Max rms voltage output with 2V rms input

[Deke609]

Quick question about voltage gain and maximum output of the BP. The BH circuit description provides that gain @1KHz is 8.3 dB.  I assume this is a voltage gain figure, not a power gain figure.

So am I correct that at 2V rms input and 8.3 dB gain, we get 5.2V rms output?

I ask b/c :

(1) the BH circuit description also states that gentle clipping begins at 18V rms output - which I find confusing. Is this only possible if someone feeds the BP with 7V+ rms input -- e.g., from a really high gain phono preamp?

(2) I am playing around with the idea of getting a custom super permalloy transformer stepped-attenuator to move volume control to the output in my BeePre -- SAC Thailand can make me one that handles 10V rms. I think that should cover me since my input signal voltage is about 2V rms or so. But I just want to check with those who actually understand these things.

cheers and MTIA, Derek

[Paul Birkeland]

I have never seen a preamp's gain given as power gain.  If you had the BeePre hooked up to a 10K load or a 100K load, the power gain would be dramatically different, but that wouldn't be readily observable.  The power gain would also change a little depending on where the fine attenuator control was positioned on the BeePre, which again complicates the usefulness of this specification.

The problem with moving the volume controls to the output will be that you limit how much signal you can send to the grids of the 300B, and an Eros with a potent MM cartridge may be able to overdrive the grid and some DACs will also put out too much voltage for this to work well. Things start to turn downhill at 3.5V RMS of signal on the 300B grid since you will be driving the grid up against the most positive end of the filament, and you really need to completely redesign the circuit in order to avoid this issue.  In reality you can drive the grid up against the filament to some degree for peak transients and likely not notice, but it's not something you would want to be constantly doing.

[Deke609]

Thanks PB. So just to be clear, as long as rms input signal voltage is below lets say 3V (just to be on the safe side), the 300B is happy and not overdriven -- and moving the attenuator is possible.  Do I have that right?

cheers and thanks, Derek

[Paul Birkeland]

Yes, then you have to figure out whether the TVCs are an appropriate load for the 300B and an appropriate choice to drive cables.

[Deke609]

Thanks again, PB.

Yes, thanks, the loading/driving issues need to be pinned down.  I've provided Kittipol of SAC Thailand with what I *think* is the relevant info: cascode CCS loaded 300B, 10uF output coupling cap, operating point (nominally 100V plate-to-cathode, with -10V cathode bias), 500 ohm output impedance, and nominal 100K input impedance of next stage (Kiaju).  Am I missing anything important?

cheers and thanks, Derek

[Paul Birkeland]

It will be interesting to see what they sell you...

[Paul Joppa]

To be clear, decibels are defined in terms of power: dB=10*log(power ratio). But power is proportional to voltage squared and the log of voltage squared is twice the log of voltage, so dB=2*10*log(voltage ratio)=20*log(voltage ratio).

[Deke609]

I haven't committed yet. I inquired about using their stock STA-522A TVC and Kittipol responded that it wouldn't work - too much signal output.  But he said he can custom build me transformers on much larger super permalloy cores that will work with the amp's operating point and output cap [edit: and the plate load] (info about which he asked for).  And they ain't cheap -- but then again almost none of the tweaks I try ever are. (Speaking of tweaks, I've ordered some Silk permalloy 11,000H 7,000H grid chokes to try in the Kaiju -- just to see).

If I pull the trigger on the TVCs I will post pics of what I get, the install and, if anyone cares to help me with measurement pointers, any test data that can be gleaned from a signal generator and scope.

cheers, Derek

[Deke609]

To be clear, decibels are defined in terms of power: dB=10*log(power ratio). But power is proportional to voltage squared and the log of voltage squared is twice the log of voltage, so dB=2*10*log(voltage ratio)=20*log(voltage ratio).

Aha! Thanks PJ. That clears up some confusion I had about why some dB formulas have 10*log and others 20*log.  It's just a log of square relationship.  Cool.

cheers, Derek

[Deke609]

I've decided to put this project on indefinite hold. Too much money for likely little to no benefit. The prices for 26+ position stereo switches are a deal breaker.  The whole project was going to cost me a grand or more -- that's crazy.

Plus - I already have a setup that pretty much does for far less money what I was hoping to do with the TVCs.  When I set the attenuator I installed at the input of the Kaiju (nominally 100K) to -15 dB, I can set the BeePre coarse attenuator to fully open when using 16R or 32R secondaries, and set the coarse attentuator to only -9 dB when using 64R secondaries. That's good enough. 

But if anyone else ever tries the TVC at output idea, I'd be really interested in hearing how it works out.

cheers, Derek

[Doc B.]

We developed the attenuator we use after we tried a whole range of existing attenuators including an S&B based transformer attenuator. PB spent a lot of hours working the math for the rather complex design and it sounded better than anything else we tried.

[Deke609]

Thanks Doc. And just to clarify: I like the BH attenuator. I wasn't looking for a better version of an input attenuator. I just wanted to move attenuation from the input to the output to maximize the signal to noise ratio. But as I mentioned above, I think I've pretty much achieved the same result by putting a -0, -6, -9, -12 and -15 100K attenuator at the input of my Kaiju.

cheers, Derek