Metrum Octave Dac - What are the Chips used

[ThorstenL]

Guys,

If you go back and correctly read what I wrote, and the context in which I wrote it...

Let's all take a chill pill and look at the facts.

In the end, if someone likes the sound from Transistor Amp's or Tube SE, from Non-Os CD or for SACD, it is their choice and there is neither the need to become defensive about one's choice if it criticised or let things degenerate into name calling.

The technical facts are well documented, but as we all know, traditional measurements and "modern"technical features do not reliably produce good sound.

So please, "de gustibus non disputandum est".!

Let's get back to observed facts.

Ciao T

 

[Salas]

Guys, chill out on the personal calls stuff. Some such posts removed.

 

[jitter]

Hey guys, don't let a little slap on the wrist keep you away from this thread...

Below, is the link to a rather interesting hobbyist review comparing the Lamipzator DAC level-4 to the Metrum Octave.

The Audio Eagle - DAC_Comparison / Does Digital Finally Rock?

I found this an interesting read. It's remarkable how much a no-frills design as the Octave can achieve for so little... I mean, a very straight forward layout without esoterics, and yet almost all the way up there with the lampizatOr according to the reviewer. And this lampizatOr too is a NOS DAC that seems to be in "overwhelming demand". I'm wondering how long it'll take the big boys to catch up on this...

Now that my Octave has had nearly a week's worth of burn in, I'm pleased to say that the sound is maturing like a good wine. The wonderful sound of this DAC keeps me digging up CDs I haven't heard in a long time, a very good sign.
I'm guessing another week of burn in and it's fully matured...

 

[Ken Newton]

Hey guys, don't let a little slap on the wrist keep you away from this thread...

It's not the wrist slap. It's that the wrist slap appears to be oddly unilateral. qusp's Comment #298, which contains the most ugly insult of any said here, inexplicably remains undeleted by Salas.

 

[Salas]

I can't be following and evaluating the whole thread each time a single report comes to the reports que. Remember guys, report what is objectionable or it may slip. The big picture contains hundreds of active threads, the mods are few plus they can't possibly be on-line all together at a given time or omnipresent.

*298's gone too.

 

[Ken Newton]

I can't be following and evaluating the whole thread each time a single report comes to the reports que. Remember guys, report what is objectionable or it may slip. The big picture contains hundreds of active threads, the mods are few plus they can't possibly be on-line all together at a given time or omnipresent.

*298's gone too.

Thanks, Salas. I hadn't considered the possibility that my comment had been reported, while the provocative comment of another here had not been.

 

[Ken Newton]

I found this an interesting read. It's remarkable how much a no-frills design as the Octave can achieve for so little...

From a features standpoint, what I find interesting about the Metrum, aside from NOS, is that the DAC chips (assuming they are, DAC8580) require no I/V or putput stages. The datasheet for the DAC8580 says it has an 18 ohm output impedance and can deliver +/- 25mA. There are four of these per channel in the Octave. Having such a robust analog output stage integrated right in to the DAC8580 may also be a key to the sound of the Octave.

 

[pieter t]

From a features standpoint, what I find interesting about the Metrum, aside from NOS, is that the DAC chips (assuming they are, DAC8580) require no I/V or putput stages. The datasheet for the DAC8580 says it has an 18 ohm output impedance and can deliver +/- 25mA. There are four of these per channel in the Octave. Having such a robust analog output stage integrated right in to the DAC8580 may also be a key to the sound of the Octave.

My Octave manual specifies output impedance of the dac being 82 ohm.
Based on this spec I doubt the dac chips being 8580's; four of these in parallel should yield an output impedance of some 5 ohm.

 

[Ken Newton]

My Octave manual specifies output impedance of the dac being 82 ohm.
Based on this spec I doubt the dac chips being 8580's; four of these in parallel should yield an output impedance of some 5 ohm.

What is typically done when paralleling low impedance sources is to place a resistor in series with each output. This is done so to suppress any tendency for the low impedance voltage sources to fight each other. With a 82 ohm net output impedance, a 310 ohm resistor would be in series with each of the four DAC8580 outputs - assuming, the Octave, indeed, does utilize the DAC8580.

 

[jitter]

My Octave manual specifies output impedance of the dac being 82 ohm.
Based on this spec I doubt the dac chips being 8580's; four of these in parallel should yield an output impedance of some 5 ohm.

The outputs of the DAC IC's are not directly coupled to the RCA connector, but through 330R resistors.

Passage from the 6 moons review of the Octave:

"These chips were designed to drive 75Ω loads so there's no need for any output buffer. In the Octave you have 4 chips per channel working in parallel. 8 chips total are connected by eight 330Ω resistors to create an output impedance of 82.5Ω."

Edit: apparently the output impedance of the DAC-IC's is disregarded in the quote. 4x 330R resistors in parallel calculates exactly 82R5.
Edit 2: with 18R+330R x4 parallel calculates as 87R. Measurement of DC resistance is hampered by a slight DC-offset (5-6 mV) on the output so I measured resistance both ways around (core/shield) and averaged the two to get 88R7. Minus measurement lead resistance of 0R6 gives 88R1. Very close to the calculated 87R, if you keep the usual tolerances in measurements and components in mind, the DAC8580 still seems to be the most likely answer.

 

[regal]

I suspect that much of the subjectively annoying aspects of CD audio stems from the dynamic interaction of the MULTIPLE (at least two) sinc filter responses across the recording/playback chain. Remove either the ADC sinc response via apodising, or remove the DAC sinc repsonse via NOS and, viola', much more natural sounding reproduction results. I would add that while eliminating one of those two (at the least) sinc responses in the chain substantially improves the sound, it may also be that removing BOTH would improve the sound to it's ultimate point. Something which could be done using high sample rate audio.

This is why I always favor studio's who properly use the PM-2 and then I playback at home with a PMD100 DAC. There is more to it than the gimmicks HDCD advertised, a lot of people missed the intellegence behind the marketing, matching the adc filter with the dac filter makes a lot of sense. But no one ever got the big picture of what KOJ did.

 

[pieter t]

But no one ever got the big picture of what KOJ did.

Who or what is KOJ?

 

[abraxalito]

Keith O Johnson - the 'Prof' : Information about Reference Recordings, "Prof" Keith O. Johnson and HDCD Recordings

 

[pieter t]

Thanks for clearing Abraxalito.

Ken, Jitter,

You're right about these load sharing resistors and the resulting output impedance.

I have experimented a bit with the input circuitry of the Octave, as there is a two dollars or so pulse transformer in it, and the digital output of my Rotel CDP also had (I removed it) a really bad quality pulse transformer (easily to check looking at SPDIF signals on the oscilloscope).
Input circuitry of the Octave looks pretty complex, and I wonder if it might be possible to make a more direct I2S connection between CDP and dac.
Ideas?

 

[abraxalito]

Thanks for clearing Abraxalito.

Welcome.

I have experimented a bit with the input circuitry of the Octave, as there is a two dollars or so pulse transformer in it, and the digital output of my Rotel CDP also had (I removed it) a really bad quality pulse transformer (easily to check looking at SPDIF signals on the oscilloscope).

In my experience (which admittedly is fairly limited) the sound quality issues surrounding SPDIF inputs relate to common-mode conducted noise, more than pulse fidelity. The route taken by the 0V return is crucial - if it just connects to the ground plane (rather than being star-earthed) then noise is conducted through the plane giving a murkier sound. So try disconnecting the 0V from the plane (if that's how its implemented) and giving it its own dedicated wire back to the PSU input 0V.

Input circuitry of the Octave looks pretty complex, and I wonder if it might be possible to make a more direct I2S connection between CDP and dac.
Ideas?

If the I2S is not isolated it has the chance of making CM noise issues worse.

 

[anton k.]

I've also ordered Metrum Octave and expect delivery around end of April.
Interestingly, one Octave owner reported sound improvement, when he substituted this cheap pulse transformer with the Lundahl LL1572:
Metrum Acoustics Octave: A NOS digital filter-less DAC - Home Source Components - www.Head-Case.org - Page 4
I intend to do the same some day.

 

[jitter]

In my experience (which admittedly is fairly limited) the sound quality issues surrounding SPDIF inputs relate to common-mode conducted noise, more than pulse fidelity.

The impulse transformer has a CMRR of 52.1 dB (100 kHz, 110 ohm), so its presence reduces CM noise.

The most important reson it's there, is to prevent a groundloop between DAC-AMP-CDP if the analogue outputs of the CDP are also connected to the amp. My Micromega DAC1 doesn't have this isolation on the SPDIF, so I was never able to connect the CDP to both amp and SPDIF without audible hum (admittedly when volume was turned up quite a bit) that wasn't there when the loop was broken.

The route taken by the 0V return is crucial - if it just connects to the ground plane (rather than being star-earthed) then noise is conducted through the plane giving a murkier sound. So try disconnecting the 0V from the plane (if that's how its implemented) and giving it its own dedicated wire back to the PSU input 0V.

The RCA connects to the transformer only as you can just see in the pic in post 211.
This way the SPDIF is floating with respect to the grounds in the DAC.

Personally I'd leave it in or replace it by a better 1:1 impulse transformer.

 

[regal]

Welcome.



In my experience (which admittedly is fairly limited) the sound quality issues surrounding SPDIF inputs relate to common-mode conducted noise, more than pulse fidelity. The route taken by the 0V return is crucial - if it just connects to the ground plane (rather than being star-earthed) then noise is conducted through the plane giving a murkier sound. So try disconnecting the 0V from the plane (if that's how its implemented) and giving it its own dedicated wire back to the PSU input 0V.



If the I2S is not isolated it has the chance of making CM noise issues worse.

I agree, and still find spdif superior in many ways to i2s direct when implemented correctly with a decent pulse transformer .

Peiter, Many consumer gear connect the primary to the secondary of the pulse transformer making it useless to get around some silly regulation, make sure you check for that short-cut.

 

[jitter]

I've also ordered Metrum Octave and expect delivery around end of April.
Interestingly, one Octave owner reported sound improvement, when he substituted this cheap pulse transformer with the Lundahl LL1572:
Metrum Acoustics Octave: A NOS digital filter-less DAC - Home Source Components - www.Head-Case.org - Page 4
I intend to do the same some day.

Thanks for posting this link. It's already made me decide to fit 75 Ohm BNC's and an industrial impulse transformer when the warranty period is over.
BTW, if you have BNCs lying around but don't know what you have, look at the white insulator, if it's thick, it's 50 ohms, if it's thin or even recessed, it's 75 ohms.

 

[abraxalito]

The impulse transformer has a CMRR of 52.1 dB (100 kHz, 110 ohm), so its presence reduces CM noise.

The CMRR only tells us how well the trafo rejects any common-mode signal relative to the differential mode signal. It does not give us an indication of how easily common mode noise enters the DAC as a whole system. For the CM noise the interwinding capacitance is a key factor, not the CMRR.

The most important reson it's there, is to prevent a groundloop between DAC-AMP-CDP if the analogue outputs of the CDP are also connected to the amp. My Micromega DAC1 doesn't have this isolation on the SPDIF, so I was never able to connect the CDP to both amp and SPDIF without audible hum (admittedly when volume was turned up quite a bit) that wasn't there when the loop was broken.

The loop is opened at DC but is still there at RF frequencies via the parasitic capacitances.

The RCA connects to the transformer only as you can just see in the pic

This way the SPDIF is floating with respect to the grounds in the DAC.

That's a helpful pic, thanks. I note a couple of things - the tracks to the Murata trafo run close to the ground plane - there will be some stray capacitance there to the plane. I don't see where the secondary of the trafo goes to - is it going to a differential input receiver? Or is one side of the trafo secondary going to the plane? If so, then that's one point to lift from the plane and wire back to the input 0V.

Personally I'd leave it in or replace it by a better 1:1 impulse transformer.

Murata doesn't tell us what the interwinding capacitance is in the datasheet. Lower is better when considering CM noise.