2Fs reconstruction filter of Sony using gyro FDNR to built.

[Hayk]

Sony CDP502 was the best CD player of the 1986. It uses a special type of filter called Frequency Dependant Negative Resistance realized using gyro opamps.
The response, I simulated is bellow.
The cut frequency can be changed by the capacitors value of 2.7nF designed for 88khz.
I am making a PCB for it in case someone would like to built it. I added an optional buffer if the source impedance is high, I replaced Sony line drivers at the output by TPA6120 for differential output with volume control and powering it by low noise regulators with 1mohm output impedance.
The opamps are designated as NE5532 as the original but using 8 SOIC, it can be replaced by LM4562 or OPA1656. I used 1/8w 1% resistors and THT capacitors for DIY friendly, I can make also with SMD to be ordered from JLCPCB fully built if needed.

 

[solderdude]

What would be the use case for such a PCB other than replacing the filter for that specific CDP of a bygone era ?

It could be used for making a 3 section fixed parametric EQ when one knows how to calculate the values but that can be done easily for free.

It might be best to put all your projects in the DIY forum instead of 'DAC reviews' forum.

DIY Audio Forum

Let us know what you have built yourself that is cool and fantastic.

www.audiosciencereview.com

 

[Hayk]

I will ask the moderator to transfer it. Thanks.

 

[Hayk]

Those who are using NOS DACs, they can make a compromise by using SRC4190 sample rate converter to upsample to 96k and use this filter.

 

[solderdude]

Those who are using NOS DACs don't want filtering though.
They think that filters reduce fidelity for some obscure reason and want filter-less for that reason.
This one starts to filter above 30kHz.

 

[Hayk]

Nos users end up with over 9th order 20khz, passive filter which reduces the transient at high frequencies. With 2.2nf instead of 2.7nf for 96k it results faster filter.
The relative phase shift at 20khz is 7° compared to 2khz. This is the compromise, to upsample inorder to use faster filter.

 

[solderdude]

They don't want to hear about any filter.... a filter 'filters' and thus 'removes' sound and they want the 'pure' sound and super fast square-wave response.

There is no reasoning with filter-less NOS people ... the word 'filter' is BAD !

Sure up-sampling seems to be allowed but then it isn't NOS but has become filter-less 2xOS.
Up-sampling comes with super sharp filters but don't tell them that there is a filter being used.

 

[Hayk]

I am developing this board to use it along the mother board

Which DACs are interesting for modest DIYs running 16bit/48khz?

I am designing a quasi universal board that can aquire many types of DACs. It has 2 selectable inputs, a ASRC 4190 board, interface to drive right and left justified DACs in differential mode and OPA1632 as differential output. The DACs are on seperate modules to be plugged in which requires...

www.audiosciencereview.com

It can eventually interest other people. It is the best sounding reconstruction filter known.

 

[Hayk]

Building the ultimate NOS DAC using TDA1541A

The end result is linear interpolation... your filtering must still be designed accordingly. :cool:

www.diyaudio.com

 

[RandomEar]

I am developing this board to use it along the mother board

Which DACs are interesting for modest DIYs running 16bit/48khz?

I am designing a quasi universal board that can aquire many types of DACs. It has 2 selectable inputs, a ASRC 4190 board, interface to drive right and left justified DACs in differential mode and OPA1632 as differential output. The DACs are on seperate modules to be plugged in which requires...

www.audiosciencereview.com

I think you developing / re-engineering an older filter design is neat. If anything, it helps you and people interested in it understand why and how Sony designed this.

It can eventually interest other people. It is the best sounding reconstruction filter known.

If this sounds audibly different than any standard linear phase fast roll-off filter, it would point to a less faithful reconstruction of the signal. I doubt it will, based on the simulations you have shared. It might sound different than pure NOS, though, because there is essentially no (average) treble roll-off in this design according to your plots. That would definitely be positive. However, it is kinda ironic that there is a need to upsample and then filter a signal for a Non OverSampling DAC

 

[RandomEar]

They don't want to hear about any filter.... a filter 'filters' and thus 'removes' sound and they want the 'pure' sound and super fast square-wave response.

There is no reasoning with filter-less NOS people ... the word 'filter' is BAD !

Sure up-sampling seems to be allowed but then it isn't NOS but has become filter-less 2xOS.
Up-sampling comes with super sharp filters but don't tell them that there is a filter being used.

I wouldn't be so sure about that. Audiophiles in the NOS-religion aren't looking at this topic objectively. Offering tham an old-school "analog" filter which is even hand made by an "audio wizard" could actually be appealing to them. Don't expect them to think this through or even understand it in detail - people in that group decide mostly based on gut feeling. If you combine the right buzzwords, I bet they will love slightly obscure stuff like this

 

[Hayk]

I think you developing / re-engineering an older filter design is neat. If anything, it helps you and people interested in it understand why and how Sony designed this.


If this sounds audibly different than any standard linear phase fast roll-off filter, it would point to a less faithful reconstruction of the signal. I doubt it will, based on the simulations you have shared. It might sound different than pure NOS, though, because there is essentially no (average) treble roll-off in this design according to your plots. That would definitely be positive. However, it is kinda ironic that there is a need to upsample and then filter a signal for a Non OverSampling DAC

I didn't say it is NOS, I said it is best compromise. The filter subtracts from direct signal from resistors what to eliminate. The opamp run at unity gain in the bandpass region.

Frequency-dependent negative resistor - Wikipedia

en.wikipedia.org

I remind that I am not the designer of this circuit.

 

[RandomEar]

I didn't say it is NOS, I said it is best compromise. The filter subtracts from direct signal from resistors what to eliminate. The opamp run at unity gain in the bandpass region.

Frequency-dependent negative resistor - Wikipedia

en.wikipedia.org

I remind that I am not the designer of this circuit.

I understand, but you mentioned it could be used with a NOS DAC. I just think it would be kind of ironic to use the suggested oversampling to 96k plus this filter to improve the signal from a non-oversampling DAC.

 

[Hayk]

Audiophiles in early 80's got confused what is oversampling and digital filter. To reconstruct the samples into continuous wave, FIR filters were the simplest way to achieve although it doesn't calculate the interpolation which needs fast multipliers. As the FIR filters needed to oversample 4 times at least just by stuffing zeros between samples, digital filter and oversampling meant the same thing.
By this NOS, is to be understood, no FIR zero stuffing digital filter.
The 44.1khz reconstruction filter is very slow and results resonant 20khz. By resampling to 96khz, this filter can be used to obtain much faster response as I showed.