Input sensitivity, Gain and Preamp matching

I am considering to buy an SET amp which has an input sensitivity of 2.5 volts (rather high) and gain is only 14 db (probably low ?). The output power is 16 watts. My choice of preamp was going to be a pure passive using a Dave Slagle autoformer or at most a single stage tube preamp with only 6db gain. But with the specs of the power amp I am not sure if a low gain or zero gain preamp would be a compromise or not. Can anyone throw some light on this subject ?

The input impedance of the power amp is a healthy 47kohms.
As I recall the amp's sensitivity needs to be about .5v. I tried a passive line stage with an amp having 1v sensitivity once with a poor result.

Just a guess that you need a higher gain pre as well, but others with weighty authority here can tell you.

I'd just scrap the idea of a passive pre altogether no matter what amp you consider. In the long run, I'm guessing you will not be happy.
I'm sure Al (Almarg) will see this. He is the master at these type of questions. Whatever he says, listen to his advice.
One variable in this is the speaker sensitivity. I assume it must be fairly high because it must be able to produce satisfying output by the time the amp has reached about 20-25% of full power or about 4 watts.

So in the average American home that means the speaker is likely about 103 db or so. But we don't know that for sure.

If the speaker is indeed that easy to drive, we are only looking for a fraction of the power of the amp so preamp gain does not need to be high. OTOH if the speaker has less sensitivity more gain will be needed.

So- what speaker is in use, how big is the room and how lively is it?
Active pre-amp would be the safe bet, even more so with a tube or SET amp, but each active pre-amp will impart a sonic signature of some sort. That's just a fact. So if teh goal is to keep that out of teh equation just don't overpay on the passive and give it a try if you must but be prepared to make tweak from there and make changes as needed. Its the usual routine. These things seldom work out just perfectly after the first crack, especially with more esoteric gear like SET amps and passive pre-amps, but things can be tweaked and refined over time. There are many ways to skin the cat which is what helps make all this stuff fun.
Thank you most kindly, Mofi. I should qualify your statement, though, by saying that at best I'm just an amateur master, while Ralph (Atmasphere) is of course a professional master :-)

Pani, Ralph frequently makes the excellent point he states above that to perform at its sonic best a SET should be operated within just a relatively small fraction of its rated power, such as 25%. Unless, and perhaps even if, your speakers are super-efficient, though, such as being in the 103 db area he mentioned, I personally would nevertheless want to have the flexibility of being able to drive the amp to full power, even if it is only a relatively few recordings having particularly wide dynamic range that might require that.

If you are using the single-ended output of a typical digital source, its maximum output won't usually be much if anything over 2 volts. And it will be less than that for many recordings. So with 0 db of preamp gain (meaning, with the volume control set at max, that preamp voltage out = preamp voltage in) you would not be able to drive the amp to full power. On the other hand I believe that some of the Slagle autoformer-based passive preamps, as in the case of the one-stage tube preamp you are considering, provide up to around 6 or 7 db of voltage gain. That would result in their output being double the voltage they are provided with when the volume control is set at max. That would probably allow you to drive the amp to full power with many and perhaps most recordings, but almost certainly not with some.

Perhaps more significantly, if I'm recalling correctly you use a high quality analog source. Most (although not all) cartridge/phono stage combinations provide significantly lower peak output levels than digital sources, for nearly all recordings (the cannon blasts on the famous and nearly unplayable early edition of the Telarc 1812 Overture perhaps being a rare exception :-)).

So it seems to me that even with 6 db of gain you would at best be introducing a significant marginality into the system, that would probably prove to be unsatisfactory for some recordings. Especially if your speakers are not in the 103 db/1 watt/1 meter area that Ralph referred to, or higher. Personally, my instinct would be to pursue a different path.

IMO. Good luck. Regards,
-- Al

Pani hi, before doing anything suggested, what are your source/s, maybe they have the output volts that's needed >2.5v as many do these days, even phono stages

Your amp is high at 2.5v, the input impedance at 47kohm is fine, if your source/s have this >2.5v then yes a 10kohm passive will work great.

But I would be wary of passive TVC's that have a gain switch, as they can "ring" if the gain part is used, in unity gain they seem ok.

Cheers George
Pani, yes as George indicated it would certainly be relevant to know exactly what your source component(s) is or are.

In the meantime, though, in regard to what I alluded to earlier about the peaks of some recordings leaving margin relative to the maximum volume capability of the medium, my curiosity was sufficiently piqued that I've just made some measurements to shed some quantitative light on that.

What I've done is to use the Sony Sound Forge professional audio editing program I have to examine the waveforms of about a dozen different tracks from about a dozen different CD's, both classical and popular, dating from very recent years all the way back to reissues of 1940's and 1950's material. The Sound Forge program can quickly tell me the level of the highest instantaneous peak on a track.

While several of the tracks essentially reached the maximum possible ("full scale") level of the CD medium, which is identified as 0 db, many others peaked at various points between -1 and -4 db, with one peaking at -6 db and another at -8 db.

A -6 db peak will result for that track in the maximum output of a CD player or other digital source being 1/2 of its rated (maximum) output. So for example a CD player rated to provide 2 volts would never put out more than 1 volt for that track.

As I alluded to earlier, my philosophy is to avoid configuring a system in a manner that might be marginal (or worse) even with just a relatively small number of recordings.

-- Al

Hey Al good to hear from you, as you stated, yes a few of the quietest early cd's have low recorded volume and your -6db seems about right for these.

So even if Pani has source outputs 2.5v this will be ample for 90% of the rest of recordings to play at full amp clipping, and if he has 4v from his source/s that will also take care of those quite -6db as well to drive the amp to clipping.
But, and a big BUT, when do we do our listening just below full output (clipping) of the poweramp anyway??? Unless we like blowing amps up to speakers.

All is moot anyway till we find out what source/s Pani has.

Cheers George

03-31-15: Georgelofi
But, and a big BUT, when do we do our listening just below full output (clipping) of the poweramp anyway???
We (or at least most of us) of course don't do that on anything approaching a continuous basis. But it should be kept in mind that the main usefulness of the upper part of the power capability of an amp is typically to support the BRIEF dynamic peaks of those recordings which have wide dynamic range (meaning a wide DIFFERENCE in volume between the softest notes and the loudest notes).

For example, listen to this excerpt from Stravinsky's "Firebird", from the 7 minute point to the end at 10:54, as the music descends to near inaudibility and then builds to a concluding note that on some well engineered minimally compressed recordings I have reaches 100 to 105 db at my listening position.

Or consider this very simple arrangement (female singer with piano accompaniment) of the old standard "All The Things You Are". That is one of the tracks I described analyzing in my previous post. Prior to the last 15 seconds of the track the RMS average level is a very low -31.1 db, with the very highest value during those 3 minutes being -13.2 db. Yet in the final few seconds a peak of -4.7 db is reached. As I'm sure you can calculate, that peak will require 437 times as much power as the average level.

My point is that unless the OP's 16 watt amplifier is being used with speakers having an efficiency in the area of the 103 db Ralph referred to, there will be SOME recordings which on brief peaks are likely to require most or all of the amp's power capability to be used. And unless his source components have output levels that are a good deal higher than average he won't be able to turn the volume control up high enough to be able to utilize that power capability, if his preamp provides little or no gain.

-- Al
My experience is based ONLY on TVC and not other types of passives. But nevertheless, I would like to share.
Your amp's sensitivity is actually very low. The way people would look at the sensitivity would be - if it requires less volts for full power, then it is high sensitivity. If it requires more volts for full power, then it is low sensitivity. For a passive to work with an amp, the max sensitivity that you should go for an amp would be 1.5V for full power. The lower the better. My amp is 1V for full power,and I use a TVC. Or else, like Al mentioned, you will not hear the "full potential" of your system. Also Ralph has mentioned about speaker sensitivity - this also plays very important role in a "passive based" system. Ideally your speakers should be 90+db.
At this point, it looks like active pre would be the way for you, unless you change some components in your system.
Most SETs have low gain because they are intended to be used with high efficiency loudspeakers (usually 98 db or more).
Milpai: My amp is 1V for full power,and I use a TVC. Or else, like Al mentioned, you will not hear the "full potential" of your system.

So long as you can get to the volume level you like to listen to IS ALL THAT MATTERS, even if a passive is full up there will be absolutely NO missing out on the dynamic peaks, so long as the listening volume is as loud as you want to go, even if full up. There is NO brick wall regarding transients.

In fact passives are usually better if they are full up, as there is less shunt resistance to ground, so the source (if a weak tube output) sees an even easier load, which equals better dynamics for them.

It's doesn't matter if you can't send you amp into overload (in fact it's a safety device) so long as the volume is right for YOU!.

Cheers George
Thanks everyone for responding on this one.

My source is mostly analog. A Naim Superline phonostage. I am sure it puts out around 3 volts because my cartridge has an output of 0.5 mv fed into full MC gain of the phonostage.

My speakers are Tannoy Turnberry SE which are only 93 db sensitive (not 103 db). Currently I use a Wavac 300B SET which produces only 9 watts. On 70% of the music I do not hear compression in my room (150 sqft). But there is still 30% of the music which I cannot listen at my desired volume levels. The amp I am considering now has parallel 300B and produces 16 watts (conservatively) per channel. I know I will better with more power but if this is going to solve the power issue with say 95% of my music I am still happy because the combination of Tannoy and 300B SET is amazing.

BTW, isnt it true that a power amp should ideally run at full gain (not full power) all the time ? That is how one can get the best dynamics at any given volume level IMO. All studio amps have a gain control which are always kept at max level. My understanding was that the input sensitivity decides the ease with which one can get the amp to full gain.

Secondly there is this concept of voltage gain amplifiers and current gain amplifiers. Does that have anything to do in this discussion ?
Pani, the Naim Superline has a specified gain of 64 db. That corresponds to a voltage multiplication of 1585. (20 x logarithm 1585 = 64). So the 0.5 mv output of your cartridge under the standard test conditions would be raised by the phono stage to 0.5 mv x 1585 = 0.79 volts. However the maximum levels of some recordings may exceed those standard test conditions several-fold, I believe by something like a factor of 5 in at least a few cases. So there may be SOME recordings on which the max level of the phono stage output might reach 3 or even 4 volts. But as far as I am aware that won't be the case with most recordings.

Also, keep in mind that the difference between 16 watts and 9 watts is just 2.5 db, which subjectively is not all that much of a difference. If you presently cannot listen to 30% of the music at desired volume levels, I would not expect that a 2.5 db increase in power would represent a solution. And even more so considering that on many recordings you may not be able to utilize much or any of that added power if you do go with a low or no gain preamp, because of the amp's low sensitivity (corresponding to a high sensitivity number, as was pointed out).

Re your question about running power amps at full gain, if the amp has a volume control and is used with a preamp having a volume control, yes, chances are that in general it would be preferable to run the power amp's control at max, which in effect takes it (and any adverse sonic effects it might contribute) out of the signal path. On the other hand, though, there may be some cases where that benefit would be outweighed by other considerations. One example perhaps being a need to increase the signal level at the output of the preamp in order to minimize the audibility of groundloop-related noise, and another example perhaps being a need to avoid running the preamp's volume control undesirably close to the bottom of its range.

I believe that your last question relates to what Ralph (Atmasphere) often refers to as power paradigm vs. voltage paradigm amplifiers, which is a separate subject relating to matching amplifiers and speakers in a manner that will result in proper tonal balance.

Good luck, however you decide to proceed. Regards,
-- Al
Pani, If you really are serious about this and plan to keep your loudspeakers, do yourself a favor and get an amplifier with significantly more power. At a minimum you are going to need 60 watts (my room is very similar dimensions FWIW).

IOW in your environment an SET is impractical. The more power an SET makes, the more limited its bandwidth- much more so than with push-pull amps!

If you really are stuck on that SET sound, get rid of the speakers and get something more efficient. The problem I think you will quickly see is there is no free lunch- as speakers become more efficient, getting full bass response gets harder and harder.

Now one of the secrets of the SET 'magic' is that they have a distortion character where as you reduce power, the distortion linearly decreases to zero (unmeasurable). Most push pull amps don't do this, but that is not to say all are like that (for example our amps share this quality with SETs). To do this generally the amp should not have a dedicated phase splitter circuit, which is one aspect that can cause distortion to rise as power is decreased below a certain level.

Secondly there is this concept of voltage gain amplifiers and current gain amplifiers. Does that have anything to do in this discussion ?

The quick answer is 'maybe'.

There are a couple of ways of interpreting this comment. Here is the first one:

Almost all amplifiers express voltage gain, whether tube, solid state or class D. Current gain amps do exist, but not in the context of audio. All speakers are driven by both voltage and current at the same time (voltage and current together are called 'power'), so if you don't have voltage gain you can't make any power.

The second interpretation:

In audio there are two paradigms of loudspeaker drive, the Voltage Paradigm which is the most common, and the Power Paradigm, which is the older of the two. See for more information. However there is a 3rd means known as current drive where the amplifier has significantly higher output impedance than the speaker. This third system never developed into a set of rules for driving speakers and remains experimental. The Power Paradigm (whether people know it or not) is fairly well accepted in the high end audio community; the Quad ESL63 is a good example as are most horn speakers, SET amplifiers and for that matter, any tube amp that does not employ negative feedback to control output impedance.
Thank Al and Atmasphere. Yes, I am darn serious about keeping the speakers and your suggestion about getting more powerful amps is well taken. My hunt is on. And thanks everyone for the caution about matching passive preamp with the low sensitive power amp. It was an important tip at the right time.