OPPO's answer ,quoted earlier in this thread ,said WIMA...... a film type cap.
Very right! That is proof design: DC coupling with no DC offset on the final stage. Everybody (hobbyist) knows to place an capacitor on the final amp output to get rid the faulty design with big offset, and having out only the usefully signal...
What about introducing (AC coupling) impedance/inductance in a wide spectre of audio frequencies, when an large electrolytic capacity is in the useful signal path? Is no doubt that one hear very clear such difference: with or without AC caps on the final stage...
Last edited:
As you can see in the picture here, It seems that the final (isolating) caps are not very special ones. I guess they are of 680µF capacity, and for sure electrolytic.
WIMA (the film red ones) caps are quite low capacity and used to decoupling the final opamps, filtering in the signal path, and so on...
Another sign of proof design should be the using of such WIMA film caps (or another ones) to parallel the isolation large capacity caps, for an much better frequency response for the output signal....At last is not to be seen in this picture...
Attachments
Last edited:
Which post number was that?
The answer I found stated they used Elna and no mention was made of the type.
BDP103 do not have ESS9018 DAC... I can not be sure about the output coupling in 103. Looking at the picture, it may be it may be not... But it seems to me that they used the same design type for both those last players.
Else such isolating caps in AC coupling in most cases are electrolytics, because the need of a large capacity at a low price...
I`m personally not worry about those caps... I think it will be a nice mod to remove it all and fix the eventual offset problem on the output of the channels, and then DC couling.
My problem is at the allocation of the DAC channels in the stereo stage. That allocation is quite bad, and I have no idea at this time how to fix this...
Else the 105 model have for sure some benefits over the 103 model (one is USB asynchronous DAC). It looks to me that I will keep on my pre order...
Oppo use Elna and WIMA caps in theirs machines. Elna for electrolytics and WIMA for film.
Attachments
I`m quite sceptic about it is happen up sampling to 32 bits inside the player... It may be possible, but nobody knows, but the producer...
Art least the specifications of 95/105 are about 24 bit (standard).
But you can do it by your self this... It take a little time, but the result is very good.
I will tell you how I do it:
I rip the CD into my editing software (Audition). Correct the level for each file to a right one (usually into -1dB). This is my standard. Up sampling the 44100/16bit file to 176400Hz /32bit (100% quality). Apply or not Channels Phase correction.
Convert the file to Flac - 176,4Khz/24 bit. Play it through my player (95).
Top result!
You can be amazed about how much information/fidelity is encoded in to the standard CD files and one do not hear it when it plays the CD at standard sampling/bit.
Coris,
Thanks for this info.
My Xindak DAC-5 upsamples redbook CD's to 24/192.
I was hoping the Oppo 95, with the 32 bit ess9018 DAC, would upsample to 32 bits.
A year ago, I phoned Oppo in California - and they weren't sure - they suggested getting in contact with ESS.
Then I phoned ESS but got an answering machine.
Here is the ESS product sheet
http://www.esstech.com/PDF/SABRE32 Reference PB Rev 0.61 120726.pdf
and it says "... DAC's 32 bit Hyper stream architecture can handle full 32 bits PCM via IS2, ..."
It says it can handle 32 bits - it doesn't say it upsamples to 32 bits.
Also, my Xindak DAC-5 had WIMA MKS 4 output coupling caps.
From listening, I am of the opinion that with the WIMA's
there was "warmth" but there was a slight roughness in higher frequencies.
However, I'm not sure if the slight roughness was due to the cap itself,
or how well the cap could mask quantization issues.
I replaced the WIMA's with Mundorf M-Cap supremes - which are huge.
They just barely fit in the Xindak DAC-5.
Without question - for whatever reason - there was less roughness with the Mundorf Supremes than the WIMA's.
I didn't try the Mundorf silver/oils because they are only rated to 55oC.
So didn't think it would be wise to mount them right beside a tube.
All the film caps, with copper plates, that I have seen are priced beyond belief.
Space is very limited inside the Oppo 105 - so maybe it would be best to try
bypassing the WIMA's with a higher end film cap.
.
Last edited:
Also, not many people get exited about small value electrolytics on a circuit board,
but, people who actually have the equipment to measure caps, are impressed by the ELNA's low ESR.
However, a potential problem maybe with the removal of the fan.
Will the Oppo 105 become an easy bake oven - and slowly cook the electrolytics ?
but, people who actually have the equipment to measure caps, are impressed by the ELNA's low ESR.
However, a potential problem maybe with the removal of the fan.
Will the Oppo 105 become an easy bake oven - and slowly cook the electrolytics ?
I seriously doubt that Oppo will come out with such details: how they handle the digital signals inside the machine.... They are of course sure about how they do it, but they can not speak out such things...
ESS9018 work with an resolution of 32 bit. If it get the informations on 32 bit, then it can handle... Is not the chip itself which up sample the stream. Such upsampling/processing it may happen in the main processor of the player. Then the stream is send to the DAC to be converted to analogue.
It may happen that the Oppo`s processor work at this 32 bit resolution for the audio signals. Else it could be quite stupid to use a 32 bit capable DAC to run some low resolution streams... But how it is in real, we can not know for sure.
If you want to get inside such details about ESS9018, the way it work, there is in this forum enough detailed informations about... Lot of to read
About the temperature inside the fanless 105 we can not say too much at this time. We have to wait the first reviews, and comments from the first owners... There are some infos about in the thread in AVS forum, but not much. We will see and feel soon...
Last edited:
Correct me if I am wrong, but, as I see the specs, all of these "32bit dac" chips from ESS & etc. are NOT 32bit dacs, but are 24bit dacs with 32bit digital filter on-board. I haven't ectually seen even a 24bit dac that's capable of genuine 24bit SNR/DNR.
Oppo's reasoning for use of 4x dac channels for headphone makes no sense. The dac chip does not see any 'extra load', as it's simply feeding a driver chip. They should have at least used, instead, 4 dac channels for the balanced outputs, which would have made good sense.
Appears that the 103 uses Nichicon caps in the dac/output stages, which I personally prefer over Elna. I find Elna's to be far too unreliable to use, even if they *maybe* sound a bit better for a given purpose(which I believe they don't, after caps are well broken in). Elna's are particularly short life in hot environments, which I expect the 105 will be.
As for the dc offset- in this sort of design, even very sloppy selection of parts will end up with totally insignificant output offset, unless the design is appallingly bad. But UL hates anything without dc blocking on output/input, so most makers use them even when they are not at all needed.
I don't know why people are at all impressed when they see Wima caps. I've known them to be terribly bad sounding caps for 25 years. Even after 1000 hours of break in, they sound hazy in signal path & grainy as ps bypass. I'd sooner use Panasonic polyesters than any Wima.
Oppo's reasoning for use of 4x dac channels for headphone makes no sense. The dac chip does not see any 'extra load', as it's simply feeding a driver chip. They should have at least used, instead, 4 dac channels for the balanced outputs, which would have made good sense.
Appears that the 103 uses Nichicon caps in the dac/output stages, which I personally prefer over Elna. I find Elna's to be far too unreliable to use, even if they *maybe* sound a bit better for a given purpose(which I believe they don't, after caps are well broken in). Elna's are particularly short life in hot environments, which I expect the 105 will be.
As for the dc offset- in this sort of design, even very sloppy selection of parts will end up with totally insignificant output offset, unless the design is appallingly bad. But UL hates anything without dc blocking on output/input, so most makers use them even when they are not at all needed.
I don't know why people are at all impressed when they see Wima caps. I've known them to be terribly bad sounding caps for 25 years. Even after 1000 hours of break in, they sound hazy in signal path & grainy as ps bypass. I'd sooner use Panasonic polyesters than any Wima.
The ESS9018 is the most advanced audio DAC chip at this time, it have an 32 bit architecture, and can process on 32 bit resolution.
There is not the meaning of allocation of the DAC channels, the load(s) on it. The chip can deliver very good performances on every channel. The point with connecting 2 or more channels together is to obtain a very high accuracy of the resulting analogue signal, a max dynamic range, and to improve the SNR in the system. The output differential channels of ESS9018 can be connected together in multiple of 2. Oppo use 2 + 2 stereo ch for XLR, 2 stereo ch for RCA out, and 4 stereo ch. for headphone. In this configuration the only possibility to switch those 2 remaining ch (after having 2 for headphone) is between the headphone out and RCA out. One may chose to double the headphone or RCA in this design. Oppo explain the allocation they used as been necessary for have an independent volume control for the XLR/RCA output and for the headphone. I suppose they fix this separation of those volume controls in the DAC firmware and it divide the outputs of the DAC in two part for this propose. Maybe they have to chose this way to have those independent volume controls. I personally mean that it should be used an DAC chip more to maintain at least the same level of quality for the output signals (at the level of 95 model) and having in the same time an independent output for the headphone. Them choice downgrade the quality level for the stereo stage in my opinion...
The debate about Elna vs Nichicon will never end...
The best is not to use at all caps in that places...I share your opinion about the DC offset. The design or the components have to be very bad to have a dangerous or uncontrolled offset at the outputs. Is not difficult at all to chose a design which can assure a very low, almost 0 DC offset. Protection for the next input is made using resistors. But the opamp chips may have internal protection anyway. AC coupling in my opinion is a downgrade from the precedent models. The modders will have so a good job to do... Right?
Of course are many (betters) alternatives when about film caps. I think using just SMD film caps is a much better solution...
It would be interesting to hear reviews from the headphone crowd on the performance of Oppos headphone stage. I suspect most of the fanatics will stick to their dedicated headphone amps so it's a pity Oppo sacrificed performance to cater to them.
A while ago I ran temperature tests on the Oppo 93 and 95. Both maintained internal temperatures around 41 - 43 degrees C. I would be disappointed if the 103 & 105 are worse than that.
A while ago I ran temperature tests on the Oppo 93 and 95. Both maintained internal temperatures around 41 - 43 degrees C. I would be disappointed if the 103 & 105 are worse than that.
I had my heart set on the Oppo 105, so the comments here are very disappointing.
IMO there is no point in using the ess9018 DAC and then using WIMA coupling caps.
The unit could be AC coupled with transformers - but the cost of audio transformers - like Lundahl is quite excessive.
I see a person on DIYAudio winds his own.
Winding transformers is not difficult - the most difficult part is acquiring the magnet wire and cores.
How about other units that use the ESS9018 DAC ?
The W4S DAC2 looks well build and has a good power supply.
The Eastern Electric DAC plus isn't as well made as the W4S DAC2.
However, the Op Amps are in sockets and the output coupling caps are reasonably large.
.
IMO there is no point in using the ess9018 DAC and then using WIMA coupling caps.
The unit could be AC coupled with transformers - but the cost of audio transformers - like Lundahl is quite excessive.
I see a person on DIYAudio winds his own.
Winding transformers is not difficult - the most difficult part is acquiring the magnet wire and cores.
How about other units that use the ESS9018 DAC ?
The W4S DAC2 looks well build and has a good power supply.
The Eastern Electric DAC plus isn't as well made as the W4S DAC2.
However, the Op Amps are in sockets and the output coupling caps are reasonably large.
.
To have a better "picture" of the 105 player you may also get known with comments/infos in this thread/forum:
Official OPPO BDP-103/BDP-105 Anticipation Thread
Some more (real) informations (review, impressions/comments) may come out when the player will be in fact on market...
if they wanted to meet some sort of analogue specification, using transformers would not be the way to attain it...they can for sure sound pleasant enough, but measured performance is shithouse with es9018 and output TX
I'm still a bit puzzled about the allocations of the dac channels to the outputs.
Coris said "Oppo use 2 + 2 stereo ch for XLR, 2 stereo ch for RCA out, and 4 stereo ch. for headphone", but that adds up to 10 channels. Others say that there's 2 for RCA, 2 for XLR, 2 for headphones and 2 to ground. Someone from Oppo said that there are 4 channels dedicated for headphones. But the specs mention that the RCA level is 2.1V and the XLR is 4.2V. This seems to indicate that 2 times 2 channels are used for XLR, 2 channels for RCA and then the remaining 2 for headphones. That would make sense.
Coris said "Oppo use 2 + 2 stereo ch for XLR, 2 stereo ch for RCA out, and 4 stereo ch. for headphone", but that adds up to 10 channels. Others say that there's 2 for RCA, 2 for XLR, 2 for headphones and 2 to ground. Someone from Oppo said that there are 4 channels dedicated for headphones. But the specs mention that the RCA level is 2.1V and the XLR is 4.2V. This seems to indicate that 2 times 2 channels are used for XLR, 2 channels for RCA and then the remaining 2 for headphones. That would make sense.
Yes, sorry. It were wrong allocation from my side.
The right whole number is 8 (channels for ESS9018).Anyway one may read the Oppo`s answer about this allocation.
XLR output is differential, and then 4,1v from only one differential channel of the DAC. Is correct. XLR is for R and L channels = 2 DAC channels. RCA is single ended and then 2,1v. This it need 2 DAC channels for stereo RCA output. It remain again 4 channels (2+2) which goes to headphone coupled in parallel.
Last edited:
Thanks for the answer.
So, am I correct in now assuming that any single
channel on the Sabre DAC can be either differential
or single ended? I had assumed that a truly balanced
output would need two DACs.
Oppo seem to make the point that "The balanced output
features a true differential signal path all the way from the
DAC to the 3-pin XLR connector", which seems to suggest
a double DAC setup, or not?
Maarten
So, am I correct in now assuming that any single
channel on the Sabre DAC can be either differential
or single ended? I had assumed that a truly balanced
output would need two DACs.
Oppo seem to make the point that "The balanced output
features a true differential signal path all the way from the
DAC to the 3-pin XLR connector", which seems to suggest
a double DAC setup, or not?
Maarten