TDA1541A Diy Pcb - "Distinction-1541 v2"

...just more spectrum analysis pics

Hi Guys,

Just a couple more pics to share.

Tonight I thought i'd try powering the 1541 PSU with a switch mode power supply (Gophert CPS-3205) to compare the noise difference with the SLA batteries as I used previously. Looks like the switch mode is quieter... Didn't expect that! Looks like roughly 5.6uV / -105dBV compared to ~ 6.3uV / -104dBV - If i've done the correct calculations.

I've also attached the +5V digital supply noise spectrum analysis. Around -100dBV / 10uV.

All measurements were taken in circuit under load while playing music.

Also, these measurements were taken with 5 point averaging, the high frequency noise may be irrelevant. All I had to do was put my hand on the USB measuring device to completely change the spread of noise in the HF (+100k). I'm weary of the accuracy of the VT DSO-2815H, I think these measurements should be taken as an indication only.

Cheers
 

Attachments

  • gophert -15.jpg
    gophert -15.jpg
    460.5 KB · Views: 943
  • digital supply.jpg
    digital supply.jpg
    460.3 KB · Views: 913
Last edited:
Diode attenuator revisit.

Hi All,

I'm slowly but surely getting through all the testing and fine tuning on my pcb.

The Diode attenuators were revisited and evaluated.

Test setup:

Probes set to 10X
LE - Diode attenuation (1V - 2V swing)
DR, DL, BCK resistive attenuation (1V - 2V swing)

Refer to picture attached.

Comparison videos:

LE diode attenuator for TDA1541A
LE no attenuation

Resistive attenuator for LE was also evaluated and found to perform poorly.

My preference for attenuation of LE in simultaneous mode:

1. Diode attenuator
2. No attenuation
3. Resistive attenuation

I'm not sure how well the comparison videos will show the sound difference given the quality of my smartphone, but to my ears its a clear winner.

Ryan
 

Attachments

  • diode atten revisit.jpg
    diode atten revisit.jpg
    434.3 KB · Views: 656
Update.

Hi All,

Attached is a picture of my latest prototype (hopefully the last).

This one is for a friend, if all tests well.

PCB features:

-4 layer design.
-Rubycon PMLCAPS.
- 22uF unity gain cap for shunt regs.
- 2.2uF MSB (pins 13, 18).
- 1.5uF (pins 12, 11, 10, 19, 20, 21).
- 100nF C0G/NP0 (pins 22, 23, 24, 7, 8, 9).
- Diode Attenuator for LE (pin 1).
- Very low output impedance (~3mΩ in the audio band for -5, +5, -15)

There are a few parts on backorder but I should be able to test this PCB within a week or two.

Ryan
 

Attachments

  • sneak peak.jpg
    sneak peak.jpg
    735.2 KB · Views: 783
1. Diode attenuator
2. No attenuation
3. Resistive attenuation

I'm not sure how well the comparison videos will show the sound difference given the quality of my smartphone, but to my ears its a clear winner.

Ryan
Hi Ryan
Great work.. You are on the home stretch. Listening to your video, it is shocking how much better the Diode attenuator sounds even on YouTube.
It would be interesting if you can post a video comparing the final V2 vs V1🙂
Walter
 
Hi Ryan
Great work.. You are on the home stretch. Listening to your video, it is shocking how much better the Diode attenuator sounds even on YouTube.
It would be interesting if you can post a video comparing the final V2 vs V1🙂
Walter

Hi Walter,

Yes finally nearly there! Thanks for your support along the way.

Yes it is interesting being able to pick up on the difference of only a few different parts in a system in person, let alone through a medium such as youtube with audio compression, etc.

I've since salvaged a few parts off my V1 PCB, but it would be a good comparison to see how far its come. It's actually hard to remember exactly what it sounded like.

Its funny. I remember only 5 or so years ago thinking - digital audio can only have one kind of sound because its digital. How very, very wrong I was. Everything about the conversion process is absolutely critical.

Cheers.
 
Nice work, any PCB's left 🙂? Do I see it correctly: you use the diode attenuator (which will give sharper edges I guess) on the LRclk and resistive attenuators on the other I2S signals?

Best regards,
Berny

Hi Berny,

The final version is almost complete (3 prototypes later).

If you could put your name on the interest list HERE ill make sure you get one.

Yes, for LE/WS im using a diode based attenuator, with the option on the PCB of using the resistive attenuator or no attenuation.

For DATA L, DATA R, and BCK im using the resistive attenuators, as these further limit the energy and bandwidth resulting in less on chip crosstalk.

Cheers.
 
Update.

Hi Guys,

Attached are some photos of Distinction-1541 V2.

It sang straight to my heart last night, you can really feel the emotion in the music, very pleased to have it im my system. I feel content to move on and finalise this design.

I've also attached some photos I took last night of how the attenuators are looking. The waveform on the right is from the diode attenuator (for LE), on the left is the resistive attenuator (for BCK, DL, DR). Probes set to x10.

My full system setup for reference:

QA550 ---> FIFO ---> I2S to PCM ---> Distinction v2 ---> CEN ---> Pass F5 Clone ---> Tang Band W8-1772 8" Neodymium Full Range Driver in Bob Brines TT200.

Today I will go over all the circuitry and make sure its behaving itself, then in the coming days upload the schematics and bom.

Cheers.
 

Attachments

  • Top.jpg
    Top.jpg
    965.4 KB · Views: 769
  • Bottom.jpg
    Bottom.jpg
    784.5 KB · Views: 589
  • v2 final atten.jpg
    v2 final atten.jpg
    751.8 KB · Views: 554
Awesome. I can imagine this is about as good as it gets. Thanks for sharing the complete audio chain. The SD into Ian's stuff as a source must be a great source.
What are you using now for the 26v source?

No problem 🙂

The SD player is great in terms of sound quality, but a bit awkward to use due to the nature of SD cards. Not a bad compromise but I will try the WaveIO when I get a chance (your old one 😛).

26V source is coming from the Lab power supply at the moment, im really surprised it can sound as good as it does with a switch mode psu. Definitely upgrade soon though. It would be interesting to try a Capacitance multiplier, or a zener + emitter follower before the zener string. It may be sufficient enough to go with a very basic schottky rectifier ---> CLCRCR.

Cheers