1/4 watt resistors... no go?

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

pdf64
Posts: 2688
Joined: Sat Mar 12, 2011 9:23 pm
Location: Staffordshire, UK

Re: 1/4 watt resistors... no go?

Post by pdf64 »

I seem to remember reading that one of the sources of noise within a resistor is at the junction between the resistor element and its end caps/leads; the higher the power ratings, the 'better' that junction and so the lower the noise generated there will tend to be.
User avatar
pompeiisneaks
Site Admin
Posts: 4222
Joined: Sat Jan 14, 2017 4:36 pm
Location: Washington State, USA
Contact:

Re: 1/4 watt resistors... no go?

Post by pompeiisneaks »

Stevem wrote:For the few times I do mods or builds with carbon comp types I always over watt them just for the sake of how it seems to help with noise hiss levels especially as these resistors age!

The heftier leads on 1 watt and up resistors also help with vibration / microphonic issues in loud combo amps !
I'm not using any carbon comp, all are carbon film, with like one or two that I couldn't find in carbon film as metal film.

Sent from my phone using Tapatalk

tUber Nerd!
User avatar
didit
Posts: 976
Joined: Thu Sep 25, 2008 3:37 pm
Location: Canada

Re: 1/4 watt resistors... no go?

Post by didit »

Stevem wrote: Fri Dec 15, 2017 11:38 am You do not want to apply more then 300 volts to a 1/4 watter on average!
Modest cost to resupply with some 1/2W or 1W recommended. Save the hassles and thinking too hard, and save those 1/4W for low voltage applications. For example, a source sufficient 1% MF with nominal 350V/700V ratings, and then relax and build & repair a bunch of amps:

--- http://www.justradios.com/reskits.html#kit6.

Best .. Ian
Mark
Posts: 2957
Joined: Sat Jan 15, 2005 8:10 am
Location: Sydney Australia

Re: 1/4 watt resistors... no go?

Post by Mark »

0.5 watt 100K plate resistors failing in Fender amps is pretty common, the reasons have been previously covered. Personally I wouldn't build using 0.5 watt plate resistors.
Yours Sincerely

Mark Abbott
User avatar
pompeiisneaks
Site Admin
Posts: 4222
Joined: Sat Jan 14, 2017 4:36 pm
Location: Washington State, USA
Contact:

Re: 1/4 watt resistors... no go?

Post by pompeiisneaks »

I don't think I recall mentioning any of them were plate resistors, I think I've got some RN65 1 watt 100k's for that purpose (not sure on the later 140 and 150 on the PI though, I'd have to check again). These are more for other random spots in the amp, and as mentioned it's a bit random on the ones I got. I just noted the 27k I ahve in the picture above is for the BIAS board, which should be in the sub 100V range and fine. (Also pretty low current i think no?)

Ordering more resistors isn't 'that' expensive, but for the shipping, especially when I just ordered a large order of that and other sundries now just a tiny order of 5$ in resistors with 4$ shipping is a royal PITA :P

~Phil
tUber Nerd!
User avatar
dreric
Posts: 962
Joined: Mon Dec 01, 2008 3:15 pm
Location: SF East Bay

Re: 1/4 watt resistors... no go?

Post by dreric »

pompeiisneaks wrote: Sat Dec 16, 2017 2:54 am
dreric wrote:I'm pretty sure that 1 watt resistors are a big part of the Dumble mojo. 183 is mostly 1/2 metal films.
You're thinking the higher watt resistors change tone? How?

Sent from my phone using Tapatalk
Having spent a fair amount of time looking at Dumble guts shots, my general recollection is that he leaned towards 1 watters. That said, I think there is much less thought put into his component choice than we tend to give him credit for. IMO, it was more based on what was available, surplus etc.

For my own builds I've always gone with 1 watt except for a Liverpool who awaits a rebuild cause it's so noisy.

Resistor noise:

http://www.aikenamps.com/index.php/resi ... -it-matter
Eric
1949 Zenith, Zenith Toggle Recoil, Zenith 55 & 440
pdf64
Posts: 2688
Joined: Sat Mar 12, 2011 9:23 pm
Location: Staffordshire, UK

Re: 1/4 watt resistors... no go?

Post by pdf64 »

Mark wrote: Sat Dec 16, 2017 10:17 pm 0.5 watt 100K plate resistors failing in Fender amps is pretty common, the reasons have been previously covered. Personally I wouldn't build using 0.5 watt plate resistors.
And the 1 watt CC plate resistors that tended to be used in SF amps rarely seem to cause problems.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 1/4 watt resistors... no go?

Post by martin manning »

Looking at a modern Ohmite CC spec sheet (for the now more PC "Little Demon" (R) series), the voltage rating goes up with the power rating, from 350V for 1/2W to 500V for 1W. If this was the case in earlier days, I suspect the 1/2 W CC plate resistor failures are likely to be driven by voltage. Data sheets for modern MF resistors show the same 350/700V voltage rating for 1/2 W and 1W, so typically there will be no extra voltage protection using 1W resistors. A typical 100K plate resistor is good for 1.58 mA at 1/4W dissipation, so a 1/2W is unlikely to be overstressed by dissipation.
R.G.
Posts: 1223
Joined: Tue Dec 02, 2014 9:01 pm

Re: 1/4 watt resistors... no go?

Post by R.G. »

Just to confuse things a bit :D it's easy enough to use two resistors in series to get even bigger power and voltage ratings. This actually probably goes up faster than waiting on high-power resistors.

I have not delved deeply enough to know, but I suspect that voltage rating is based on physical size above some minimum value. CCs are a different technology, but I think it may apply here too. Resistors are not homogenious chunks of proto-resistor atoms. They are conductive areas connected with connections. The film resistors come closest to homogenious, because you can keep thin films more consistent than chunks. But in any case, the inner structure is connecting areas and resisting areas. You don't want internal arc-overs starting, so the max voltage will likely be set by how physically far apart conductive areas are, compounded with the insulation stuff between them. The voltage across a resistor is ... V = I*R (!) so the voltage across the resistor is likely limited by current flow at lower resistances overheating the resistor by P = I*I*R and by arc-over at higher resistances. It's unlikely that you're going to find a packaged 10R resistor of any nominal wattage that will stand 300V across it. That's 9,000 watts. Things get easier for the 1M type. 300V on a 1M is only 90 mW.

So let's play with the calculator a bit. What power does 300V across 100K make? Easy - 900mW. So for good design practice, you'd use a 2-watter. >>If << you have 300V dc average across that resistor. A 1W resistor there would live, but it's surface temp would be near 150C (gotten by eyeballing resistor data sheet dissipation vs temp charts) and would live a short life because high temps speed up degradation in general for all processes except nuclear fusion. :shock:

Good practice usually says to keep the surface of resistors under 100C, and that translates back to keeping internal power under half the rated max. You can get longer life by derating further, down to nearly shelf life if you put enough dissipating ability in to keep the temperatures down.

Temperature also reduces the effectiveness of insulation, and speeds its decay and ultmate failure.

So - if you want long life, use very low dissipation. That relation applies up to the point where you're not dissipation limited, and start working into the voltage break-over region. In the voltage breakover region, you can get single bigger devices, skip to a different technology (i.e. MOX, CC, WW, bulk foil, etc.) or use more than one smaller device and divide the dissipation and voltage by Ohm's law. I'm a fan of using multiple 1/4W devices where needed because they're cheap, and it makes stocking easy. You build up power rating by putting resistors either in series or parallel, and voltage capability by putting them in series. The only problem with this approach is that you need some way to tell when to stop replicating, so you're back at having to know dissipation and the breakover point where the voltage rating is more important than the dissipation.

ACK!! Back to the spreadsheet. It's probably useful to have the spreadsheet calculate the voltage of 1/2 the rated dissipation for each of the EIA values, and the voltage producing full dissipation. This lets you know pretty quickly for any resistor if you were nearing its dissipation cutoff or its voltage limit.

I'm blathering on about this mythical spreadsheet. I suspect it's only fair if I go generate it. I'll try to get that running in the next little bit.,
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: 1/4 watt resistors... no go?

Post by martin manning »

Here's a shot of the spreadsheet I used to generate the log-log plots above, with the addition of an input for a particular value (ohms) to calculate limits for that resistance at several common power ratings, assuming dissipation at half of rated power.
You do not have the required permissions to view the files attached to this post.
R.G.
Posts: 1223
Joined: Tue Dec 02, 2014 9:01 pm

Re: 1/4 watt resistors... no go?

Post by R.G. »

Nice, Martin.

Here's another view of the same thing, just a different format and without the calculators. It helps me to have a diagram with colors (and blinky lights, etc. :lol: ) to just contemplate the relationships.
Image
R.G.
Posts: 1223
Joined: Tue Dec 02, 2014 9:01 pm

Re: 1/4 watt resistors... no go?

Post by R.G. »

The chart is the voltage that generates the power in the column heading for each resistance in the row. For instance, it takes 141.42Vdc to make a 10K resistor dissipate 2W.

I hid all the rows between decade values below 20K, because you hit dissipation limit before you hit the voltage limits. The blue entries are the values where the voltage to generate that power is between 200V and 250; then 250-300, 300-350, 350-400, 400-450, 450-500, and over 500.

It's likely that the first line anyone here looks at is the 100K line. 100K needs 111.8V to generate 1/8W, 158.11V for 1/4W, 223.6 for 1/2W, 316 for 1W and 447 for 2W. So for a 100K plate resistor with 100V across it, can you use a 1/4W?

Sure. 100V makes it dissipate less than 1/8W, is under the nominal 200V that all manufacturers state for their resistors (I copied a lot of manufacturers' stuff on that on another sheet). You really start needing a 1/2W if you get much above 100V, and at over 150, you ought to use a half watt just on dissipation and longevity.

The table could be generated on a voltage basis, and show the power generated (or power rating required) for each voltage increment, etc. Or could be made into a line chart, etc, depending on what makes you happier and is more easily useful.
User avatar
pompeiisneaks
Site Admin
Posts: 4222
Joined: Sat Jan 14, 2017 4:36 pm
Location: Washington State, USA
Contact:

Re: 1/4 watt resistors... no go?

Post by pompeiisneaks »

OH wow those tables are outstanding, thanks! Now I can see what voltage drop I can use for a 1/4 watt quickly. I'm still thinking that a few of the areas like that 27k in the bias circuit may still be fine as they don't impact the signal chain directly, but anything in that path sounds like it should be over spec to just reduce noise, per previous comments, etc. So I'll be likely ordering some new resistors.

~Phil
tUber Nerd!
R.G.
Posts: 1223
Joined: Tue Dec 02, 2014 9:01 pm

Re: 1/4 watt resistors... no go?

Post by R.G. »

Bias networks are special things. I'm all for massively overdesigning bias networks. Consequences of component failure are so much more severe.
User avatar
pompeiisneaks
Site Admin
Posts: 4222
Joined: Sat Jan 14, 2017 4:36 pm
Location: Washington State, USA
Contact:

Re: 1/4 watt resistors... no go?

Post by pompeiisneaks »

oh absolutely, I get that, I just didn't do the math, now that I do that math, it should likely be 1W if you ask me:

at 100V, with a 27k resistor, 100/27000 = .0037A, which when multiplied by the 100V = .37W, so I guess that's over 1/4 watt a bit :P and really close to 1/2 watt so I'd be better with 1W.

Edit: had to update Amp to Watt.

~Phil
tUber Nerd!
Post Reply