[hq136185925]

Dear Schiit
I modified the Bifrost Multibit and then I found the most satisfying sound. I hope Schiit engineers try this out.

Here is my PCHIFI with Schiit Bifrost Multibit
1. Modified the original I/V OPAMPs to ADA4625-2
2. Changes USB to UNISON
3. Under Windows10. set sound as 16bit 176400Hz (Match your DAC precision if you has a NOS DAC. If you has a Delta-Sigma DAC, then choose the highest bit precision and 44100Hz) (Bifrost Multibit has a huge problem with downsampling from 32bit to 16bit and 48kHz series, let windows do this job)
4. Under Windows10, set system properties -> advanced -> performance -> Processor scheduling -> adjust for best performance of Background services.
5. Use the Quick CPU unlock CPU park (edit power plan manually on software if possible)
6. Use foobar2000 listen to music and install foobar Ramdisk and WASAPI components
7. foobar playback ReplayGain source mode none, Processing none
8.set the foobar2000 output as WASAPI(push) and buffer length 50ms (increase slightly if sound boom), output data format 16-bit (for Audio, Match your DAC precision if you has a NOS DAC. If you has a Delta-Sigma DAC, then choose the highest bit precision the DAC does support)
9.set the Potplayer output as WASAPI 2.0 output data format 16-bit Dithered (for Video)
10. Set foobar2000 Advanced Decoding Tone/sweep sample rate 705600, Playback WASAPI Hardware buffer push mode: enable high worker process priority

Advanced:
11. Use WEISS Saracon Convert audio file to 64bit 176.4Khz and then playback (don't forget to use Sonic Foundry 64bit Wave format)
12. Use WEISS Saracon Convert audio file to 16bit 176.4Khz dithered with powr3 (it depends on DAC precision )
sorry above said Bifrost has downsampling problem, actually, it was DAC chip precision not enough (AD5547 has 16bit only, in conclusion, do not down or upsampling the audio file that does not match your R-2R DAC chip bit precision, if bit precision not high enough, just dither them)

BTW:
1. ada4625 does not have a large current for output, it is best to work with Asgard3 as a pre-amp.
2. and always use foobar2000 Ramdisk function
3. 32bit audio provide really good sound atmosphere.

Actually, AD5547 is a greater chip than the other 5781 etc, because it has faster setting time and higher frequency. three bottlenecks of the multibit Bifrost first-generation, 1, the I/V opamp should replace to ADA4625, 2, the output of the SHARK DSP frequency should higher. 3, the audio file need to upsampling to 32bit and then downsampling to 16bit, the 32bit upsampling will provide huge sound image and music atmosphere, but the Bifrost multibit has a huge problem with downsampling (when using UNISON eat 32bit data) and output frequency (ad5547 has 6.8Mhz bandwidth! why not feed it in full) according to my listening experience.

as the result, Bifrost multibit 1 has two big problems on the sound, 1, the sound image is mess and noise, 2, some sound should appear at back but always appear at the front. the first problem can be solved by replacing the I/V opamp to ADA4625 and with UNISON USB or cleaner digital source. The second problem should be solved by re-design your digital filter (add Dither).

If AD5547 can run at 6.8Mhz, this would be revolutionary, this would take the advantages both R2R and delta-Sigma. but, you need very good I/V (ada4625 is the best choice) and best digital filter design.

actually, AKM is using 64bit upsampling, do you know the AK4191?

In Result: After using the above settings, the overall sound quality is almost perfect, The naturalness of the sound, the space of the sound, and the power of the sound can reach a very high level. It does not exquisite or smooth like Delta-Sigma (if Schiit can feed ad5547 in high speed, would be smooth) but overall sound quality reaches the perfect.

 

[hq136185925]

Since the Weiss USB dongle is hard to buy, here is a mysterious version

 

[manueljenkin]

This is a beautiful write up. I hope to try it someday if I get the bitfrost. Keep doing the great work

 

[hq136185925]

This is a beautiful write up. I hope to try it someday if I get the bitfrost. Keep doing the great work

cheers

 

[manueljenkin]

I started going through Weiss and the opamps you mentioned. You've opened me new portal to enter.

I would like to seek guidance from you regarding amplifiers. I wish to build a diy amplifier that can put out a little over 500mW into 40-80ohm loads (I wouldn't be bothering for loads lower or higher than that, since the headphones I use are almost always within that range). I want it to be battery powered and common wisdom tells me that I'll need +/-6V rail to rail at the minimum hence two 7.2v batteries (or even two 9v batteries regulated to about 7.5 or 8V to be on safer side) But how should I go about the amplification part. Ad8637 seems to be able to push out 310ma current which would cover my requirement, but the amps I've seen using them like Amb mini 3 are like 200mw into 32ohm which is a far cry from what I desire. Wonder if you can guide me. I don't want it to be crazy expensive.

Power aside, the Amb mini 3 is very interesting, the active ground etc to help get proper biasing currents for the ad8637.

 

[hq136185925]

I started going through Weiss and the opamps you mentioned. You've opened me new portal to enter.

I would like to seek guidance from you regarding amplifiers. I wish to build a diy amplifier that can put out a little over 500mW into 40-80ohm loads (I wouldn't be bothering for loads lower or higher than that, since the headphones I use are almost always within that range). I want it to be battery powered and common wisdom tells me that I'll need +/-6V rail to rail at the minimum hence two 7.2v batteries (or even two 9v batteries regulated to about 7.5 or 8V to be on safer side) But how should I go about the amplification part. Ad8637 seems to be able to push out 310ma current which would cover my requirement, but the amps I've seen using them like Amb mini 3 are like 200mw into 32ohm which is a far cry from what I desire. Wonder if you can guide me. I don't want it to be crazy expensive.

Power aside, the Amb mini 3 is very interesting, the active ground etc to help get proper biasing currents for the ad8637.

I can not find any datasheets regarding ad8637, maybe its AD8397
I searched Mini 3 uses OPA690 or LMH6642 as entering amplifier, try to replace them as ADA4625-2 or AD8512 or OPA1656 or OP275 or ADA4898

 

[manueljenkin]

My apologies, meant to mention Ad8397. The high power edition of Amb mini 3 uses ad8397 for amplification and opa690 for ground and bias stability (active grounding). I would likely need about 150-200mA of current headroom into 40-80 ohms to drive my headphones reasonably well.

I can not find any datasheets regarding ad8637, maybe its AD8397
I searched Mini 3 uses OPA690 or LMH6642 as entering amplifier, try to replace them as ADA4625-2 or AD8512 or OPA1656 or OP275 or ADA4898

 

[hq136185925]

Use WEISS Saracon Convert audio file to 16bit 176.4Khz dithered with powr3 (it really depends on DAC precision )

sorry above said Bifrost has downsampling problem, it was DAC chip precision not enough (AD5547 has 16bit only, in conclusion, do not down or upsampling the audio file that does not match your R-2R DAC chip bit precision, if bit precision not enough, just dither them)

 

[hq136185925]

The biggest problem of Schiit DACs is bit precision not enough, Schiit should give a solution with Dither, not just make the sound noise to cheating on auditory
for higher bit hope I can get a http://www.soekris.dk/index.html since all product has been sold out.

The benefit can get from Soekris is more bits, But I would worry about the precision of the discrete resistors. Because AD5781 has ±0.5 LSB INL, ±0.5 LSB DNL which is an 18bit DAC, and this is the best of the chip industry. Soekris did not give the INL/DNL of the DAC, we cannot verify the linearity of this DAC. According to my listening experience, the sound is not accurate, in other words, it is not natural enough.

 

[hq136185925]

My apologies, meant to mention Ad8397. The high power edition of Amb mini 3 uses ad8397 for amplification and opa690 for ground and bias stability (active grounding). I would likely need about 150-200mA of current headroom into 40-80 ohms to drive my headphones reasonably well.

the one i used is schiit Asgard 3. Im not familiar with headphone amp

 

[hq136185925]

Btw How can I make Jason Stoddard have a look this thread

 

[manueljenkin]

I think Schiit has a thread, you could possibly request him to have a check by providing the link.

 

[ev666il]

I think you might stand a better chance by name tagging him in this thread, like this: @Jason Stoddard

 

[hq136185925]

I think Schiit has a thread, you could possibly request him to have a check by providing the link.

Cheers

 

[hq136185925]

I think you might stand a better chance by name tagging him in this thread, like this: @Jason Stoddard

Cheers