Is it too much gains for my FP III?

[audiblesoundwave]

Hi,

Can my Bottlehead FP III work with the amp with the following spec.?
Will it have too much gains that the noise would be heard from the speakers with no music playing?

Power amp spec:

Power 36 Watts per channel
Frequency Response 20Hz - 85kHz +/- 0,2dB at rated power; 4Hz - 110kHz -3dB @ 1W
THD with AABB <0,1% @ 1W; <1% @ 10W; max 2% at rated power into resistive load
S/N Ratio 89 dB (unweighted) ref 1W
Input Impedance 100 kOhm
Input Sensitivity 600mV rms input for rated output power
Gain 28x (29dB)

Thanks,
Milton

[Paul Joppa]

What is the sensitivity of your speakers? Your question can't be answered until we know that number too.

Your amp is quite sensitive, so unless your speaker are of low sensitivity, you will probably need to attenuate the amp's input to optimize the signal-to-noise ratio. It's not that hard to do.

Meanwhile, if you have not read my white paper on signals and noise (linked off the Bottlehead "community" page) you might give that a shot. I'm sorry that it's fairly complex - we'll help!

For a full analysis, you must also know the output voltage of your signal source(s).

[audiblesoundwave]

Hi Paul,

My speakers sensitivity is 87db!

Thanks,
Milton

[Grainger49]

Milton,

This is a FAQ, many folks need lower gain with the FP III.  That is why you received a set of resistors that can pad your inputs with the kit.  Look on page 24 of the manual and you will see what to do.  PJ will probably give you the resistor to use and if you amp needs a voltage divider on the input of your amp he will most likely describe that.

[Paul Joppa]

Hi Paul,

My speakers sensitivity is 87db!

Thanks,
Milton

OK, that's good. You amp will just make 102dB with those speakers - that's my usual estimate of enough power. So you will need a peak output of 0.6v, the amp's sensitivity. The Foreplay with the stock 33K input resistors has approximately unity gain at 3 steps down from maximum gain, and a CD player will usually put out 2 volts, so you will probably want to attenuate further by a factor of 0.6/2.0 - about 10dB. Use the 182K resistors (12dB) and you should be good. I expect you will be able to hear some tube rush with your ear next to the speaker, but not from your listening position.

It would be a bit better to put the attenuation at the amplifier input, but that's more difficult to implement. In this case I don't suggest you do that unless you really need to, for example if you get speakers that are much more sensitive.

[audiblesoundwave]

Thanks Paul!

I feel I am slowing learning more.  Thank you for your patience.

One more question.  In your reply, you mentioned "....so you will probably want to attenuate further by a factor of 0.6/2.0 - about 10dB."

How do you map the factor of 0.6/2.0 to about 10db?

Milton

[Paul Joppa]

The formula is dB = 20*log10(voltage ratio).

For what it's worth, a decibel is (of course) a tenth of a Bel. (The Bel in named for Alexander Graham Bell.) A Bel is the base-ten logarithm of a power ratio, so every Bel represents a ten-times increase in power. Power is proportional to voltage squared, so a Bel is the log10 of a voltage ratio squared, or twice the log10 of a voltage ratio. All very logical if you know the whole history of it, but confusing as heck if you don't.

[audiblesoundwave]

Thanks Paul!

Things are slowly sinking in!  All these is to manage the gain structure and to match the pre-amp and the amp better for a given set of speakers. :-)

One more thing!  I am using Goldpoint 10K attenuator and I remember I read that there would be approx. 2.5db gain reduction.

What does this do to the result?

Thanks again,
Milton

[Paul Joppa]

...
One more thing!  I am using Goldpoint 10K attenuator and I remember I read that there would be approx. 2.5db gain reduction.

What does this do to the result?...

You may need to go to two steps below maximum instead of three, to get the highest level you want. But individual variation is large (music genre, room acoustics, personal taste, etc.) so I think it's time to try something and see how it works. Engineers call it "successive approximation"...  :^)