the pic above is from first watt M2
Hello everyone. I am DIYing a resistor-ladder DAC. it's similar to MSB platinum DACs.
I am a FPGA enginner, have no experience in analog circuit design.
Here's the thing. I have built a voltage divider to generate the signal voltage. But I do not want to follow an opamp as the output stage. I would love to try the most musical solution, just like first watt does.
Therefore, I am asking for help.
1.Pure class A, no feedback.
2.MOSFET probably.
3.signal input is 0-5V. signal 'zero' is 2.5V. final output is about 2.5V peak to peak.
I have no idea what MOSFET to choose, or how to calculate the values of nearby resistors and caps.
Please feel free to give any suggestion or idea. Thanks!
Transformer coupled? What transformer do you intend to use?
Purists typically don't like audio transformers and designers avoid them unless they offer a significant advantage. What advantage does this transformer coupled circuit offer over a typical modern class A buffer?
Use an active biased class A emitter or source follower or just (GASP) use an op amp.
Or if you're just trying to be cool, use a cathode follower buffer. There are even modern lower voltage tubes, and PC mount tubes. But using these devices might cost you snob factor.
Purists typically don't like audio transformers and designers avoid them unless they offer a significant advantage. What advantage does this transformer coupled circuit offer over a typical modern class A buffer?
Use an active biased class A emitter or source follower or just (GASP) use an op amp.
Or if you're just trying to be cool, use a cathode follower buffer. There are even modern lower voltage tubes, and PC mount tubes. But using these devices might cost you snob factor.
Thank you, Eddie.
What I intended to show the figure is the simplicity. It is a Nelson Pass's work. I need no transformer.
Since a DAC analog output does not consume large power, so I probably go for a source follower. The problem is I do not know how to design this. To be specific, which MOSFET, values of the resistors or caps, slew rate and etc.
What I intended to show the figure is the simplicity. It is a Nelson Pass's work. I need no transformer.
Since a DAC analog output does not consume large power, so I probably go for a source follower. The problem is I do not know how to design this. To be specific, which MOSFET, values of the resistors or caps, slew rate and etc.
Well you showed a circuit with a transformer. I thought you wanted to use that cirucit. I would never use a circuit like that for a buffer.
So just to be clear, you are familiar with digital circuit design but not analog circuit design, right?
I need to know what your input impedance requirements are for the buffer, and what kind of power supply you have for the buffer.
There are a lot of buffer circuits available here and on a google search. Which one is for you depends on your answer to the above questions.
I'm not a SPICE guy, and I have never used schematic software either. I come from way back when transistors were new in hi fi equipment. That being said, I like to use op amps extensively in my designs. Despite the woo, op amps can outperform typical discrete circuits for many audio applications. Simple discrete line level buffers seem kind of snobby to me when you can make an excellent buffer with a couple buck$ worth of parts and a cheap op amp.
Member
Joined 2009
Paid Member
Hi iceyo,
I am not clear what your plans are - is it a commercial DAC in a box, or do you have a DAC and want to interface to the DAC chip ?
I believe a transformer is a very valid option and I don't agree with Fast Eddie on it not being the choice of purists - there are many purists who build vacuum tube amplifier for whom transformers are the purest of all approaches.
Sowter make a transformer specially for coupling to DAC chips which many people say is a very musical solution.
If you already have a DAC with line output then you are asking about a power amplifier. If this is your first project then I would suggest picking a design where the parts values are known and published. If you want a Nelson Pass design, there are many and you should go to the Pass Forum, not this one, and ask there for advice - you will get good support on building Nelson's published designs there.
Good luck - DIY is fun !
I am not clear what your plans are - is it a commercial DAC in a box, or do you have a DAC and want to interface to the DAC chip ?
I believe a transformer is a very valid option and I don't agree with Fast Eddie on it not being the choice of purists - there are many purists who build vacuum tube amplifier for whom transformers are the purest of all approaches.
Sowter make a transformer specially for coupling to DAC chips which many people say is a very musical solution.
If you already have a DAC with line output then you are asking about a power amplifier. If this is your first project then I would suggest picking a design where the parts values are known and published. If you want a Nelson Pass design, there are many and you should go to the Pass Forum, not this one, and ask there for advice - you will get good support on building Nelson's published designs there.
Good luck - DIY is fun !
Purists in the US, perhaps. But it's completely the opposite in Japan. Just sayin'.
Designers don't like them because they are usually quite expensive. But sometimes they are the proper answer to the question.
Thank you. There is no DAC chip. the D to A module is built from a resistors network (R2R). I can choose an opamp as output buffer, but I do want a dicrete one. I have 5V,+12V,-12V on board. I think the input impedance of a MOSFET is large enough.
I am a SPICE guy and I respect experience, but I try to be a know-how guy. I will simulate the circuit.
If you haven't discovered already, one of the major problems with R2R DACs is glitching. That's a reason why DAC chips for audio don't use R2R in an unmodified form. You've chosen a very steep path to become an experienced analog engineer, I wish you a lot of luck
Vincent Brient has done this incidentally - http://www.vishaypg.com/foil-resistors/case-studies/study/totaldac_1/
Last edited:
Thank you. Copying one's circuit and make it sings will not satisfy me. I start to learn basic circuits and simulate.
If you haven't discovered already, one of the major problems with R2R DACs is glitching. That's a reason why DAC chips for audio don't use R2R in an unmodified form. You've chosen a very steep path to become an experienced analog engineer, I wish you a lot of luck
Vincent Brient has done this incidentally - Vishay Precision Group - Foil Resistors - Case Study: Developing a Multi-Channel DAC for High-End Audio Applications
thanks.
1. I am not trying to be an analog expert, just for a deep understanding of a voltage buffer circuit.
2. Clearly, you misunderstand the R2R's problem. glitching can be eliminated by clock and registers. As your link says, the problem comes from the accuracy of the resistors.
Well if my misunderstanding is clear to you, please explain where I'm wrong. How do clock and registers eliminate timing mismatches in switches?
Incidentally as you're in China, you can find a Chinese published version of this book much more cheaply than anyone in the West. I have it and its excellent. Amazon.com: Wide-Bandwidth High Dynamic Range D/A Converters (The Springer International Series in Engineering and Computer Science) (9781441940261): Konstantinos Doris, Arthur H.M. van Roermund, Domine Leenaerts: Books
Incidentally as you're in China, you can find a Chinese published version of this book much more cheaply than anyone in the West. I have it and its excellent. Amazon.com: Wide-Bandwidth High Dynamic Range D/A Converters (The Springer International Series in Engineering and Computer Science) (9781441940261): Konstantinos Doris, Arthur H.M. van Roermund, Domine Leenaerts: Books
Last edited:
For example, clock 1 is 1Hz, clock 2 is also 1Hz with 0.5s delay, glitches happen at clock1, 1s,2s,3s,...and so on with a range, let's say +-0.1s. but the final conversion happens at clock 2, 1.5s, 2.5s, 3.5s... and so on.
See, simple.
Well of course you need a transformer for a tube output circuit. People love their tubes for various reasons, and I like them too.
I'm not a hater and will be the first to admit that for some applications a transformer is a viable option. For example, a microphone or phono cartridge input is a good place for a transformer; you can have a balanced input and voltage gain in one package. The advantages of using a transformer in these applications should be obvious.
I didn't know that. Do you have a quick link?
FYI, I'm not up to speed on digital audio circuits. I'm working on that.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.