Why do all attenuators produce buzzy harmonics with beam tetrodes?

[Wallington]

Hi Guys,

This is a long-running bug bear of mine that I've not been able to get to the bottom of - I'm hoping the fair folks of the forum may be able to shed some light where others haven't?

In essence, the problem is this - run any guitar amp with a beam tetrode based power section into any type of attenuator and when the amp begins to break up and reasonably heavy attenator, a harmonic buzz is produced from the bass-end notes which sounds like it's phased.

Any beam tetrode he says? Take an amp like a Komet 60 that can run EL34, 6L6, 5581, 6550 etc - play it unattenated and whilst it's likely to get me arrested due to the volume, there's no nasty artefacts in the tone

Now let's run it through a re-amp (Power Station, Two Notes) a reactive load (Rock Crusher, Radial) or a passive load (Kock loadbox or Airbrake) - the EL34s still sound awesome, but any of the beam tetrodes buzz like someone is accompanying me on a Moog Synth

So it must be the amp then? .. OK, let's try two Komet 60s side by side with 2 Power stations, swap EL34s and 6L6s between them and then swap the attenators - and follow the buzz)

So it must be the circuit then or the mismatch of the OT Primary? ..try and amp with a Primary impedance tuned to one tube type? OK, so let's put a 5E3 with an 8k Primary or a 5F6 in there - same issue

Just to the record, i've systematically ruled out speakers, cabs, cables, guitars, power supplies and even the room itself

I currently have 2 Power Stations, a Rock Crusher and an Airbrake and I could predict 100% of the time when this is going to happen and it's never heard with Pentode power tubes and always there with beam tetrodes

I've gone as far as I can with this - I've even tried to EQ it out with a graphic and tried snubber caps pre-PI and on the plates with no effect

If you look at the difference in the tube design, then this has to be something to do with secondary emissions, but I've no idea how to practically prevent it from happening?


So Gentleman, what's going on here and how do I fix it?

[Jazz P Bass]

Any chance that you could possibly provide a 'scope shot'?

I am thinking screen supply/ screen resistor.

[olddawg]

Have you tried a Weber type attenuator with a speaker motor design? Imhe, ALL attenuators sound pretty crappy if used more than 25% or so.

[Mick Bailey]

In a truly reactive attenuator the loading would be no different than connecting to a speaker - you'd get the back-emf and damping that a speaker produces. The tubes would see no difference in loading and frequency-dependent impedance variation.

So you're assertion is that any attenuator produces artifacts with any beam-tetrode equipped amp? Not in my experience. I had an early Marshall JTM45 with KT66 tubes running into an Aracom attenuator and it didn't have any problems whatsoever.

[R.G.]

There is a set of data that you need to produce before we can help you much. Two of the three responses you've already received are tied up in that.

You have asserted that

1. there is an audible buzz
2. with ALL amplifiers
3. ... which use beam power tetrodes for outputs
4. ... with ALL attenuators

It's very tall order first to prove that there is a buzz to start with, no matter than you hear it every time, as third of the three responses notes. It is remarkably hard to prove that ALL listeners will hear a buzz, as we've seen with many audio phenomena, just to add another ALL to the list.

Taken together, the assertion behind the question is very difficult to accept to start with. Not that it does not exist, just extraordinarily hard to prove. And it's a truism in the sciences that extraordinary claims require extraordinary proof.

Please - show us that the buzz (1) exists, perhaps with scope traces and/or spectrum analyzer traces (2) with all attenuators (3) in all amps with beam tetrodes and NOT in (4) ALL amps with power pentodes. When we get that, we'll be armed with the base knowledge needed to dig in and go help you identify what's causing this problem.

[pdf64]

Given the same amp, I find that 6L6 tend to sound a bit harsher, and have a stronger bottom end, under overdrive than EL34.
But that's with speakers, I don't use amp attenuators.
I put it down to screen grid sag; pentodes tend to draw a larger Ig2/Ia than beam tetrodes, and that ratio tends to increase with signal level.
Hence with the saggy HT supplies of typical guitar amps, the 'sliding screen' action may be more pronounced with pentodes than tetrodes.
So it may just be a 'more overdrive compression/lower overdriven power output' (pentode cf beam tetrode) kind of thing.
That it isn't noticed when unattenuated may just be a power dependent hearing / speaker compression etc.

[Dave H]

Here's an example. I have a 30W amp that can use either 4 x 6V6 or 2 x EL34. You can see the difference in the plots below. The 6V6s clip harder and have more high order harmonics than the EL34s. The load was a 10 ohm resistor.

[Mick Bailey]

It would be interesting to compare the plots with a reactive attenuator, and with an actual speaker.

[teemuk]

It would be interesting to compare the plots with a reactive attenuator, and with an actual speaker.

I don't know if there would be a drastic difference.

These plots are obviously for resistive load. Assuming moderately poor damping factor of the amp (otherwise the comparison is more or less useless), with reactive load there would be "flyback" transient at every sharp corner of clipping (the plotted clipping is probably too soft to display such) and different phase of current and voltage would move crossover distortion's "kink" to some other point in the waveform than the usual cross over at zero volts. With moderately poor damping factor the crossover kink would create yet another "flyback" transient.

Heavy ovedrive would create waveform similar to one in the bottom in this picture:

...Where any apparent "square wave" -ish clipping gets hidden by flyback transients and amp's overall non-linear response to reactive loads. Crossover distortion gets that flyback "kink" and usually the kink doesn't appear at zero volts but some other voltage value (because load current and voltage have different phases).

[Wallington]

Hi Gents

Thanks for the replies and apologies for the slow comeback - the terrorist attack in London on Saturday was pretty close to home for me

I'm afraid that I don't have a scope, so I can't offer any data on that front. An amp builder who came over to my studio and used a phone app to scope the noise reckoned it to be IRO 2KHz based on an open low E string

I have a small mpeg audio file where you can hear it quite clearly, but I can't immediately figure out how to insert it in my reply here.

What I can say is that from my own experience over 2 years of investigating and roughly 10 amps and 6 attenuators, is that I consistently get this at the edge of breakup in lower frequencies and it's worse with HBs

It's probably also worth noting that the issue is worse on something like a Power station which uses a reactive load down to line level and re-amplifies than a Rock Crusher, which is attenuating down say 12dB