Tube Amp Design Confusion

[apehead]

Hello,

It's been a while sine I've last posted, but I had a question concerning the gain between tube amp stages.

If I had a single triode stage biased for a gain factor of 10, and I injected a 1Vpp signal, the output would be 10Vpp. Now if I coupled it with another stage biased for a gain of 10, wouldn't the grid be be pushed positive far too often, destroying the tube? I guess what I'm trying to ask is how amp designers couple multiple stages together while avoiding extremely low grid biases/plate voltages without cherry-plating tubes?

- Jack

[Hardtailed]

I believe the important thing is that the average signal is roughly 0V (or whatever bias voltage is applied) since an audio signal is a wave centering on 0V after it's been through a capacitor to block DC.

[GibsonLover]

There's a good application for Tube CAD. The Vth data calculated is the input threshold at at which the output will clip. You may be feeding it 10Vp-p, but a 12AX7 may start to clip at 1.3V (for instance, depending on biasing). The 10Vp-p = 5Vpk. 20*log(5/1.3) = 10.4dB of overdrive. If cascading stages for the purpose of distortion, you want to adjust your interstage attenuation to overdrive 6-9dB, depending on your taste. (Trying to get all of your 'drive' by overloading one stage sounds like crap.) You're only going to red-plate if you exceed the current or voltage specs, and those are biasing parameters. The inter-stage attenuation is what allows you to cascade stage after stage, and still leave the amp at a modest level for an FX loop. I've been flamed for mentioning Tube CAD before, but you can get close to biasing and adjusting your gain staging on paper before you let your ears at it. I guess some folks like to copy Fender, do what they always do, or just change parts all day. You can't beat the education for $40 either. I wasted more than that at the bar last night, and didn't learn a damned thing...

[mooreamps]

Hello,

It's been a while sine I've last posted, but I had a question concerning the gain between tube amp stages.

If I had a single triode stage biased for a gain factor of 10, and I injected a 1Vpp signal, the output would be 10Vpp. Now if I coupled it with another stage biased for a gain of 10, wouldn't the grid be be pushed positive far too often, destroying the tube? I guess what I'm trying to ask is how amp designers couple multiple stages together while avoiding extremely low grid biases/plate voltages without cherry-plating tubes?

- Jack

No. Be-cause of the loading caused by the next gain stage. Let me see if I can try to make this seem clear.

ahm, lets say on the first gain stage you have a stage gain of 10, and lets assume the plate impedance of the 12ax7 is 65 K ohms, and then lets assume the total loading presented by the next gain stage is also 65 K ohms. You will drop half of your stage gain on the plate impedance of the first gain stage, and have only half left over for the second gain stage. Now, a most real life ratio is plate impedance of 65 K feeding an input impedance of 1000k, assuming the grid leak of the second gain stage is 1 meg. But, as a second order of magnitude loss, you will drop some of your gain off the plate impedance from the first gain stage.


Secondly, as a first order of magnitude, designers use interstage attenuators between gain stages to further drop the level of the signal into the second gain stage. An interstage attenuator could be a pair of tapped resistors in series, or a tone stack.

Lastly, as a third order of magnitude, you will drop some signal from the plate coupling cap, but it's value will determine the low frequency cut-off, with respect to the output impedance it's seeing, in parallel with the input impedance of the second gain stage. Hence, as a third order.. . .



:}





-g