P3 Adjustment for the average Hobbyist

Official Court Jester
Joined 2003
Paid Member
as Pa said ..............

50% of ppl prefer dominant 2nd over 3rd

other 50% prefer dominant 3rd over second

then you can divide both groups again in halves , depending who likes positive vs. liking of negative phase of dominant harmonic

:devilr:

I know where I belong , mostly ........... but when gadget in question is exquisite , I'm not exclusive , at all
 
I am a bit unsure if we are able to check if the phase is negative or positive with the equipment we have (assuming that it is 2i2 and ARTA software). But we can see from 6l6 video in which direction of P3 settings is negative phase and which is positive phase. It will be interesting to know what we like best. It is interesting that high-end amp stuff is designed in such details…...as to look if the 2nd harmonic is in phase or not with the fundamental. Think the designer of cheaper amps does not think in these details other then maybe have the overall distortion as low as possible to be able to show fine data.
 
Member
Joined 2016
Paid Member
Focusrite 2i2 2nd gen loopback and attenuator teaser

I can only give a small teaser here, because I will be going on vacation with the family tomorrow morning for the next two weeks ...

This is my plain loopback spectrum:

attachment.php


I got the best performance at -3 dB, and sampling set to 48 kHz.
I used a plain TS mono cable, no balanced / SE conversion.
This is the cable:
Cordial CCFI 0,6 PP – Thomann UK

I also built myself an attenuator box to a) protect the Focusrite, and b) adjust input signal to be able to measure power amp output as well and not overload the inputs.
The attenuator I built follwed a design given by Dan Joffe of Akitika,
1 kHz Oscillator
(look to the bottom of the page, where it reads "Building the Switchable Attenuator")
https://www.akitika.com/documents/BuildingTheAttenuatorRev4.pdf

I'm sure I got the pointer to this attenuator from somewhere here, but can't remember at the moment in what thread I picked it up ... :scratch:

Anyway, the attenuator works fine, but even at the setting with the least attenuation (-6 dB), in my implementation at least, the loopback through the attenuator gives not as good performance as straight loopback:

attachment.php


At -18 dB:

attachment.php


This doesn't bother me much, because the distortion and noise figures of the devices I have measured so far are quite a bit higher ... :p

Here a few pics, at least from the outside - I put several input options in to be able to use a variety of probes:

attachment.php


attachment.php


More photos from the inside, if somebody is interested, after family vacation ... so far, I have measured distortion of an Aikido pre-amp and my M2, and it works fine.

Best regards,
Claas
 

Attachments

  • Spectrum Loopback optimized 180515.jpg
    Spectrum Loopback optimized 180515.jpg
    271.8 KB · Views: 442
  • Spectrum Loopback Attenuator -6 180513.jpg
    Spectrum Loopback Attenuator -6 180513.jpg
    275.1 KB · Views: 456
  • Spectrum Loopback Attenuator -18 180513.jpg
    Spectrum Loopback Attenuator -18 180513.jpg
    277 KB · Views: 438
  • Attenuator and 2i2 small.jpg
    Attenuator and 2i2 small.jpg
    421 KB · Views: 447
  • Attenuator and 2i2 top small.jpg
    Attenuator and 2i2 top small.jpg
    386 KB · Views: 439
Just another Moderator
Joined 2003
Paid Member
Interesting you are getting very good distortion performance with the TS loopback cable! Is yours a second gen 2i2? I'll have to try with 48Khz sampling, I've only been testing with 96Khz...

I assume you increased the gain on the preamp when using the attenuator to get the same -3db reading. Try doing it again with arta showing -9db and the performance may be better. I noticed that both the input gain and the output level affect the distortion even when within limits.

Tony.
 
Just another Moderator
Joined 2003
Paid Member
I've done a few more tests this morning on my 1st gen focusrite 2i2 at 48Khz sampling rate, and can confirm that the distortion performance is better at 48Khz for the gen1 as well.

However it is still much better if the Sleeve is isolated. I also did the calibration in ARTA this time. Voltage output at -3db is around 1.6V. All measurements below with exact same settings on 2i2 volume knob at max, ARTA set to -6db on generator, inout gain set to read as close to -6db as I could get it.

First the loopback with TRS loopback cable.
attachment.php


Next loopback using Tip and Ring only, no sleeve connection.
attachment.php


finally loopback using a fender guitar TS cable (I don't have a short TS loopback cable)
attachment.php


So at least for my 2i2 it is quite important to not connect sleeve.

I will do a test today with passive means measuring my active crossover, and check results against with the balanced to SE converter.

Tony.
 

Attachments

  • bal_loopback_48Khz_-6db.png
    bal_loopback_48Khz_-6db.png
    26.6 KB · Views: 457
  • tr_loopback_-6db.png
    tr_loopback_-6db.png
    27.8 KB · Views: 442
  • TS_loopback_-6db.png
    TS_loopback_-6db.png
    25.5 KB · Views: 442
Just another Moderator
Joined 2003
Paid Member
Well the plot thickens!! When measuring my active crossover (unfortunately the best line level device I can try) I get some very odd results!! Note that measurements were taken with ARTA set to -6db as before, and all controls left as they were.

Result with passive only TR cables looks ok, but distortion was higher than I was expecting. I moved outside as the bal2se results inside were MUCH worse than outside, I suspect noise pickup from HTPC and other electronics in the lounge room due to the breadboard.

So the result with the TR cables on the HP circuit. The crossover frequency is roughly 3rd order electrical at 270 hz, and the HP circuit has around 6db of attenuation as well.
attachment.php


Now the result with the balanced to SE converter and then TR for the input. pretty ugly!
attachment.php


Finally with the Bal2SE then HP crossover and then SE to balanced circuit.....
attachment.php


Oddly quite a bit more hum pickup (oddly because it is differential in and out) but the distortion is an order of magnitude lower! the hump in the noise floor is much reduced as well. I think I need to do a bit more experimenting with values in the converters. The Bal2SE currently has 2.2K resistors in it which may be too low.

This is only a test with one (probably quite unusual) piece of equipment, but it shows that loopback tests with different options alone are not enough to determine what will be best when measuring real world devices.

Unfortunately my BA3 boards have been gathering dust for quite a while. I have all of the fets for a build but that is as far as I got, so unfortunately I don't have a working BA3 to do tests on (and unlikely for quite some time).

edit: note I changed the sine frequency to 998 hz to move it away from the 100Hz harmonics, which I guess the breadboard is picking up from the PS in the crossover.

Tony.
 

Attachments

  • bal2se_HP_se2bal.png
    bal2se_HP_se2bal.png
    25.3 KB · Views: 440
  • bal2SE_through_hp_outside.png
    bal2SE_through_hp_outside.png
    26.2 KB · Views: 435
  • se_loopback_hp_outside.png
    se_loopback_hp_outside.png
    25.7 KB · Views: 437
Last edited:
Just another Moderator
Joined 2003
Paid Member
All in line mode. I did a test in instrument mode and it has much higher gain and is not appropriate for line level measurements (specifications show this as well).

I re-checked the documentation and it is a bit ambiguous. It does mention that inst mode is for instruments and it expects a TS cable in that mode (I use a TS cable all the time with my mic preamp without any issues). It doesn't say specifically that you can't use a TS cable in line mode. On the outputs it says you can use either TS or TRS cables.

These results could be specific to the gen 1 2i2, or even possibly to my unit, just something to be aware of if you don't seem to be getting the results you would expect :)

I did some more measurements with 6.8K and 15K resistors in the Balanced2SE circuit and that improved things. Note also that the SE to balanced effectively doubles the voltage hence why it is showing 6db higher than the other measurements. I put a 6.8K resistive divider before the buffer and it then gave the same levels as when inputting with the TR connection.

6.8K resistors below. The only thing that changed with 15K resistors was that the 50 hz spike went up, however I suspect that was me moving cables and not the resistors.
attachment.php


Next image is with 15K resistors instead of 6.8K and a 6.8K voltage divider on the input to the SE2Balanced circuit (removing the resistor to ground that was there, which I didn't actually check, but suspect was 220K).
attachment.php

The noise floor rize between 200 and 700 hz is all but gone.... Note that it was the 6.8K resistive divider that got rid of the noise floor bump, the change to 15K did not affect it (though the change from 2.2K to 6.8k did seem to have some effect there).

Tony.
 

Attachments

  • bal2se_6.8k_HP_se2bal.png
    bal2se_6.8k_HP_se2bal.png
    25.2 KB · Views: 608
  • bal2se_15k_HP_6.8kVD_se2bal.png
    bal2se_15k_HP_6.8kVD_se2bal.png
    25.3 KB · Views: 633
Last edited:
Interesting…...I need to play a lot with 2i2 and ARTA software. There is a lot of new stuff. When I get my TRS og RCA converters I will play with loop back first to see what I get on the screen.
Do you know if it is possible …..in the spectrum analyzer mode to view both channels at same time?
There is a mode in ARTA called "dual channel freq. response".......but.....we will se what is does :)
 
Just another Moderator
Joined 2003
Paid Member
Attached are snippits from the 2i2 gen1 and 2i2 gen2 manuals on inputs, outputs and specs.

This page has more info. Guide to Setting up your Studio Part 1 – Focusrite Audio Engineering

Specifically it says that outputs can be simply TS type but that inputs should be TRS with T connected to the tip of the output and R connected to the shield of the unbalanced output).

I originally made up TS cables as some article I found said that was what you should do, but it would seem the best solution for unbalanced to balanced cables would be to use a TRS jack for the balanced end. A two wire sheilded cable with all three connected at the balanced end, and just the two internal wires connected at the unbalanced end with shield floating at that end.

Tony.
 

Attachments

  • scarlet_2nd_gen_specs.png
    scarlet_2nd_gen_specs.png
    80.6 KB · Views: 79
  • scarlet_2nd_gen_outputs.png
    scarlet_2nd_gen_outputs.png
    83.4 KB · Views: 86
  • scarlet_2nd_gen_inputs.png
    scarlet_2nd_gen_inputs.png
    49.4 KB · Views: 64
  • scarlet_specs.png
    scarlet_specs.png
    80 KB · Views: 81
  • scarlet_outputs.png
    scarlet_outputs.png
    75.8 KB · Views: 93
  • scarlet_inputs.png
    scarlet_inputs.png
    62.5 KB · Views: 273
Member
Joined 2003
Paid Member
you can see few crappy tricks , over in Babelfish M25 , AKA M2 on steroids , AKA M2-XA25 bstrd

though , don't count on too much help from me , regarding settings of sound card ..... I simply am lacking in that sort of elaborative approach , resorting to that only when I'm desperate

:)

I saw that and commented for more info...is that some kind of DIY interface? Are you using ARTA or some other method? Trying to get my head around this. Looks like something that would help here. What's the hookup details mighty Zenmod? Or is this something totally different? Resulting screen shots look like what those here are getting.

Thanks,

Russellc
 
Member
Joined 2003
Paid Member

Just noted this is the same diy sound card probe as mentioned in the 6L6 thread I posted above. The "Visual Analyser" is way over my head, I assume the Techmind software is used in the same way others are using ARTA? Are there any pros/cons to the use of either for purposes of this thread?

Thanks for any use info you may have, much appreciated!

Russellc
 
Last edited:
Member
Joined 2003
Paid Member
Zenmod uses EMU 1212 as part of his setup, (from this thread: Babelfish M25 , AKA M2 on steroids , AKA M2-XA25 bstrd) looks like a card that goes in computer. Looked it up on musican's friend, looks like it is no longer available...this led me to EMU 1616, which, blast it...is also no longer available. If I can find a new version I will post about it...Altough I have no idea what this does. ZM said once at his PC, he would add a little clarity here.


Russellc
 
Last edited: