I/V stage question (explanation needed)

Status
Not open for further replies.
Hi,

I have a rather newbish question for you guys and I hope I can get some info on I/V stages.

I have seen the reference to an I/V stage at DAC outputs and in pre-amps (after a volume control). What does that stage do exactly?

I got something like, it converts current to voltage, but I am not quite clear on how or why and how does it function.

Can anyone explain a bit on the I/V stage and some design principles and where it is applied? I have seen implementations with transistors and op-amps.

Maybe a good schematic pic with an explanation of what the signal does would be really helpful.

Thank you!
 
Many DACs have current output - see, for example, the datasheet for PCM1794A. These DACs require a very low (ideally zero) impedance load, and will supply current on the order of a few mA peak-to-peak into that load. The important part is that the voltage drop on that load must be small in order for the DAC to meet its specifications.

Your normal amplifier has voltage input with reasonably high impedance (say 10 to 100kOhm) and require reasonably high voltage (say 100mV to 2V) to operate.

An I/V converter provides low impedance load for the DAC and converts the DAC's output current into voltage for the following stages.

The simplest converter is a resistor, and there are designs that use just that. An example is ezDAC (search the thread for the schematic; the original developer's website doesn't seem to be working anymore). However, the DAC's output current flowing through a resistor produces an unwanted voltage drop.

A more popular converter is an inverting opamp, which has virtual ground at its inverting input. A DAC's output is connected directly to that virtual ground, the DAC's output current flows through the opamp's feedback resistor, and the opamp works to maintain the virtual ground by providing a voltage swing at the output. An example schematic can be found in the above referenced datasheet for PCM1794A.

There are other designs, such as Zen I/V.
 
Last edited:
Alex thanks a lot for the explanation! This really helps! The links are most helpful too!

I have been looking at the PCM1794 DAC for my future project. It looks like it is a really good DAC and it's used in some high end stuff.

I guess it is all in the datasheet once again hehe. But I do have one more question regarding the following form the PCM1794 page.


"
The current of the PCM1794A on each of the output pins (I​
OUTL+, IOUTL–, IOUTR+, IOUTR–) is 7.8 mA p-p at 0 dB
(full scale). The voltage output level of the I/V converter (Vi) is given by following equation:
Vi = 7.8 mA p–p
× Rf (Rf : feedback resistance of I/V converter)
An NE5534 operational amplifier is recommended for the I/V circuit to obtain the specified performance. Dynamic
performance such as the gain bandwidth, settling time, and slew rate of the operational amplifier affects the audio

dynamic performance of the I/V section."

How do I figure out what my voltage output level of the I/V converter should be? Any general rules for a DAC? 120mV @10mA?

Thanks,
Ivan
 
How do I figure out what my voltage output level of the I/V converter should be? Any general rules for a DAC? 120mV @10mA?
Well, the usual CD player output level is 2 Vrms (5.66 Vpp) for 0 dBFS, so the signal out of your I/V should be no more than 10 to 20 dB away from that level. This gives about 220 ohms or so. It must be high enough for noise to be dominated by the DAC rather than opamp voltage noise.

Normally DAC manufacturers also show an evaluation circuit that should work well (look out for the docs of an eval board if necessary). In this case the datasheet doesn't go into much detail, but there's an eval board named DEM-PCM1794 for which a detailed schematic is provided.

On the eval board they're even using a less-than-unity gain amp after the I/V to get 2 Vrms.
 
Well, the usual CD player output level is 2 Vrms (5.66 Vpp) for 0 dBFS, so the signal out of your I/V should be no more than 10 to 20 dB away from that level. This gives about 220 ohms or so. It must be high enough for noise to be dominated by the DAC rather than opamp voltage noise.

Normally DAC manufacturers also show an evaluation circuit that should work well (look out for the docs of an eval board if necessary). In this case the datasheet doesn't go into much detail, but there's an eval board named DEM-PCM1794 for which a detailed schematic is provided.

On the eval board they're even using a less-than-unity gain amp after the I/V to get 2 Vrms.

Thanks for the info! Good tip on having the DAC dominate the noise levels. Need to look more into what your saying. Btw you mean after the I/V stage the amplification op-amp should give me 10-20db gain to get to that 2Vrms output? Therefore the I/V stage has a very small gain? (I probably need to do the 1794 reference circuit op-amp gain calculation to see it)

On the PCM1794 datasheet, see Figure 24 for TI's recommended output stage. The LT1028 at the output could probably be whatever opamp you liked.

Yes, I see it, I am starting to understand it more. I dont get why they use an fc=162kHz and fc = 217 kHz? Seemsa bit high? Can we lower it down to say 80k or 60kHz or will it affect the sound? I mean this is the analog part already, pretty much up to 20kHz.

Here's a cheap tweak. Try using the circuit in the app notes but also load
the current outputs of the DAC with resistance to ground. Pick a value
that cuts into the output signal slightly.

😎

Ok, that would be easy to do I guess, maybe something in the kOhms (i guess without trying) to bring the Vout to say 1.97Vrms....... buttt why would I want to do that? What's the benefit of the tweak?

tnx guys
 
Here's a cheap tweak. Try using the circuit in the app notes but also load
the current outputs of the DAC with resistance to ground. Pick a value
that cuts into the output signal slightly.

😎

about
Digi Scoop 1
you will find the appropriate instructions concerning the resistor value

BTW - by post #38 about
http://www.diyaudio.com/forums/digital-source/205464-tda-1541a-s1-vs-modern-dac-4.html
I've compiled various threads from this forum concerning IU resp. IV converter topologies.
 
Last edited:
about
Digi Scoop 1
you will find the appropriate instructions concerning the resistor value

BTW - by post #38 about
http://www.diyaudio.com/forums/digital-source/205464-tda-1541a-s1-vs-modern-dac-4.html
I've compiled various threads from this forum concerning IU resp. IV converter topologies.

Thanks a lot! I will need to dig trough those when I have a little more time, some really good info.

Idk why but it seems to me that some people really don't like the op-amp I/V stage... power supply noise? Hard to find a good opamp?
 
In Zen I/V article referenced above, Nelson Pass briefly discusses why opamps may be out of their element when it comes to I/V conversion. A quote:

Without wanting to get into an argument about the pros and cons of op amps in audio circuits, I must point out that this sort of usage of an op amp places it at the greatest possible disadvantage – operation at unity gain with a mix of audio and high frequency noise.

However, I cannot remember one current output DAC datasheet which would recommend anything else but an opamp I/V to obtain the specified performance.
 
Last edited:
In Zen I/V article referenced above, Nelson Pass briefly discusses why opamps may be out of their element when it comes to I/V conversion. A quote:
Without wanting to get into an argument about the pros and cons of op amps in audio circuits, I must point out that this sort of usage of an op amp places it at the greatest possible disadvantage – operation at unity gain with a mix of audio and high frequency noise.
However, I cannot remember one current output DAC datasheet which would recommend anything else but an opamp I/V to obtain the specified performance.


This would be very interesting to do: a dual implementation of the Zen I/V and an op-amp topology and run a test.

Since the schematics are already there for the Zen that implementation would not be too demanding to make, it will be a straight copy for the most part.

Now the op-amp implementation might require some reasearch to replace the NE5534 from the PCM1794.

I have already done some looking into and it looks like the op-amp BANDWIDTH, Slew Rate, are important factors here. Also unity gain stability is a big one as well.... and many other parameters.

Here are some articles on Slew Rate and Bandwidth and how they affect the audio signals:
Slew Rate Article
Slew Rate Explanation Forum Topic and some opinions



The NE5534 has a slew of 13 V/us and 10MHz bandwidth. My theory is that we can look inot a higher slew rate to bandwidth ratio. We dont want an increase of bandwidth necesserely but a better slew rate with a nice rolloff might be the desirable factor here.

Maybe a couple of us can make some requirements for what we want and start digging around datasheets and find an opamp with what might be the best performanc for the stage (on paper anyways).... I'm sure someone has done that before but if people are interested we can do some combined research.

Off of first glance OPA140/OPA2140 looks pretty good at unity gain stable (according to the datasheet) at 11MHZ with a slew rate of 20V/ms and some nice noise characteristics.

NOISE​
Input Voltage Noise
f = 0.1Hz to 10Hz 250 nV​
PP

f = 0.1Hz to 10Hz 42 nV​
RMS

Input Voltage Noise Density e​
n

f = 10Hz 8 nV/​
Hz
f = 100Hz 5.8 nV/
Hz
f = 1kHz 5.1 nV/
Hz
Input Current Noise Density I
n

f = 1kHz 0.8 fA/Hz
 
Status
Not open for further replies.