Hi all,
I've been looking into the whole IV conversion thing and came across two different methods for using an opamp as an IV converter. I have attached an image of the two configurations.
The top one is the traditional active way.
The bottom one uses the input resistor to do the IV conversion and the Opamp simply amplifies the passive conversion, so its passive with a gain stage.
It seems to me that the second way is intrinsically superior. This is for a few reasons;
-the signal is passively converted.
-the input can be filtered so that the opamp has none of the hash to deal with
-following on from this the opamp can be freed from slew rate considerations
This is the main configuration in video amplifier applications where they have similar slew rate issues to contend with.
Would anyone care to cast some pearls of wisdom before this swine.
Shoog
I've been looking into the whole IV conversion thing and came across two different methods for using an opamp as an IV converter. I have attached an image of the two configurations.
The top one is the traditional active way.
The bottom one uses the input resistor to do the IV conversion and the Opamp simply amplifies the passive conversion, so its passive with a gain stage.
It seems to me that the second way is intrinsically superior. This is for a few reasons;
-the signal is passively converted.
-the input can be filtered so that the opamp has none of the hash to deal with
-following on from this the opamp can be freed from slew rate considerations
This is the main configuration in video amplifier applications where they have similar slew rate issues to contend with.
Would anyone care to cast some pearls of wisdom before this swine.
Shoog
Attachments
Of course, the second method does not provide the ideal short-circuit load to the DAC. And designing it so that the noise floor is not compromised is a decidedly non-trivial task.
I guess I'm not sure why passive conversion is "better." High speed fast settling opamps are pretty common and reasonably inexpensive. And if one objects to using a chip, a discrete circuit is not all that complicated.
I have done a bit of research on this and it seems generally that very few people are happy with the sound of the traditional opamp implementation. I have read in numerous places that the best results are achievable with passive conversion.
It seem that many people feel that the input overwhelms the opamp which then has slewing problems with the high frequency residuals in the signal. Placing the compensating cap in the feedback loop means that the damage has already been done before you try to correct for it.
So the answer is, passive is widely considered to sound better, but has issues with output level, and in some cases with placing the DAC's output out of its specified voltage range with resultant distortion. It also makes a very poor line driver. The implementation I am looking at works reasonably well with a voltage range up to 0.8V on the output, and I was thinking of running the output into a 100R resistor to keep that a lot lower. I believe you are right to say that ground noise maybe the enemy. If so it is not a problem to raise the resistor to about 2K and reduce the noise.
The cap I presume is the reason for your concern about short circuit behaviour. I suppose a zobal could be used.
Shoog
It seem that many people feel that the input overwhelms the opamp which then has slewing problems with the high frequency residuals in the signal. Placing the compensating cap in the feedback loop means that the damage has already been done before you try to correct for it.
So the answer is, passive is widely considered to sound better, but has issues with output level, and in some cases with placing the DAC's output out of its specified voltage range with resultant distortion. It also makes a very poor line driver. The implementation I am looking at works reasonably well with a voltage range up to 0.8V on the output, and I was thinking of running the output into a 100R resistor to keep that a lot lower. I believe you are right to say that ground noise maybe the enemy. If so it is not a problem to raise the resistor to about 2K and reduce the noise.
The cap I presume is the reason for your concern about short circuit behaviour. I suppose a zobal could be used.
Shoog
How people "feel" has nothing to do with the reality of circuits and devices.
This may have been true 20 years ago for some players, but the reality is that any issues from the inability of the I/V to track and settle would be easily measurable. If they're not there, then the "problem" is imaginary (imaginary problems are not uncommon in high end audio). And indeed, it's trivial to slap in an opamp that will not have slewing issues.
I have no idea what the second sentence means.
http://members.chello.nl/~m.heijligers/DAChtml/Analogue/IV.html
From someone who measured many of the options.
And a lengthy discussion of why most opamps fail to deliver from this forum;
http://www.diyaudio.com/forums/showthread.php?threadid=34324&pagenumber=2
Hope that clarifies where I am coming from.
Shoog
Where would I find the measurements? Not at the link you provided.
Interestingly, Gary Galo's article on the Benchmark DAC (AX, 1/09) with a good assortment of measurements by forum member chascode showed no indication of I/V slew rate or settling deficiencies. The I/V converter was a plain old 5532. Galo's open listening "test" writeup was a rave about the sound quality.
The first sch (upper one), is "standard" OP IV transimpedance mode, with virtual ground... output signal is in phase. This topology is fashion in almost all comercial units...
*
the other sch is R-IV (with C close to DAC, R too...) phase opposite V, then amplification
with OP inverting mode, so output of V is in phase...
*
?
*
the other sch is R-IV (with C close to DAC, R too...) phase opposite V, then amplification
with OP inverting mode, so output of V is in phase...
*
?
I would be interested to know if anyone had ever built both versions and compared them for sound. I suppose that is the proof of the argument. My thoughts are a product of the generally low regard people have expressed about their encounters with opamp IV converters, and the high regard I have heard about passive conversion. I won't be in a position to try any of this out until about a months time so I was hoping for some "experience" to chew on. We all know that the final result cannot always be judged from the figures.
Shoog
Shoog
it is interesting that so many audio DAC app circuits presumably trying to showcase their distortion numbers still use the NE5534/2
I have clearly seen on my 'scope the asymmetric step response into typical I/V feedback due to the poor PNP performance of the older NE5534 op amp technology
a more modern complementary process op amp or a speedy buffer in the loop will improve the step response symmetry
for that matter even the old "Class A" bias trick with a resistor or CRD to the negative rail improves the look of the NE5534 response a lot - but I don't recall ever seeing this on modern audio DAC application note circuits
it is actually less of a problem now with near universal use of internal oversampling/interpolation that limits modern audio DAC output step size with CD source - NOS DAC sound probably is more sensitive to I/V electronics implementation
I have clearly seen on my 'scope the asymmetric step response into typical I/V feedback due to the poor PNP performance of the older NE5534 op amp technology
a more modern complementary process op amp or a speedy buffer in the loop will improve the step response symmetry
for that matter even the old "Class A" bias trick with a resistor or CRD to the negative rail improves the look of the NE5534 response a lot - but I don't recall ever seeing this on modern audio DAC application note circuits
it is actually less of a problem now with near universal use of internal oversampling/interpolation that limits modern audio DAC output step size with CD source - NOS DAC sound probably is more sensitive to I/V electronics implementation
Following on from this, I am going to try the "other way of doing opamp IV" with a little OPA627 buffer I have adapted. All I need now is the DAC to get going on, but it will have to wait a little.
Following on from this, does anyone understand the concept of "mixed feedback" as proposed by Lundahl and discussed by Walt Jung. I don't unfortunately have access to the Jung book and the critical page is missing from Googler books. Anyway I have a LL1517 which I am just itching to use and wondered if the attached diagram was a functional implementation of "mixed feedback" or whether it would produce runaway positive feedback.
Shoog
Following on from this, does anyone understand the concept of "mixed feedback" as proposed by Lundahl and discussed by Walt Jung. I don't unfortunately have access to the Jung book and the critical page is missing from Googler books. Anyway I have a LL1517 which I am just itching to use and wondered if the attached diagram was a functional implementation of "mixed feedback" or whether it would produce runaway positive feedback.
Shoog
Attachments
I found a copy of the Walt Jung article and the way I was trying to achieve the circuit requires a DAC sense resistor of just 9.2R which would have given an output signal of just 0.01V *7. Not quite good enough. I worked out the correct implementation which is given here.
The only real downside is the brutally large cap in the return leg.
Shoog
The only real downside is the brutally large cap in the return leg.
Shoog
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.