Honey, I Shrunk the Resistors

[bob p]

I just ordered a bunch of stock from Newark. wow, was I surprised by the variations in size of 1/2 watt metal film resistors.

Take a look at the photo. All 3 groups of resistors are 500 mW, 1% metal film resistors that came from different manufacturers.





LEFT Vishay/BC Components, SFR16S0 series. As I recall they're manufactured in Brazil and Thailand. They're diminutive in size. Being the size of a flea, its hard for me to imagine that these things truly have a 500 mW power rating, but here's the data sheet:

http://www.vishay.com/docs/28722/sfr16s25.pdf

MIDDLE Wellwyn, made in the UK. They are about the size of what I think of when I think of a 1/2 watt metal film resistor.

RIGHT MultiComp, made in China. They are HUGE, and oversized for what I think of as a 500 mW resistors. Their spec sheets are spotty. I only bought them when I couldn't find the required value from another supplier who provided more complete data sheets.


Can anyone provide feedback for me on the Vishay product that's located in the left side of the photo? Although things seem to look OK on the spec sheet (and the spec sheet does offer far more complete data than you'll find in most resistor spec sheets), their diminutive size worries me. Maybe I'm just old enough to be prejudiced to the idea that size matters more than it should. I know that metal film resistors are more volumentrically efficient than other types, and that manufacturing technology is making things smaller, but geez...

My gut feeling is that those little flea-sized resistors will get really hot, and that they will just burn up, or they will burn the board, or they will burn my fingers.

Does anyone have experience with them?

[Chuck H]

They're diminutive in size. Being the size of a flea

For starters I never want to see fleas that size.

And an entertaining pic. It's especially fun that they're all the standard light blue color so the differences in size seem more strange. Fun editorial.

I'm not sure how to interpret every thing on the data sheet but it looks like those little ones are 5% resistors. It also looks like they might be less graceful with pulses of power, get hotter and do have faster de-rate than the more ordinary sized resistors. The "noise" spec was different too but I can't say for better or worse as I didn't understand the nomenclature. Probably worse since HEAT would be part of the formula for "thermal noise"

I've bought a couple of different types of 1/2 watt MF resistors from Vishay. Big brown ones (about the size of your big ones in the photo but cylindrical, not peanut shaped like normal) and the normal sized blue ones. I probably wouldn't trust the little ones as plate resistors. But if Vishay makes a similar series in 1 watt rating it might be worth a look. Maybe not as good as other 1 watt resistors but probably better than any 1/2 watt resistor in the same size package.

Chuck

[bob p]

I thought it was interesting that the Vishay resistor data sheet offered a noise specification. The typical resistor data sheets tend to omit noise specifications. I interpret the omission of important data like that to indicate that the manufacturer isn't proud of the specification, so they purposefully omit it from the data sheet. The data sheet for the MultiComp product from China, for example, lists no noise speicfication. I think its obvious that if they're not listing their noise specification that they don't want to talk about it!

The Vishay data sheet is very complete, as far as resistor data sheets go.


Just to clarify, here's the part number for that little flea-sized resistor on the left, and how it decodes according to the data sheet:

Part # SFR16S0003010FR500

"SFR16S0" decodes to the smallest body size. Body length is 3.5 to 4.1 mm. Diameter 1.9mm. Even though the part looks like a flea, the specs say that it's the proper body size.

"0" means neutral variant
"0" means standard material
"3010" means 301 ohm
"F" means 1% tolerance
"R5" refers to packaging
"00" means standard part

It is indeed a 1% resistor. That's a brown band that indicates 1%.

The colors may be hard to see in the photo, but the five color bands are:
orange, black, brown, black, brown.





Here's a handy decoder for the 5-band codes:

Resistor color code calculator - 3, 4 and 5 band resistors

[dai h.]

might be from a different maker but I have a bunch of metal films I got surplus that look like the leftmost little ones (mine--on the ones that came with some original bits of packaging, IIRC were 1/6 or 1/4W). I've also seen resistors that appear similar (to those small ones) in my MXL603S condenser mic. On the miniaturized Rs, it seems to be typically the voltage rating is lower (data sheet above shows 200V max. vs. 350V for the two 0.5W listed).

(Chuck, those all look like 1%--check out the brown band on the ends.)

edit:

mine are apparently these (slightly different dimensions)

http://www.koaproducts.com/pdf/old/SN.pdf

[Enzo]

Don't read too much into missing specs. I don;t assume the lack of a noise spec means they are hiding somethiing. I assume they don;t want to have to characterize the part that way. If they do, then they have to have consistent noise behavior over time and stuff the engineers want to know for demanding applications. They are no longer free to change the processes in manufacture. If they just spec resistance and wattage, then they have more room to make changes as needed. These are general use parts, you want the noise characterized ones? Buy our advanced line at a penny or two more.

Imagine describing your PA system to some potential gig. You might tell them you never exceed 100db SPL on the dance floor, and your subs go down to 40Hz. You have a really good PA system, but you might not really want to have to chart out the freq response completely, and give them the radiation polar patterns of your speakers. And so on. Not that there is anything missing from your system in those areas, but it is more documentation than you care to provide. For instance you can't promise the rediation patterns you show will hold in ALL venues you might set up in. They want more details, then those would be spelled out in a rider and you'd be asking a higher price for services.

I been buying some resistors lately, and some brands show "1/4 watt rated to 1/2 watt." meaning essentially, here is our 1/4w, and we guarantee performance as a 1/4w resistor, but these also will work as 1/2w parts.


Did you look at voltage specs? The film around the ceramic core is cut in a spiral (helix?), and the space between "windings" is a determinant in voltage that can be placed across the resistor. That could affect size.

The part is on a certain size core, and it is spec'd at some resistance. There is no standard for how thick the film should be and how many turns around the core a give resistance should make. SO different makers can make the same resistor different ways.

Wattage is about how much current can flow before it burns out, but are the temeprature specs different for those types? All those sizez are probably fine at 1/4 watt, but how about stability over a wider or narrower range of temps?

[Chuck H]

+++ on the 1%. Thanks, I actually don't need a resistor color code chart. Just more patience and a pair of reading glasses

Chuck

[bob p]

All those sizez are probably fine at 1/4 watt, but how about stability over a wider or narrower range of temps?

I hope those sizes are all fine at 1/4 watt, as they're all 1/2 watt resistors.

My experience with the Chinese metal film resistors has been that they are exceptionally noisy in the higher resistance values. I try to avoid them unless no other part is available. I don't think its a coincidence that they don't publish noise specs.

[R.G.]

I'll post some more on this when I get a chance, but here are some things to chew on.

If you heat anything, its temperature (measure of thermal "voltage") rises until the temperature difference between the heat producing region and the surroundings causes enough power (thermal "current") to flow out of the device to equal the power input. The thing stabilizes at a temperature where heat out equals heat in.
You can make temps over 1000C with 300mW of power inside a type 6833 bulb rated for 5V/60ma. That's because the glass casing and gas filling it make it impossible for the heat to get out of the surface of the very fine wire filament at lower temps. A "1/4W" resistor will be sizzling hot but not glowing at 300mW, a 1/2W will be hot but not dangerously so at the same 300mW. What is different is the surface area to let heat out.

It's easy to get a higher power rating on a resistor: just make it out of stuff that will not burn up when it gets hotter. The resistor sizes we know and expect were based on phenolic casings filled with carbon granule goo. That material set limited the external and internal temps. So the outsides of the resistor had to be made big enough to have enough square inches of convecting/radiating surface to let the heat get out below the temperature which degraded the materials.

That being the case, the only thing which causes a 1/2W resistor to be a given size is how hot you let it get. Make it out of materials which can stand a higher temperature, you can rate it for a reasonable (maybe!) lifetime in a smaller physical size.

As a fr'instance, iron and copper do not "go bad" until several hundred C. Make the insulations right, and you can fry eggs on a transformer and the trannie will be perfectly fine. It's the insulation that breaks down. Same for resistors. Make the resistive element out of a metal film or wire which can go to a few hundred C and glaze/insulate it with high temp stuff, and it may brand you, but it won't fail.

Then there's the whole world of advertising to be considered.

[Chuck H]

What about thermal noise as it applies to the above considerations???

Sorry, but it just popped into my head upon reading. I should do my own research first, but hey... Just thinking out loud.

Chuck

[R.G.]

What about thermal noise as it applies to the above considerations???

Sorry, but it just popped into my head upon reading. I should do my own research first, but hey... Just thinking out loud.

That's a good point. Thermal noise goes up with temperature above absolute zero, and with resistance. So high temps make for more noise.

I was focussing on thermal rating. Noise goes up too. That may or may not be an issue depending on whether the 273 degrees above absolute zero of our normal ambient temperature plus the temperature rise of the resistor *and* the criticality or not of the resistor noise in the amplifier has any measurable effect on the amplifier's performance. Input resistors are **critical** to noise performance. Power resistors in the output stage are not. Input resistors often have microwatts of dissipation, so self heating may not be an issue. First-plate resistors may have some significant dissipation, so maybe the self heating could heat it up enough to be significant. Gotta do the math.

[bob p]

Thanks everyone for your feedback.

One thing that impressed me about that Vishay data sheet is that it has to be the most complete data sheet that I've ever seen for a resistor. It contains all sorts of information that you never see in resistor data sheets. (I wouldn't like to have been the engineer who got to put the data sheet together -- Yawn.)

As complete as the data sheet is, they only rated the noise specs as a function of resistance. They didn't rate thermal noise specs.

The noise specs did seem rather low.

One thing that is really growing on me is that resistor's diminutive size. I've got some SS amplifier circuits to work on that have very high parts density and relatively low voltages (compared to tubes), and the more compact size is a nice feature. I guess its a good thing that the boards are made out of G10/FR4.

[Enzo]

Well then the question is how much dissipation is that 1/4w resistor being ASKED to dissipate? A 1/8w resistor won;t get much hotter than a 10w resistor if nothing is flowing through it.

[bob p]

A 1/8w resistor won;t get much hotter than a 10w resistor if nothing is flowing through it.

you're absolutely right. i can honestly say that as those resistors sit on my desktop, they are indeed cool to the touch.

to me the only interesting question is how hot the resistors get at their rated dissipation. nothing below that really matters, does it? if they're being used at an operating point that's well below their dissipation rating, then the question isn't worth asking.

[R.G.]

The manufacturer's data sheet will actually include a chart of surface temperature versus power dissipation, or the inverse of that, a maximum temperature and a power derating chart.

In general, 200C is the max surface temp for many resistor families based on ceramic resistor forms like wirewound and metal film. I believe carbon film does this as well. I'd have to go back and look for the max surface temp on carbon comp. But getting a scar-tattoo from a resistor working normally at over 100C is a possibility if you're running it at max dissipation.

[Mark Hammer]

(Sorry to be so slow, Bob. Didn't check for incoming messages.)

In some respects, the package size of a component dictates not only heat dissipation, but space requirements, and inter-component spacing. If I have no plans to push the envelope with respect to power handling, then I have the luxury of using a smaller package that permits closer spacing. If I want to more effectively manage heat (and other things that accompany it, like thermal noise), then I might wish to use a package that forces me to space things further apart.

Keep in mind as well that contemporary robotic/automated assembly techniques would likely lend themselves to use of a single resistor size as much as possible. So, a "more-or-less-1/4w" in a 1/8W package size lets you get close enough to a 1/4W spec for rock-n-roll, but tool up for the 1/8W that's good enough for all the other needed resistors.

Incidentally, is it just me, or are the leads on resistors getting thinner these days? I'm not complaining, just commenting.