Rossl,
I've spend quite a bit of time looking through this thread with great interest - thanks for initiating it.
I have opened up my OPPO 970 and certainly agree with your assesment of MT1389 DVD controller pin spacing being far to close to attempt soldering.
But, I was curious if you or anyone else had considered looking at tapping off the digital signals near the input of the CS4360 DAC instead (Assuming that there aren't problematic signal modifying components in the path)?
I have found three small "circle pad holes" with better soldering spacing for the SCLK, LRCK and MCLK inputs for the CS4360.
Another question...
You had mentioned in your 11/30/07 posting that;
"It also looks like they swapped the data lines on the 970. Data 2 on the DAC is the main stereo signal"
By this were you refering to what is labeled the "SDIN2" on the CS 4360 DAC as the equvalent to the Data 2 signal, or am I missing something else here? I am unsure about which of the three SDINx lines (or possibly even another pin) I should try and trace back for access for use with 2 channel (based upon signal names on the CS4360 pin descriptions).
I've spend quite a bit of time looking through this thread with great interest - thanks for initiating it.
I have opened up my OPPO 970 and certainly agree with your assesment of MT1389 DVD controller pin spacing being far to close to attempt soldering.
But, I was curious if you or anyone else had considered looking at tapping off the digital signals near the input of the CS4360 DAC instead (Assuming that there aren't problematic signal modifying components in the path)?
I have found three small "circle pad holes" with better soldering spacing for the SCLK, LRCK and MCLK inputs for the CS4360.
Another question...
You had mentioned in your 11/30/07 posting that;
"It also looks like they swapped the data lines on the 970. Data 2 on the DAC is the main stereo signal"
By this were you refering to what is labeled the "SDIN2" on the CS 4360 DAC as the equvalent to the Data 2 signal, or am I missing something else here? I am unsure about which of the three SDINx lines (or possibly even another pin) I should try and trace back for access for use with 2 channel (based upon signal names on the CS4360 pin descriptions).
And do you have any pictures?
music2diy4 said:
The CS4360 has SDIN1 on pin 2, SDIN2 on pin 3, SDIN3 on pin 4. Those correlate with the 1, 2, and 3 pairs of single ended outputs .
The Oppo 981 uses SDIN1 as the front speakers and the other two for the rear, center and sub. I use the SDIN1 in the ASRC mod. It has been a long time since I looked at the 970, but I am pretty sure that the 970 has mapped them differently. It may be that 3 is the front. To figure out which one is the front, trace the signals back from the rear panel rca jack to the outputs.
You might have to scrape off some of the green solder mask and expose bare copper in order to solder to those plated through vias. There is mask down in those holes and that makes it hard to solder to them. That is probably a better plan than trying to solder to the IC pins.
If you trace out the data and clocks from the MT1389 to the CS4360 you will also find some very small resistors on the lines.
I have tried to solder to these resistors and it is very difficult also. It sounds like it would be a simple task, heat up one end of a resistor, poke a wire in to the molten solder and then let it cool. That is not what happens. The parts are so small that both ends of the resistor melt and the resistor moves off the pads when you poke the wire to it. Either that or the resistor comes off completely and sticks to the soldering iron and it burns up into nothingness.
I did take a few photos before closing the lid, but I haven't downloaded them yet. I enjoyed my first taste of SACD. It sounded very good, but I haven't yet done any serious listening or comparison to standard CD.
Unfortunately, it appears that I wasn't perfectly successful in the installation, as the player began periodically locking up, sometimes in response to tapping on the unit. So I hauled it back out and opened it up. I think one of the clock leads to the Oppo board was not solidly soldered, leading to loss of clock signal. Maybe that explains why the unit would stay powered on but would stop playing and responding to inputs? Anyway, I've soldered the clock leads back down, and the unit played through an entire SACD disc without problem here on the workbench (I don't have it hooked to anything to listen to it). I even lightly tugged on the clock wires a little bit to see if I could trip it up. So hopefully it will all be good when I put it back in the system tonight!
Ted
Unfortunately, it appears that I wasn't perfectly successful in the installation, as the player began periodically locking up, sometimes in response to tapping on the unit. So I hauled it back out and opened it up. I think one of the clock leads to the Oppo board was not solidly soldered, leading to loss of clock signal. Maybe that explains why the unit would stay powered on but would stop playing and responding to inputs? Anyway, I've soldered the clock leads back down, and the unit played through an entire SACD disc without problem here on the workbench (I don't have it hooked to anything to listen to it). I even lightly tugged on the clock wires a little bit to see if I could trip it up. So hopefully it will all be good when I put it back in the system tonight!
Ted
Beware of SACD clipping in the Oppo player!
Long time lurker, first time poster here. I received my Twisted Pear SPDIF transceiver board, installed it in an Oppo 981, and got it working with very little effort. That's definitely one of the easier mods to do!
So I've been enjoying SACD at 88.2kHz, and the sound is very nice in most cases, but some recordings are so hot that they get clipped by the MT1389 chip. Apparently, the MT1389 does not gain scale the decimation filter to handle SACD's loudest peaks, and about half of my SACDs end up clipped, as analyzed in Audacity or Wavelab. I know the problem is real because the CD layer does not exhibit the problem. So ironically, a hybrid SACD may be more distorted on the SACD layer than the CD layer when played through your Oppo!
Of course, the SACD mastering engineers may be to blame here too, since it's possible to go over "full scale" in the SACD spec without clipping the signal.
Also, for you polarity sensitive folks out there, note that the data on the 88.2kHz output is inverted. This can be cured by passing the SDATA through an inverter before it goes to the SPDIF transceiver.
Ultimately this is a bit frustrating because, in some cases the clipping is actually audible. The CAN "Future Days" SACD probably peaks at 5dB over full scale on the Oppo's 88.2kHz output! This is a problem through both the analog outputs and the modded SPDIF output. So I consider this to be a serious design flaw in the MT1389 chip. Perhaps it can be addressed in a firmware update?
Long time lurker, first time poster here. I received my Twisted Pear SPDIF transceiver board, installed it in an Oppo 981, and got it working with very little effort. That's definitely one of the easier mods to do!
So I've been enjoying SACD at 88.2kHz, and the sound is very nice in most cases, but some recordings are so hot that they get clipped by the MT1389 chip. Apparently, the MT1389 does not gain scale the decimation filter to handle SACD's loudest peaks, and about half of my SACDs end up clipped, as analyzed in Audacity or Wavelab. I know the problem is real because the CD layer does not exhibit the problem. So ironically, a hybrid SACD may be more distorted on the SACD layer than the CD layer when played through your Oppo!
Of course, the SACD mastering engineers may be to blame here too, since it's possible to go over "full scale" in the SACD spec without clipping the signal.
Also, for you polarity sensitive folks out there, note that the data on the 88.2kHz output is inverted. This can be cured by passing the SDATA through an inverter before it goes to the SPDIF transceiver.
Ultimately this is a bit frustrating because, in some cases the clipping is actually audible. The CAN "Future Days" SACD probably peaks at 5dB over full scale on the Oppo's 88.2kHz output! This is a problem through both the analog outputs and the modded SPDIF output. So I consider this to be a serious design flaw in the MT1389 chip. Perhaps it can be addressed in a firmware update?
I haven't noticed that, and I have about 25 or 30 SACDs.
What firmware version are you using?
Have you changed any of the defaults in the speaker setup page?
Can you just turn down the volume? I normally leave it full up for the best dynamic range. I'm sure I would notice clipping if it was a widespread problem.
The DSP in the MT1389 can really mangle the music if you change the tone controls or bass managment and speaker settings. I have all that set to default.
Oh, and are you sure you have the board set up for 24 bit, left justified?
What firmware version are you using?
Have you changed any of the defaults in the speaker setup page?
Can you just turn down the volume? I normally leave it full up for the best dynamic range. I'm sure I would notice clipping if it was a widespread problem.
The DSP in the MT1389 can really mangle the music if you change the tone controls or bass managment and speaker settings. I have all that set to default.
Oh, and are you sure you have the board set up for 24 bit, left justified?
So there is a digital volume control on the Oppo? I didn't even think to look for that.
I will double check my speaker settings, and also the serial data interface. I thought the correct interface was 24 bit I2S? I tried all 4 interfaces, and LJ mode seemed loud and distorted. Hmm, time to bust out the 'scope.
Thanks for your suggestions. I'll report back.
I will double check my speaker settings, and also the serial data interface. I thought the correct interface was 24 bit I2S? I tried all 4 interfaces, and LJ mode seemed loud and distorted. Hmm, time to bust out the 'scope.
Thanks for your suggestions. I'll report back.
Yes the volume control is digital, but I'm not sure how well it is implemented. I haven't ever used it, other than playing with it to see what it does. A decent digital volume control would just shift the bits down. If the DSP recalculates every sample, then it is probably full of errors at thr lower bit levels.
There is an Oppo 981 thread over at AVS forum with over 4600 posts. I have read most of them, and this is the first complaint I have heard about the SACD implementation.
There is an Oppo 981 thread over at AVS forum with over 4600 posts. I have read most of them, and this is the first complaint I have heard about the SACD implementation.
Yes, the volume control - should be left at full level, and set all the settings to off or default. Also the DVD-A/SACD should be set to 2-ch mix-down so you get all the music. Any dynamic or EQ should be set off, etc.
Definitely has to be Left-Justified: I2S will give you lots of distortion.
not sure about the inverted signal? I have no trouble with SACD, HDCD (20 bits output) even 192kHz DVD-A is very nice.
Definitely has to be Left-Justified: I2S will give you lots of distortion.
not sure about the inverted signal? I have no trouble with SACD, HDCD (20 bits output) even 192kHz DVD-A is very nice.
SACD clipping persists
I got out the 'scope and confirmed that the serial data is left justified. I actually did have the correct serial data format set on the Wolfson part (AIFCONF[1:0] = 10).
I tried reducing the volume control but the tops of peaks were still clipped. The signal was just lower in volume. I also played with the speaker level trims, but the results were the same as with the volume control. Increasing the trim levels beyond 0dB is a bad idea, period. I'm surprised Oppo allows this. Speaker level matching should only be accomplished with reductions in trim levels as needed.
I suspect what I'm seeing is that some recordings treat a 50% DSD modulation index as "full scale", while other recordings ignore this guideline and allow the delta sigma modulation to go higher, perhaps as high as 70% to 80% modulation. The MT1389's DSD decimation filter probably gain scales a 50% modulation index to 0dBFS, so when hotter SACD recordings are played through it, they get clipped.
It's important to note that in the NPC SM5819 datasheet, the DSD decimation filter has two gain settings, one for 50% MI and another for 100% MI. It can be found on page 15 of this pdf:
http://www.npcamerica.com/pdf/SM5819.pdf
So my confidence that this is a legitimate problem is increasing. Next I will capture a regular CD through the modded SPDIF output and compare it to a rip of the same CD. This will show conclusively that the SPDIF mod is implemented correctly, and that we're dealing with a DSD decimation filter issue.
I got out the 'scope and confirmed that the serial data is left justified. I actually did have the correct serial data format set on the Wolfson part (AIFCONF[1:0] = 10).
I tried reducing the volume control but the tops of peaks were still clipped. The signal was just lower in volume. I also played with the speaker level trims, but the results were the same as with the volume control. Increasing the trim levels beyond 0dB is a bad idea, period. I'm surprised Oppo allows this. Speaker level matching should only be accomplished with reductions in trim levels as needed.
I suspect what I'm seeing is that some recordings treat a 50% DSD modulation index as "full scale", while other recordings ignore this guideline and allow the delta sigma modulation to go higher, perhaps as high as 70% to 80% modulation. The MT1389's DSD decimation filter probably gain scales a 50% modulation index to 0dBFS, so when hotter SACD recordings are played through it, they get clipped.
It's important to note that in the NPC SM5819 datasheet, the DSD decimation filter has two gain settings, one for 50% MI and another for 100% MI. It can be found on page 15 of this pdf:
http://www.npcamerica.com/pdf/SM5819.pdf
So my confidence that this is a legitimate problem is increasing. Next I will capture a regular CD through the modded SPDIF output and compare it to a rip of the same CD. This will show conclusively that the SPDIF mod is implemented correctly, and that we're dealing with a DSD decimation filter issue.
No problem. I've confirmed some clipping in a wav editor on the following:
1. Can: Future Days (also confirmed that there's no clipping on the CD layer)
2. David Bowie: Scary Monsters
3. Stevie Ray Vaughn: Couldn't Stand the Weather (very slight, in only a few places)
Others I suspect have clipping due to the recorded signal level, but haven't confirmed clipping in a wav editor yet:
1. Michael Jackson: Thriller (this one runs very hot on the signal meter!)
2. The Police: Synchronicity
3. The Pixies: Doolittle
Then there are recordings I've confirmed have no clipping or are perfectly peaked to fit within the available dynamic range of the Oppo's 88.2kHz output:
1. Peter Gabriel: Us
2. Boston: Boston
3. Bob Dylan: Love and Theft
4. Keane: Hopes and Fears (this one is a surprise because the recording is pretty compressed, i.e. a victim of the loudness war).
If none of these match your collection, I have many more pop/rock SACD titles (Dylan, Rolling Stones, Byrds, etc.), so name a few titles that you have and I may be able to check for clipping.
1. Can: Future Days (also confirmed that there's no clipping on the CD layer)
2. David Bowie: Scary Monsters
3. Stevie Ray Vaughn: Couldn't Stand the Weather (very slight, in only a few places)
Others I suspect have clipping due to the recorded signal level, but haven't confirmed clipping in a wav editor yet:
1. Michael Jackson: Thriller (this one runs very hot on the signal meter!)
2. The Police: Synchronicity
3. The Pixies: Doolittle
Then there are recordings I've confirmed have no clipping or are perfectly peaked to fit within the available dynamic range of the Oppo's 88.2kHz output:
1. Peter Gabriel: Us
2. Boston: Boston
3. Bob Dylan: Love and Theft
4. Keane: Hopes and Fears (this one is a surprise because the recording is pretty compressed, i.e. a victim of the loudness war).
If none of these match your collection, I have many more pop/rock SACD titles (Dylan, Rolling Stones, Byrds, etc.), so name a few titles that you have and I may be able to check for clipping.
OK, I have Stevie Ray and Synchronicity. I will test them out tomorrow, don't have time for it right now.
SACD does have a defined maximum level, and it is well above 50% modulation.
see the third page of this document:
http://ieeexplore.ieee.org/iel5/79/27334/01226728.pdf?arnumber=1226728
If we can reproduce the problem, then maybe Oppo can fix it in a firmware update.
It would be nice to get a picture of the analog output with a oscilloscope to send to them. My scope is an old HP analog scope. I would have to set the trigger, wait for a peak, and then take a picture.
SACD does have a defined maximum level, and it is well above 50% modulation.
see the third page of this document:
http://ieeexplore.ieee.org/iel5/79/27334/01226728.pdf?arnumber=1226728
If we can reproduce the problem, then maybe Oppo can fix it in a firmware update.
It would be nice to get a picture of the analog output with a oscilloscope to send to them. My scope is an old HP analog scope. I would have to set the trigger, wait for a peak, and then take a picture.
Sorry for the belated update, and for interrupting the current discussion, but I believe my Oppo is working now. After resoldering the clock signal wires to the Oppo board, I haven't had the persistent lock-ups I was before. I'm getting digital output from CD, SACD, and DVD discs.
I've attached a couple pics of the Oppo with the new board...
Ted
I've attached a couple pics of the Oppo with the new board...
Ted
Attachments
Hi Ross!
Thanks again for offering this mod!
So far I've only got one SACD in my library - Prokofiev's Romeo & Juliet from Telarc. It sounded good to me; perhaps Telarc would avoid mastering a SACD with the potential to cause clipping? For the discs that have clipping, is it the fault of the disc mastering or the player?
I hope to add more SACD's to my inventory over time, so I'll be interested to see which titles have this issue. I suppose DVD-A's are fine?
Ted
Thanks again for offering this mod!
So far I've only got one SACD in my library - Prokofiev's Romeo & Juliet from Telarc. It sounded good to me; perhaps Telarc would avoid mastering a SACD with the potential to cause clipping? For the discs that have clipping, is it the fault of the disc mastering or the player?
I hope to add more SACD's to my inventory over time, so I'll be interested to see which titles have this issue. I suppose DVD-A's are fine?
Ted
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.