L0rdGwyn's DIY Audio

[goldwerger]

So @goldwerger , I think the issue is you are mis-equating voltage gain from how much current is being demanded of the amplifier from the transducer, which is a more objective translation of the term "drive" that audiophiles use. Let's estimate roughly how much power and current is being utilized by each of the two transducers, your B&W 805 and the Hifiman Suvara.

Solid state power amplifiers typically do not have impedance matching functionality, they have a fixed voltage gain, in this case 31dB for your Moon power amp. How much current goes into the load is determined solely by the impedance of the load. So, for the same voltage output, a higher impedance will draw less current.

For the sake of this example, let's see how much voltage the amplifier is outputting for your 805 speakers. They are 88dB/W efficient at 8ohms. Based on Ohm's law, Power = Voltage^2/Impedance, so 1W into 8ohms requires 2.828V.

With your amplifier's 31dB of gain, we have to do some math to see how much voltage is coming from your source to get 2.828Vrms. So, output in dB = 20log(Vout/Vin). 31dB=20log(2.828V/Vin), doing the arithmetic, Vin = 0.08V. That is what is going into your amp from the source to get a 88dB listening volume from your speakers at 1m, which is closer to 84dB at 9-10feet away from the speakers, which is more typical listening position (not sure where you sit, I am taking some liberties here).

Okay, so 0.08V. And you say that you have to increase the volume from your source +8dB when listening to the Susvara. So, we can then calculate how much voltage is going in to listen to the Susvara using the same logarithmic equation. 8dB=20log(Vout/0.08), which results in a Vout of 0.2V.

Putting 0.2V into an amplifier with 31dB of gain, doing the same math, results in an output voltage of about 7V. So as you would expect, more voltage in means more volts out. HOWEVER this 7V output is now going into a 60ohm load for Susvara, as opposed to the 2.828V going into the 8ohms of the 805 speaker.

The 805 as we said are utilizing 1W into 8ohms, more math by Ohm's law says that the speaker is taking roughly 350mA.

For Susvara, we are seeing 7V into 60ohms, that is roughly 0.811W drawing around 116mA of current. That is an impressive current draw for a headphone and would seem to suggest Hifiman's 83dB sensitivity (which does not specify at what power output or voltage) is misleading. Their specification is very vague... Around 0.8W is roughly what I am seeing in terms of listening volume using the Aegis too.

So, despite the fact that the output voltage of your amp is higher for the Susvara, the current being drawn by them relative to the 805 is around 33%.

When I am talking about "drive" and the demands placed on an amplifier from a design perspective, I am talking about the current draw, and the 805 is clearly a more difficult load for an amplifier output stage. The amplifier itself will be generating less distortion and have better bandwidth when powering the Susvara.

Now don't get me wrong, that is still extremely inefficient for a headphone, but my point is it is still easier to design around, you just need an amplifier with voltage swinging headroom.

This type of reply is why I love these threads so much. That’s just awesome! (I wish there was a “save” or bookmark button).

Now a follow up. And I don’t know this can be answered (as I don’t have this info for Susvara):

Would the impedance curve of these two very different creatures likely be very different and would you guess that would make the comparison above more or less severe as a result ?

 

[L0rdGwyn]

This type of reply is why I love these threads so much. That’s just awesome! (I wish there was a “save” or bookmark button).

Now a follow up. And I don’t know this can be answered (as I don’t have this info for Susvara):

Would the impedance curve of these two very different creatures likely be very different and would you guess that would make the comparison above more or less severe as a result ?

Oh yes definitely, the impedance curve of a planar magnetic headphone is relatively flat, meaning the current draw is theoretically similar across the frequency spectrum. Dynamic transducers have nonlinear impedance curves, the amplifier has to have a low enough output impedance (and by definition, enough current on tap) to ensure these dips in frequency response are not heard by the listener. The demands on the amplifier will be higher at these frequencies to generate a similar volume.

Here is the 805 D3 impedance curve from Stereophile, it is 8ohms nominal but dips to around 4.6ohms at around 180Hz.

 

[JazzVinyl]

This type of reply is why I love these threads so much. That’s just awesome! (I wish there was a “save” or bookmark button).

Load this page and bookmark it via your browser:

https://www.head-fi.org/threads/l0rdgwyns-diy-audio.921105/page-391#post-17389821

On each message there is a Bar at the start of the message - on the right hand side is a link for just that one message, as above.

 

[goldwerger]

Oh yes definitely, the impedance curve of a planar magnetic headphone is relatively flat, meaning the current draw is theoretically similar across the frequency spectrum. Dynamic transducers have nonlinear impedance curves, the amplifier has to have a low enough output impedance (and by definition, enough current on tap) to ensure these dips in frequency response are not heard by the listener. The demands on the amplifier will be higher at these frequencies to generate a similar volume.

Here is the 805 D3 impedance curve from Stereophile, it is 8ohms nominal but dips to around 4.6ohms at around 180Hz.

Makes sense!

This would the actual math even more extreme than in actual comparison than to the average nominal

Thanks for taking the time to work through the math!

Load this page and bookmark it via your browser:

https://www.head-fi.org/threads/l0rdgwyns-diy-audio.921105/page-391#post-17389821

On each message there is a Bar at the start of the message - on the right hand side is a link for just that one message, as above.

Thanks!

 

[L0rdGwyn]

PCM63 DAC is coming together, have almost everything mounted, then lots of wiring to be done. We'll see if I have the energy to finish it today, might break it up into two days of work.

 

[L0rdGwyn]

PCM63 DAC is done, still have some tinkering to do but it is up and running! Whatcha think?







It's a little tricky with the transformers mounted on top, have to keep the leads pretty long to be able to remove the top plate.







It can use 5687, EC182CC, and 7044 tubes.



Front LED is amber.



Next thing to do is experiment with secondary loading of the output transformers, but first some lunch and a break.

 

[ColSaulTigh]

How does it sound?

 

[raindownthunda]

PCM63 DAC is done, still have some tinkering to do but it is up and running! Whatcha think?







It's a little tricky with the transformers mounted on top, have to keep the leads pretty long to be able to remove the top plate.







It can use 5687, EC182CC, and 7044 tubes.



Front LED is amber.



Next thing to do is experiment with secondary loading of the output transformers, but first some lunch and a break.

Impressively clean build! The internal layout looks super organized and like the sleek black look. Looks great Keenan. After seeing your 6336 OTL and other builds, those transformers look almost comically small. Is this a DAC for ants??!

 

[L0rdGwyn]

How does it sound?

Super holographic and clear. Just getting started though, I've only listened to it for an hour! And as I said some more tinkering to be done, I'll have a better idea tonight when I compared it to my other two DACS, DIY PCM56 and Audial S5, both NOS designs.

I will say this one isn't going to win any measurement awards, using the tube output stage I did and old school tube preamp transformers, it is going to have a 2nd harmonic dominant THD pattern. But what you get in return is a super clean signal path, all transformer coupled, fixed bias, no caps, something that doesn't show up in the measurements IME but makes a night and day difference.

Impressively clean build! The internal layout looks super organized and like the sleek black look. Looks great Keenan. After seeing your 6336 OTL and other builds, those transformers look almost comically small. Is this a DAC for ants??!

LOL good one, yeah they are small! Little preamp transformers, only need to be gapped for 30mA DC and low power obviously.

 

[L0rdGwyn]

I've been A-Bing the new DAC against the Audial S5, a boutique NOS DAC that uses the Philips TDA1541A.

I really like the Audial S5, it's been in my headphone system for over a year, but it's not even close, the PCM63 DAC has a larger soundstage, more body and girth to the music, more tactile and impactful bass, and better imaging and clarity. It's really better than it in just about every aspect, at least that's my first impression. Oh but the PCM63 DAC most definitely measures worse lol just goes to show ya, measurements aren't everything.

Going to be interesting to compare to my PCM56 DAC later, but I can already tell you this is the best DAC I have ever owned.

 

[triod750]

PCM63 DAC is done, still have some tinkering to do but it is up and running! Whatcha think?







It's a little tricky with the transformers mounted on top, have to keep the leads pretty long to be able to remove the top plate.







It can use 5687, EC182CC, and 7044 tubes.



Front LED is amber.



Next thing to do is experiment with secondary loading of the output transformers, but first some lunch and a break.

When do you receive the Clarity blue bells for the transformers?

 

[L0rdGwyn]

When do you receive the Clarity blue bells for the transformers?

No thanks

 

[L0rdGwyn]

I made my final edits. I ended up loading the output transformer secondary with a 510K resistor, making it effectively an open secondary. That increased the gain of the output stage and allowed me to reduce my I/V transformer secondary resistor to 1.5K, both improved the sound further.

So it is now done done, I closed up it up for good. Last pic, here with the RCA 7044 which sound really, really good.



I know I've said it before, but here is everything that went into this DAC. I bought the chips March of last year, it has been a lot of work to get this thing done over the past year.

- 1:1 isolation transformer with the secondary wired for balanced AC, as such the DAC circuit is running off balanced mains
- Dual mono boards, one DAC PCB per channel with two PCM63 run in parallel
- Separate cascaded series regulated power supplies for all four DAC power supplies (+analog, -analog, +digital, -digital)
- JLSounds I2SoverUSB digital front end, also powered by cascaded series regulated power supply
- Sowter 1465 I/V step up transformers for converting the current output of the parallel PCM63 into small signal voltage
- HEXFRED solid state rectified B+ supply with transformer snubbing circuit and current-limiting resistance
- Maida style regulated B+ supply with separate regulators for left and right channels
- E182CC / 7044 / 5687 output stage, battery biased and loaded with Sowter 9705 line output transformers
- Ultra low noise regulated DC heater supply for the E182CC / 7044 / 5687 output tubes

 

[triod750]

When you put a lot of effort into something and it comes together beautifully, well, isn't that a good way to get satisfaction? Should be.
Next challenge will be to get that feeling stay alive.
Good luck!!

 

[L0rdGwyn]

When you put a lot of effort into something and it comes together beautifully, well, isn't that a good way to get satisfaction? Should be.
Next challenge will be to get that feeling stay alive.
Good luck!!

Absolutely! It is very satisfying. To come up with an idea and see it realized is why I am addicted to DIY. That first listen to a new piece of gear that takes the listening experience up another level is intoxicating