Of course. The programmers of the other software packages did the same work as Virtins did. My point is that adapting the software to any new / changed interface is possible only if the full documentation is made available to everyone, not just Virtins.
Separated ADC chips
Yes it is. Under certain conditions (measuring ultra low distortions by means of a notch filter) cross-talk is a real problem. Please see this post,
in particular: "I also repeated these experiments with an RTX6001. Using a DAC and ADC for synchronization: THD = 119.4dB. Using the "bypass" and at measuring times of 43, 427 and 4268 seconds, THD was -143.3, -151.0 and -161.4dB respectively."
Given the large drop in distortion (-161,4 vs -119.4dB) completely separated ADC's would be very welcome.
Cheers, E.
Hi Jens,
Yes it is. Under certain conditions (measuring ultra low distortions by means of a notch filter) cross-talk is a real problem. Please see this post,
in particular: "I also repeated these experiments with an RTX6001. Using a DAC and ADC for synchronization: THD = 119.4dB. Using the "bypass" and at measuring times of 43, 427 and 4268 seconds, THD was -143.3, -151.0 and -161.4dB respectively."
Given the large drop in distortion (-161,4 vs -119.4dB) completely separated ADC's would be very welcome.
Cheers, E.
Any plans to use the newer DACs, preferably the AK4499?
Or would that require too much new development?
Best wishes
David
The AK4499 could be an interesting option. Others have a wish for fully isolated generators, which will require 2 DAC's. 2 x AK4499 plus supporting circuits would increase the cost significantly.
I am preparing to do some tests with another option, which is much lower cost, but still with a lower distortion and lower noise than the present solution. It does have some small drawbacks though, so I will need to get all the results of the test and make a choice.
I have also tested the AK4493. It gives a solid improvement over the AK4490 and the risk of downsides is limited. The cost is only slightly higher than the cost of the AK4490. And surrounding circuits are very similar.
I am preparing to do some tests with another option, which is much lower cost, but still with a lower distortion and lower noise than the present solution. It does have some small drawbacks though, so I will need to get all the results of the test and make a choice.
I have also tested the AK4493. It gives a solid improvement over the AK4490 and the risk of downsides is limited. The cost is only slightly higher than the cost of the AK4490. And surrounding circuits are very similar.
Lots of options-
One AK4493
Double, isolated AK4493
Mystery "lower cost...lower distortion" option
One AK4499
Double, isolated AK4499.
Looks like some nice improvement from any of those.
Very interested to see how this works out and curious about the "mystery" option, if it's not proprietary.
Best wishes
David
What order was the improvement you observed with the AK4493 vs AK4490? >, = , or < 3dB.
Have you run into any challenges with isolation and RF ingress or emissions?
I ask since even the existing isolation of the USB interface presumably required the USB port shell to not connect to chassis. A new model with isolated XLR outputs would then not have Pin 1 / shell tied to chassis? I've never checked on my unit but I just assumed they were. Seems like a compromise there is no way around.
The "mystery" option is a device, which, as far as I know, has not been discussed much on this forum. It is the Cirrus Logic CS43198.
The distortion at high levels (up to 0dBFS) is very low (below -130dB at 1 kHz). The noise is also very low in the audio band. There is some out of band noise though, when used at 192 kHz sample rate. That is one of the drawbacks. The specified temperature stability is worse than for the AK4490.
See the attached figure.
It compares the AK4490 (with different supply voltages and op-amps) with the AK4493 (with different op-amps).
The orange curve is an arbitrary production unit of the RTX6001. It should have been like the red curve, but there is some vspread in performance.
The good results are the ones with AK4493 and either LME49720 or OPA1656. The result with the OPA1612 is still a mystery to mee. Perhaps there is something like the "ESS hump" here ?
I have not designed it yet. And yes, the pin 1 / shell og the XLR is a challenge. I have not found the perfect solution to that at the moment. Ideas are welcome !
Another challenge is of course to make the isolated (most likely switch mode) supply without getting a lot of noise.
Attachments
Have you considered the old, faithful AD797?
You probably know that it is based around a distortion elimination idea which is theoretically attractive.
All the measurements I have seen indicate that the theory does indeed pay off with excellent results.
It seems an obvious candidate, I am a bit surprised it hasn't been mentioned.
Maybe it's just my attraction to a clever use of theory.
Best wishes
David
Last edited:
It has been confirmed several times that the AD797 is not a good op amp for audio. The LME or OPA is more suitable. Every application needs certain op amps.
By whom?
I would need to see some very solid evidence, Samuel's tests of the AD797 were pretty impressive.
Why?
Best wishes
David
Last edited:
I can't always collect the measurements that were published here. Please search for yourself. Among other things, the distortions are higher than with the LME.
@diyralf
I know that application note.
If I connect pin 1 to the chassis, which is what should normally be done, the output is not really isolated. But maybe I should just keep the outputs as they are on the RTX6001.
The AD797 is a good op-amp, but since it is also very expensive, I would minimize the use of it to places, where it really matters. The input stage could be one such place, but for now I have the LME49990.
I don't think it would make a big difference in the DAC filter/buffer.
Stability can be a bit tricky with the AD797.
I know that application note.
If I connect pin 1 to the chassis, which is what should normally be done, the output is not really isolated. But maybe I should just keep the outputs as they are on the RTX6001.
The AD797 is a good op-amp, but since it is also very expensive, I would minimize the use of it to places, where it really matters. The input stage could be one such place, but for now I have the LME49990.
I don't think it would make a big difference in the DAC filter/buffer.
Stability can be a bit tricky with the AD797.
The usual blabla, not backed up by facts.
That which can be asserted without evidence, can be dismissed without evidence. (Cristopher Hitchens)
Last edited:
Of course.
Yes, I think there's no dispute the stability requires more care.
With the End Of Life of the LME49990 I just wondered if you wanted to stay committed to it.
If you have sufficient stock of LME49990 for the foreseeable production run then it makes sense to stay with them.
Thank you for the explanation.
Best wishes
David