How important is the pre-amp?

I wont add too much redundancy to what has been said here. MillerCarbon made me chuckle.

But an overlooked fact is that in the majority of cases the volume and balance controls (essentially two sequential attenuators, so the distortions are doubled keep in mind) are one of the most non-linear and distorting parts of the preamp. Most are in fact pretty bad.

In fact, my current design project is a preamp that eliminates traditional volume potentiometers and replaces them with one of two technologies at different price-points, both are substantially better, and both lend themselves to very nice remote control.

Let’s put aside volume control in the digital domain which has an entire raft of issues, but under the right circumstances can be very good. 99.999% of people have never heard those right circumstances.

Your basic volume (or balance) control is a carbon powder strip with a wiper. It is noisy, prone to dust and moisture, and does not track evenly left to right.

The next step up are Noble/ALPS style, which are still carbon powder, but in a plastic carrier. They are smoother, largely sealed and less prone to issues. They are expensive and in doing so track better left to right. They are still very imperfect.

At the top of the food chain are pairs of discrete, precision, metal-film, low noise resistors. You switch in two matched pairs of these for each volume level - maybe 64 pairs. A smaller number suffice for the balance. Stepped attenuators can cost a HUGE amount due to the labor and the 128 position switches, and are large in size, don’t play well with remote control. BUT computers and relays can come save the day ( as i am doing).

In the middle are what boil down to chips with 128 pairs of resistors that can be switched in and out. They are chips resistors, so imperfect, but i speculated that they are better than the lousy alternative. They can be, but if you use them normally or per the data sheet they are truly awful. It took me many weeks to figure out how to us them correctly. (and a really great chip engineer at the fab firm that made me his fun project).  These are a little complex but overall can be cost effective in somewhat reasonably priced designs - the biggest cost being yet another very quiet power supply (or else.....).

So, the little things are way more complicated than you might imagine. Now that i have prototypes floating around, i can’t go back.

By the way, switches have their own issues, and a preamp does have an amplification stage. If its a Phono (RIAA) it has a very complicated amplifier with up to 2,000X amplification which makes noise a (expletive omitted) nightmare.

So yea, all you need to do is be perfect and it works great.

@cheeg raises some interesting points.

Point #1: Wikipedia URL with comment "you can see its not quite that simple". Actually, it is that simple. he wikipedia article manages to make a simple topic complicated by badly explaining older dB references in the context of impedance matching. Impedance matching is, in fact, one of the key jobs of a preamp. Sometimes it’s needed, sometimes not. If only high end products adhered to some standards or conventions (they often don’t) the problem would be far simpler. In fact that article points out just how wildly out of spec most DAC outputs are today. nominal line level - under a volt rms (caveat I always thought it WAS updated to 1V rms, but no matter). Many DACs now put out 2, 3 and even 4V this means a 15V swing at the INPUT to your preamp. Are they mad? Which brings us to point #2.....

No, they are not mad. Just narrow minded (IMNSHO) or maybe lazily practical for heir own product anyway. If, as a manufacturer, you wish to offer a DAC with volume control and claim it's basically a control center for all-digital systems, you need sufficient gain (output level) to drive any amp. And they have gains that run the gamut from low to high.

So what do you do? You put out a crazy high voltage level and then attenuate it. Its a preamp. Let’s say that again, it **IS** a preamp, right there at the output of your DAC. The only question is "is it a good one"? Some are, some are not. Volume control in the digital domain is fraught with issues; so that may be compromised. With R2R DACs its largely impossible except via DSP interpolation. Next, too often the output stage (line circuitry) is an op-amp chip. Unimpressed. To be fair, some are excellent, discrete, even tubes. but most are not.


So the idea that the preamp solves the impedance matching problem, if an output stage exists (even an op-amp) is not really valid. More likely the answer to why this "simpler" chain is not in fact superior is:

1. it’s as simple as you might have thought
2. The volume control is compromised (Nice word for "sucks")
3. The line circuitry is compromised
4. They really DID simplify the signal path, by leaving out the line circuitry and placing the volume control at the output and the impedance is now high (yep seen it)

OTOH i have used high quality digital volume (done in the DAC at 32-bits resolution, with all truncation below the 18/20 bit threshold) and a decent (not all that impressive output stage - essentially what one would have with a preamp too) and it soudns very, very good.

And then there’s the elephant on the table. Most tube components ADD euphonic distortion. We like it. this is not really a bad thing - its also what a Piano’s sounding board does. but in that case technical arguments go out he window,and we pivot to music theory.

And, I kid you not, i have a couple of Pianos to try out.

happy listening, and maybe, playing.
G

A few comments on the "transformers are the answer" topic.  I agree that especially when you have an impedance mismatch, particularly the wrong way (high into low), that needs **transformation** rather than gain or attenuation (fractional gain),a transformer could be very good.

Its that "could be" part.  They have many issues - several notes above, several not. And like most problems, they can be mostly solved with money, lots of money, and few have attempted it.   No market maybe, or very very hard.   An ideal, variable, volume-controlling transformer must be VERY expensive.

Ditto resistors.  They have issues.  You can buy better ones.  But.... I actually think even modestly good (name brand 1% metal film) resistors are pretty transparent. And with resistors there is a mature market for VERY good ones, and while they too are expensive (like 10-50X the cost of modestly good ones), even at 10-50X they are still less than half a buck (Caddock, DALE mil-spec noise controlled, non-inductively wound, blah blah) or were the last i checked which i admit was years ago.  To be clear, you need a dozen to hundreds of pairs depending on how you implement the solution.  I use 11 pair.  Oh, and a micro-controller  to control the mess.


But look a the big picture: we are now worrying about two more resistors in the signal path, one series one shunt. There are likely dozens in your signal path already. Are you replacing all of those too? Or removing them and inserting.... what?

I firmly believe that for 99% of applications, a true, discrete, stepped attenuator using a minimal number of resistors, great contacts, and good (not awesome) resistors is pretty darn transparent. I can say for sure its revelatory compared to what nearly all of us have heard - and I can do A-B-C side by side tests since i have them all, independent of the preamp circuitry, with a switch to do the rest evaluations.

For just 1 added component, saya CD player, we only need 1 out
So you are saying a simple DIY 44 step attenuator + some resistiors, = beats high priced pres?
Yes??

Well....sorta. Mine is intended to go into volume production in products so it s a bit more complex. It was never intended to be independent from a preamp or integrated amp, and i have never used it stand-alone.  Mine involves 4 inputs, 32 x 2dB steps, a bunch of relays (not cheap if you want good ones to control all this stuff and get you down from 128 to say 44 resistors) a display, a micro-controller, an IR receiver, and a TON of code. The code is a big challenge - especially controlling the IR, synchronizing actions, displaying where you are, etc. but the results, yes, are clearly superior to my monolithic ladder chip solution (also intended for volume production, in fact sooner) and that is superior to an ALPS or Nobel POT (IMO).


You will also have to design and fab some fairly complex mixed analog and digital circuit boards.


A project like this is not for the feint of heart.

There is a dutch guy selling a complete attenuator kit (two units, dual mono i believe).   I think his are intended to go into an amp, but you could build a box and a power supply.  No balance to the best of my knowledge (which, along with mute, is a logical nightmare BTW since both are stateful and you don’t want to swap states wrong). 


For DIY i’d look at the Dutch kit. I suspect his day job is with Philips. Can’t recall his name, Google is your friend. I really wanted remote control that was as good as the best rotary pots. I suspected they could be better (even the monolithic ones) and they are. See my post way way above.

Noble is a good place to start…..

???? for what?

It’s about the size of the signal.
The signals from sources are very small and delicate...

Truer than most realize - gain from low output MC to line is about 1500X.

But i believe this thread would benefit form a very clear distinction between a **phono stage** and a line stage preamp with control circuitry. he question is clearly about the latter.

That said, its the existence of the phono (mixed analog and digital) that sometimes makes a preamp totally mandatory rather than a convenience and sonic choice.

If I have just a power amp hooked up I know you need one.Then it comes to tube or soild state.I have tried both and use many different brands.I still like switching things around.

The point of this thread is ... no you don;t. But is it beneficial?  You can use a DAC with volume control straight into your amp. Or a regular DAC with Roon and DSP based volume control. or .....  But as has been discussed extensively (read for the details that need not be lowed yet again) there are issues.

a key word that seems to have been overlooked. IMPEDANCE.

Actually both MillerCarbon and I went into it.  I did very specifically.  And basically, yes, a hgih output Z and low input Z is a bad combination. Plus some (very long) cables may have some issues being driven by a high Z output Z due to capacitance.

... should have continued.  But a DAC **can** have a solid, low-output impedance.  That doesn't mean they all do, or that they are all good ( the vast majority of chip opamps, sigh).  So again, there is no right answer, no simple answer. You need to look into what you have, and evaluate if its lacking. Sorry, math.