Transformer I/V stage, pros&cons.

[Steve Eddy]

Re: Re: Re: Re: Re: Re: Re: Re: Re: Transformer I/V stage, pros&cons.

The traditional current to voltage conversion when applied to
a DAC referres to the fact that the DAC feeds an ideal virtual
ground and there is NO voltage swing at the DAC OP.

Yes, for DACs whose outputs are by way of current sources and when you use an opamp whose feedback changes its output voltage in order to keep the voltage across its inputs zero.

This was how traditional R2R ladder DACs were designed and attained best linearity, with NO voltage swing. In it's ideal form, the I-V converts ONLY current with NO voltage swing (at the - opa IP) to a voltage at it's OP.

Huh? An R2R is a resistive divider network that uses a voltage reference and outputs a voltage. Current output DACs use binary-weighted current sources and outputs current.

With a transformer I-V, the conversion to voltage from the DAC's OP current HAS ALREADY HAPPENED before, with or without the transformer. The transformer IS NOT responsible for the conversion to a voltage, the low value R at DAC OP is.

I wasn't looking at it in terms of a low value resistor at the DAC output. As you say, if you're using a resistor on the output, you've already converted the current to a voltage.

I was thinking in terms of driving the transformer with the current source. Though I realize that ultimately won't work. Since an ideal current source has an infinite output impedance, the secondary would also have to be driving an infinite impedace, but this infinite impedance would be reflected back to the primary in which case no current could flow in the primary.

So a transformer is V-V with an intermediate I, ultimately making it a V-I/I-V converter. And given what I said above, I now don't see it as being capable of I-I.

The transformer does NOT reflect a load back to the DAC OP which is solely responsible for conversion to voltage, it merely amplifies that existing voltage that is already at the IP.
The only load reflected in a so called "transformer I-V" will be
that of a possible snubber network and the following stage.
These are relatively light loads compared to the I-V resistor at
the DAC OP.

Yes, but I wasn't looking at it in terms of simply using a step-up transformer to kick up the voltage across a small resistor across the the DAC's output but rather driving the transformer straight from the DAC's current source output.

I hope this clarifies things.

Not entirely. But I find it more nourishing than "Go read a book."

Thanks.

se

 

[Terry Demol]

Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Transformer I/V stage, pros&cons.

Huh? An R2R is a resistive divider network that uses a voltage reference and outputs a voltage. Current output DACs use binary-weighted current sources and outputs current.

Early (traditional) current OP DACs such as PCM63, PCM1702, PCM1704, AD1862 etc had R2R architecture. You can download
PCM1704 data sheet for a good explanation here:
http://focus.ti.com/docs/prod/folders/print/pcm1704.html

DAC manufacturars have since moved away from this
architecture to get better THD/DR and lower manufacturing costs
as the R2R requires very accurate laser trimming of R's,
although some DACs still have unity weigted R based OP
topologies (NPC SM5865/66). Most other newer DACs are open
drain type.

I wasn't looking at it in terms of a low value resistor at the DAC output. As you say, if you're using a resistor on the output, you've already converted the current to a voltage.

I was thinking in terms of driving the transformer with the current source. Though I realize that ultimately won't work. Since an ideal current source has an infinite output impedance, the secondary would also have to be driving an infinite impedace, but this infinite impedance would be reflected back to the primary in which case no current could flow in the primary.

Actually, this WOULD work with a finite secondary load reflected
back to primary. However it wouldn't work so well, as all high
performance trannies (like Jensen etc) are optimised to work
from a certain source Z. Exceptions are very high Z types
(10k : 10K) but these would be totally unsuitable for lo Z DAC
applications.

There is 1 transformer type which could present a true
virtual gnd to DAC, wich is called a zero field transformer.
However, a zero field tranny actually works into a virtual
gnd (short) and so needs an I-V following it. These transformers
have the lowest distortion of any audio transformer as there
is theoretically no voltage swing. As a matter of ineterest I have
tried these for DAC applicatins.
See Lundahls website for more details.

So a transformer is V-V with an intermediate I, ultimately making it a V-I/I-V converter. And given what I said above, I now don't see it as being capable of I-I.

For interests sake, I recommend reading the literature on
Lundahls website WRT zero field trannies. They have an OPA
based circuit which generates negative resistance equal to
transformers winding R and so creates a true I-I scenario.
These transformers virtually eliminate core related distortions
completely, because there is virtually no magnetisation of the
core (almost no voltage).
I believe Rupert Neve designed some products for Amek with
ZFT IP stages, but I am not aware of any other usage in audio.

Cheers,

Terry

 

[Jocko Homo]

Say what, Steve????

Huh? An R2R is a resistive divider network that uses a voltage reference and outputs a voltage. Current output DACs use binary-weighted current sources and outputs current.

Terry beat me to it. You need to know what you are talking about before shooting form the hip. There are countless examples of DACs, and not just ones for audio, that use this architecture. AD use to have a good explanation of how this worked.....somewhere......in one of their data books. Don't remember which one though.

Jocko

 

[jan.didden]

Re: Re: Re: Re: Re: Re: Re: Re: Re: Transformer I/V stage, pros&cons.

Because it is a current to current AND voltage to voltage
converter.

The traditional current to voltage conversion when applied to
a DAC referres to the fact that the DAC feeds an ideal virtual
ground and there is NO voltage swing at the DAC OP. This was
how traditional R2R ladder DACs were designed and attained
best linearity, with NO voltage swing. In it's ideal form, the I-V
converts ONLY current with NO voltage swing (at the - opa IP) to
a voltage at it's OP.

With a transformer I-V, the conversion to voltage from the DAC's
OP current HAS ALREADY HAPPENED before, with or without the
transformer. The transformer IS NOT responsible for the
conversion to a voltage, the low value R at DAC OP is.

The transformer does NOT reflect a load back to the DAC OP
which is solely responsible for conversion to voltage, it merely
amplifies that existing voltage that is already at the IP.
The only load reflected in a so called "transformer I-V" will be
that of a possible snubber network and the following stage.
These are relatively light loads compared to the I-V resistor at
the DAC OP.

I hope this clarifies things.

Cheers,

Terry

As much as I hate to admit (just joking!), I'm with Steve Eddy here. The formula for induction is B*i*L. That's what causes the transformer to work. Unless I missed something big time, "i" in my book still stands for current.
The fact that ultimately the current may be *caused* by an impressed voltage *in this case* doesn't make a difference.

Jan Didden

/There's nothing so practical as a good bit of theory.

 

[Fred Dieckmann]

Deja Vu all over again......

"I was thinking in terms of driving the transformer with the current source. Though I realize that ultimately won't work. Since an ideal current source has an infinite output impedance, the secondary would also have to be driving an infinite impedace, but this infinite impedance would be reflected back to the primary in which case no current could flow in the primary.

So a transformer is V-V with an intermediate I, ultimately making it a V-I/I-V converter. And given what I said above, I now don't see it as being capable of I-I."



The impedance seen at the primary is the impedance at the secondary divided by the square of the turns ratio. When you step up the voltage by a factor of 2 the impedance seen at the primary will be 1/4 of that loading the secondary.

The intermediate coupling between the primary and secondary is the magnetic flux in the transformer. Transformers do current, voltage, and impedance transformation.

Yes there are current transformers.

http://www.crmagnetics.com/newprod/ctransformersg.asp

 

[Fred Dieckmann]

There's nothing so practical as a good bit of theory.... yes...

"Unless I missed something big time"

I am afraid you did.

 

[jan.didden]

OK Fred, I know, I asked for it.

Thought experiment # 1: wind a transformer with a primary with infinitely thin wire, giving infinitely high resistance. Connect it to a voltage source. Does it work? No, there is no current to exite a flux, although there is voltage.

Thought experiment # 2: wind a transformer with an infinitely large wire cross-section, giving zero resistance. Connect to a current source. Will it work? Yes, there is current to exite a flux, although there is no voltage.

What's wrong here?

Jan Didden

 

[jewilson]

Point what the point

I believe the point Fred was making is we don't get something for nothing. If we have a pluse, power transformer kind of step-up devices say 1:4 we cannot expect to have the more power at the secondary than at the primaries. If I am wrong, there would be supplies with greater than 100% efficiency. Now that would be cool.

 

[Pedja]

The point

We got nice discussion on “how do the transformers work” topic.

The point here is can you put certain current into transformers primary and get certain voltage across its secondary? My guess is you can’t…

Pedja

 

[Fred Dieckmann]

We got nice discussion on “how do the transformers work” topic.

No what we have here is a train wreck of misinformation and confusion.

I would recommend a good introductory electronics text for an introduction to what the relationship between the turns ratio and the current, voltage, and impedance transformation between the primary and secondary as a function of the turns ratio.

For impedance:
http://www.edcorusa.com/gadgeteer/tech_notes/tn12.htm

"The point here is can you put certain current into transformers primary and get certain voltage across its secondary? My guess is you can’t…"

Then don't guess, learn the basic relationships. Of course you can figure out the current by knowing the turns ratio and the load impedance on the secondary. It is very simple algebra.

"OK Fred, I know, I asked for it.

Thought experiment # 1: wind a transformer with a primary with infinitely thin wire, giving infinitely high resistance. Connect it toa voltage source. Does it work? No, there is no current to exite a flux, although there is voltage.

Thought experiment # 2: wind a transformer with an infinitely large wire cross-section, giving zero resistance. Connect to a current source. Will it work? Yes, there is current to exite a flux, although there is no voltage.

What's wrong here?"

Number one, these describe the loss terms and have nothing to do with the impedance transformation for an ideal transformer.
You have to understand that before moving on to study the losses. Do thought experiments with real numbers. An infinite series resistance is neither an ideal or real example and will confuse you. Pick a turns ratio, a secondary load impedance, and a primary source impedance. The primary side is where the energy is applied.

It is unfortunate when people start a discussion with no reference to basic equations that describe the circuit, expound their
theories as facts with little to back it up, and make generalizations about things like DACs without looking at a few data sheets to see how the DACs being used for audio today work. It seems to be the same few people over and over again who must thrive on humiliation. I would think most of the regular members would know who they are and stop being sucked into this totally confusing discussions that end up ****ing everyone off including the originator of the thread.

I don't want to hear the personalities card played. I have gone out of my way not to mention names and there are several. The personalities subterfuge comes along after you disagree with there proposed idea and try to present reasons based on technical data they didn't grasp. One has two choices sit back and ignore stuff that is obviously wrong while it confuse many, or try to explain why it is wrong and get called names for it. I would usually prefer to take a few lumps in the interest of education of people who really want to know about a subject.

 

[Steve Eddy]

Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Re: Transformer I/V stage, pros&cons.

Early (traditional) current OP DACs such as PCM63, PCM1702, PCM1704, AD1862 etc had R2R architecture. You can download
PCM1704 data sheet for a good explanation here:
http://focus.ti.com/docs/prod/folders/print/pcm1704.html

Yes, but as far as I can tell, their output isn't from the R2R ladder as it would be in a straight R2R ladder type DAC, rather the R2R ladder seems to be used for controlling the current sources, rather than the output being from the R2R ladder which I took your previous post to mean.

Actually, this WOULD work with a finite secondary load reflected back to primary. However it wouldn't work so well, as all high performance trannies (like Jensen etc) are optimised to work
from a certain source Z. Exceptions are very high Z types (10k : 10K) but these would be totally unsuitable for lo Z DAC applications.

Not in any practical sense. And certrainly not in the ideal sense.

Let's say you have an ideal 1:1 transformer with a 100 ohm load across its secondary. Sure, that 100 ohm load will be reflected back to the primary and the current source would see a 100 ohm load. But the output impedance of an ideal current source is infinite, and that infinite output impedance would be reflected to the secondary. Now, if the output impedance of the secondary is infinite, then there could be no current flowing through that 100 ohm resistor in which case the voltage across that 100 ohm resistor will be 0.

Even if you look at practical current sources with finite output impedances, they're still high enough that you wouldn't get much voltage to speak of across that 100 ohm resistor.

There is 1 transformer type which could present a true
virtual gnd to DAC, wich is called a zero field transformer.
However, a zero field tranny actually works into a virtual
gnd (short) and so needs an I-V following it. These transformers
have the lowest distortion of any audio transformer as there
is theoretically no voltage swing. As a matter of ineterest I have
tried these for DAC applicatins.
See Lundahls website for more details.

Sure. There are all sorts of tricks you can do with active circuits and negative feedback. But I'm just focusing on the basic behavior of the transformer itself.

For interests sake, I recommend reading the literature on Lundahls website WRT zero field trannies. They have an OPA
based circuit which generates negative resistance equal to
transformers winding R and so creates a true I-I scenario.
These transformers virtually eliminate core related distortions
completely, because there is virtually no magnetisation of the
core (almost no voltage).
I believe Rupert Neve designed some products for Amek with
ZFT IP stages, but I am not aware of any other usage in audio.

That technique is also used by Audio Precision. You might find Bruce Hofer's patent, 4,614,914 interesting.

se

 

[Steve Eddy]

Re: Say what, Steve????

Terry beat me to it. You need to know what you are talking about before shooting form the hip. There are countless examples of DACs, and not just ones for audio, that use this architecture. AD use to have a good explanation of how this worked.....somewhere......in one of their data books. Don't remember which one though.

But the outputs of the DACs terry mentioned apparently aren't R2R ladders, rather the R2R ladders are used to control the current sources. I took Terry's post to mean that the R2R ladder was the output.

se

 

[SY]

Now, if the output impedance of the secondary is infinite, then there could be no current flowing through that 100 ohm resistor in which case the voltage across that 100 ohm resistor will be 0.

You're going to have to explain to me how the voltage won't be the sourced current times 100 ohms.

 

[Fred Dieckmann]

I will try again.........

"Let's say you have an ideal 1:1 transformer with a 100 ohm load across its secondary. Sure, that 100 ohm load will be reflected back to the primary and the current source would see a 100 ohm load. But the output impedance of an ideal current source is infinite, and that infinite output impedance would be reflected to the secondary. Now, if the output impedance of the secondary is infinite, then there could be no current flowing through that 100 ohm resistor in which case the voltage across that 100 ohm resistor will be 0.

Even if you look at practical current sources with finite output impedances, they're still high enough that you wouldn't get much voltage to speak of across that 100 ohm resistor."


Man I really don't get why this is so confusing.

Yes, for a 1:1 transformer the load impedance of the secondary is reflected to the primary. The DAC develops a voltage across the reflected 100 ohms.
The energy is being developed by the DAC. Its output current times the reflected impedance from the secondary will determine the voltage across the primary and, by the relation of the turns ratio, the voltage across the secondary For a 1:1 transformer this will be the same voltage as on the primary. Yes, looking into the secondary you will see the very high output impedance reflected to the secondary. The impedance looking across the secondary will be this reflected high impedance in parallel with the 100 resistor, The current into the primary will be the same as that from the secondary, transformers convert current as well as voltage. The circuit looking into the transformer secondary will be 100 ohms connected across the secondary. The impedance seen by the DAC will be 100 ohms. Use the Math Luke, use the Math.

 

[Steve Eddy]

Re: Deja Vu all over again......

The impedance seen at the primary is the impedance at the secondary divided by the square of the turns ratio. When you step up the voltage by a factor of 2 the impedance seen at the primary will be 1/4 of that loading the secondary.

So? You also step up the source impedance by a factor of 4. What's your point?

The intermediate coupling between the primary and secondary is the magnetic flux in the transformer.

Yes, that magnetic flux ultimately being produced by current flowing in the primary, which, by Faraday's Law, induces a voltage in the secondary. Again, what's your point?

Yes there are current transformers.

Isn't a current transformer really just a secondary winding, the primary winding effectively being the conductor which you want to monitor the current through?

se

 

[Fred Dieckmann]

Re: Re: Deja Vu all over again......

So? You also step up the source impedance by a factor of 4. What's your point?



Yes, that magnetic flux ultimately being produced by current flowing in the primary, which, by Faraday's Law, induces a voltage in the secondary. Again, what's your point?



Isn't a current transformer really just a secondary winding, the primary winding effectively being the conductor which you want to monitor the current through?

se

My point is that you understand the concept of impedance transformation you will continue to be confused.

My second point is that a transformer is not a voltage to current to current to voltage converter. Magnetic flux is not the same thing as current.

Current transformers are still a transformer. There is no secondary winding without a primary. For a clamp on current transformer the current carrying wire that you are trying to measure is the primary winding.

Learn the basic laws for transformers first and stop tossing Faraday around. We all all impressed that you know who he is, but would be more impressed if you knew how transformers work first.

You want to disscuss the origion of language before learning your ABCs.

Do you want to learn or argue. You sure are not teaching at any rate.

 

[Pedja]

Are you talking to me?

Then don't guess, learn the basic relationships. Of course you can figure out the current by knowing the turns ratio and the load impedance on the secondary. It is very simple algebra.

It is indeed a part of the basic relationship and that relationship is even logical by nature. It is not any algebra, be it simple or complicated, that tells us the load (impedance) is not a part of the transformer. When you define the load, you can start to talk about how I put into the primary by the current source relates to the V across the secondary (and thus also across the primary). But only if you have the load. Not without it. Not by the transformer itself.

So, my point was you can’t convert I to V by the transformer itself. And hello, should I say I’ll fight for that to get that seriously? It is my guess! The end!

Peace on Earth.

Pedja

 

[tbla]

 

[Bricolo]

Re: Re: Say what, Steve????

But the outputs of the DACs terry mentioned apparently aren't R2R ladders, rather the R2R ladders are used to control the current sources. I took Terry's post to mean that the R2R ladder was the output.

se

Steve,


What DACs are true R2R ones?

 

[fdegrove]

Hi,

Somehow I think someone's on the phone with Jensen calling them names for giving him the wrong brainwash....

Transformers aren't all that difficult once you understand the basics....once you understand the basics...once you understand the basics....once you understand the basics...

Errrr, this is the Duracell add, isn't it...isn't it....isn't it...isn't it ?

Seriously, when it comes to xformers and DACS you have a number of ways to deal with problems...theorising about it without a solid basic knowledge is not going to help anyone.

So, wouldn'it be wise do leave xformers to do what they do best?

Have the I/V conversion at the primary side and be done with it?

Cheers,