Why does this amp sound so good?

[wizard333]

WHO says it sounds good?

Well, I do of course. But then I like a warm, mid-range-y fat M type tone. If you like a scooped/bright/brittle kind of thing, then you wouldn't like it.

[nickb]

Tested it through a few, the same few I run about 20 other amps through (or have run about 100 or so other brands/models of amps through) so the speaker is not the cause of any difference.

OK. Do you like the sound of it clean, cranked or both?

[wizard333]

Both are ok; by "clean" or "cranked" I assume you mean the pre-amp. Pushing the pre-amp hard (but even with master fairly low so the power tubes aren't pushed hard) it sounds as I described; of course when you start to hit those EL84s hard it adds more high end, but still sounds very good.

[nickb]

Both are ok; by "clean" or "cranked" I assume you mean the pre-amp. Pushing the pre-amp hard (but even with master fairly low so the power tubes aren't pushed hard) it sounds as I described; of course when you start to hit those EL84s hard it adds more high end, but still sounds very good.

I think you hit the important circuit variations pretty well in your original post. The only thing I can suggest is that maybe it's a rather poor output transformer with lot's of leakage inductance. If you have a scope you could test the idea by comparing the PI input amplitude to the speaker output and vary the frequency.

[wizard333]

That's an interesting idea. From everything I read, his transformers are quite up to the task and people generally only change them if they are hooked on internet hooey that says "you have to have a Mercury/Heyboer/etc."

That doesn't mean the supplier couldn't have slipped in a bad one though, so I'll check it.

[darkbluemurder]

I add a link from the SLOclone forum where the 470p II 470k RC filter before the grid of a tube is discussed (you must be registered and logged in to see it). www.slocloneforums.com | Login

So maybe the 1 meg II 500pf combination before the second gain stage adds more to the final tone than expected.

Cheers Stephan

[Mike Sulzer]

If you want to cut bass (300 Hz), why not just use a smaller coupling capacitor, which I think would give the same effect on blocking distortion also?

I add a link from the SLOclone forum where the 470p II 470k RC filter before the grid of a tube is discussed (you must be registered and logged in to see it). www.slocloneforums.com | Login

So maybe the 1 meg II 500pf combination before the second gain stage adds more to the final tone than expected.

Cheers Stephan

[darkbluemurder]

The way I understood it it's not a bass attentuator since it is not wired as a voltage divider (there is no resistor to ground after it). Instead it is a bass boost below the -3db point (ca. 300 hz in this case), the strength of it depending on the output impedance of the preceding stage and the gain characteristics of the gain stage it's going into.

The discussion linked was on the Bogner Exstasy and it was noted that the 470p II 470k before the cold clipping stage (unbypassed 10k cathode resistor) would not have much of an effect. In the TMB 36 it's before the 2nd gain stage, and that has an unbypassed 820R cathode resistor so as I understand it could have more of an effect.

Cheers Stephan

[Mike Sulzer]

I see. During linear operation, the effect of the capacitor would be to prevent high frequency roll off due to Miller. When you hit the stage hard, the effect is complicated, but I think the resistor increase the time it takes to charge the coupling capacitor. Maybe this delays the onset of blocking distortion?

The way I understood it it's not a bass attentuator since it is not wired as a voltage divider (there is no resistor to ground after it). Instead it is a bass boost below the -3db point (ca. 300 hz in this case), the strength of it depending on the output impedance of the preceding stage and the gain characteristics of the gain stage it's going into.

The discussion linked was on the Bogner Exstasy and it was noted that the 470p II 470k before the cold clipping stage (unbypassed 10k cathode resistor) would not have much of an effect. In the TMB 36 it's before the 2nd gain stage, and that has an unbypassed 820R cathode resistor so as I understand it could have more of an effect.

Cheers Stephan

[wizard333]

I can't see the discussion there as it requires a log in, but if you think about what the typical M style 470k/470pf combo does:

A series resistance lowers signal level and attenuates high end frequencies. How much depends on the impedance driving it but we don't care about absolute values here, just the effect. Now if you bypass that series resistance with a cap, you're effectively shorting that resistance for high end frequencies bypassed by the cap.

So what you have is a passive high pass filter, or a bass attenuator, not a bass boost. The larger the resistor, the more high frequency you are attenuating (ignoring the cap for a second) and the more overall voltage reduction you have across the resistor (V=IR). The bigger you make that resistor, the more you drop voltage, EXCEPT for the frequencies passed by the cap, which "don't see" the resistor. So a 470pf/470k series combo will attenuate lower frequencies less than a 500pf/1M series combo. That would tend to make the amp brighter and more brittle, which is the opposite overall effect of this circuit vs. standard Plexi circuit, so there is something more going on here that more than compensates for the smaller couplers and aforementioned larger series resistance bypassed by that cap.

[darkbluemurder]

A series resistance lowers signal level and attenuates high end frequencies. How much depends on the impedance driving it but we don't care about absolute values here, just the effect. Now if you bypass that series resistance with a cap, you're effectively shorting that resistance for high end frequencies bypassed by the cap.

I agree to that.

So what you have is a passive high pass filter, or a bass attenuator, not a bass boost. The larger the resistor, the more high frequency you are attenuating (ignoring the cap for a second) and the more overall voltage reduction you have across the resistor (V=IR). The bigger you make that resistor, the more you drop voltage, EXCEPT for the frequencies passed by the cap, which "don't see" the resistor. So a 470pf/470k series combo will attenuate lower frequencies less than a 500pf/1M series combo. That would tend to make the amp brighter and more brittle, which is the opposite overall effect of this circuit vs. standard Plexi circuit, so there is something more going on here that more than compensates for the smaller couplers and aforementioned larger series resistance bypassed by that cap.

This confused me a bit. Without a cap bypassing it, the 1 meg would attentuate more treble than the 470k (down to about 1 khZ vs. about 2.2 kHz w/ 470k). By bypassing it with a 500pf cap, the resistor for AC purposes is out of the circuit until the -3db point of that combination (which would be 300 hz for the 1m/500pf and 720 hz for the 470k/500pf). So the treble attentuation should be offset in both cases. What should remain is the attentuation effect in the bass of the RC combo going into the tube but seeing the tube grid as an almost infinite resistance I am not convinced it matters. But on the other hand, if it does not matter why did Ceriatone put it in there? Funnily enough, it's not in the 18W model of the same maker.

I guess I have to take one of my homebrews and try it for myself. I have an amp with a two stage preamp with a gain control between the two stages. Currently there is nothing between the gain control output and the second stage's grid. I will insert a 1m/500pf combination and see whether that does something or not. The second gain stage has a 1k cathode resistor which can be bypassed with a 2.2uf cap so I can also check if that affects something. I will post the results when I get to it.

Cheers Stephan

[Mike Sulzer]

The bigger you make that resistor, the more you drop voltage, EXCEPT for the frequencies passed by the cap, which "don't see" the resistor.

Looking into a grid is essentially infinite impedance at low frequencies; so making the resistor bigger does not drop low frequencies. High frequencies would be attenuated by the Miller effect, but the cap takes care of that.

[wizard333]

A resistor drops voltage if there is current flow. The cap makes the resistor essentially transparent for those frequencies that pass it.

[darkbluemurder]

Finally I got around to the empirical experiment and put the 1M/500pf network before the grid of the second gain stage in my Hiwatt inspired preamp. First gain stage is two sections of a 12AX7 in parallel, Ra = 110k, Rk = 1.5k, no Ck, coupling cap 22nf into a 500k gain pot, then to the 2nd gain stage, Ra = 100k, Rk = 1.0k, Ck = 0.047uf.

The result is rather subtle in this amp when the gain control is fully dimed. Below the full up setting I don't hear any difference. With the full up setting, the tone appears to be a bit more refined and a bit less splatty in the low end. It is an improvement but not a night and day difference.

Cheers Stephan

[SoulFetish]

I will say I have seen their kits built and I have seen factory wired Ceria units, and they appear to be good quality units.

I would agree with this. in my experience, their work is clean and high quality.