably it'll be just fine.

I would certainly not use it close to 85C electrolytic caps.... 

[bobbydazzler]

I made a similar post a few weeks ago about some 1w Vishay resistors as well, my 1w resistors could actually dissipate 1w without scorching or melting its own solder.
https://www.eevblog.com/forum/beginners/1w-resistor-in-a-0-25w-resistor-footprint-how/msg3544799/#msg3544799

[floobydust]

Welcome to the modern ripoff in resistor specifications.  I got screwed that way with Stackpole, Yageo, Panasonic resistors as well.
1/8W is the new 1/2W it seems, all are literally rated 2-4X what they used be rated years ago.
Parts are good to 155°C which is stupidly unrealistic as the rest of the board burns up.

SFR16S "Rated dissipation 0.5W P₇₀" The graph shows a temperature rise of 85°C at 0.5W, so 25°C ambient it's going to run at 100°C.
Larger SFR25 rated less at 0.4W part, graph shows a temperature rise of 80°C at 0.4W, so with a 25°C room it's going to run at 105°C.
Miraculous SFR25H (extra 0.1W in the same package, same dia. leads) somehow runs 20°C cooler, again defying the Laws of Physics.

Physical size matters as far as practical power dissipation, there are many "mini" power resistors being marketed that again are due to better materials that can run hotter- but FR-4 PC boards cannot last at high temps.
I was shopping for 2W or 3W resistors and the size differences are just crazy between manufacturers and models.

[CaptDon]

We design for 20 year cycle life and those resistors just won't work at that power level in my 20 year world!!! I don't want ANY part on my boards to rise more than 35C above ambient!!! We test in our chambers at 85C ambient which means my worse offenders could still hit 115C. I have found that most ethernet ferrite transformers die at around 105C as they seem to saturate or something. Also the SMPS magnetics get wonky at about 110C. If I am going to burn around .5 watts in an SMD resistor it better be 0805 at least.

[Kleinstein]

For power resistors it is normal that they can get quite hot. I remember some wire wounds that said it would reach up to 300 C and needs extra care (e.g. longer leads) not to melt the solder joint.

The power rating of the resistors is not all - choose by size and thermal resistance, not so much the power rating.
I also got hit by this one - a 1 W resistor in what used to be 0.4 W size.


Another issue is with SMD parts - the power rating depends on the copper area needed and this can be quite substantial. With just thin lines the useful power rating is way lower.

[Siwastaja]

This is engineering, you need to know what you are doing, if not, don't throw a tantrum, admit your lack of skill and learn. (This comment not being directed to the OP as much, but to those describing this as a "ripoff" what it clearly isn't.)

ALWAYS when you are dissipating non-negligible amount of heat, do thermal analysis, this is true for every component type in existence, for example you can't pick a MOSFET based on maximum drain current specification, not even some derated version of that, instead you calculate RthJ-A, you need to know max ambient temperature, then calculate max power dissipation, or calculate what heatsinking you need.

Similarly, this is just power rating for the resistor. I'm sure it holds. It doesn't have derating in it; datasheet has a derating curve.

They obviously make sure to use materials that can take the higher temperature due to less surface area available to dissipate the same power.

The difference in surface area is larger than the difference in cooling capability because a lot of the heat escapes through the leads into your circuit; this part is unchanged assuming lead thickness is unchanged. So now the resistor is putting slightly more heat through the leads but the difference is likely less than you think.

The issue is heating up the surroundings. FR4 is slowly scorched, traces lift, solder joints become brittle. Yet, you can find metric shitton of devices that have failed exactly that way even when using Good Old Resistors with larger packages per power rating; in fact, people discussed them being "ripoffs", instead of fixing their designs. So it doesn't matter; the circuit designers failed thermal design then, they will fail it now, too. Except those who know how to design. Nothing really changed, resistors were good and still are.

Vishay can't know what you are going to do with it. It's not a claim that you can use it in any (or most) applications at that power level; it's a description of that resistor alone, not in a circuit. You are not a consumer, you are a designer. The parts are for professionals. Suck it up and stop wasting Digikey's time when it's clear their parametrization matches with the datasheet. Ask for Vishay for datasheet clarifications.

[floobydust]

These are "miniature" through-hole resistors that several manufacturers offer nowadays. They are basically 1/2 to 1/4 the size package but with an unrealistic power rating. So a 1/4W part is the size of 1/6W, 1/2W the size of 1/4W etc. and it extends well into 3W power resistors as well. They are mixed in with the "normal" sized parts on the datasheet that have the same power rating. Examples, general purpose metal film with normal and mini sizes:
Yageo MFR, MFRS
Stackpole RNF, RNMF
KOA Speer MF, MFS

The mini parts run much hotter to achieve the same power spec, which I call a ripoff as it's hidden in the datasheets or not given at all (see Yageo datasheet, no thermal data).
P₇₀ "Power Rating at 70°C" {ambient} Vishay SFR16S rated 0.5W runs at 155°C or 310°F  Hurrah you've got a high temp epoxy nobody would use the part that hot, thanks for the completely useless spec.
For those people that sort by a column (Power) and wrongly use that as a figure of merit, it's a trap.

First glance at the rabbit hole of "thermal engineering", a smart person would notice the temperature-rise graphs never contain information about the test - was the resistor mounted to a PC board, how much copper etc. or was it in free air and with what lead configuration, axial, radial, steel or copper leads etc? This makes the graphs misleading. IEC 60115 $350 to buy it, yet parts are tested to only a few clauses anyhow. You can't really do a proper thermal design.

For the noobs, in school it was taught to always derate a resistor by 50% of rated power but that's no longer a rule of thumb nowadays. I'm not sure what the Art of Electronics advises.
Resistor temperatures for long life, low drift I'm not sure are taught anywhere.

[SilverSolder]

These days, anyone can use a thermal camera to look for things getting too hot -  and then fix them, and check that the fix worked?

Even if you are not a good thermal designer, this approach should work...

[dmills]

You ALWAYS got to read the datasheet (And remember that the real thing starts with the first page of graphs, up to there it is often contaminated to the point of uselessness by marketing).

I can entirely see use cases for a 155c rated resistor, and it generally does NOT involve FR4 boards, if Vishay had called this thing a 1/8W, high temperature resistor specified to 1/8thW @ 125c, there would be nobody commenting, but extend the curve downward to the point that something else becomes likely to be the failure (The board) and suddenly it is a problem?   

[TimFox]

Given their choice, the marketing department will always choose the highest rating that can be truthfully printed in the data sheet.  The obvious example of this is the headline power rating of power transistors.  I learned my lesson in 1969 when I ran an Ohmite hollow-core wire wound resistor (on the supplied stand-off brackets) at its full rated 50 W power and burned myself accidentally. 

[vk6zgo]

These days, anyone can use a thermal camera to look for things getting too hot -  and then fix them, and check that the fix worked?

Even if you are not a good thermal designer, this approach should work...

Not everybody has a thermal camera, even now.

Usually, EEs blithely design boards with high dissipation resistors snugly sitting against the PCB, & "call it good!"
Manufacturers love such boards as they are easier to stuff with components----- besides, they look pretty!

After a few years, the resistors connection to the board fails, a long suffering Tech comes along, removes the wreckage, repairs the board, fits new resistors standing off with longer leads.(& before someone has conniptions about lead length & inductance, these problems normally occur on low frequency PSU circuitry)

It no longer looks " pretty", but will probably finish its required multi-decade life span with no further problems.

[Siwastaja]

Prototyping and lab testing works as a good substitute when you lack input data and/or skills of doing the analysis.

Lack of thermal camera can be compensated with a drop of ethanol/IPA or water to see if it boils.

Ghetto-style accelerated life testing is easy. Lifetime roughly doubles/halves when temperature changes by 10degC. So want to see what happens aften 1000hours at 40degC ambient? Run the prototype at 80degC ambient for 1000/2^4 = 63 hours. Not accurate but if it works in such conditions it won't be a disaster.

Good old instruction of derating resistor power by 50% works usually quite well, I'm sure you are equally fine if you derate the miniature resistors by 55%.

The resistors are just better and can take the temperature.

Power is still dissipated mostly through leads, even with the larger parts, so it's as it always was; the heatsinking you provide (through PCB tracks/fills) is the critical factor, and the resistor manufacturer isn't very helpful here (never was).

Many old failed devices show that the resistors itself are fine but the solder joints are brittle and the tracks lifted near the joints because the resistor didn't dissipate the heat but brought it to the PCB through the leads.

Heat transfer by radiation has exponent of four for the absolute temperature, I wouldn't be surprised if the increased resistor surface temperature completely makes up for the decreased area. I wouldn't be surprised if the consequence is lower lead temperature and thus smaller risk of solder joints failing. Who knows, do a test if interested.

If you dissipate 0.5W, most of it is going to end up in your PCBA in any case, partially through leads partially through radiation and convection from the resistor body. Smaller and hotter resistor does NOT mean more heat into the PCB, no, the heat is the same, namely the power you dissipate. It all ends up in your hands.

Through leads, it causes solder embrittlement and FR4 scorching near pads; through body heat transfer it leads to FR4 scorching under the resistor. Both are bad but the former usually more destructive for the circuit.

Now when the resistor designer plays with the resistor surface temperature, what they do is they change the ratio of power transfer through leads to power transfer through body. No one can offhand say in which direction this ratio changes with said modern types, and it completely depends on the circuit if it helps or hinders.

Hope this helps.

[Siwastaja]

Given their choice, the marketing department will always choose the highest rating that can be truthfully printed in the data sheet.  The obvious example of this is the headline power rating of power transistors.  I learned my lesson in 1969 when I ran an Ohmite hollow-core wire wound resistor (on the supplied stand-off brackets) at its full rated 50 W power and burned myself accidentally.

Yes, but the point is, any engineer have learned this early on, like you in 1969 or me in 1990's. It is surprising to me this comes as a surprise to anyone here, excluding the actual beginners asking for help, of course.

Obviously, transistor max power and max current ratings are mostly useless to engineers, yet they are still printed year after year, decade after decade, despite the fact the target audience doesn't look at said ratings. The question whether the numbers are there by marketing is interesting but in the end, it doesn't matter.

Engineering is difficult and requires a lot of knowledge around little details, but the power and current ratings by component manufacturers is some of the most widely known and easiest to work with.

[floobydust]

OP's resistor SFR16S is the only one in the world rated 0.5W in so tiny a package. Blame engineering? I say no, this is marketing wank taken too far.

I compared over 50 different TH resistors to figure out what is going on. Because a long time ago, making PCB footprints for a "1/2W" resistor became a nightmare because they vary in size from 3.5-10mm long- due to their specs being all over the place. Heat dissipation is largely related to surface area, so a 1/2W part is typically ~10mm long.

P₇₀ is rated power dissipation in a 70°C ambient, not sure of the exact test conditions i.e. mounting.
SFR16S max is 155°C and this works back to their "1/2W" rating, so the part is at its max. 155°C in a 70°C ambient for their spec.
Other manufacturers i.e. Stackpole RNMF14 same size, same 155°C max - the P₇₀ for the part is 1/4W. KOA Speer MFS1/4 is 1/4W as well. The 3.5mm package used to be called 1/8W or 1/6W back when marketing was out of the picture.

How does Vishay claim twice the power? The part comes with a little man inside waving a bullshit fan? Datasheet typo causes shitstorm?

Engineers deciding the (power) derating factor, or the manufacturer giving a number - it's a mess now. It could be antique IEC/DIN standards being exploited- but other manufacturers are giving a lower power spec yet same size, even those with higher 175°C max rating.
I found resistors that are "professional" verses "commercial" as yet another marketing wank term to confuse people.
Vishay also gave two applications for a part: "Power" and "Reliability" with mystical derating numbers out of their hat. You can do all the thermal design and still need to know the magic number of what operating temperature to use.

I think we've reached a point where component manufacturers now give absolute maximum values to compete with those shopping based a parameter column at Mouser or Digi-Key. There was a time when manufacturers gave practical maximums for their parts.

[tooki]

Welcome to the modern ripoff in resistor specifications.  I got screwed that way with Stackpole, Yageo, Panasonic resistors as well.
1/8W is the new 1/2W it seems, all are literally rated 2-4X what they used be rated years ago.
Parts are good to 155°C which is stupidly unrealistic as the rest of the board burns up.

SFR16S "Rated dissipation 0.5W P₇₀" The graph shows a temperature rise of 85°C at 0.5W, so 25°C ambient it's going to run at 100°C.
Larger SFR25 rated less at 0.4W part, graph shows a temperature rise of 80°C at 0.4W, so with a 25°C room it's going to run at 105°C.
Miraculous SFR25H (extra 0.1W in the same package, same dia. leads) somehow runs 20°C cooler, again defying the Laws of Physics.

No, it’s not defying the laws of physics: right on the datasheet, they clearly state the thermal resistance of each model. The SFR16 and SFR25H both have significantly lower thermal resistance than the SFR25, meaning that the heat created can escape more easily. I don’t know exactly what they do differently to achieve this, but I imagine it could be things like a different material for the body (normally they’re ceramic, which isn’t a great thermal conductor), different metal in the end caps, or even different paint that’s less of a thermal insulator.

Assuming the SFR16 and SFR25H are made of the same materials, the larger surface area of the SFR25H explains why it can run coolest of all.

P.S. 25°C+85°C=110°C

[madires]

OP's resistor SFR16S is the only one in the world rated 0.5W in so tiny a package. Blame engineering? I say no, this is marketing wank taken too far.

Nope, they are just adopting Chinese Watts.

[floobydust]

I don't see any outrageous claims by anyone else for OP's 3.3mm part, china uses skinny all steel leads (use a magnet to verify) which makes them run hot because of the higher thermal resistance conducting heat to the PC board.

[Siwastaja]

How does Vishay claim twice the power? The part comes with a little man inside waving a bullshit fan? Datasheet typo causes shitstorm?

The fact you don't understand the reason does not mean it's a ripoff or a conspiracy.

I'm not an expert in resistor design and manufacturing or componnet thermal analysis, but I can offhand say a few possibilities how a certain sized resistor can dissipate twice the power compared to the competitors;
* Temperature handling of materials
* P_radiated ~ T^4
* Thermal conductivity of leads (material, thickness)
* Bonding of the leads
* and so on and so on

Do you understand that by claiming this can't be true, you are the one to prove such claim.

Really, resistors dissipating power, power having to go somewhere, and temperature being related to both power dissipation and dissipating surface area, these are fundamental basics any designer needs to know, always have been. I'm glad miniature high-temperature resistors exist, you have a choice! I might use them in some specific applications. It's not Vishay's fault they come up in Digikey's parametric search when you don't want them. BTW Digikey has Size/Dimension parameters so why not use that. I admit though the list is long because diameter and length are not separate parameters as they should be.

Vishay also gave two applications for a part: "Power" and "Reliability" with mystical derating numbers out of their hat.

This isn't nice but isn't unheard of; tantalum capacitors for example have always (or for a long time) rated for a completely imaginary operating voltage which basically results in some tens of hours of operating life. Recommended derating numbers are supplied in often separate documents (typically Vrated * 60%).

There was a time when manufacturers gave practical maximums for their parts.

Wat, really? Was there? I'm sceptical. Maybe you are just old and grumpy and feel like everything was better in the past.

I think I have seen the opposite. Especially in discrete semiconductors, go back 20 years and you often only have absolute maximums. Today we see helpful "recommended" sections, it's more common today even to see a realistic or even pessimistic Id rating for a MOSFET, rated at high-ish ambient temperature and no heatsinking assumed, so a "100A" part may have an alternate 10A rating available (and you can get like 40-50A out of it if you know how).