Does a DAC need a pre-amp

[sandyk]

12 minutes ago, Speed Racer said:

My experience is that a quality active preamp sounds better than no preamp or a passive preamp.  

 

 WTF !!!

We are in complete agreement on this occasion.

 

A decent Preamp is usually better able to drive modern Power Amplifiers, which these days often have lower Input Impedances of typically around 10 Kohms, in order to assist with reduced D.C. offset at their Output etc. They also have a much lower output impedance and increased current driving capabilities, as well as often being able to properly drive a quality pair of headphones from their dedicated headphone output jack.

A few CD/DVD/BR players however,do also have a proper Buffer stage (low impedance) at the output which is better suited to driving the later Power Amplifiers, as well as making them less susceptible to Interconnect vagaries.

 

Quote

. The output of a simple pot, may (or may not) mess with the amp's frequency response. The exception is if, instead of a pot, the control is a true "T-Pad". Then there would be no impedance mismatch, but these are few and far between. If you decide to try a passive volume control, my advice is to use the shortest connection between the control and the amplifier possible.... George

 

As George has recommended, use the shortest possible cables from a Passive Preamp to reduce HF rolloff.

[sandyk]

1 hour ago, Speed Racer said:

I could put a system together with very low distortion numbers that sounds like crap. I can put a system together with much higher distortion numbers that sounds sublime. 

In which case they aren't measuring the correct things.

Perhaps partly because we still don't always know what to look for ?

[sandyk]

11 minutes ago, Speed Racer said:

Class D amps? Yuk.....

Yes.

But Dennis luvs 'em !

[sandyk]

7 minutes ago, Cornan said:

 

Wrong. Mytek Brooklyn DAC have both digital and analogue volume control.

 

 Is this the one with the optional 12V PSU ?

If it is, have you tried a very low noise12V PSU with it ?

I noticed that the one I looked at said a "Worldwide PSU"

Is it SMPS, and if it is have you tried the earthing trick"?

[sandyk]

8 minutes ago, esldude said:

His batting average is the same at all times.  This is just one more swing and a miss among many.  

 

 Yes.

It's in the same league as suggesting putting Analogue Out through a freaking A-D Converter.(no matter how good it is !)

ANY added conversions can only result in some degradation.

That's not Rocket Science either. See what George has to say on  the subject.

[sandyk]

33 minutes ago, GUTB said:

Okay so, there’s this thing called a “digital potentiometer” — a little IC package that costs a few dollars that does a stepped resistor ladder. Is this what the Mytek uses?

 Perhaps you are thinking of something like the PGA2310PA I.C.

 It was used as part of the design of the attached magazine article page.

 

 

[sandyk]

2 hours ago, mansr said:

There are also devices like this one: http://www.ti.com/product/pga2311

That's basically a variable resistor combined with an opamp to produce a variable-gain device with fixed input and output impedance.

 

I mentioned the PGA2310PA back on the previous page.

Attached is a simplified diagram from Silicon Chip magazine.
However, this device is a little different in that the gain of it's opamp is also digitally programmable.

[sandyk]

Just now, esldude said:

Or do we call that analogical?

 

Sounds like a good description provided that when you say it, you have a small pause between the "a" and the "l" at the beginning.

[sandyk]

George

 I agree with 4est.

A good Preamp can  have distortion that is not accurately measurable without using special techniques.

Once you get close to 4 Zeroes in the distortion figures you need to use measuring techniques similar to those suggested by Douglas Self. A good Preamp will also be able to drive today's more recent SS amplifiers with input impedances around 10K ohms far better than any Passive Preamp, even when they use VERY short connecting leads to reduce HF rolloff.

Many people will also tell you that a very wideband Preamp sounds better .

My own DC  coupled Class A Preamp has distortion figures of very close to 4 Zeroes, and can even drive a terminated 75 ohm line without problems. It has a gain of only 3.2 times, so that excessive attenuation is not normally needed.

 

 According to a U.S. friend who also built one, it is :  

 

Quote

Freq. Response Flat 10hz to 500khz; falls off 2db at 1.6mghz. (my generator goes up to 3mghz).
Output w 50ohm load= 12volts rms= 2.9 watts
Output w 100 ohm load= 12 volts rms= 1.45 watts

 

He didn't measure S/N as "there is none to be heard with my CD player connected & the gain wide open."

 

P.S.

 Have you read any of the Amplifier design books by Douglas Self ?

[sandyk]

2 minutes ago, gmgraves said:

have several amplifier design books including one by Doug Self

 George

 Is it "Audio Power Amplifier Design Handbook 5th ed - D. Self" ?

 

Alex

[sandyk]

59 minutes ago, esldude said:

So what is the difference in the analog output stage of a DAC/pre and an analog pre-amp.

 

 Basically, it's usually Current Output capabilities, and more often than  not, a discrete output stage using higher supply voltage rails with the Preamp. Your average DAC would be lucky to have  + and -15V supply rails using a generic opamp for their output, sometimes even using a Dual opamp, and most likely also using 100 ohms series output resistors,(or higher)  as many I.C.s don't like driving capacitive loads of >100pF as with some Interconnect designs. 

Some better DACs MAY also use an output Buffer I.C. which makes them less sensitive to Interconnect vagaries.

 Many discrete Preamps have > + and - 20V supply rails and a markedly more sophisticated PSU, most likely not SMPS as with many DACs either. Neither do they usually need to share power supplies with a Digital area either.

With my own DIY Class A Preamp, you will see it is of a dual mono construction with separate sophisticated PSUs for each channel, and it is fed low voltage A.C. by 2 separate external transformers. This Dual Mono type construction also results in a considerable improvement in channel separation, especially in the HF area.

(Click twice on the image for a much larger image.)

[sandyk]

6 minutes ago, esldude said:

Many, many expensive preamps now use the same op-amps as the DAC/pre units.

 Hi Dennis

 You will see in the photo that there is only 1 OPA134PA opamp per Channel PCB. This is only for DC offset correction as this Preamp is Direct Coupled.  The Class A Bias is adjusted for approx. 100mA in each channel, hence it's power output capabilities of :

Output w 50ohm load= 12volts rms= 2.9 watts
Output w 100 ohm load= 12 volts rms= 1.45 watts

Several other people have constructed my design , some for use as a Headphone Amplifier where the only real difference is the output resistor value., which is chosen to match the headphones used.

 100 ohms is used in the Preamp version, and has been found by several constructors to be ideal with most 1M long Interconnects.

Alex

 

 Attached is a photo of a C.S.I.R.O friend from Brisbane's H.A , where he managed to squeeze in 2 small transformers as well.

[sandyk]

8 hours ago, 4est said:

It has been said that one of the issues is that analog engineers are not always the best digital engineers and vice versa.

 

+1 

[sandyk]

6 hours ago, Panelhead said:

 

   Very nice design. One item that MAY improve it is using shaft extenders and moving all the switches and pots to the back. If the chassis is tall enough the shafts can pass over the top of the circuit board. 

  This also allows moving the power supply to the front and placing the amplifying circuits next to the inputs and outputs.

  I find this is an audible improvement. This way the signal wires can be a few inches in length. 

  

 The Chassis  is a 1U rack case. This makes implementation of the shorter signal paths harder to implement.

Yes, I did consider that possibility originally.

However, I have not noticed any audible degradation using the existing layout.

The PSU lead lengths have a much greater impact,  when they are optimised as I mentioned., especially equalising the  0 volt PSU lead resistances !

 I also have a larger, beautifully made ceramic selector switch that unfortunately, I can't use.

I have compensated for the different lengths of PSU leads by using thicker gauge wire for the + and -  of the more distant PCB, and doubling up of the o volts leads on this as well ( 2 green wires)

I could have used relay switching at the RCA input sockets, but didn't want to draw current from just one PSU PCB. The copper traces of the Relay Switching PCB may also degrade HF channel separation!

I use a different layout for my headphone amplifier which uses the same channel PCBs which were designed to my requirements by collaboration among members from another forum.

 The H.A. version now also uses a DACT2 Attenuator.

Incidentally, the width of the main PCBs makes it difficult to further optimise the layout when using this 1U rack case.

[sandyk]

5 minutes ago, mansr said:

So get an expert in each and let them build a DAC together.

 

 Perhaps you should collaborate with Marc ?

No, I am not being facetious !

I don't doubt that either of you have vast experience in your individual areas of specialisation.