New DISPRE preamp, successor to previous popular version

[PMA]

Pavel, have you checked the datasheet and app note for the BUF634? Something there was mentioned about instability under certain conditions.

Per-Anders, thanks, I will recheck it. I did read the datasheet carefully, but long ago.

 

[scott wurcer]

I would like to ask Scott Wurcer if he or his team has experienced UHF local oscillations in output of opamp stages, especially diamond buffers. Thanks for ev. answer.

Of course we have experienced all types of oscillation, hopefully before release. 😀

 

[PMA]

Thanks for your reply 😀

 

[Jan Dupont]

PMA

Thanks, Jan. I have nothing to hide. And I also have no reason to debate with everyone who only irritates (meaning not you, of course).

I know 🙂

Sigurd;
I have the EMU Tracker Pre and find it very good when making measurements using a laptop

 

[Salas]

You can run the Tracker Pre on a separate battery if you wish.
I do not know about power consumption - check datasheet.




Sigurd

There is a 5V/1A DC adapter input, I know, but I was looking to avoid extra PSU noise and have a simpler set up when mobile. Since there is no USB consumption info in the data sheet I wanted to know the toll on the laptop battery by someone who is already using the Pre. I will have to find out when I get it I guess. Thanks for your loop measurement of the Pre and your answer anyway.

 

[john curl]

PMA, I see that the major technical input has moved here. Lots of luck with SYN.

 

[bogdan_borko]

PMA is there a major differences between real measured distortion and MicroCap simulation?

 

[PMA]

In case that transistors are matched to less than 1%, no. No significant difference, in case of this circuit. Not valid for every circuit.

 

[PMA]

PMA, I see that the major technical input has moved here. Lots of luck with SYN.

Welcome and thanks for your input, John.

I reply constructive ideas and questions and ignore irritations 😉

 

[scott wurcer]

PMA is there a major differences between real measured distortion and MicroCap simulation?

PMA you may want to check this out, fine scale distortion simulation with normal JFET models is hopeless. They are "primitive" to say the least.

 

[PMA]

Scott, thanks for your input, I should have mentioned that I simulated the BJT preamp, and BJT models are sometimes not 'that' bad.

However, I failed in simulations of output current limitation (shrt), simulation showed total nonsense (no Vbe zener etc.) -

Hereby the measurement on output Re resistor (2R7), output shorted to gnd. Input is 8Vp-p, gain 4x approx.

BTW, I studied enthusiastically Wong and Ott during my university studies.

 

[Sigurd Ruschkow]

Pavel - did you try to do Monte Carlo analysis in MC and check how the distortion sims correlate to the real amp?

Amazing what MC can do!
I use it almost every day.



Sigurd

In case that transistors are matched to less than 1%, no. No significant difference, in case of this circuit. Not valid for every circuit.

 

[PMA]

Sigurd, thanks for the hint. I haven't tried yet.

 

[Sigurd Ruschkow]

Here's a measurement made witht he Tracker Pre in unbalanced connections with no special consideration taken for
ground issues or signal conditioning or battery power. Just connected the Tracker Pre to a Technics SUA6 line amp.

Spectra is not as clean as it can be but the noise floor is still
at 140 dB.
This time I used ARTA for software.


Sigurd

Line out -> Line in balanced mode is like the AP self test. The problems are starting when connecting to the outside world (ground loops, EMI radiation, etc...). Otherwise said, I'm not surprised by the intrinsic noise floor of the sound card but by the low noise of the entire measurement setup. In post #338 http://www.diyaudio.com/forums/showthread.php?postid=1704731#post1704731 while measuring the preamp, the noise floor is -132dB, and this is not even in a balanced configuration! That's why I'm asking about details on the measurement configuration.

What sound card are you using? I have tried last year a Lynx L22 and I was unhappy with the results. Of course averaging helps but it took forever to average down to a reasonable floor.

A question I was never able to get a good answer for: if soundcards + software are that good, how come AP is still in business?

 

[syn08]

Here's a measurement made witht he Tracker Pre in unbalanced connections with no special consideration taken for
ground issues or signal conditioning or battery power. Just connected the Tracker Pre to a Technics SUA6 line amp.

Spectra is not as clean as it can be but the noise floor is still
at 140 dB.
This time I used ARTA for software.


Sigurd

Thanks. Can you change the FFT window type (Hanning, Flat, etc...)?

What chipset is this audio card using? What type of opamps?

How many averages are you using? How long is a measurement cycle?

 

[PMA]

- usually Blackman (Blackman-Harris), sometimes Hanning
- 64K or 128K FFT record at 48kS/s or 96kS/s; 192kS/s my colleague
- averaging 1 to 100

 

[Sigurd Ruschkow]

Yes, one can change to several FFT windows.
I highly recommend to try out the software ARTA (incl STEPS and LIMP) and also SpectraPlus.
ARTA is only about 100 Euro. SpectraPlus is a professional SW.
Both can be found as demos.

The Tracker Pre uses these converters: AK5385 and CS4392.
If you want better ones, the EMU 1616m uses AK5394S and CS4398 and the 1515m has even lower THD and better SNR.

I do not know what OPamp the Tracker Pre uses.

Measurement time is dependant on FFT size and nr of averages. At 128k samples with ARTA each sweep takes about a second.
At <16k it is real time.

I use between 1 and 100 averages depending on what type of measurement I do, and what accuracy I want.



Sigurd

Thanks. Can you change the FFT window type (Hanning, Flat, etc...)?

What chipset is this audio card using? What type of opamps?

How many averages are you using? How long is a measurement cycle?

 

[estman]

Not sure the JFETs sound as good as MOSFETs for output stages, though.

Sigurd, thanks for your opinion. Do you have real experience with both of them - measurement and/or listening? Sim?

 

[PMA]

Well, in my experience the sound is not much a case of output devices. Many many times the JFET or MOSFET link stage output (regardless if push-pull or SE) was beaten by bipolar in our tests, and SE MOSFET owners were quite surprised. IMHO the result depends on design, and for sure the ouput stage must work in class A, which is no problem for the link stage.

 

[estman]

PMA, please explain, what is ''link stage''.
Class A, yes certainly - it's like magic wand and most distortion disapears 😉