Oppo's BDP105 - discussions, upgrading, mods...

no, you dont need 2, each ES9018 (yes just 1x S guys please) has internally 8 parallel balanced dacs, no it does not have any single ended output, you need to perform conversion to single ended output

also the ES9018 does not output 4.2v, it has differential output of 3.05vac point to point, can be slightly higher or lower depending on the AVCC supply voltage. maybe they are adding gain? 4.2v is a pretty hot output, even for balanced, I reckon some preamps will not like that
 
Last edited:
Member
Joined 2009
Paid Member
Yes, there is gain after (I/V) DAC. The 2,1v/4,2v are measured/specified at the output pins of RCA/XLR.
I think is nothing wrong with this. 2,1v/4,2v is a kind of standard line output.

(I have my self 28ppV on my outputs (opamps)... it goes directly in a potmeter input...) ;)
 
Member
Joined 2009
Paid Member
New pictures published by Oppo. It seems that the final caps (AC coupling) have an 4 figures µF capacity and 16v. Producer of the caps: SILMIC... I could see something like 400 (last 3 figures) on one of those... It may be 1400µ... 2400µ...
It could be enough for an low end of the audio spectre... But is not possible to be seen an paralleled cap for better high end of the spectre... I can not agree with this kind of design on the output line signal path, coming from an ESS9018 DAC...
 

Attachments

  • 105EU-7.1-audio-hires2.jpg
    105EU-7.1-audio-hires2.jpg
    753.3 KB · Views: 992
  • BDP-105EU-hires-XLR-pcb.jpg
    BDP-105EU-hires-XLR-pcb.jpg
    628.6 KB · Views: 945
Last edited:
Yes, there is gain after (I/V) DAC. The 2,1v/4,2v are measured/specified at the output pins of RCA/XLR.
I think is nothing wrong with this. 2,1v/4,2v is a kind of standard line output.

(I have my self 28ppV on my outputs (opamps)... it goes directly in a potmeter input...) ;)

no 2.1v is a standard, 4.2 is not. some complain of the already reasonably hot level without any gain.

so you are trying to squeeze best performance out of the ES9018 dac and you are adding a heap of gain then using a noisy potentiometer to throw it away? pick your gain structure, use the minimum you can and use ess excellent digital 40bit volume for slight adjustment. 28v is a huge amount of voltage gain for most systems and any pot will add more noise than the digital control as long as you dont need to attenuate too much
 
Last edited:
Member
Joined 2009
Paid Member
4,2v is specified for XLR output (2,1 x 2)

About my output on 28 Vpp (on max dynamic signal), I may say that is not here much gain involved. Is something about 1,2 (I/V is 1)... The output is like this because the chips are powered with +/-15v on final... and the slew rate is also very high.
This huge output voltage gave me a very good SNR through lines/cables (if you can believe it, not shielded, but not very longs) even though I have to use a analogue (high quality) potentiometer to adapt the line to the power amp...
Maybe this way is not very usually, but I assure you that I have absolutely not any problem with "noise" in the system (which is not audible at all), the dynamic is on top, and everything is working just excellent in my (customized) system...
But this subject is a little bit different than what is supposed to be discussed in this thread...
 
Last edited:
perhaps you should add that to wikipedia.... 4.2 even if it is 2 x the 2.1v (which is actually supposed to be 2.194vpk if its referring to ARD, or 2.1vrms) is not an XLR standard. the professional audio standard which is XLR is +4dbu which is 1.228vrms, 1.737Vpk, 3.474Vpp. 4.2v is too hot for any kind of audio standard. the ARD level would put it at 4.384vpp, but thats got nothing to do with XLR, it began at a broadcaster in Germany.

ive found a couple of references on audiophile manufacturers sites just now as 2.1 and 4.2v RMS, but ive never seen such high voltage on balanced lines outs on any pro recording gear ive used and its not mentioned anywhere on the standards pages. it does make more sense I admit if its double the power for balanced, hardly any of these places stick to a standard..



4,2v is specified for XLR output (2,1 x 2)

About my output on 28 Vpp (on max dynamic signal), I may say that is not here much gain involved. Is something about 1,2 (I/V is 1)... The output is like this because the chips are powered with +/-15v on final... and the slew rate is also very high.
This huge output voltage gave me a very good SNR through lines/cables (if you can believe it, not shielded, but not very longs) even though I have to use a analogue (high quality) potentiometer to adapt the line to the power amp...
Maybe this way is not very usually, but I assure you that I have absolutely not any problem with "noise" in the system (which is not audible at all), the dynamic is on top, and everything is working just excellent in my (customized) system...
But this subject is a little bit different than what is supposed to be discussed in this thread...
 
Last edited:
Member
Joined 2009
Paid Member
Coris, what are your thoughts on how best to improve the output stage?
Simple capacitor bypass?
Transformer?

Well, transformers to output the DAC`s signal is also an idea I have since a time ago... I really do not know how it works (sounds), but it seems that they who are using so, are satisfied. Is a really drop of signal in this way to do it, but I think is worth to be try it. Technical speaking, specially in 105 model to use transformers is very difficult (to have the player enclosure completely closed after...). Not at least, this is a very expensive choice... And is more suitable for stereo output... I think one may drop this idea.
Before conclude to bypass/remove the final caps, one may take some measurements, to appreciate better such improvement. Not least, one may hear first the sound of the player hot it is out of the manufacturer.
My main idea is to bypass those caps, but how the mod is to be done in details, is a little to earlier to say. Examining the board and see what kind of circuits/design were chosen by Oppo, is a priority, before find the best way.
One may see first what about the DC offset before those AC coupling caps, and how the final stage works.
I can see also on the boards pictures that Oppo still use some RF filters right before RCS/XLR connectors, an divider after the AC caps, and so on...
The main thought here is "less components in signal path is best". Why they chose to load this output signal path with all kind of filters and ferrite beads, caps and so on, is enough difficult to understand. The signal it may look well on the oscilloscopes displays, but is for sure not best sound out of it...
In my own design I have not any filters at all after I/V stage, and the only caps are some of few pico on the opamps feedback. I succeed to lower to 32mV the residual noise from the DAC chip (very far from audible spectrum), while audio output signal it swinging from few Vpp up to 28Vpp.

So, let`s wait the player out on marked first... and on ones each rack second...
 
Last edited:
Member
Joined 2009
Paid Member
perhaps you should add that to wikipedia.... 4.2 even if it is 2 x the 2.1v (which is actually supposed to be 2.194vpk if its referring to ARD, or 2.1vrms) is not an XLR standard. the professional audio standard which is XLR is +4dbu which is 1.228vrms, 1.737Vpk, 3.474Vpp. 4.2v is too hot for any kind of audio standard. the ARD level would put it at 4.384vpp, but thats got nothing to do with XLR, it began at a broadcaster in Germany.

ive found a couple of references on audiophile manufacturers sites just now as 2.1 and 4.2v RMS, but ive never seen such high voltage on balanced lines outs on any pro recording gear ive used and its not mentioned anywhere on the standards pages. it does make more sense I admit if its double the power for balanced, hardly any of these places stick to a standard..

You may have right. Honestly I did not get in this fields details... My judgement were that so far the 2.1Vpp is a standard on SE, then 4,2 it may be a standard on differential (XLR). But I will not argue on this, if you say is a quite high output voltage. My personal opinion is that this not so big problem, and it seems to work for the most of the users of the player (BDP95 so far...). I could`t read any claim so far about this...
 
Last edited:
above reads odd I missed an edit. obviously the 2.1vRMS is a standard, but the balanced output of gear is all over the shop, 4.2 is somewhat logical but its not really a standard, thats all i'm saying and its RMS, not pp as you wrote.

anyway shouldnt you guys be waiting to actually have a unit in hand before suggesting or requesting mods? just bypassing the caps I think is a bad move

I agree Coris, soz, lets not argue over details like this. but I do suggest that you take a look at your gain structure.

since you appear to be tweaking for highest performance, adding gain and then taking it away with a pot is not only a bit odd and expensive if you use quality parts, its going to give you worse performance than measuring the voltage across your speakers at your max listening level and assembling your gain stage either in your amp, or in the dac to achieve that level, then use the digital control to adjust for lower listening. Digital volume controls used to be crap, but these days there really isnt a better method, the one in the ESS is particularly good.

28v across the terminals on all but pretty inefficient speakers is a huge amount and whether you hear actual noise or not, you are adding noise if you are attenuating; even a 1k resistor adds noise compared to the performance the es9018 is capable of. your pot, if you are running balanced, will also not have perfect channel matching, or matching phase to phase, so the distortion will be higher as well. you will be surprised at how excellent the inbuilt digital control is as long as you structure your gain appropriately.

is your poweramp just a buffer? how much gain in your amp on top of the 28v?
 
Last edited:
Member
Joined 2009
Paid Member
anyway shouldnt you guys be waiting to actually have a unit in hand before suggesting or requesting mods? just bypassing the caps I think is a bad move

Fully agree that one may wait to have first the unit in hand!
There are not here about "suggesting/requesting" mods, but only speculations on eventual (necessary) mods... :)

You know, I will measure again this about output on my mod. I`ve did it actually many times before, because it was surprising for me too to have a such output. You can see in one of my last picture of this final stage, that I just have an 1 gain. But I have in the same time this huge swing on outputs... I`m running SE for moment. I may precise that those 28V are pp and measured on max amplitude of the signal from DAC. Usually the (average level) signal is somewhere around 15-18Vpp. This signal goes directly in to an 100Kohm pot (the amplifier input), and then in the one stage amplifier which it have some gain of course on it. I can not run the pot of the amplifier more up than first quarter of its whole (log) scale. In this way I have 10-20W output (from approx 80W, and running on max 100W speakers). Is working so just perfect I can say. But all the time is little more place for better... ;)
Anyway, I will double check soon the measurements and I will come back about this...
 
Last edited:
Fully agree that one may wait to have first the unit in hand!
There are not here about "suggesting/requesting" mods, but only speculations on eventual (necessary) mods... :)

haha but until you know the circuit exactly its only guessing isnt it? I reckon this is going to be a tricky unit to mod properly with all that annoying routing theyve done. splitting the channels like that between HP and Line out, even if you manage to bypass all of that and run all of the channels together; I figure the act of doing so will lower performance so its not worth it.

the use of silmic and wima is obvious pandering to the public, wima are not particularly useful for HF decoupling, ceramics are far better, but they are a pretty red colour....
 
Member
Joined 2009
Paid Member
Fully agree again about ceramics to HF decoupling. I just use this way. Oppo`s choice in this player is wrong in my opinion. They could use SMD film caps and not those WIMA big (old fashion) caps. Having such big components on the boards is just bad for performances and for the place they need for implement their design...
My serious concern is the same as you could see: the implementing and the mess with the channels of the stereo DAC. Here it will be difficult to fix something. But one may take a closer look on it later on... The most evident mod/improvement one can already see form the published pictures, are about those AC coupling caps. We will see how it goes, after having the unit on the table...
 
Last edited:
Coris said:
I can not run the pot of the amplifier more up than first quarter of its whole (log) scale.

then you have far too much gain, thats what I mean. you should really be sitting at about 80% volume at least for your normal loudest listening level (not 100% you need to save some for low volume recordings) a quarter turn of a log scale is not going to be 20W either, (edit: depends on the amp I guess) the ideal is no attenuation ie the pot turned up 100%, but thats not realistic because different recordings are different volumes. You need a little bit of play above your max to account for that, but at 1/4 you are both adding noise from the voltage gain in the first place and then adding more noise and lowering SNR by attenuating that much.

this is what gain structure is all about, you should never have much more gain than you actually use. consumers usually just have to live with this sort of thing, they rely on manufacturers, but we as DIYers can control these things. by using it at 1/4 you place most of the pot in the signal path, as well as using it at the beginning of its range, where its channel matching is usually worse.

rather than modding the output stage, you might be better off designing a small PCB and sending all of the channels to that, tapping them all from vias or something. forget playing with their mess, just bypass it altogether.
 
Last edited:
Member
Joined 2009
Paid Member
I will try to find a better way to explain what about is happen in my system, and I will come back with more accurate measurement results.
Anyway I do not have any noise problem, because of that potmeter attenuation, or something else. My amplifier is not an consumer one, and the dynamic is very good in the system. But I can remark that at low volume recordings I feel the need to go a little bit up with the pot volume... And I do it...The good dynamic compensate also much (of this volume up need), so at last the resulting audition it steel very comfortable, even at low volume recordings...I can of course go quite much up in volume, but then is just uncomfortable to listen music (even rock).
I`ve learned quite long time ago, that not the power of an amplifier, or high volume level is the clue to have all out of recordings. This apply to bad/cheap (consumer) systems... and to an "uneducated" listener... A very high fidelity signal, high dynamics and the most important, high as possible slew rate of every component of the audio system, it bring out everything and all from (even low quality) recordings.
 
Member
Joined 2009
Paid Member
I found those my pictures published earlier in this forum. It were measured on the RCA output of (mod) BDP95.
One it show 21Vpp (high dynamic sound/orchestra) and another one 5Vpp (80Khz sinus played through DAC, from digital file)...

But I will measure again and I will present it here soon.
 

Attachments

  • normal high dynamic sound.jpg
    normal high dynamic sound.jpg
    197 KB · Views: 815
  • the output sound1.jpg
    the output sound1.jpg
    177.2 KB · Views: 797
I'm a bit concerned about the presence of elco's in the output of the 105. They might be Elna Silmics, but they cannot be anything near as good as film caps.

Could it be that these are actually polarised, I mean non-polar caps with a voltage put across them? Would that make sense? And, if so, how much better would film caps then still be in the same location?
 
Member
Joined 2009
Paid Member
I'm a bit concerned about the presence of elco's in the output of the 105. They might be Elna Silmics, but they cannot be anything near as good as film caps.

Could it be that these are actually polarised, I mean non-polar caps with a voltage put across them? Would that make sense? And, if so, how much better would film caps then still be in the same location?

Yes, is right! Those AC coupling caps should be at last unpolarised. To have there film caps is good for the high pass of the spectrum, but at such large capacity (needed for low pass) is very difficult and it will be very big... Those caps shouldn`t be there anyway.... This design (AC coupling) is wrong! I just can not understand (with the informations I have for moment), why they chosen this way to output the signal. It were very well about this aspect, how it were in 95 model... I can not see this as an improvement at all.
Let`s wait for the player come out and reading the reviews and impressions from the owners...