Does the denon 3930 also have the poor pll-clock with 450ps jitter as the 2910 and the 3910 do?
I was considering to buy one secondhand, but only if they did a better job with regards to the clock.
I was considering to buy one secondhand, but only if they did a better job with regards to the clock.
The jitter specs of the pll of the 3930 is much better (about 50ps). I'm puzzled why it is fed with the output of a RAM-chip, instead of directly from a crystal.
For best audio change the clock of the dsp-chip and improve the power supply.
For best audio change the clock of the dsp-chip and improve the power supply.
The chip that does the al-24 processing has the 24...Mhz clock and also does the spdif out. So it's a good candidate for reclocking.
27Mhz is used for al the video stuff.
dsp is clocked with 25Mhz
27Mhz is used for al the video stuff.
dsp is clocked with 25Mhz
The music streem could go from the main processor through the FPGA and directly to the DAC if im not using the al24 processing.. right?

Professional modders considder the 24...Mhz clock the most important for audio.
The fpga is connected to the "al-24" chip. In the blockdiagram you can considder them as one.
According to denon, all audio is al-24 processed.
The fpga is connected to the "al-24" chip. In the blockdiagram you can considder them as one.
According to denon, all audio is al-24 processed.
I am curious where did you get the 450ps value for the jitter?
The jitter on audio part is based on the 24MHz one when you are in PCM mode and on the 24.576MHz only in SACD DIRECT mode. Both are good clocks.
The jitter on audio part is based on the 24MHz one when you are in PCM mode and on the 24.576MHz only in SACD DIRECT mode. Both are good clocks.
Last edited:
The pll of The 3910 (sm8701) and the 2910 has a jitter-spec of 450ps with fs-frequencys higher then 48Khz (so 88,2K, 96K, 172K and 192K)
The AL24 DSP seems to run at 25MHz
Some companies is swapping the 27MHz clock
Some recommends the 24.576 MHz clock!
But if all music media sources is processed by the AL24chip DSP then maybe the 25MHz one is the one!? Sorry for my lack of knowledge, i just get confused when i get three different suggestions from people who knows...
Same block schematic as above but in two parts here, left and right side:
Some companies is swapping the 27MHz clock
Some recommends the 24.576 MHz clock!
But if all music media sources is processed by the AL24chip DSP then maybe the 25MHz one is the one!? Sorry for my lack of knowledge, i just get confused when i get three different suggestions from people who knows...
Same block schematic as above but in two parts here, left and right side:


I think the 24.576 MHz clock is the most important for audio. With the 27Mhz as second.
But there are a lot of chips with a lot of clocks. A professional modder did them all in the 3910.
If your good with a soldergun buy two or three xo's from tentlabs, build a low noise ps for them and feed the important chips with them. This wil cost you about 100,-/120,- euro
But there are a lot of chips with a lot of clocks. A professional modder did them all in the 3910.
If your good with a soldergun buy two or three xo's from tentlabs, build a low noise ps for them and feed the important chips with them. This wil cost you about 100,-/120,- euro
I already bought a clock from Guido Tent with complete PCB.
I will mount it, install some JLH shunts, build a new I/V converter and analoge stage (ZEN I/V from Nelson Pass), and finally add capacitors to various digital chips incl the PLL supply, maybe also lowering the PLL feedback frequency to reduce jitter.
I will mount it, install some JLH shunts, build a new I/V converter and analoge stage (ZEN I/V from Nelson Pass), and finally add capacitors to various digital chips incl the PLL supply, maybe also lowering the PLL feedback frequency to reduce jitter.
What is feedback frequentie and how can you reduce that?
Does it have anything to do with termination?
Does it have anything to do with termination?
SACD doesn't get the AL24 treatment, it goes straight to the DAC's. Clock for the DSD decoder chip is the 27 MHz via PLL in SM8707 - 33.8688MHz. That is derived from the 27MHZ XTAL via the MN2DS chip...
The 24.576MHz one looks like it is used only for communication via Denon Link (it generates RX4 signals, that is the reverse channel of the DLIII)?
The 24.576MHz one looks like it is used only for communication via Denon Link (it generates RX4 signals, that is the reverse channel of the DLIII)?
Last edited:
I think you are right about the dsd. Maybe there are settings in the menu that control dsd to pcm conversion if at all.
Maybe the 24.576Mhz clock is responsible for upconverting to 24bit/192Khz?, that would make it more important.
Maybe the 24.576Mhz clock is responsible for upconverting to 24bit/192Khz?, that would make it more important.
The DSP runs at 25 MHz, it needs to execute some operations in the 192KHz datastream.
25MHz - 24,576MHz is 424kHz
Since 424 kHz is 2,208 times faster than 192 kHz music stream, could that be the spare time where every bit is executed?
Is it logical that the DSP has to run faster than the music stream?
24,576 is 192kHz*7
It´s 7 channels in this player in total, FL FR C SW SL SR MIXL MIXR the two latest is 2 channel direct audio.
Maybee I shall hook up the scope to that PCM1796 dac tomorrow [🙂]
25MHz - 24,576MHz is 424kHz
Since 424 kHz is 2,208 times faster than 192 kHz music stream, could that be the spare time where every bit is executed?
Is it logical that the DSP has to run faster than the music stream?
24,576 is 192kHz*7
It´s 7 channels in this player in total, FL FR C SW SL SR MIXL MIXR the two latest is 2 channel direct audio.
Maybee I shall hook up the scope to that PCM1796 dac tomorrow [🙂]
The DSP is not running syncron with the datastream, it has internal RAM, software FIRs... It's like approximating the 192kHz signal with some 550-1100 intermediate samples if you want.
The 24.576MHz clock I can see that is only responsable with the RX4 channel from the DenonLink - that is the communication (for encryption of DSD and high res PCM?) from the external DAC (receiver) and also I think the external DAC can send a sync signal thru that so the PLL will lock to the outside clock.
If you don't use Denon Link it doesn't matter (I do with a Denon AVR 3805 receiver).
The 24.576MHz clock I can see that is only responsable with the RX4 channel from the DenonLink - that is the communication (for encryption of DSD and high res PCM?) from the external DAC (receiver) and also I think the external DAC can send a sync signal thru that so the PLL will lock to the outside clock.
If you don't use Denon Link it doesn't matter (I do with a Denon AVR 3805 receiver).
Yes! RX4 comes from the FPGA and is going IN to the IC107 LC89057W-VF4A which is clocked by the 24,576MHz X-tal.
So i guess you are right! The 24.576clock is not the one to change then...
Leaving the 27 and 25 MHz clock left..
So i guess you are right! The 24.576clock is not the one to change then...
Leaving the 27 and 25 MHz clock left..
- Status
- Not open for further replies.
- Home
- Source & Line
- Digital Source
- denon 3930 pll