System Gain, Volume, and component matching?

I'm facing a bit of an unknown regarding proper gain matching of components and could really use some help in understanding this. First, with my current setup I typically listen at ~60 out of 99 on my Rowland Pre (maybe 90db or so). Here is what I believe my component gain structure is right now:
APL3910 4V output
Rowland 20db overall gain
JC1 amps 1V=28.3V out (56.6V max power)
Magnepan MG3.6 85db at 2.83V (4ohm load)

Am I correct in saying that to reach 85db at 1m on a continuous test tone, my pre would need to feed the JC1's 100mV?

If this is true, it seems that my pre would have to attenuate that 4V output from the APL a huge amount? But given the Rowland Pre's 20db overall gain, doesn't that suggest that: 1.) at max vol setting, Vout/Vin = 100? 2.) With APL outputting 4V, at max vol setting, the pre would be outputting 400V (no chance). 3.) If the volume knob is linearly calibrated, that would suggest that 4V per vol point (0-99) (no chance).

Could someone help me understand this? As I said earlier, I usually listen at ~60 out of 99 on the Rowland. I'm replacing the Rowland with another Pre that has 8db of gain (Rowland has 20db), and I'm sending my APL back to have a linear power supply upgrade which will drop output from 4V to around 1.5 to 2V. If I used the mathematical approach above, I wouldn't be worried about a thing. However, from a practical point, I currently listen at 60 out 99 on the Rowland, will be decreasing Pre gain from 20db to 8db, and will be reducing APL CD Player output from 4V to possibly 1.5V.

So, with those changes, am I going to be craving a pre that will "go to 11"? Will I run out of gain? And finally, I've heard people mentioning problems with too high of gain in a system causing certain distortions and very little (e.g. 7 to 9o'clock) control of volume, but is it possible to have too little gain when source is only a line level source with outputs in the 1+V range (No phono in system)?

Lots of questions in there I'm sure. I really appreciate any help in understanding this matter.


Hi Jordan,

I believe the formula for dB is 20 * log(Vo/Vi), so I think instead of 400V you get 40V max out of the Rowland, which is still not likely. I'm guessing that the Rowland limits the input (clips?) in some way. My understanding is that the Redbook standard sets 2V as the max output level of CD players, so it seems reasonable from a design standpoint to limit the input signal.

It is true that having too much gain in the preamp can limit fine volume control. There are people striving for a pure signal that omit the preamp completely and use passive volume controls. So, I'd think the 8dB gain of your new preamp would more than suffice, though the 20dB of the Rowland is not excessive. It seems the issue is with the CD player.

I have a friend with the JC-1s running a passive attenuator. He's looking for an active preamp. Would you explain what you didn't like about the Rowland and what preamp you are replacing it with? Thanks.
By comparison, the JC1's are moderately sensitive with the range from 0.75 to 1.5. Lower being louder. The Maggies are on the low end of the scale and the CDP is twice the standard. You already know all that. What you might not be taking into account is the logarithmic scale. Still, 60% of the max seems unusual for that setup for normal listening levels. Must be a big room... I said, IT MUST BE A BIG ROOM :)

I went from a pre with 20 db gain to a passive and the volume setting went up marginally. Not nearly double. That 20 db of gain has probably never been used.

I also changed a jumper in a DAC that doubled it's ouput from 3V to 6V and that did halve the knob rotation. Even with the passive, that was way too much. I'm dealing with a more senstive amp and speakers though, 0.8V and 91db. I'm also using balanced cables, which may be an option as those outputs can be double.
Of course the output of the CD player depends on how loud the recorded music is when you measure it. Two volts is the maximum rms voltage from a CD player when all 16 bits of the CD media are used, and this is called the 0 dB level. 20 dB down from this is an easier-to-use "-20dB reference" level.

A Denon test CD (CO-75046) is very useful for evaluating signal levels. I have been doing this just recently to set up my Behringer DEQ2496 digital equalizers. These have 24 bit A/D and D/A, but if the signal level going through the DEQ is too low you may only be using 8 bits or so...not good. I now have my gains adjusted so that when I play a typical disc the peak level (overall for the whole disc which is loged by the DEQ) is in the range of -6 to -3 dB. In other words, very high but no clipping. This is a lot like setting level for a tape recording.

One track on the Denon CD (#39)is a siren at 997 Hz and recorded at the maximum 0 dB level. My Denon 2900 player puts out 2 volts rms for this track. During most musical passages the player output is only a few hundred millivolts.
Thanks for the responses so far!

Bob, I think the conversion is Bel=log(Vo/Vi), so dB=10log(Vo/Vi), right? So, 20db gain with Vi = 4V from APL should be: 20=10log(Vo/4) => Vo=400V. Isn't this right? I love the Rowland BTW, it is very smooth, clean, and articulate...even dimensional. I'm replacing it with a Cary SLP-05, a vacuum tube preamp.

NgJockey, my room is 19'x21'x8', so it's not huge. The maggies are ~85db, BUT, that's a 4ohm load @2.83V, not 1w. I also use balanced cables from the Pre to Amp.

I just got out my radio shack spl meter and measured an average loudness (c-weighted, slow) around 83-85dB at the listening position (11' from speakers) while listening to "Turn me on" by Norah Jones at "60" on my Rowland Pre. For this CD, that's about normal listening for me. This means that at 60/99 the Rowland must be putting out something over 100mV, right? I don't the relationship between distance and loudness for a dipole.

Eldartford, I thought when they say it outputs 2V rms that it means that is the average output, not the max, right? Maybe that's where I'm confused?

Thanks for the help, the general consenus that I should be fine with a 1.5-2V output player and only an 8dB gain preamp?


Jordan, I can't comment on your precise situation, but I have an APL Marantz which, after a recent mod to get the AKM DACs and the new upsampler, has a very low output (around 1.2V I believe, which is less than you will have). I am using it into a Rowland Concerto pre and then into either Rowland 201s or a Primaluna Prologue 5. I'm listening right now (as I type) at a VERY modest level and the Rowland is set at 60.0. If I had less preamp gain, I think I'd be hosed. My advice is to have the APL mod done before you sell the Rowland, just to be on the safe side.
Hi to you like your APL Marantz with the AKM DACS and new upsampler? Also, are the Rowland 201's and Primaluna amps competitive in quality? I'm going to take your advice and hold on to the Rowland until I'm sure I have enough gain...although on paper, I have plenty. ugh!

I think there might be something wrong with Rowland's specs, or they have a highly non-linear volume control. For example, to drive the JC1's (~14.5dB gain) to full power (~800w into 4ohm) would take: 14.5=10log(56.6/Vin) => Vin =~2V. So, technically, I could drive my amps to full power with a unity gain preamp and a 2V output player? And given my amp and speakers...that would be around 110dB!! I haven't even had peaks that get close to that number!



Germanboxers...2 vrms is what you get when the CD that you are playing is maxed out. Music has wide dynamic range so an average value would not make much sense. Now, if the player has been mod they may have made gain changes up or down. Increased gain might be intended to enable you to drive the power amp directly.
Jordan, from the Wikipedia encyclopedia for decibel, you'll find this: "the decibel ratio of two voltages V1 and V2 is defined as 20 log10(V1/V2), and similarly for current ratios."

You are correct that you should be able to drive the JC-1s to full power with a standard CD player. There's also the issue of impedance matching between the preamp and amp, but the JC-1 has a high input impedance so it would be a poor tube preamp that would ever give you any problems.

For example, the Cary SLP-05 specs state an output impedance of 400 ohms, but they don't state at what frequency. I'm guessing your Rowland is at most half that large.
Bob, looks like you're right...dB=10log(Po/Pi) and since P is proportional to V^2, it becomes dB=2*10log(Vo/Vi). Thanks for the correction! Ok, that's an order of magnitude difference: 400V to 40V. This makes much more sense, though 40V is still way too high to feed an amp with 29dB of gain. Can you think of any reason why I'm using so much pre Vol control for such modest listening levels?


Jordan, the APL Marantz is less than a week into break-in after the upgrade but it is spectacular. Honestly, I can't believe the improvement. All the usual superlatives -- in another league, etc. -- apply, and then some. The man's a major talent.

I recently changed from Thiel to Harbeth speakers. I bought the 201s for the Thiels. The Harbeths don't need that kind of power and, with their more intimate presentation, I was curious to try a tube amp. Hence the Primaluna (which is not quite enough power, but it works). Both amps are great, but it's definitely apples and oranges.
Jordan, well two things come to mind. What you consider modest levels (80 - 85dB) is still quite high, maybe not rock concert level high, but high. I wear ear plugs when doing speaker calibration at 85dB. Your listening distance (11 ft) is quite far away and I believe sound follows an inverse-square law of decay. How acoustically absorbant is your room? Nothing "fancy" with your interconnects between preamp and amp? How long are they?

Wish I had $$$ for either the Rowland or Cary preamp. Sounds like you've got an excellent system.
Thanks Bob, since the maggies are dipoles, they do not follow the same inverse law of decay that a point source would, but I'm not sure what relationship applies? I have a Rives test disc and I agree that 80dB is VERY loud with a continuous tone, but with normal music, it's not continuous and with slow response, C-Weighted on the RadioShack SPL, I don't think it's super high. If I listen for more than 30min, I'm listening at slightly lower volumes, but on some songs, I'll turn it up to around 85dB.

I am using a 7m balanced pair of Purist Audio Museaus IC's between pre and amp. Room is carpeted and I have two 16" tube traps, and two 11" tube traps, plus bookcases, etc to diffuse the sound.

Thanks for all the help!

SPL vs Distance for Maggies vs point source is something I just recently looked at. From anpther thread...

Here are some SPL measurements.
Planar is a MG1.6. PS is a Dynaudio Gemini (small MTM).
Signal is pink noise generated by Prepro (intended for setting levels).
Distance is feet.


The first PS column shows the SPL leveling out at 77 dB. I think this reflects room effect. For the second PS column I increased the volume so that the SPL at 12 feet was the same as the planar, 90dB. This required the near field SPL to be a lot higher, 108 dB at 1 foot.

I don't think anyone listens at a distance of 3 feet, although that is close to the 1 meter distance used for the spec. At a more reasonable listening distance, between 8 and 12 feet, the Planar SPL falls off by 1 dB over this distance range, while the PS falls off by 3 dB. To produce equal SPL at 12 feet the PS speaker would need to be about 3dB more efficient at 1 meter. I think that these results are generally in agreement with theory.
Can't comment on the Rowland. Never been fortunate enough to own one. It could be nonlinear for obvious reasons. I can think of a few, that I know of, like that. However, standard practice for actives is to amplify and then attenuate. There would be some loss going from 20 db to 8 but mostly at the end of the scale.

Volume is directly related to voltage. For example; on resistor based passives, the selected resistor will reduce voltage but current remains the same. Reducing the output of the CDP will reduce volume significantly, all other things being equal.

Eldartford - There are many CDPs/DACs that have higher than standard output stages. Consider the AA Capitole or many of those with volume controls that max out at 6V. Wadia has settings to tailor the output. Many preamps such the SFL-2 specifically state not to use more than 2V supplied to RCA inputs (4V XLR) to avoid damage.
Ngjockey...Having a standard, like 2 vrms max, assures compatability. Good thing for the industry. When standards are not observed problems may arise. In this case a high signal level problem can be dealt with quite easily using a fixed attenuator, so that the preamp volume control range is better. I understand why a CDP with a volume control might have outputs with extra gain (so that you can drive a power amp directly) but I think that it ought to include a second set of outputs, fixed at the standard level.
I agree. Absolutely. No argument here.
Hey Jordan, I could not find any specs on the Purist Audio Museaus IC's, but based on the specs for the newer products, which I found in their PDF catalog, the cables have higher capacitance (over 3 times) than Belden 1800F (13 pF/ft). They don't publish a nominal impedance, so there's no way to tell how much of a mismatch there is with the 110 ohm balanced connector. It's my understanding that low capacitance is desirable for ICs.