I wanted to try a corrective filter (As Seen in RCA in my TW Rocket clone, which has a less than optimal OT.
A 10k/5W in series with a 2.2n across the OT primaries can be found in the Dr. Z Carmen Ghia and Suhr Badger 18, so I want to give it a go.
The original article says that the resistance should be about 1.3x the plate-to-plate impedance, which is about 10k for an typical 8k primary (assumed to be in both the aforementioned 2xEL84 amps). I have a 4K primary with 4x EL84s, so I'm thinking 5k would be fine.
My questions is about how to "size-up" the 2.2n for the two extra tubes in parallel. Is it necessary? Would I simply double the capacitance to keep the same filter?
Is that what it's called in the RCA book? I recall several discussions over at el34world and AX84 - google "conjunctive filter" and you'll get some good hits.
I'm thinkin' that such a zobel could be applied at other spots in the amp, such as a Vox "cut" control across the plates of the output drive tube, or even early in the chain, after all what's a tone control in a guitar?
In the case of having this "corrective" filter applied across the primary of an OT = output tube plates, one has to wonder if it's there to damp ultrasonic ringing of the OT. If the intention was to simply affect the tone, there are other spots in the circuit where that could be done, and employ components that need not have to stand up to the high voltages across the OT primary.
Thanks, Bill, I see now in the Maven Peal article. It makes sense that the resistor needs to be smaller to accommodate the extra parallel tubes, and the simple RC filter just needs to be adjusted for that. So it looks like your intuition was right, matt h.
The amp in question has VVR on the output tubes and PI, and at very low setting gets this annoying swirly fizz on top of everything when the gain is cranked up. The amp sounds normal at less than whisper volumes.
I did notice a small leading edge on the scope, but nothing that looks fuzzy like a parasitic, and it only appears at the output, therefore no low pass filter before the output get rid of it.
I hope Ian (Gingertube) jumps in on this thread. IIRC he may have been involved in a fairly in depth filter discussion over at the now defunct powerscaling.com site.
BTW, regarding the swirly PI, I'm guessing you are over driving it. As you are scaling the PI down, the signal feeding it needs to be decreased as well. Think of an open window getting smaller and smaller as it is closed, while the signal remains the same size. In fact, the feeding signal may even increase as the power amp is unloaded.
Thanks, John. I would be great to hear Ian's 2c. I've actually tried a few things to pin down the problem already...
First I added a large grid stop to the PI, but that didn't help. Next I added a pre-PI master to dial down the signal, but it didn't get rid of what I was describing. I also tried just scaling the output tubes, but that made the swirly fizz even worse.
I'm pretty sure the leading edge on the waveform is unrelated, and what I'm hearing is just the result of a lot of overdrive (especially EL84 overdrive (at really low volumes.) Seems like the corrective filter could kill two birds with one stone, however.
FWIW, I've played around with the lead dress, and it is really tidy in the output stage with short grid wires and 10k stoppers on the power tubes. Nothing I did affected the sound, so I've ruled out all the obvious wiring stuff.
Reeltarded wrote:Zobel? So that is what i am doing when i dump treble to ground with a cap and a resistor. Zobeling?
Yep. Keep it up and you may win the zobel prize. Or igzobel prize. Depending...
For extra credit add an inductor.
If you've seen any of JBL's latest passive crossovers, they're paved with them - "correction circuits", lots of little coils. Thank heaven for little coils.
Gaz, Sounds like you have approached it logically. I remember working on a 2204 build that had the swirlies that I could never fully get rid of. It may be related to filtering / grounding. I remember Kevin O'Connor saying that power scaling exposes all weaknesses in grounding.
Structo wrote:I remember Paul Ruby did a article about buzzies or fizzie, in the 18w amps.
This PDF shows scope shots and good info how to cure this stuff.
thanks for sharing that article, I have a pretty well known booteek amp that looks like this on the output and have been wondering where to start with it. Lifting the cathode bypass cap helps considerably. It is also biased super cold but didn't make much difference with a more appropriate cathode resistor. [IMG577]http://i1233.photobucket.com/albums/ff388/randalp3000/amps/IMAG0481_zps009f95dd.jpg[/img]
My problem isn't the paul ruby thing. It's just a matter of all the really high harmonics becoming more present at very low volumes with the scaling. I'm also using Merlin's grounding to a "T" (basically the same as Kevin's "galactic ground")
I tried the the corrective filter and it definitely fixes it, by simply evening them out to the level of the other frequencies. Now, however, the very low volumes are kind unusable because they are so smooth. I guess I can't have my .1W and play it too.