MDAC First Listen (part 00101000)

[Steven Toy]

The slot loader will be an optional extra.

My question 'including' referred to when it would be ready. Would there be any delay if the slot loader was specified?

Put me down for this optional extra please.

XSQ as well please, I'm a subjectivist.

The Rega Apollo R has just been fitted with a brand new mech so I should be able to sell it without a problem.

 

[davidjt]

I'll opt for the full width chassis (+ XSQ version) but if they don't happen to coincide, I'd rather install the board myself than wait.

 

[JohnW]

Hi John,

I'd be interested to learn what devices you are using for the MOSFET version? Low power P-channel devices can be hard to find...

Cheers, Jack.

Jack,

I happened to have samples of Vishay Si1901DL and Si1906DL which are 20V parts with Qg of 350pC each - sadly they are no longer manufactured - but I will try parts from Fairchild and NXP.

The Vishay Si1901DL and Si1906DL are dual SOT363 (SC-70) devices and are pin compatible with the BC846BS / BC856BS dual transistors I originally designed the input stage for.

MDAC2 does not require Gain from its analogue stage and I also paralleled x16 devices in each input pair to reduce noise. The MDAC2 is also driven directly from an low impedance DAC array (less then 100 ohms) so the MOSFETS "relatively large" input Qg has little effect.

I'd not consider using such MOSFETS for your Power amplifier design as you typically require x26 Gain which will magnify the IP stage noise & will be driven via a high impedance source so the Gate Qg would start to be an issue - especially as you will need to paralleled devices to improve the Noise performance.

 

[JohnW]

Unless it may be possible to have spdif outputs from the master mdac2, in which case an original mdac could potentially be used

Its possible - but synchronization between channels (total group delay) could be an issue (OK for the rear channels).

When I say "possible" it depends in Dominiks work load, and I expect we will have many higher priority "software projects" such as the FPGA modulator design.

 

[PhilCTTE]

Full width case please with xsq, whatever this means (resistor choice?)

Just checking how much Mdac2 slaves are and how many I,d need for multi channel audio. 6 channels would be 1 master 2 slaves. Then I'd have to think about what I'd power them with to do a proper job.

 

[Sibbers]

Full width with xsq makes sense to me too. If there's really no delay added from going full width and you can fit stuff in easier, bar an extra hundred quid or so, this seems like a no brainer where everyone wins. I can get the MDAC to use elsewhere too, which would be great.

 

[JohnW]

Full width case please with xsq, whatever this means (resistor choice?)

Just checking how much Mdac2 slaves are and how many I,d need for multi channel audio. 6 channels would be 1 master 2 slaves. Then I'd have to think about what I'd power them with to do a proper job.

Phil,

I have a basement full of brand new 8200MBs (Silver or Black) - I could do these for GBP800 a pair + Shipping, I'll also modify them so they sound better

With your VFets for the main speakers....

With the Oppo digital extraction card you would have a nice AV system

 

[JohnW]

bar an extra hundred quid or so, this seems like a no brainer where everyone wins. I can get the MDAC to use elsewhere too, which would be great.

Its GBP100 to cover the development costs + the Chassis 'At Cost'.

I'll have an idea of the chassis cost during my trip.

I'd like to see the Full width Chassis + HQ PSU & Front panel display for lessthen GBP400...

 

[JohnW]

Full width case please with xsq, whatever this means (resistor choice?)

Phil,

The XSQ option (eXtended Sound Quality) sacrifices ultimate Dynamic range performance for what I believe will be better Sound quality by using MOSFETs in the analogue input stage.

http://www.pinkfishmedia.net/forum/showpost.php?p=2650118&postcount=834

http://www.pinkfishmedia.net/forum/showpost.php?p=2650261&postcount=839

 

[JohnW]

My question 'including' referred to when it would be ready. Would there be any delay if the slot loader was specified?

Hopefully not - it looks like I'll need to spend more time in Asia to insure a smooth production run as soon as I've updated the Analogue stage PCB design with all the modifications + the ADC section

 

[Superior]

MDAC2 'slaves' are offered at a discounted development cost when combined with a Master MDAC2 (as you have already helped finance the MDAC2's development with the first MDAC2 'Master' unit development payments).

Hallo John,
how much would the Slave MDAC2 cost for us members?

Please put me also on the list for the full width chassi.

Is the TDAC still on your to do list, or would there be no essential benefit against the MDAC2 with the Mosfets Output Stage?

Greetings

 

[JohnW]

Superior,

The MDAC2 Slaves are sold at a discounted development cost you basically save GBP100 (on the L2 or GBP200 on the L3).

There is a MDAC2 Slave payment option on our outdated webpage:-

http://mdac2.lakewestaudio.com/

 

[JohnW]

Is the TDAC still on your to do list, or would there be no essential benefit against the MDAC2 with the Mosfets Output Stage?

The TDAC was originally 'conceived' as I'm uneasy about using transistors in the Analog stage's Frontend circuit due to the level of RF injected from the DAC Array.

Using the MOSFET's should close the performance gap between MDAC2 XSQ and the TDAC - and whose to say how long it will take me to complete the TDAC design - just look at the MDAC2 development times!

 

[adamdea]

I gained an extra 1dB dynamic range, so the XSQ MOSFET version of MDAC2 PCB's Awtd dynamic range balanced is 126.5dB (There is a good chance the MDAC2 HDR version will now be around 134dB dynamic range).

https://dl.dropboxusercontent.com/u...7.19Vrms BAL 126.5dB Awtd 123.5dB Non Wtd.jpg

You can see above that the MOSFET the noise floor is elevated (not flat) at LF, but is still nice and distortion free.

https://dl.dropboxusercontent.com/u/86116171/MDAC2 MOSFET ESS 0dB Bal THD 0.00045% 5Vrms REF.jpg

MDAC2 XSQ THD at 0dB Balanced is 0.000045% so still ultra linear

I'm not sure I'm following this John- Is the first pictures the FFT of the OP amp m-dac output stage and the second one the MOSFET output stage? The second does seem to have additional distortion components albeit at a pretty low level. If not could you possibly show a comparison for the two different versions

Well it will be a "Battle" between the subjectivist's who 'only' listen and Technocrats who only go by measurements - I suspect I'll be going for the XSQ version - IF it performs as I expect...

It will be interesting as the ONLY difference would be the use of Transistors or MOSFETs in the input differential stage.

The problem is that if one doesn't have the opportunity to listen oneself before choosing then it's actually likely to be a choice between faith in your measurements and faith in your ears. Tricky.

 

[JohnW]

I'm not sure I'm following this John- Is the first pictures the FFT of the OP amp m-dac output stage and the second one the MOSFET output stage? The second does seem to have additional distortion components albeit at a pretty low level. If not could you possibly show a comparison for the two different versions

Both test results are from the MOSFET version.

The first is a -60dB test tone for the Dynamic range test (using MOSFETs)

The second is 0dB THD test (Again using MOSFETs)

 

[Steven Toy]

The TDAC was originally 'conceived' as I'm uneasy about using transistors in the Analog stage's Frontend circuit due to the level of RF injected from the DAC Array.

Using the MOSFET's should close the performance gap between MDAC2 XSQ and the TDAC - and whose to say how long it will take me to complete the TDAC design - just look at the MDAC2 development times!

In two years' time I might be ready for the next DAC upgrade. It was two years ago almost to the day that the MDAC L3 Toy arrived here.

It's still great but I'm looking forward to what comes next.

 

[adamdea]

Both test results are from the MOSFET version.

The first is a -60dB test tone for the Dynamic range test (using MOSFETs)

The second is 0dB THD test (Again using MOSFETs)

Thanks John. Any chance of seeing the comparison of the standard version. Also is the difference between the noise floor at the two levels expected?

 

[Steven Toy]

I'm not sure I'm following this John- Is the first pictures the FFT of the OP amp m-dac output stage and the second one the MOSFET output stage? The second does seem to have additional distortion components albeit at a pretty low level. If not could you possibly show a comparison for the two different versions

The problem is that if one doesn't have the opportunity to listen oneself before choosing then it's actually likely to be a choice between faith in your measurements and faith in your ears. Tricky.

It's a gamble for sure but I think I trust John's ears enough having spent a bit of time with him in the lab almost 3 years ago whrn we were comparing the Sovereign MDAC with the standard version.

 

[banjoman]

Great to catch up on the ongoing developments here! I'd definitely go for the XSQ version

 

[adamdea]

I'm a bit concerned that the dynamic range advantage of having two sabre dac chips may be wiped out. It's a shame you can't switch as you can with filters. Whether the difference would get past the power amp is another question.