I ran a search and read some threads on resistors in this forum. It seems that a significant number of members build their projects with resistors that don`t have a working voltage rating that is adequate for tube voltages, probably unknowingly. So, is this a case of a spec that isn`t critical, or an important spec that is missed by less knowledgeable builders?
Well I used to buy those cheap Xicon .5w metal film resistors from mouser but have noticed two failures in the past. One failed and had a resistance that was a few meg ohms while another simply went noisy. The noise was of course intermittent making it a real pain to identify 😡
Lately I have been buying KOA carbon resistors and have not seen any fail so far. I do agree with your opinion of unknowingly using underrated parts. The specs are there for a reason 😉
Lately I have been buying KOA carbon resistors and have not seen any fail so far. I do agree with your opinion of unknowingly using underrated parts. The specs are there for a reason 😉
I am in favor of the Riken carbon composite resistors. Unfortunately, most of the common value ones are not available any more.
I am really getting lost what type of resistors I should use for my DIY audio projects. The Tantalum ones are too expensive and the Kewame ones don't look nice to me.
Any recommendations??
Johnny
I am really getting lost what type of resistors I should use for my DIY audio projects. The Tantalum ones are too expensive and the Kewame ones don't look nice to me.
Any recommendations??
Johnny
I use underrated (voltage) resistors all the time.
One time i wired an amp with CAT5.. which is rated 48v or something... with 400V B+.. rock n' roll.
huh, just looked it up... apparently cat5 is rated 300V rms.. so i was ok anyhow 😉
One time i wired an amp with CAT5.. which is rated 48v or something... with 400V B+.. rock n' roll.
huh, just looked it up... apparently cat5 is rated 300V rms.. so i was ok anyhow 😉
If you need a particularly high voltage rating the answer is to series them. The ratings sum. If dissimilar ohmic values are used be sure the drop across any one resistor is not too high.
Ive see bare wire run at 500KV, just need to space it right.😉
Wirewounds for anode loads.
Ive see bare wire run at 500KV, just need to space it right.😉
Wirewounds for anode loads.
It is important to watch voltage ratings. In any of my designs you will occassionally see me specify use of a 2W Metal Film resistor. That is often for the voltage rating rather than the amount of power that resistor will actually dissipate.
I use MRS25 0.6W 1% Metal Films as my "standard" resistor BUT I keep in mind that these have a 350V rating. Where that is insufficient I either use 2 resistor of 1/2 value in series or specify the 2W metal film instead. The MRS25 DO (from experience) go open circuit if you subject them to over voltage.
I experimented a lot with Rikken Ohms. Substituting Rikken Ohms often made an amp sound better. I eventually realized that they have a significant voltage dependence in their resistance value and what they do is introduce 2nd harmonic distortion. That is great for masking all sorts of nasties in the amp. Once this idea permeated the skull I have always found that it is a much better idea to fix those nasties that the Rikkens were masking and once that was done the sound was much better with the Rikkens removed.
Similarly I am not a fan of Carbon Film resistors even though they generally have good voltage rating. They are just too noisy for me.
Appologies to anyone whose "sacred cow" I've just taken a shot at.
Cheers,
Ian
I use MRS25 0.6W 1% Metal Films as my "standard" resistor BUT I keep in mind that these have a 350V rating. Where that is insufficient I either use 2 resistor of 1/2 value in series or specify the 2W metal film instead. The MRS25 DO (from experience) go open circuit if you subject them to over voltage.
I experimented a lot with Rikken Ohms. Substituting Rikken Ohms often made an amp sound better. I eventually realized that they have a significant voltage dependence in their resistance value and what they do is introduce 2nd harmonic distortion. That is great for masking all sorts of nasties in the amp. Once this idea permeated the skull I have always found that it is a much better idea to fix those nasties that the Rikkens were masking and once that was done the sound was much better with the Rikkens removed.
Similarly I am not a fan of Carbon Film resistors even though they generally have good voltage rating. They are just too noisy for me.
Appologies to anyone whose "sacred cow" I've just taken a shot at.
Cheers,
Ian
No problems with carbon-film noise here. In fact, I just, as an experiment, built a PP 6L6 amp with all-Radio-Shack carbon-film resistors, from an assortment pack. I did make sure the tolerances were good (and they were AMAZINGLY tight- not one was more than 2% off, even though they were rated 5%... most were less than 1% off), but other than that, out-of-the pack.
With the amp connected to 93dB sensitivity Tannoy Kensington speakers, you couldn't even hear any noise with your ear inside the tweeter horn!
And boy, it sounded good, too. Simple circuit that worked quite well...
Regards,
Gordon.
With the amp connected to 93dB sensitivity Tannoy Kensington speakers, you couldn't even hear any noise with your ear inside the tweeter horn!
And boy, it sounded good, too. Simple circuit that worked quite well...
Regards,
Gordon.
gingertube said:
How exactly are you doing this? Are you making sure that your resistor or paired resistors have a voltage rating that meets B+, or some higher factor for plate load resistors, for example?
GordonW said:
Try that with a phono stage where the average signal amplitudes are several orders magnitude lower or more and you will be amply acquainted with the excess noise that carbon films generate relative to the much quieter metal film or wirewound types. I would not expect any problem with power amplifiers of moderate sensitivity..
Johnson (thermal) noise is constant for a given resistance value regardless of construction, (but it is a function of the resistive material temperature - cooler resistors make less Johnson noise) however excess noise is a function of the materials used and the current flowing through them. This is one reason why fans of carbon comps often use 2W resistors in noise sensitive locations because the lower current density in the resistor generates less noise, and the larger body dissipates heat more efficiently reducing heating and.. Johnson noise to some extent. The quietest resistors of all are wire wound types which have very little excess noise.
As to the goodness of RS resistors they're mediocre in terms of excess noise performance and the linearity of their end terminations. They also have a pretty large voltage coefficient like the Rikken Ohm mentioned in a previous post. FWIW I have observed similar sonic effects with the Kiwame as well, but did not investigate to determine the exact cause.
The older Holco resistors also had noise issues due to the nature of their construction, above a certain voltage, and well below the rated voltage they would become noisy due to voltage gradients along the spiral cuts causing local breakdowns. This issue might also have been a symptom of process control problems - the problem is moot however as the construction has been changed and the current version also has linearity issues.
On the subject of linearity there are a lot of resistors that generate harmonic distortion and in some cases the mechanism is a variation in resistance due to changes in the voltage across the resistor - this is distortion generated by the voltage coefficient of the resistor. In other cases it is due to the nature of the end terminations and the way they are fitted to the ceramic as well as the dissimilar materials acting as minute diode junctions. I've actually encountered resistors that generated substantial thd with just a small ac signal impressed across them. (Some surplus Mepco Electra types I have comes to mind.)
kevinkr said:
This I'm aware of. The amp I built started out as an experiment, to see just how good I could build a design to work, even with "garden variety" components such as RS resistors, normal mylar and inexpensive poly caps, and such. Must have been a success, I guess... a fellow heard it while I was testing it out, and offered me more for it than I could say no to! 😀 I do have to say, for a totally "old school" design (modified paraphase splitter, no CCS's, 12AX7 driver tube, among other so-called "no-nos"), it was quite surprisingly satisfying to listen to, and measured well too (less than .1% average distortion, mostly second and third harmonic, at 1/2 power with a speaker load connected)!
I guess the point is- it's all about the application. Low noise resistors only matter when the noise is significant to the signal level...
Regards,
Gordon.
For power amps, an interesting compromise is the metal glaze resistor. They are non-inductive, more like a carbon comp in characteristics, and can handle large pulses and high voltages. Their noise is higher than a metal film, but lower than a carbon comp. You can get 3 watt military spec examples for 30-40 cents each. For a tube amp, I'd prefer a varied mix according to where in the amp the R's going to go.
I have asked this before...
How exactly do you exceed a 350V limit on a resistor? Sure you could do it with a 2W 100K unit, but try it with anything smaller and the power dissipation rating is going to be hit well before the 350V rating is.
what am I missing? Does this rating assume the body will be grounded?
dave
How exactly do you exceed a 350V limit on a resistor? Sure you could do it with a 2W 100K unit, but try it with anything smaller and the power dissipation rating is going to be hit well before the 350V rating is.
what am I missing? Does this rating assume the body will be grounded?
dave
The Vishay RH/NH series shows the working voltage calculated by (PxR)^0.5.
Is this calculated based on the resistor wattage max rating? As in, a 4K, 50W rated resistor,
(50 x 4000)^0.5 = 447V?
Is this calculated based on the resistor wattage max rating? As in, a 4K, 50W rated resistor,
(50 x 4000)^0.5 = 447V?
Last edited:
Your assuming a nice stable DC system..
The real world isn't like that, you can get transient pulse and back EMF.
This is taking into account that sometimes you get a fault that will clear very quickly. The components will either survive or they blow open or short circuit. OK your going to say how can this happen..well self healing coupling caps is one such issue. tube flash over is another. And this doesn't mean that suddenly all the components fail.
Another example is using a pile of parallel half watts in a tube cathode instead of something like a 10W wire wound. They work great then one day you switch on an bang they blow apart ...so the tube is no good now (wrong the fault has cleared and if you had put a 10watt in you wouldn't even have noticed it before the tubes warmed up).
What you have to remember is that you can look at voltage drop within a working circuit but you still need to assume some margin of reliability.
Take a look at mains supplies and the transients riding on the sine wave. The components on that supply must be rated at worst case scenario which may also compromise safety should they fail.
You wouldn't want a DC bias supply to fail with 8 output tubes..so you overrate the components that are critical and also put some kind of fail safe in place.
Another example is when you power off a transformer or inrush currents at power up.. you can get a back emf that will exceed working voltages and with inrush and warm up times the voltages across components can and do change if you exceed the working voltages eventually they will fail.
That's where you can fit MOV and inrush suppression etc.
Regards
M. Gregg
Last edited:
All components and devices have more than one failure mode. Getting too hot is just one failure mode. Having too much voltage across is it another failure mode for most things, even if the heat generated is not enough to fry the item. For a resistor, too much voltage can cause whatever is creating the resistance to modify itself so it has less or more resistance. Exactly how it does this will depend on the type of resistor.
Brief transients are a separate issue.
Brief transients are a separate issue.
There are actually two voltage ratings on a resistor, and trust me they are real. I have blown resistors in half due to a flashover.
A resistor will have a rating referring to the voltage from end to end. Put 650 volts across a 1 Meg 1 watt bleeder resistor rated for 350 volts and you will get a very big BANG and a black stain where the resistor used to be.
A resistor will have an insulation breakdown rating, which may be different than the maximum voltage rating. This is important where two resistors may be touching each other, or one resistor is touching the chassis, or exposed plating on a PCB. It is also very important when you touch it to see if it's too hot and find the chipped spot in the coating!!!!!!
A resistor will have a rating referring to the voltage from end to end. Put 650 volts across a 1 Meg 1 watt bleeder resistor rated for 350 volts and you will get a very big BANG and a black stain where the resistor used to be.
A resistor will have an insulation breakdown rating, which may be different than the maximum voltage rating. This is important where two resistors may be touching each other, or one resistor is touching the chassis, or exposed plating on a PCB. It is also very important when you touch it to see if it's too hot and find the chipped spot in the coating!!!!!!
- Status
- Not open for further replies.