dave g wrote:It's not the only way to design a preamp, though. The Dumble preamp only clips a single stage (the 4th). However, the magic of that preamp is in the interaction between all four stages through the power supply. In response to an input signal from your guitar, the current draws of the 2nd and 4th stages are in phase, while the 1st and 3rd stages are in phase with each other but out of phase with the 2nd and 4th. Ever try to measure the amplitude of the signal on each grid under normal conditions? It's low (~100mV) on the grids of the 1st and 3rd stages, and much higher (on the order of a volt to several volts) on the 2nd and 4th stages. If you draw the load lines of each of the 4 stages and then inject a test voltage to the grid of the 1st stage, you'll see that the current through the 1st and 3rd stages drops about .1mA in each stage given a 100mV input signal, while the the current through the 2nd and 4th stages rises by >1mA. So you're talking about at the very least an additional 2mA being drawn by the preamp in response to this signal. Big deal, right? Except for the fact that there's a 15k-22k dropping resistor that will drop the preamp voltages 30-40 volts in response to that extra current draw!!! You never see a preamp power supply dropping resistor that big in a Fender This in turn drags the load lines closer to the origin on all 4 stages - basically preamp "sag"! This creates that extra compression the Dumble preamp is known for...that's why you get the dynamic, singing sustain when you dig in, but it's so clean and clear when you pick lightly or roll your volume knob down.
I think this is the source of the "bloom" - faced with a large input signal, the load line of the clipping stage will be pulled towards the origin, and the maximum voltage swing will actually decrease. As the input signal decays, the current draw of stage 2 shrinks, allowing the load line of the clipping stage to expand outwards again, which increases the signal swing seen on that stage's plate.
I don't know about this being 100% accuate. I just looked at one of my Dumble topology amps that happened to be open on the bench.
I see the voltage swing increasing at every stage, not low on 1 and 3 then high on 2 and 4 like you described. Plus I see a rounding off of the sine wave at the grid of stage 4 indicating that stage 3 is producing some distortion. I have my trimmer set at 24K to ground as is commonly done.
Interesting about the 22K rail resistor dropping the plate voltage on V1/V2. That's more obvious on V2 but I did see a 20-40 V drop on stage 3 and 4 as I increased the drive to stage 4.
Are you using an HRM or non-HRM? The signal on the 3rd stage's grid will be considerably lower on an HRM...
Even on my #1 non-HRM, my 3rd stage doesn't distort very much. Of course, it all depends on the circuit variation and your settings...I was just giving rough numbers in my description, I do need to take measurements but like I said I'm off at school and away from my workshop.
Regardless, my main point is that a lot of the D sound is due to sag in the preamp power supply.
Last edited by dave g on Thu Nov 18, 2010 9:20 am, edited 3 times in total.
I don't like using tube rectifiers so much because I'm cheap and have tons of 1N4007's. I've found that a large dropping resistor in combination with a cathode biased output stage gives a nice squishy sound without having to work with a tube rec. You don't get the squishing effect like in a tube rec/power stage setup but I'm personally not a huge fan of this always anyways.
dave g wrote:Are you using an HRM or non-HRM? The signal on the 3rd stage's grid will be considerably lower on an HRM...
This one's an HRM, I haven't compared to the non-HRM amp but I agree that the 3rd stage is pushed harder on the non-HRM.
Even on my #1 non-HRM, my 3rd stage doesn't distort very much. Of course, it all depends on the circuit variation and your settings...I was just giving rough numbers in my description, I do need to take measurements but like I said I'm off at school and away from my workshop.
"Much" is the operative word here. You're right, it doesn't distort much, but it does show some.
Regardless, my main point is that a lot of the D sound is due to sag in the preamp power supply.
In an amp with very basic design, one gain stage, a paraphase inverter, no global feed back
the inverter provides the gain, limiting the number of stages.
It is power side distortion and the inverter type dictates the harmonic content
of the distortion, also the power tube type colors the tone.
All tube circuits have feed back, without global feedback you have to use
degenerative feed back, and be more critical of bypass cap values.
I favor a simple amp, I prefer the tone of the distortion, a wreck takes that to the next step...
Pre amp dirt is different, you don't necessarily experience the same power side response.
I like the "response" and how a simple amp reacts when its "pushed".
You can blow away nearly any MV type amp for tone with the right circuit
and just the right watts for the venue.
I wouldn't call it "holy grail tone", you have to know what to do with it.
Some high gain rigs I just dont know how to play right because I don't get
that "feel", what I've come to expect from a well voiced simple amp.
The opposite can said as well. The player makes the amp, not the other way around.
Another thing about output and PI clipping is that it pretty much symetrical distortion as both of these stages are biased to be close to symetrical. They thus produce a lot 2nd and 3rd harmonic content.
Once you get back to stage 3, any clipping is usually asymetrical as the stages are biased off center of the available power supply voltage and one phase clips before the other. This pushes harmonics out to the 4th and 5th primairly with less 2nd and 3rd.
If the 2nd stage goes into clipping, it will compliment the 3rd stage and again bring up 2nd and 3rd harmonics in addition to it's 4th and 5th harmonics.
TheGimp wrote:Another thing about output and PI clipping is that it pretty much symmetrical distortion as both of these stages are biased to be close to symmetrical. They thus produce a lot 2nd and 3rd harmonic content.
This is only true for single-ended outputs. Push-pull outputs cancel even-order harmonics.
Wife: How many amps do you need?
Me: Just one more...
TheGimp wrote:Another thing about output and PI clipping is that it pretty much symetrical distortion as both of these stages are biased to be close to symetrical. They thus produce a lot 2nd and 3rd harmonic content.
Once you get back to stage 3, any clipping is usually asymetrical as the stages are biased off center of the available power supply voltage and one phase clips before the other. This pushes harmonics out to the 4th and 5th primairly with less 2nd and 3rd.
If the 2nd stage goes into clipping, it will compliment the 3rd stage and again bring up 2nd and 3rd harmonics in addition to it's 4th and 5th harmonics.
For a sinusoidal input, symmetrical clipping produces NO even order harmonics, only odd order. The harder you clip, the more your higher harmonics start to appear in relation to the fundamental.
Asymmetrical clipping is kinda of a weird term because it refers to multiple things. If your signal is clipping on only one side then it will produce nearly all even harmonics (really mostly second order and a hair of third). If it is clipping on one side more than the other, THEN you will get the nice mix of even and odd harmonics.
These plots may shed a little light on the subject:
What I've posted is spectrum analysis snapshots of a marshall plexi circuit;
The first is a sine wave on the amp input. (~100 to 200 mV IIRC?) The output is taken at the output of the tonestack. So all you're seeing is the preamp distortion spectrum.
The second is a CLEAN (1 volt?) sine wave fed to the PHASE INVERTER input; in other words, bypassing the preamp. The output is taken across the speaker load. So all you're seeing is the power amp distortion spectrum
The third is feeding a 100-200mV sine to the preamp input, then looking at the output taken across the speaker load. So this one shows how the whole amp distorts.
Now I realize that linear analysis fails when we're talking about distortion, BUT the signal coming from your guitar is dominated by a high energy fundamental so it's "close enough" to a sine wave...
Anyway, my point is that at least in this specific example, what we see is indeed cancellation of even order harmonics when you're only getting distortion in the power amp section...the 2nd is almost 35dB down from the 3rd. And then notice how you don't get nearly as much distortion in the preamp alone, but it's predominantly even order, and the interaction between the preamp distortion and power amp distortion fills in the spectrum quite nicely.
You do not have the required permissions to view the files attached to this post.
TheGimp wrote:I think that cancelation is a strong word.
Minimized? Maybe.
Canceled? No. Cancelation is in this sense is a relative term used in comparison to a SE amp, not in absolute terms.
The 2nd harmonic may be down 6-9dB from the 3rd, but I believe it will still be there.
We wouldn't want a completely cancel even-order harmonics anyways. The misbalance in push-pull pairs is part of the magic of tube amps. This is why I don't like tightly matched pairs in the power stage, unless I'm going for that ultra-clean sound (in which case I'll probably use heavier negative feedback and maybe 40% screen taps).
This in turn drags the load lines closer to the origin on all 4 stages - basically preamp "sag"! This creates that extra compression the Dumble preamp is known for...that's why you get the dynamic, singing sustain when you dig in, but it's so clean and clear when you pick lightly or roll your volume knob down.
