• Disclaimer: This Vendor's Forum is a paid-for commercial area. Unlike the rest of diyAudio, the Vendor has complete control of what may or may not be posted in this forum. If you wish to discuss technical matters outside the bounds of what is permitted by the Vendor, please use the non-commercial areas of diyAudio to do so.

Buffalo II & transformers

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Who says you can't use a transformer in 'current mode' ?
T
What I said was:
Russ White said:
In my opinion it is simply is not possible to run directly (even with resistors) into *any* transformer and keep the DAC in "current mode".

The Directly part is important. :)

Sure if you add some feedback from an active or passive device you can do it, but then you might as well not. :cool:

It is my understanding that you would also most likely need amplification after the ZF transformer in any case. :)
 
Last edited:
No, there are no added 195R resistors, these represents the internal impedance of the ES9018 when used in Stereo mode.

I'm afraid that I'm quite a newbie on the topic of I/V stage or transformer stage design for ES9018 DAC chip. May I ask you some primitive or basic questions? I'd be very happy if any of you could kindly answer to my questions.

We always find +1.65V and 195R in Joe's schematic.
1. I understand that the value 1.65V is calculated as AVCC/2 (where AVCC=+3.3V). When we use Buffalo II with Russ' standard shunt type AVCC sub board, the actual voltage is a little bit higher. Is my understanding correct?
2. I understand that the internal output impedance value, "195R" is a result of summing up of four analog output sections in one side of one phase in the DAC chip.
- How was the "195R" originally derived?
- Is the value applicable even to "pseudo differential" mode output where only + and - outputs are involved? Is it also valid for "9 bit quantizer configuration"?
- Can we actually measure and confirm the "195R" by any means? Is the constant value "195R" valid for all the frequency range 10 - 200 kHz?
 
I'm afraid that I'm quite a newbie on the topic of I/V stage or transformer stage design for ES9018 DAC chip. May I ask you some primitive or basic questions? I'd be very happy if any of you could kindly answer to my questions.

We always find +1.65V and 195R in Joe's schematic.
1. I understand that the value 1.65V is calculated as AVCC/2 (where AVCC=+3.3V). When we use Buffalo II with Russ' standard shunt type AVCC sub board, the actual voltage is a little bit higher. Is my understanding correct?
2. I understand that the internal output impedance value, "195R" is a result of summing up of four analog output sections in one side of one phase in the DAC chip.
3. How was the "195R" originally derived?
4. Is the value applicable even to "pseudo differential" mode output where only + and - outputs are involved? Is it also valid for "9 bit quantizer configuration"?
5. Can we actually measure and confirm the "195R" by any means? Is the constant value "195R" valid for all the frequency range 10 - 200 kHz?

1. AVCC/2 is the more correct way to state this.
2. Ok
3. This is covered in numerous places even the datasheet.
4. Yes. Pseudo differential does not have any effect on the analog characteristic output impedance of the device. It is the same for any quantizer mode and pseudo or true diff as long as you parallel all four output channels. This is very easy to demonstrate on the Buffalo III with IVY-III or Legato. Just play something and switch between modes. You will notice the output volume stays the same.
5. Absolutely, and quite easily. Just play a signal of known amplitude across a fixed resistance and observe the output amplitude. The impedance of the DAC is the same regardless of frequency, but keep in mind that the DAC has a filter that could come into play but that is completely orthogonal to the output impedance. :)
 
Last edited:
Dear Russ,

I greatly appreciated and I am satisfied with your clear-cut answers!

Bunpei

No problem. Very glad to help.

One thing to remember is that while the output impedance is always exactly the same, the output signal amplitude at the + and - outputs (referenced to AVCC/2) will be different between modes. :) The thing to always remember is that when the two are properly summed (by the output stage) the differential result is the same amplitude in any case. :cool:

So the key point here is the practical difference between amplitude at any single output and output impedance. Output impedance never varies, yet the individual components of the output signal can, but when summed it all works out the same in the end. :)

Cheers!
Russ
 
Last edited:
Hi Terry

Shows you can do a lot with feedback. But as a wise man once said "all things may be possible, even lawful, but not all things are advantageous." Seems the guy's name was Saul or Paul or something. :)

Hi Joe,

It depends. In the ZFT's case, you don't necessarily have to use an opamp
with FB to get very low impedance at the DAC OP. You just need the ZFT to
feed a low impedance and the ZFT itself to have a very low winding
resistance. Can be sub 1 ohm if that is required.

They can be quite useful for various apps, Studer used to use them for
summing bus applications in consoles.

I thought about using a ZFT for Sabre but came up with better circuit
which uses finite resistance of DAC to cancel distortions in the I-V.

Come around and listen to the Oppo 95 here, puts a smile on my face. Need two clocks though, makes a real difference.

Cheers, Joe R.

Thanks, am very busy ATM. Maybe we should blow froth off a few for xmas :)
WRT 2 clocks are you referring to 27MHz and audio div freq for the second?
Does Oppo 95 use 9018?

cheers
 
Hi Joe,

I thought about using a ZFT for Sabre but came up with better circuit
which uses finite resistance of DAC to cancel distortions in the I-V.

Ahhh, sounds interesting, care to elucidate? :D - Suppose not.

Maybe we should blow froth off a few for xmas :)
WRT 2 clocks are you referring to 27MHz and audio div freq for the second?
Does Oppo 95 use 9018?

Got some nice German "bier" in the fridge right now. :)

Oddly enough, they have gone for 25MHz on the digital motherboard into the Meditek chip. There is also a buffer there from USB/eSATA ports (I have been able to 'trigger' it when no activity from the port as the HDD has gone into sleep mode) that is clocked (re-clocked really) from the 'master' clock. Our VSE guy in Munich, Ned says it almost certainly does the same off the transport.

Yes, uses ES9018 and have 80MHz on it.

Then transformer coupled 1:1 to XLRs, but has two 330R to ground from both phases of the DAC (stereo mode - 4 parrallel) and get a decent 1.3V RMS. Without the resistors in simple voltage mode, it sounds light in balance, but then firms up nicely.

Got 600+ albums, some 24 bit stuff as well, on a 2TB hard drive and yet only filled 1/6th of its capacity.

But watching movies on the 95 is a real hoot. Mate, the realism and IMPACT it has is HUGELY enjoyable. I am worried, is this thing also turning me into a videophile? Shock, horror! :D:D:D

Cheers, Joe
 
Last edited:
What I said was:

"In my opinion it is simply is not possible to run directly (even with resistors) into *any* transformer and keep the DAC in "current mode"."

The Directly part is important. :)

You said "directly into *any* transformer".

To me that reads a transformer connected 'directly' to the DAC OP with
nothing in between.

I don't see any mention whether any additional circuitry should / not be used
after the transformer.

Sure if you add some feedback from an active or passive device you can do it, but then you might as well not. :cool:

a) The zero field transformer can reflect a sub 1 ohm impedance to the DAC
without the use of mixed (as they call it) feedback.

b) How can you add some feedback from a passive device?

c) Your "might as well not" suggestion is counter to the DIY spirit in this
forum.

My suggestion was to make people aware that the transformer itself can be
used in 'current mode' and hence reflect a very low impedance. I've done
it and it works.

T
 
Check this out:

http://www.lundahl.se/pdfs/datash/6404.pdf

Note, it uses positive feedback and "C" needs to a rather high value, hence electro?

Cheers, Joe R.

Yes I am well aware of that schematic. It is a ZF trafo followed by an I/V stage with some positive feedback. Wow isn't that novel. :)

I think Brian was asking for the "purely passive" version with nothing active after it. I am not saying its not possible. Maybe it is.

I personally can't yet see any good way to use it without an active stage after the transformer. Basically nullifying the need for the transformer. Might as well go directly to a low impedance I/V stage IMO.
 
Last edited:
Could you share a schematic or other details?

I used the ZFT without + feedback to compensate for winding resistance, (as
Joe has shown in Lundahl link).

So the ZFT is merely inserted between the DAC and the virtual gnd of the
following stage. I recommend using a 1:1, low resistance transformer.

The tricky part is optimising transient response, as with all transformers.

The best way to do this, in this case is use a 10kHz square wave gen -> fast
opamp balanced buffer -> 195R / phase buildout R's -> ZFT -> post
transformer stage.

A snubber will have to be fashioned on either prim or secondary of TX for best
square wave response at OP with minimal overshoot. But this should be done
for any transformer application.

T
 
I used the ZFT without + feedback to compensate for winding resistance, (as
Joe has shown in Lundahl link).

So the ZFT is merely inserted between the DAC and the virtual gnd of the
following stage. I recommend using a 1:1, low resistance transformer.

The tricky part is optimising transient response, as with all transformers.

The best way to do this, in this case is use a 10kHz square wave gen -> fast
opamp balanced buffer -> 195R / phase buildout R's -> ZFT -> post
transformer stage.

A snubber will have to be fashioned on either prim or secondary of TX for best
square wave response at OP with minimal overshoot. But this should be done
for any transformer application.

T
Meh, you might as well omit the trafo in this case(because you would get closer to 0R without positive feedback). But I am listening....
 
Last edited:
You said "directly into *any* transformer".

To me that reads a transformer connected 'directly' to the DAC OP with
nothing in between.

I don't see any mention whether any additional circuitry should / not be used
after the transformer.
T

Feedback into a transformer is a form of indirection.

Beside that *if* you put -something- (especially an active something) after the transformer the transformer is no longer the output stage but merely something between the DAC and the output stage. :)

Not saying that is good or bad, just saying that this is no longer a "direct" use of a transformer. :) The transformer is completely dependent on some later stage. We can't omit talking about that later stage out of convenience. It is important. Actually in many ways (as Terry's stipulations prove) more important than the transformer. :)
 
Last edited:
Yes I am well aware of that schematic. It is a ZF trafo followed by an I/V stage with some positive feedback. Wow isn't that novel. :)

I think Brian was asking for the "purely passive" version with nothing active after it. I am not saying its not possible. Maybe it is.

I personally can't yet see any good way to use it without an active stage after the transformer.

It is clearly stated that the ZFT needs to feed a further stage to enable
voltage swing at the OP. This has been stated from the outset.

Basically nullifying the need for the transformer. Might as well go directly to a low impedance I/V stage IMO.

The ZFT does a few things that you may / may not find useful:

a) It sums the 2 DAC phases - a prerequisite for all optimum I-V's used on
Sabre - you know this.

b) If you are using an opamp OP then because of a) you can use a single
opamp be it monolythic or discrete. Many designers more successful than
both of us prefer discrete, for whatever reasons, these can be complex so
1 versus 3 is an advantage.

c) The ZFT, or any transformer, offers galvanic isolation and with it the
ability to DC shift the transformers secondary. This opens up more
circuit possibilities for I-V using direct coupling for the more creative
designers.

In the end it's just another tool in the toolbox - and sometimes those odd
tools are the ones that work, ie; sound good. It's not just the usual
predictable plain vanilla designs that give the best results.

In the past they were used with success in mix console summing bus
applications, I believe mainly for galv isolation of grounds.
 
In the end it's just another tool in the toolbox - and sometimes those odd
tools are the ones that work, ie; sound good. It's not just the usual
predictable plain vanilla designs that give the best results.

In the past they were used with success in mix console summing bus
applications, I believe mainly for galv isolation of grounds.

I agree with you in spirit.

I was just making the point that this is in every respect an active output stage. Just with a transformer in front of it. :)

I heartily welcome this sort of discussion as it is nothing but positive. At the same time I reserve the right to form my own opinion about what sounds "better or best" as will everyone else. Results are so subjective in this realm.
 
Last edited:
a) It sums the 2 DAC phases - a prerequisite for all optimum I-V's used on
Sabre - you know this.

I agree the phases need to be summed. My personal opinion is the very best place to sum these signals (at least for hi-fi) is at the loudspeaker. My ideal is to keep things fully symmetrical until that point. This of course does not hold true in some cases where fidelity is not the first concern because in many cases (such as live performance) raw power and reliability are far more important.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.