some people just can't resign the game when the time comes... Anon. vs Anon.: B02 Alekhine Defense: Brooklyn Variation • lichess.org
1) All those shown figures, like a bucket turned upside down, have no meaning when you don’t tell:
- what they are, simulations or measurements
- what was measured resp simulated, i.e. topology of the used circuit
- how it was measured, to convince us of the reliability.
2) One one hand you don’t believe simulated opamp I-V figures that are below so called measured figures of your discrete I-V circuit.
On the other hand you say opamp THD figures are so low that they can hardly be measured.
So please stop contradicting yourself and accept that opamps can easily match your discrete design.
3) I mentioned that in all cases, either discrete or opamp, simulated THD figures below 120dB are to be disregarded, because:
- most of the time they are unreliable
- even more important, they do not play any sensible role in our hearing perception.
Hans
Hans,
The circuit is essentially a (complimentary) grounded base stage -> folded cascode -> buffer with no global FB or inter-stage FB, just local. I think it's a very cool circuit.
If we consider just the GBS and in it's simplest zero feedback form and measure the OP current distortion it will be lowest when fed from a very high impedance, true CCS DAC.
As the source (DAC) impedance lowers the distortion will go up. The same can be said for a GBS with local feedback (CFP etc) but obviously the distortion will be much lower.
In this case the circuit uses CFP arrangements but fed with CCS from rails instead of the usual resistors, this dramatically lowers distortion further again (I've used it -> powerful design trick)
so it may take a very low DAC OPZ for this distortion to show up.
This I-V could be a good candidate for ES9038 pro in Mono mode. It would be interesting to simulate with 9038's (Mono) OPZ and current swing - see how
it handles the heat!
TCD
Last edited:
Hans, no one said that such distortions can be heard.
But there are places where it is necessary for measuring equipment, for example.
low Distortion Oscillator GS 5534-8 (NE5534)
low Distortion Oscillator GS 5534-8 (NE5534)
But there are places where it is necessary for measuring equipment, for example.
low Distortion Oscillator GS 5534-8 (NE5534)
low Distortion Oscillator GS 5534-8 (NE5534)
Not all operational amplifiers in monolitic integrate circuit form are compensated for unity gain, you find some of them compensated for 5x and 10x as well , but I read your other topic and i know what you mean.It may not be a problem using those circuits at fixed frequencies in state variable filters, i'm not an expert in that , but with the usual DAC audio range you get some even higher range of frequencies than just the audio frequency range to deal with and for that you need stability at least in the audible range and up to more than 44kHz which is a different perspective.Did you run some stability tests as well?
Last edited:
Hi,
I'd like to repeat my Q from #6 as it was meant a serious question ... what do You expect to achieve with such complex designs?
For a audio DAC intended for home HiFi there are numerous simpler designs that sound excellent and that show more than sufficiently low thd figures.
I remember a single power MOSFET iv-design (was a lengthy thread, from iIrc a ausralian guy) with extremly low thd-figures. The MOSFET was biased quite hot and the circuit was quite inefficient power wise.
That should answer Your Q.
Now it'd be kind if You answered my Q.
The circuit in #10 appears overly complex also -I'd rather call it low elegance- as unipolar output DACs like the PCM179x series lend themselves to singleended designs
jauu
Calvin
I'd like to repeat my Q from #6 as it was meant a serious question ... what do You expect to achieve with such complex designs?
For a audio DAC intended for home HiFi there are numerous simpler designs that sound excellent and that show more than sufficiently low thd figures.
I remember a single power MOSFET iv-design (was a lengthy thread, from iIrc a ausralian guy) with extremly low thd-figures. The MOSFET was biased quite hot and the circuit was quite inefficient power wise.
That should answer Your Q.
Now it'd be kind if You answered my Q.
The circuit in #10 appears overly complex also -I'd rather call it low elegance- as unipolar output DACs like the PCM179x series lend themselves to singleended designs
jauu
Calvin
There is the OP27 (1); OP37 (5) and there was OP47 (uncomp). 1979?
Was discontinued because of no demand
It depends on what the actual sources of distortion are.
There are certainly cases where at G=1000, distortion is not 1000 x G = 1 in magnitude. A good example is some CFB mic pre amp chips.
TCD
Hi,
the Q was not about if it´s hard to design ... the Q was about why you design it with that high degree of complexity.
You repeatedly claim that one couldn´t reach similar´parameters´ with integrated OPAmps or simpler designs.
What You have shown so far are some tables which are hard to read and almost impossibe to interpret in first place, and that seemingly only contain THD-figures.
What You haven´t shown so far are the setups You used to derive that single parameter.
You also haven´t shown us other parameters which may play a more important role depending on the application, like noise, offset, input bias, supply rejection, temperature coefficient etc. etc.
If you claim Your discrete OPAmp be superior, than provide us with more parameters and a list of suggested applications, otherwise all further exchange is moot and won´t lead to a decent and fruitful discussion.
jauu
Calvin
the Q was not about if it´s hard to design ... the Q was about why you design it with that high degree of complexity.
You repeatedly claim that one couldn´t reach similar´parameters´ with integrated OPAmps or simpler designs.
What You have shown so far are some tables which are hard to read and almost impossibe to interpret in first place, and that seemingly only contain THD-figures.
What You haven´t shown so far are the setups You used to derive that single parameter.
You also haven´t shown us other parameters which may play a more important role depending on the application, like noise, offset, input bias, supply rejection, temperature coefficient etc. etc.
If you claim Your discrete OPAmp be superior, than provide us with more parameters and a list of suggested applications, otherwise all further exchange is moot and won´t lead to a decent and fruitful discussion.
jauu
Calvin