You probably know my first ultimate DAC built here is the non-oversampling TDA1541A DAC that built to emphasis on better overall layout, better components and better power supply design. It sound great by the way. Then, all my next DAC built is always having the same principal as that Ultimate DAC built in mind.
Today DAC build is another story. This is started when I'm read so many discussion on the most popular DIY audio forum on internet about building a DAC with mostly all in the same guideline. Great DAC should this and should that said the "experienced" forum members. This makes me restless. It is seems there is a strict line guides created by the members on how to make a great DAC. On the other hands, they killing the creativity. Like this example, why the non-over-sampling DAC will be sound more natural and better? If so the over-sampling would be sound worse? Why the IV stages should be using as simple as a resistor per channel and avoid the op-amps? Is the op-amps really that bad?
And so on, and so on.
That is why, this ultimate DAC build will emphasis on what everyone (mostly) on the internet avoiding at, but what I personally think it is better, theoretically. Please read on.
The first things that I will share in this part-1 is the digital receiver.
We are spoiled by CS8412 or CS8414 for decades. Don't go wrong with me, I also really like this CS8412 CS8414 as the simplest digital receiver IC to build. This IC also have so many features, especially for the serial digital output mode can be choose from I2S, 16 bit and 18 bit two complementary. But the CS8412 CS8414 is long discontinued. It is long gone and hard to find the original now. I have bought one few years ago, but sadly it is not operate properly and I guess it is fake.
On my hands, I have this original Yamaha YM3623 digital receiver dismantled from the old unit. This YM3623 is the most popular digital receiver IC among the audio manufactures in late 80's to early 90's before the CS8412 take the places. But this YM3623 is not so popular among do it yourself just because the CS8412 CS8414 are more easy to build. But do YM3623 will not sound as good as CS8412 CS8414? I also don't know if I am not using it right?
One reason why this YM3623 is not popular because the serial output from this IC is only 16 bit two complements. Then mostly the input to the DAC IC are 16, 18, and 20 bit with two complements serial input. The example is AD1862, AD1865, PCM63, PCM58, PCM1702. They also require separate DATA input for right and left channel. With that reason, this YM3623 should go first to the over-sampling digital filter IC before go to the DAC IC's like above. The connection between YM3623 to the digital filter over-sampling IC will be share later in the part 2.
So in this post, I will share the schematic for the SPDIF to digital input PIN 28 of YM3623 which is quite different compares to the CS8412 CS8414 digital input. This is maybe one of the reason why YM3623 are not so popular.
YM3623 digital input is on PIN 28. This input require TTL input, unlike the CS8412 CS8414 that you can directly connect the SPDIF input by using only 10 nF capacitors for both input RXP and RXN.
When using this YM3623 as a SPDIF receiver, you need to convert the input signal to TTL compatible output first. There is few ways to do this actually. The one I share here is the most popular method. It is the most simple method to convert the digital SPDIF signal to TTL output to PIN 18 YM3623.
You can see from the schematic above, it only using a 74HCU04 as un-buffered hex inverter. One IC contain 6 separate hex inverter and this schematic is so simple that we only need 2 of them to convert from SPDIF to TTL compatible data output. This is the most popular schematic that you can find in most all DAC in the late 80's. But I am not using this schematic.
Let's we going to look to the second schematic for this YM3623 digital input.
The schematic above is taken from the original Kinergetics Research KCD-55 Ultra I have here
As you can see this is a different approach from the first schematic. It is using SN75157 as dual differential line receiver. For this application, it only using one. Then the output from SN75157 directly connect to 74LS153 as the input selector and buffer signal. I am not completing all the schematic of this Kinergetics Research digital input, respecting they proprietary design. But you can figure out on how it works.
SN75157 is a dual differential line receiver with TTL compatible output. This really suitable for the PIN 28 input of YM3623 receiver. You can substitute this IC with others similar function like I am using on my previous DAC PCM1702 project here using SN75176 or UA9637. They have similar function but the IC's pins arrangement are all different. You should check to the datasheet first.
Then for this ultimate DAC build here, I am using the schematic bellow.
My approach here is also simple and using the same dual differential line receiver SN75157. Then the signal going to the popular 74HC04 as buffered hex-inverter. Because it invert the signal, so I need two in series to keep the signal non-inverted.
The most significant different in my design is by using the SPDIF input transformer. Using this transformer for digital input could be the avoided by the others, maybe because this transformer is quite expensive and using capacitors is just as simple as that. This 1:1 input transformer is the galvanic isolation between the source to the DAC receiver. I will using it in this DAC build since this is theoretically better.
So that's all for this part 1.
On the next post, I will share about the components and the connection from digital receiver to the digital filter.
Disclaimer: Any statement and photos in this article are not allowed to copy or publish without written permission from the writer. Any injury or loss from following tips in this article is not under writer responsibility.
When I first used the YM3623B S/PDIF input circuit I found it was extremely jittery. I found a way to reduce the jitter though - once the signal is locked, disable its internal oscillator. Even with this fix, I think other receiver chips (Crystal for example) are lower jitter.
ReplyDeleteYes, agree. YM3623 is not the best receiver IC compares to CS8412/4. But finding the original CS8412/4 is hard, alternatively using YM3623 is way better
Delete