VAS CMCL
(regarding post #2479 )
By the way, talking about the 'simplest solution' , in principle, the Iq of a high gain complementary VAS can be controlled by a CMCL comprising only two additional trannies. (not recommended for utmost performance)
As for post # 2480, I have a lot of "subconscious agenda's": Iq stability, distortion, headroom. phase margin, clipping behavior, etc.
.
Cheers,
E.
(regarding post #2479 )
By the way, talking about the 'simplest solution' , in principle, the Iq of a high gain complementary VAS can be controlled by a CMCL comprising only two additional trannies. (not recommended for utmost performance)
As for post # 2480, I have a lot of "subconscious agenda's": Iq stability, distortion, headroom. phase margin, clipping behavior, etc.
.Cheers,
E.
I have a lot of "subconscious agenda's": Iq stability, distortion, headroom. phase margin, clipping behavior, etc.
Cheers,
E.
+1
Edmond, any news regarding the TIS?
Cheers
Stein
/OT
Hi Stein,
Not yet. I'm still busy with perfecting my PC based distortion analyzer (DiAna). One of the problems is how to calculate the exact frequency (and other variables) based on sampled data of an external (not synchronized) sine wave source. With 'exact' I don't mean within a few decimals. No, I mean 16 decimals behind the comma. On the web I couldn't find anything about this topic, so I have to figure it out all by myself. Not an easy job!
I've also tested a couple of sound cards. Regrettably, most of them were cr@p, in particular the ones from Creative.
By the way, I just downloaded a new C++ compiler from Walter Bright (for free!), which support the Intel 80 bits FPU 'long doubles'. Just what I was looking for.
Cheers,
E.
+1
Edmond, any news regarding the TIS?
Cheers
Stein
Hi Stein,
Not yet. I'm still busy with perfecting my PC based distortion analyzer (DiAna). One of the problems is how to calculate the exact frequency (and other variables) based on sampled data of an external (not synchronized) sine wave source. With 'exact' I don't mean within a few decimals. No, I mean 16 decimals behind the comma. On the web I couldn't find anything about this topic, so I have to figure it out all by myself. Not an easy job!
I've also tested a couple of sound cards. Regrettably, most of them were cr@p, in particular the ones from Creative.
By the way, I just downloaded a new C++ compiler from Walter Bright (for free!), which support the Intel 80 bits FPU 'long doubles'. Just what I was looking for.
Cheers,
E.
Hi Guys
For a computer-based distortion analyser to be of any true use, the interface has to have more than 20kHz bandwidth, which pretty much leaves common and even exotic sound cards in the bin. I've heard of a 32-bit converter that is very good but hard to get - and if you can there are no drivers for it. Having that kind of resolution seems important for measurement, otherwise you diddle around trying to keep generated signals and sampled signals as close to full-scale as possible to have the lowest THD out and lowest resolution errors in. But what you also need is 200kHz bandwidth.
I'm sure you've checked out the Dr. Jordan site and his descriptions of his software. For the software literate, the descriptions should give away enough for you to see what path he took.
Could we be more off topic? At least Bob's book has a chapter about audio measurements. Personally, I found ch.3 to be the most informative.
Have fun
Kevin O'Connor
For a computer-based distortion analyser to be of any true use, the interface has to have more than 20kHz bandwidth, which pretty much leaves common and even exotic sound cards in the bin. I've heard of a 32-bit converter that is very good but hard to get - and if you can there are no drivers for it. Having that kind of resolution seems important for measurement, otherwise you diddle around trying to keep generated signals and sampled signals as close to full-scale as possible to have the lowest THD out and lowest resolution errors in. But what you also need is 200kHz bandwidth.
I'm sure you've checked out the Dr. Jordan site and his descriptions of his software. For the software literate, the descriptions should give away enough for you to see what path he took.
Could we be more off topic? At least Bob's book has a chapter about audio measurements. Personally, I found ch.3 to be the most informative.
Have fun
Kevin O'Connor
Last edited:
Last edited:
/OT
The bandwidth of my software (DiAna) is virtually unlimited and independent of the sampling rate. The BW of the whole setup is only limited by the anti alias filter in front of the ADC.
> Dr. Jordan (Win Audio MLS)
Come on Kevin, you can't be serious.
> keep generated signals and sampled signals as close to full-scale as possible to have the lowest THD.
The majority of sound cards produce more distortion at full-scale.
Cheers,
E.
The bandwidth of my software (DiAna) is virtually unlimited and independent of the sampling rate. The BW of the whole setup is only limited by the anti alias filter in front of the ADC.
> Dr. Jordan (Win Audio MLS)
Come on Kevin, you can't be serious.
> keep generated signals and sampled signals as close to full-scale as possible to have the lowest THD.
The majority of sound cards produce more distortion at full-scale.
Cheers,
E.
Hi Guys
Sure I can be serious - but why not have brevity? Life is short.
The Dr.Jordan stuff was the best I could find that presented any description of performance and tests possible. His software allows mining out -140dB distortions and more, just as a lot of software and mathematical manipulation can do, and they used to have some graphs showing the results with the best sound cards available at the time.
Everything else I could find seemed to be designed by guys who thought some arbitrary (usually quite high) level of THD was "good enough", or they simply did not do THD tests, or did not test to 20kHz
Most info I found back when I was looking at such things also suggested that to make the best use of standard sound cards, one has to make best use of the voltage range allowed. To measure a power amp output at anything above milliwatt levels requires attenuating the amp output in a precise way but keeping the level to the card on the high side. You know better than I about the vagaries of the card performance and software compromises to make it "good" but "affordable", Your statement that the "majority" of cards have high THD at high levels suggests that there are some exceptions that work much better? If that is the case, then they would be the units to employ.
Audio Precision set the standard to which all THD measurement systems aspire. They have priced themselves into a very stratified domain which makes them a target as much as an icon. The Dr.Jordan software was the only one I found that claimed similar possible performance. I would hope there are others available now, or other actual measuring systems that are priced more sanely than AP.
Have fun
Kevin O'Connor
Sure I can be serious - but why not have brevity? Life is short.
The Dr.Jordan stuff was the best I could find that presented any description of performance and tests possible. His software allows mining out -140dB distortions and more, just as a lot of software and mathematical manipulation can do, and they used to have some graphs showing the results with the best sound cards available at the time.
Everything else I could find seemed to be designed by guys who thought some arbitrary (usually quite high) level of THD was "good enough", or they simply did not do THD tests, or did not test to 20kHz
Most info I found back when I was looking at such things also suggested that to make the best use of standard sound cards, one has to make best use of the voltage range allowed. To measure a power amp output at anything above milliwatt levels requires attenuating the amp output in a precise way but keeping the level to the card on the high side. You know better than I about the vagaries of the card performance and software compromises to make it "good" but "affordable", Your statement that the "majority" of cards have high THD at high levels suggests that there are some exceptions that work much better? If that is the case, then they would be the units to employ.
Audio Precision set the standard to which all THD measurement systems aspire. They have priced themselves into a very stratified domain which makes them a target as much as an icon. The Dr.Jordan software was the only one I found that claimed similar possible performance. I would hope there are others available now, or other actual measuring systems that are priced more sanely than AP.
Have fun
Kevin O'Connor
the AP boxes are really good deals for what you get
you aren't going to diy yourself into that level of reliability, usability, have built in the benefits of decades of experience serving diverse customers thousands of units in the field...
you can cobble together stuff to beat maybe one or two parameters in a dedicated test
but by the time you box up the rigs, fixtures, wrote software to do even 10% of what a modern AP does you would have saved time working minimum wage to buy the AP
you aren't going to diy yourself into that level of reliability, usability, have built in the benefits of decades of experience serving diverse customers thousands of units in the field...
you can cobble together stuff to beat maybe one or two parameters in a dedicated test
but by the time you box up the rigs, fixtures, wrote software to do even 10% of what a modern AP does you would have saved time working minimum wage to buy the AP
Hi Guys
Yes, AP products are great. I believe they discontinued the portable unit that does just the basic audio measurements most people want, but haven't been to their site for a while. I am fully aware of their history and expertise.
I have zero interest in 90% of the capabilities of the AP systems since I only build analogue audio gear.
Have fun
Kevin O'Connor
Yes, AP products are great. I believe they discontinued the portable unit that does just the basic audio measurements most people want, but haven't been to their site for a while. I am fully aware of their history and expertise.
I have zero interest in 90% of the capabilities of the AP systems since I only build analogue audio gear.
Have fun
Kevin O'Connor
Bob Cordell's differential shunt resistor is probably the simplest solution to complementary differential inputs with current mirrors. Bob selected a 47K resistor to handle most component variation.
I have been using THAT transistor arrays for the input stage current mirrors and cascode transistors. The tight array transistor tolerances allow a much higher differential shunt resistor value. The higher input stage gain provided by current mirrors does require adding a series R+C differential shunt to control high frequency oscillations. This R+C was not necessary with low gain simple resistor differential loads, and some golden ears favor resistor load complementary differential stages for this reason.
Agree on that part.
100K will work as well without degrading perfs and it slightly increase the gain..
Increasing the compensation seems more cautious imo.
It could be that golden ears like distorsions of the worse kind since only current mirror loading allow radical reduction of IMD , at least in my tries.
