V4 & V5 100 ohm resistor tolerance

[jackhadley]

Long time Lurker, first time poster...

This is my first 18watt amp build, and first amp build ever. Lots of pedal builds, but this is my first amp.

I got a StewMac 74x kit (which I'm pretty sure is a rebranded Mojo Tone), original term circuit (not TMB). I noticed that the 100 ohm carbon comp 1w resistors that go on pin 9 of the EL84 sockets are pretty far out of tolerance. One is reading 130 ohms, the other 110. I haven't soldered these yet because this variance gave me pause.

I have the advantage of my buddy's exact same build right beside mine for reference as I build, and when I measured those resistors (after draining the caps) one was 115.6 ohms and the other 119.6 ohms. These are obviously out of tolerance too, but the sound of that amp is good to me, and if my build sounded like that I would be happy.

My questions are:
- Should I be concerned about this being 30% out of tolerance?
- Should I find one much closer in tolerance?
- If so, is there a different type of resistor I could consider?
- Does the resistor need to be 1w?
- Can it be 1/2w? (I can get these pretty easy)

Thanks to everyone for this amazing resource.

Cheers and Happy New Year!

[JMPGuitars]

Hi,

Go higher in wattage, not lower. I would use 3W 1K wirewound resistors for the screens. 100 ohm is not enough.

Thanks,
Josh

[crgfrench]

Nice choice for a first build!

- Should I be concerned about this being 30% out of tolerance?
NO but a 1k would do more for the circuit. 100R is almost like, why bother?

- Should I find one much closer in tolerance?
NO but go up to 1k.

- If so, is there a different type of resistor I could consider?
YES Try the 3W reduced mass wirewound Dralorics.

- Does the resistor need to be 1w?
YES (AT LEAST) but 2W or 3W would be safer here.

- Can it be 1/2w? (I can get these pretty easy)
NO.

[TriodeLuvr]

Those resistors serve as parasitic suppressors (stoppers). They must be non-inductive at RF frequencies, which is the reason they are carbon comp. No wirewound resistor of any type meets that requirement, nor do most carbon film or metal film resistors, due to their spiral construction. The two resistors don't need to be matched, and their value isn't critical. You could buy a batch and hand-grade for exact values, but you won't hear a difference. About the power dissipation, each screen might draw 25mA occasionally. That calculates to less than 1/10W. So yes, 1/2 W resistors are fine in this application.

[JMPGuitars]

Those resistors serve as parasitic suppressors (stoppers). They must be non-inductive at RF frequencies, which is the reason they are carbon comp. No wirewound resistor of any type meets that requirement, nor do most carbon film or metal film resistors, due to their spiral construction. The two resistors don't need to be matched, and their value isn't critical. You could buy a batch and hand-grade for exact values, but you won't hear a difference. About the power dissipation, each screen might draw 25mA occasionally. That calculates to less than 1/10W. So yes, 1/2 W resistors are fine in this application.

That is not correct. Those resistors are there to protect the screens, and keep the voltage potential lower than the plates. 100Ω is really low, 1k is better.

You can learn more by reading Aiken's Q&A regarding screen grid resistors: https://www.aikenamps.com/index.php/technical-q-a

Merlin also discusses it at the bottom of this page: http://www.valvewizard.co.uk/gridstopper.html

Thanks,
Josh

[TriodeLuvr]

Well, let's go through it point by point.

First, these resistors aren't normally used to control voltage on the screens. The screen grid doesn't need to be lower than the anode. (I think that's what you meant to say, not "lower than the grid." Correct me if I'm wrong.) Below is a compilation of application data from the EL84:



And here is the EL34, which in many cases operates with the screen elevated above the anode:



Regarding parasitic oscillations, the discussion at Valve Wizard mentions this in relation to UL operation only. However, I have seen this occur in pentode mode as well, and I always use stoppers on the screens for that reason.



You are correct about larger resistors protecting the screens from over-current, as also mentioned in the Valve Wizard article. However, that's a different issue and not the one to which I was responding. I was only considering the purpose of the 100 ohm resistors and the fact they are not quite the same. At that value, they only serve one purpose - parasitic suppression - so that was the basis of my response.

Jack

[JMPGuitars]

You are correct about larger resistors protecting the screens from over-current, as also mentioned in the Valve Wizard article. However, that's a different issue and not the one to which I was responding. I was only considering the purpose of the 100 ohm resistors and the fact they are not quite the same. At that value, they only serve one purpose - parasitic suppression - so that was the basis of my response.

Jack

Yeah, I meant plates. I corrected it.

Those resistors are there for what I stated above, and that value isn't high enough to have any real affect on oscillations.

For tube safety, and tone, it's a good idea to keep the screen voltage below the plate voltage. The voltages you're referencing are well below the tube's limits. In that case you're safer anyway. I run tube traces all the time with equal screens and plates, but that doesn't mean you want to run an amp that way. And when I go above a certain threshold, I drop the screen voltage in those tests.

I should also mention that if you read those datasheets, you'll see that the upper limit for the plate voltage on an EL84 is 300V. But these are typically running 40 or 50 volts higher than that.

When you start running the plate that much higher, the screen needs to be lower. If you don't believe me, go try it out. Setup an amp with a set of EL84s with the plates at 350V, and the screens at 375V. With 100Ω 1/2W screen resistors. Let me know how good it sounds, and for how long. I take no responsibility for what happens to your amp or anything else. Test at your own risk.

[TriodeLuvr]

I don't know how much current EL84 screens draw under those conditions, so I can't speak to that. It's easy enough to measure though, and I would hope the kit supplier and/or amp designer have done that, as well as reliability testing.

As for parasitic suppression, 100 ohms is more than sufficient. It's purpose is not attenuation per se, but rather a broadband reduction of grid circuit Q. This same value is used in high-power RF transmitters to tame tubes that are much more sensitive to this phenomenon than the little pentodes we use. Control grids in audio tubes generally need a higher value stopper than screens or anodes due to their higher operating impedance.

Jack

[crgfrench]

"This resistor should be at least 1k Ohm, 2W or more."

[JMPGuitars]

I don't know how much current EL84 screens draw under those conditions, so I can't speak to that. It's easy enough to measure though, and I would hope the kit supplier and/or amp designer have done that, as well as reliability testing.

Hahahahahahahahahahaahaha! Most kit suppliers supply things based on what was okay 60 years ago, before anybody knew better / before better components were available...but more importantly, do a search for mojotone on this site. Their kit is a nightmare. They make good cabinets, and sell some good parts, but their 18w kits are trash.

As for parasitic suppression, 100 ohms is more than sufficient. It's purpose is not attenuation per se, but rather a broadband reduction of grid circuit Q. This same value is used in high-power RF transmitters to tame tubes that are much more sensitive to this phenomenon than the little pentodes we use. Control grids in audio tubes generally need a higher value stopper than screens or anodes due to their higher operating impedance.

Jack

I think the key factor there is "tubes that are much more sensitive to this phenomenon than the little pentodes we use." - Since we know they aren't very susceptible to this, even if that value would work, it likely doesn't serve that purpose, or at least isn't there specifically for that purpose. Though we do know that the correct value there supports other important things. Like safety, and IMO/E, better tone.

Happy New Year

Josh

[TriodeLuvr]

It's not unusual for commercial amps to use a screen resistor value around 100 ohms for EL84 designs. Some omit them altogether. I wouldn't do that because I've seen the problem with parasitics, but it's obviously dependent on various factors. The resistors do no harm, and they help to ensure amplifier stability. On the subject of limiting screen current, here's a Marshall 20 watt operating the screens at +370V. No resistors.



Happy New Year to you too. I'm sure hoping 2021 is better!

Jack

[JMPGuitars]

It's not unusual for commercial amps to use a screen resistor value around 100 ohms for EL84 designs. Some omit them altogether. I wouldn't do that because I've seen the problem with parasitics, but it's obviously dependent on various factors. The resistors do no harm, and they help to ensure amplifier stability. On the subject of limiting screen current, here's a Marshall 20 watt operating the screens at +370V. No resistors.

Happy New Year to you too. I'm sure hoping 2021 is better!

Jack

Thank you, me too!

lol, if you search for that 20W amp here you'll see where it was recently problematic (including red plating), and one of the solutions was to add 1K screen resistors.