THAT 2150 - 2181 sonic differences

I was discussing a channel strip project design with a fairly respected set of "golden ears" who suggested it'd be cool to be able to select between the two.  Unfortunately, he's not paying me to build him one of these things I'm working on, otherwise I'd say "sure, that'll cost extra."  ;D    But pressed further, he was incredibly oblique about the difference -- more of a "I liked the old ones better."

Has anyone A/B'ed those two VCAs in the same device?  What subjective difference do you perceive, and is there anything about the 2150 that one might construe as "mojo" lost in the newer chips?

-Bob

The 2150 and 2180 have the same schematic diagram, except for one resistor, which is now already built-in. The fab has changed (2150 was subcontacted, 2180 is their own fab); as a result, the 2180 has a little less distortion, a little less noise, and a little more high-current linearity. The differences are very small, but may be audible in some circumstances (typically splitting hairs, not in real world situation).
I made the switch from 2150 to 2180 about 5 years ago; at the time I evaluated the results, and since specs were better and the subjective difference was so thin, I never cared to advertise the fact to our customers. Nobody ever noticed!
I suggest you blind-test your Mr golden ears; I'd be surprised if he could tell which is what.

abbey road d enfer said:

I suggest you blind-test your Mr golden ears; I'd be surprised if he could tell which is what.

Click to expand...

Heh heh.   Now there's a valid reason to go to the trouble of putting them both in one box with a 12-position switch to randomly pick one or the other!

You mention "high current linearity"  what do you mean by high current?  Hot output level into low impedance?  If so, how hot and how low?  I wonder if that non-linearity is a 2nd-harmonic and that's what he's hearing as being more sonorous.  - (though one would think he would have mentioned that in his description.)

I can't address the subjective differences, but objectively the newer parts are fabricated on a better semiconductor process so the primary difference will be missing errors that were present in the earlier IC parts. I suspect there are even greater differences looking back to the first generation all discrete transistor VCAs.

Gary Hebert of THAT corp wrote a 30+ page technical paper for the AES Journal "An Improved Monolithic Voltage-Controlled Amplifier" in  1995. The paper goes into pretty good detail on the improved part, with lots of performance graphs.  I doubt it will satisfy the hunt for missing magic since it doesn't do side by side comparisons.

One difference with the newer process is somewhat closer GBW/speed/beta/etc between the NPNs and PNPs that alternately handle the positive and negative half swings so it seems logical less distortion from that one mechanism. Of course there is more going on than just that one and other subtle differences that could interact with specific implementations.

Of course they improved upon other known shortcomings. I can't imagine the part being worse in any way.

I would expect that envelope modulations due to gain control ripple to be orders of magnitude more audible than basic VCA performance differences.  But of course the customer is always right...  (even when he's wrong). 

JR

I recently swapped dbx2151As for THAT2180LBs in the dynamics cards on my amek BCII mixer.  I did a very brief A/B, and it seemed to me that the 2180 was cleaner and clearer, with a better spatial sense and perhaps less distortion on sibilant vocal consonants.  Because I mainly record jazz this appealed to me, and I stuck with the 2180.  The 2151s did sound slightly warmer but less detailed, and I can see that these qualities might be appreciated in some situations.

I should add as a caveat that my a/b was brief, and my impressions confirmed what I was expecting to hear, so perhaps I am not a truly objective witness.

Josh

stickjam said:

abbey road d enfer said:

I suggest you blind-test your Mr golden ears; I'd be surprised if he could tell which is what.

Click to expand...

Heh heh.   Now there's a valid reason to go to the trouble of putting them both in one box with a 12-position switch to randomly pick one or the other!

You mention "high current linearity"  what do you mean by high current?  Hot output level into low impedance?  If so, how hot and how low?  I wonder if that non-linearity is a 2nd-harmonic and that's what he's hearing as being more sonorous.  - (though one would think he would have mentioned that in his description.)

Click to expand...

I'm talking of only about 1mA peak current. This current is sum of Iin + Iout. This is related to the core current, which is fixed by the resistor that goes to -V. Usually they recomment 3.9k for -15V, so ca. 4mA. Increasing the core current improves somewhat the peak current capability, but also increases noise and cannot be made significantly larger, due to thermal problems. The 2150 was capable of ca. 1mA with 4mA core current, the 2180 is close to 1.5mA. As JR mentioned, the improved pairing of N's and P's results in lower 2nd order distortion, which may be the reason for the subjectively less euphonic perception... Following JR's line "envelope modulations due to gain control ripple to be orders of magnitude more audible than basic VCA performance", you may deliberately introduce some asymetry in the control voltage, this is guaranteed to produce much more significant doses of 2nd harmonic

Hi,


  in a G-SSL, i have 2151, 2181, 2001 and 202c's. They all "sound" ever so slightly different. Action is identical, I'd say. Out of 2151 and 2181, the 2151 is more "clacky" and punchy, and the 2181 is subjectively cleaner. Both are great in different applications, 2151 for drum sub-bus, 2181 for mix-bus. If you want to hear vca doing it's business, I'd go for the 2151. If you want less artifact, use the 2181.

  I tried swapping 2181 for 2151 in my Sony MXP2000 faders. I preferred the original 2151's, but then I am rather used to the sound.


    Just my 2p's worth


  ANdyP

abbey road d enfer said:

you may deliberately introduce some asymetry in the control voltage, this is guaranteed to produce much more significant doses of 2nd harmonic

Click to expand...

I am not sure I follow your suggestion. Are you talking about the symmetry of the control voltage driving the VCA, or intentionally adding an offset at the symmetry trim port?

The performance plots provided in the AES paper were generated driving only one of the two control voltage ports so in my understanding an asymmetrical control voltage. While they only characterized THD with plots for + and - 15 dB gain, I don't expect significant THD from this, since they surely presented their device in it's best light. Note: their plots were THD+N and N dominated for several of the wideband measurements over most of the audio band. Not like the old (VCA) days.

My understanding is that using both control ports is most useful when commanding large amounts of gain or attenuation for better DC operating points of the internal VCA devices in the context of hundreds of mV total CV. Of course nothing is free so driving both CV ports adds another potential noise/distortion source. Source impedance and noise at these control ports have consequences.

Noise at the CV ports introduces a modulation noise that will show up on THD measurements, but will not be very audible in use. Source impedance at these control ports introduces another distortion mechanism (due to mismatch between beta of NPNs and PNPs). You also need to keep source impedance at the control port low to prevent instability. You can shunt these ports with caps to ground to keep the impedance low at HF. I would suggest keeping them low at all frequencies, but I prefer to keep my audio paths as clean and linear as possible.


JR

I haven't made myself clear AT ALL  :-\
I was referring to the symetry adjustment on the 2252 RMS detector. If you deliberately mis-adjust it, its output will have some residual 2nd harmonic content, which, applied to the control port of the VCA, will introduce some nice euphonic distortion.

abbey road d enfer said:

I haven't made myself clear AT ALL  :-\
I was referring to the symetry adjustment on the 2252 RMS detector. If you deliberately mis-adjust it, its output will have some residual 2nd harmonic content, which, applied to the control port of the VCA, will introduce some nice euphonic distortion.

Click to expand...

There might be the difference - maybe his comparisons were between a misadjusted 2150 versus the 2180 which comes pretrimmed--presumably at the lowest distortion sweet spot--rather than the trimmable 2181, which I was asking about in this thread.  Might actually be a nice feature to bring out to the front panel in the form of a center-detent pot to permit a certain degree of adjustment/misadjustment at will.  Label it "Mojo" or with a -/0/+ scale.  ;D

I think I need to order some 2181s and do a bit of breadboarding.  Are 2150's obtainable anymore for comparison?

-Bob