In the internal pictures above, you can see the black and silver chassis sides of the 9000 & 9100 respectively.  The 9100 is in the top picture and the 9000 in the lower.  I’m not too worried about the cap types, as these affect the analogue audio and video inputs, while I only plan to use a handful of the digital audio & video inputs on the left hand side.  Still, it’s a good pointer to Sony’s approach to component selection, and might give a few pointers to what the rest of the amp is like.      

I’m going to take this opportunity to describe why I’m so enthusiastic about all-digital amps, and this one in particular. I’ve been thinking about this for a while, and I wanted to get something down in print.  Consider a conventional stereo or multi-channel system that typically has a disc player, a DAC / pre-amp / processor, and an amp.  This system has a lot of processes and interfaces that can’t help but degrade the sound:   

1) I2S to SPDIF or HDMI conversion in the player.  
2) Digital audio connection to DAC / processor.   
3) SPDIF / HDMI conversion to I2S in DAC / processor.   
4) Digital to analogue audio conversion.   
5) Current to voltage conversion.    
6) Analogue audio switching    
7) Analogue audio volume control   
Analogue audio gain stage     
9) Analogue output buffer.    
10) Analogue audio connection to amp.   
11) Analogue audio input buffer in amp.   
12) Voltage gain stage.   
13) Power gain stage.  
14) Class AB amp cross-over distortion.    

In my system, with the DA9100 and electrostatic speakers, I effectively get to side-step all those processes, and add a few more tricks:

1) DAC-master configuration with i-link    
2) 24-bit digital audio chain throughout
3) No global or local negative feedback  
4) No analogue speaker cross-overs.   

Analogue systems and interfaces can work very well, but looking at all the main processes within pretty much any system, there’s a lot of hurdles to be overcome.  Depending on whether you view a digital amp as a DAC without an amp, or an amp without a DAC, there has to be a D to A or an amplification stage somewhere in the system. 

But there’s got to be a lot of advantages to getting rid of everything else.  I’m especially thinking of analogue outputs that have voltage drivers with low output impedance, that connect to input buffers with high input impedance.  Each of these buffer stages has to provide power gain, even if there’s no voltage gain.  But gain is gain, and its being thrown away everywhere you look.  I don’t think the DA9100 will be as good at being a DAC as the best DAC, nor as good an amp as the best amp, but as a system I’m hopeful about the end result.    

Nickfficeffice" />>>

I've been having a close look inside both amplifiers, and this has been quite interesting.  From a distance they look the same, apart from the HDMI board and the additional video interface board. There are small differences in the structure and cooling vents of the 9100, which are slightly upgraded.       

However, I always imagined the 9100 was essentially a 9000 with the HDMI board from the 7100, and like Sony CIS, imagined that it might be possible to modify a 9000 with HDMI inputs.  Looking closer, I doubt this would be possible.     

The power supply section does appear to be identical, but otherwise ALL the boards in the 9100 seem to be different.  The power amp board seems to use the newer S-master processor from the 7100, and the key components are different, including the power FETs and the output filter capacitors.  Many of the I/O cards are similar, but it’s clear that the active devices – decoders, processors etc are different, and this even applies to the existing digital audio interface board.     

Comparing the schematics for the 9000 & 7100, there’s a big difference in the internal architecture, and I presume the 9100 follows the 7100.  The latter uses entirely PCM internal processing, whereas the 9000 converts all audio inputs to DSD before converting everything back to PCM for lip-sync adjustment IIRC and S-Master processing.      

I had simply hoped that the 9100 would simply retain the performance of the 9000 with legacy sources and add the LPCM capability.  It now seems like the 9100 is effectively an upgraded 7100 instead, so I’ve no real idea what to expect with regards to sound quality.  I had assumed that I would keep the 9100 and sell the 9000, but I think I’d rather keep both bases covered until I see for myself.  I’m slightly worried about the possible impact of all the additional high-speed digital audio and video processing that the 9100 has, which may leave it sounding better with Blu-ray but worse with everything else.  And it could be that apart from simply updating the design with more up-to-date components, Sony may have taken the opportunity to reduce manufacturing costs, rather than moving things forwards.        

It’s rather a surprise that they have made so many changes for a model that can’t sell in great quantities and bewilderment that having spent so much development time, effort and money on the new model, that they haven’t tried to keep it more up to date.  I think the proof will be in the eating though, and I’m looking forwards to firing it up shortly.  Here’s a couple more pictures.    

Nick
fficeffice" />>>

>>

>>

> >


 

On 2010-07-25 10:40, Welwynnick wrote:
I did notice a difference in the main power supply though.  The power amp supply is regulated, and the 9000 mother board has provision for two pairs of regulators - the large ones that go on the biggest heat sink at the front.  The 9000 doesn't actually have them fitted, but the 9100 does.  This is curious.  They appear to have been designed-in, but maybe Sony took a cost-cutting decision with the 9000 to leave them out?

On 2010-07-25 17:05, jehill wrote:

On 2010-07-25 10:40, Welwynnick wrote:
I did notice a difference in the main power supply though.  The power amp supply is regulated, and the 9000 mother board has provision for two pairs of regulators - the large ones that go on the biggest heat sink at the front.  The 9000 doesn't actually have them fitted, but the 9100 does.  This is curious.  They appear to have been designed-in, but maybe Sony took a cost-cutting decision with the 9000 to leave them out?

That's interesting, because the STR-DA9000ES Mother Board schematic clearly shows two regulated outputs to the amp board, one from Q807 and one from Q808.  However, there are typos on the schematic.  The schematic indicates a 6.8V output from each.  I suspect that the actual output is 68V.  The input voltage to the regulators is 85.6V.


-----------------
Nick

From the reverse PWB layout on p98 you can see that these connections are fully parallel-up with necessarily heavy-duty tracks and through holes.  The duplicate regulators are fully populated on the 9100, and the devices appear to be identical.       

I’m wondering if this just removes an element un unreliability from the 9000, or whether it actually increases the headroom with the 9100?  It can’t be a bad thing.       

Nick<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>

It’s taken a long time, but I’ve finally made some progress on the DA9100.  I bought a big variac, so that I could get 100V mains, but I’m not happy using step-down transformers with power amps.  After a lot of searching and chasing, I’ve found a company that will make a small quantity of toroidal transformers to the spec needed for the Sony.                  

This search also indicated to me that in spite of Sony’s trumpeting, the transformer used by the DA9000/9100 is only about 750VA in size.  That’s probably equivalent to everyone else’s 1500VA when you consider the efficiency of class D, which is actually pretty good.  However, it didn’t escape my greedy thoughts that this might be a good opportunity to upgrade slightly, and there IS quite a bit of space in there.  Big transformers aren’t that much larger or more expensive, so lets see what we can do.                

The Xformer winder finally came back to me with a quote that I can’t resist.  They suggested using two Xformers, a small 80VA unit for the display and interface boards, and big 1500VA unit for the main supply.  They want $200 ea for a short run of these xformers, which is a fraction of what a spare 750VA Sony unit costs.  At that price its worth replacing a perfectly good stock DA9000 transformer, which I’ll do with mine, and wondered if anyone else was interested?                   

Nickfficeffice" />>>

On 2010-11-02 05:01, mykyll2727 wrote:
The DA9kES serves as a pre/pro and is connected to an external power amp which has alot in common with yours. It's a Verastarr SSA644 which is basically a modded Cinepro 3k6SE gold and I absolutely love it. IMO it's just a truly magnificent MCH amp.

On 2011-07-19 15:44, jehill wrote:
I doubt that the CXD9856N A/D used in the STR-DA7100ES is used in the STR/DA9100ES.  According to the "ES Series Receivers Technical Background" (link below), "The 32-bit upgrade is as follows:

<FONT face=Arial>
<P align=left>"Where the acclaimed STR-DA9000ES took advantage of 24-bit S-Master Pro processing, Sony's latest design boasts the superior precision of Sony's 32-bit processing. Sony proudly presents a masterpiece of integrated circuitry, Sony's CXD9773Q Large Scale Integrated circuit (LSI)."</FONT>
http://www.docs.sony.com/release/ES_STR_05_Final.pdf  
I suspect that the A/D processing in the STR/DA9100ES is the same as in the STR/DA9000ES, namely, A/DSD conversion followed by DSD/PCM conversion.

On 2011-06-29 10:10, mykyll2727 wrote:

So if I got a 9100ES I could swap the Xformer from my 9000ES into it? Would that also switch it to run on 120v AC?_mykl

The 9000 and 9100 have an AC board and a DC board.  The AC board has the AC inlet and outlet (soldered to the board) plus filters, relays and the standby power supply.  The DC board has separate power supplies for both the power amp, and power amp control.  IIRC, the AC & DC boards (and the neighboring speaker outlet boards) are the only things that the 9000 and 9100 have in common. 

If anyone’s worried about the mains connector, why not solder a captive cable direct to the AC board?  Having become so intimately familiar with the 9100 this year, I admit I’m slightly concerned by the sheer number of connectors.  In the power supply chain, there are connectors for AC board outlet, DC board inlet and outlet, and motherboard inlet and outlet.  In the output chain, there are connectors for power amp outlet, filter board inlet and outlet, followed by some long cables to the speaker board inlets and outlets and sub-board inlets.  There are RF chokes all over the place, and the only soldered connections are the AC board inlet, DC board outlet and the speaker outlets.  Soldering all those high-current connectors would be a nightmare. 

It feels like the core of the amp is the 7-channel S-Master Pro board itself, which Sony must have taken a long time to develop (awesome achievement with no NFB BTW).  The rest of the amp was then built around that board. 

In my dreams, the ideal architecture would have the high power parts on (two) symmetrical boards running down the sides of the amplifier, inside vertical heat-sinks, with the DC inlets on one end and the speaker outlets at the other.  That’s 3 / 4 channels with all the regulators, amplifiers, filters & relays on each board.  I doubt we’ll ever see that of course, though its not unlike the elegant, single, DA5000 board. 

The other thing I’ve been thinking about is what could be done with the output filter caps?  I believe caps are the most critical of all components, and though the rest of the 9000 is all-digital, there’s no escaping the LC output filter.  There must be some opportunities to improve things with better caps…? 

Sorry for rambling!

Nickfficeffice" />>>

On 2011-06-28 05:11, Welwyn Nick wrote:
I may have to have to change the soft-start resistors again.  The stock 9000/9100 has 3 x 3.3ohm / 5W power resistors, and I upgraded those to 3 x 8.2ohm / 10W.

Well I just made a big discovery, and I’m going to sleep a lot easier tonight. 

The rest of the World has known this all along, but I found out the hard way.  I’ll describe the learning (blundering) process, and see if you spot it.  Problem is that I actually had three problems…...

I thought I had been overloading the main transformer and damaged the secondaries or something like that.  It was strange that once I took the transformer out of the amp, it seemed to be completely healthy.  Just having it sit in the amp, out of the frame, was OK.  The problem seemed to be when it was clamped into the mounting frame - it was installation dependent, but difficult to reproduce.  There was something wrong with the installation; it wasn’t robust or reliable, and it would be difficult to trust it. 

When something goes wrong, I step right back until I have something that does work, the work forwards one step at a time.  I started with the transformer on the bench, vertical and horizontal, which was fine.  I put the frame on top, upside-down, then rested the transformer on the base of the frame, right way up.  It was also OK with the transformer sitting in the bottom of the amp, either on the floor, or resting on the frame.  In this case, I powered the transformer directly from the mains (though I couldn’t power it through the amp’s AC board unless I disconnected it from the DC board, which was a red herring). 

Back on the bench, I could fit the stud through the transformer, even with the rubber bushes fitted, which were quite tight.  I wondered if there was too much pressure on the core, that was putting stress on the windings, but that didn’t seem to be it.  It was pointing towards the pressure of the mounting frame stressing the windings.  Were they that sensitive?  The funny thing is the first, 75V, transformer seemed to behave the same way as the second, 65v one.  Too much of a coincidence? 

Next, I part mounted the transformer, with the stud running through one of the mounting holes on the frame, and part of the way through the transformer core.  No problem.  The next stage was to complete the mounting, but now the buzzing the burning small came back.  I worked back, by taking the bolts off the stud, so as to relax the pressure on the transformer.  Guess what – no difference!  But with the nuts off I could now see the problem.  I withdrew the stud from the second hole of the mounting frame, and the buzzing disappeared.  With the stud sitting through both of the mounting holes, there were sparks at the metal joint between the stud and the frame.  An induced current was flowing round the loop created by the stud and the U-shaped frame – a big current!  

But wasn’t this effectively the same installation configuration as the stock amplifier?  I was using the same mounting frame, and the stock transformer also had a nut and bolt holding the (plastic) transformer mountings together.  Those mountings were screwed to the frame, but there was no continuous electrical connection from one side of the frame to the other.  So all I had to do was break the electrical circuit…?  I put a rubber glove over the end of the stud, and pushed that through the mounting hole in the metal frame, and the buzzing and smoking magically disappeared. 

It felt like I have been discovering what all consumer electronics manufacturers must have known for decades.  All I had to do was insulate the stud from the frame on one side.  The stock configuration would do that, but there wasn’t room for the plastic spacers around the larger transformer.  I think the solution is to open up one of the mounting holes and run a longer rubber bush right through it. 

Like I said, I’ve been learning things the hard way, and it’s been unraveling here in pretty much real time.  NOW I’m confident of being able to put it all back together again. 

Nickfficeffice" />> >

There are quite a few implications from all this.  In hindsight it should have been easy to anticipate the underlying cause.  The mounting frame and the stud constitute a conducting loop around the toroid – effectively an additional, single-turn, secondary winding in itself.  It doesn’t matter that it stands-off from the transformer, the loop still cuts all the magnetic flux generated within the transformer.  Being a single turn, the voltage generated will be low, but being a short circuit, the current generated will be correspondingly high.  This short circuit output constitutes a high electrical load, one that isn’t switched-in by the usual start-up sequence – the full load will be seen by the AC board as soon as power is applied. 

That explains the failure of the soft-start resistors (even the 10W ones).  The start-up sequence carries on regardless, and then the DC board sees the full force of the in-rush.  This explains a lot of things, like how disconnecting the secondary inlet to the DC board made no difference, nor did powering the transformer directly from the mains – all very confusing at the time. 

It also suggests there’s not much wrong with the original 70V transformer – the main regulators would just have to dissipate a few more volts – but I’m happy that the 65V one is the right one, and the lower voltage means it has proportionately more current capacity. 

Although I got some over-rated soft-start power resistors, the 10W jobs may actually be sufficient.  I can get them at a local shop, and I’d rather keep them on the AC board, where they fit nicely.  My plan was to mount the power resistors on the rectifier heat-sink (on the DC board), meaning I’d have to un-solder them whenever I wanted to remove the AC board.  So although I want the new installation to be robust, I’d also like it to be as close to stock as possible. 

I haven’t received the new diodes (from ffice:smarttags" />lace w:st="on">Hong Konglace>) yet, but I’ve got a much warmer feeling about this.

Nickfficeffice" />>>

On 2011-07-17 17:45, Welwynnick wrote:
I now have the 9100 side-by-side with the old faithful 9000 for some audio comparisons.  Which are proving to be very interesting...........

Nick

"Where the acclaimed STR-DA9000ES took advantage of 24-bit S-Master Pro processing, Sony's latest design boasts the superior precision of Sony's 32-bit processing. Sony proudly presents a masterpiece of integrated circuitry, Sony's CXD9773Q Large Scale Integrated circuit (LSI)."

http://www.docs.sony.com/release/ES_STR_05_Final.pdf 

I suspect that the A/D processing in the STR/DA9100ES is the same as in the STR/DA9000ES, namely, A/DSD conversion followed by DSD/PCM conversion.