I meant to say if he removes the cap and pot the 470k will set the input impedance, thank you for clarifying.
I was thinking with a high output impedance on the preamp and including unknown capacitance of the interconnect cables there could be treble rolloff issues if the input impedance isn't sufficiently high on the power amp.
I agree my thoughts were confusing since everyone else was talking about the RC network in question and it's effect on the corner frequency for bass. My only advice on that was if it isn't necessary then remove it and let everything through.....the less parts in the signal path the better IMHO.
Bear in mind that since the output stage has been changed to triode operation, the open loop gain and output impedance
have decreased and increased, respectively.
The amplifier may benefit from a feedback network redesign more suited to the present circuit.
The 10k feedback resistor seems much too large. The shunt capacitor value would change as well.
It is very unusual for a tube amp to have a 20k Ohm input impedance. There seems to be no reason for it this case.
Last edited:
Tweak value of R21 for the amount of global feedback, higher value for less feedback and lower value for more feedback. Without a scope it's hard to determine value of C17 though. Any idea how to find the right value on that cap?
Normally you would have a gain of around 10 or 20 (20-26 dB) in the amplifier, so the resistor might be 2k-4k.
The capacitor value across the feedback resistor could be set experimentally with a scope, square wave generator,
and a dummy resistor load approximating your speaker impedance. Adjust it for the best square wave shape
at a low power level, a watt or two.
There is also a series RC network across the plate resistor of the input pentode. This was used to set the transient behavior
of the closed loop amplifier as it was previously designed. This RC interacts with the the feedback network, so if this RC is changed,
the feedback network may need to change as well.
Last edited:
Thanks all for the, very logical, solution. I rewired the amp and its working in triode. Bias was still right. Should the 200r grid stoppers be changed? I've been using 100r for a while now and they are in there right now so I wonder if they now need to be higher.
So far i dont hear much difference, but its late at night and the amp volume cant be very high....neighbours lol
So far i dont hear much difference, but its late at night and the amp volume cant be very high....neighbours lol
Last edited:
So... I now have a speaker inpedance mismatch? (Bear in mind the amp is using 6L6GC not EL34)
Last edited:
There's always some degree of mismatch with speakers, since they vary in impedance at various frequencies.
You should try each of the output taps to see which you like best with your particular speaker.
Last edited:
Well, listened some more and its not an improvement. Dont know why some many people usig this amp raving about how much better it sounds doing just this simple triode mod without redesigning the rest. They must be having hearing problems...
The amp becomes way less tight and less refined, highs become shrill, mids are upped and you lose low end push.
Put it back to UL and she sounds way more controlled and hifi.
But its fun to try ;-)
Is there something to be gained by changing or removing the nfl? Little more high end sparkle?
The amp becomes way less tight and less refined, highs become shrill, mids are upped and you lose low end push.
Put it back to UL and she sounds way more controlled and hifi.
But its fun to try ;-)
Is there something to be gained by changing or removing the nfl? Little more high end sparkle?
Last edited:
Amplifier Compensation.
Only guide I found about this. It looks complicated, but that's the way I did it in a recent build and it works!
You don't have much feedback with the large resistor value, 50k.
The open loop bandwidth is set to around 13kHz by the RC network on the pentode's plate.
Last edited:
But, since UL allready has a neg feedback loop in the output stage why also use one connected to the input stage? There really should not be a need for it...
The UL connection could be viewed as a type of local nfb, since the plate signal is being applied to the screen grid, or instead as just an "equivalent triode" tube, but this connection does not control the entire circuit. Most tube amps include the input stage and the output transformer in the feedback loop.
You might be interested in some background reading.
http://www.tubebooks.org/Books/crowhurst_basic_3.pdf
http://www.tubebooks.org/Books/Atwood/Crowhurst%20Cooper%201956%20High%20Fidelity%20Circuit%20Design.pdf
The opposite is true. If there is high shunt capacitance then a low input impedance will help, but by pulling everything down so swamping the capacitance.famousmockingbird said:
Something has to set the LF rolloff. Early in the chain is better than late in the chain.
Ok... I've been looking at different schematics of 6L6 UL amps and they all seem to use 12k to 24k feedback resistor value. The 10k listed for my amp is for EL34.
I've been told on this forum that using 6l6gc tubes you might benefit from a change in NF. I know from working on vintage guitar amps that less feedback often means better high end response.
I'd like to know what values to try...
I think I see........like for RF transfer the load is equal to the previous stage' s output impedance?
Wouldn't most output transformers do that anyway? I guess early in the chain would be better to maximize power efficiency.
You can't just look at one resistor value because there needs to be another resistor to form the voltage divider. Usually when we speak of feedback we talk about it in db's.
Last edited:
Agreed, but I think that is because most vintage guitar amps use a pentode output stage, giving an approximate 'current source' drive, which responds to higher speaker impedance at higher frequencies by giving greater output voltage. (In other words they have high output impedance - which is reduced by global negative feedback.)
- Status
- Not open for further replies.