Slew rate . . . .

[strangeandbouncy]

Dear PRR,

Thank you so much for your posting. Just to clarify a few things . . . I use this La4a daily as vocal compressor. Simple, I hardly ever adjust it. Easy set-up for both recording and mix. It is just "plug n' play". Many different artists visit our studio, and many different so-called "engineers". My particular combination of Neumann U48, Consant-impedance attenuator, Helios Mic-pre and La4a is about as universal as it gets. I can set it for visiting peeps the night before, and spend my day off at home! One size fits all. I can recreate every single vocalsession everytime, and with an awesome sound(for most vocalists, anyway). My Protools inputs are 882's giving -14dB equals 0dB ref +4db - ie 0dB - on my vu's!- please excuse me if i have transgressed the correct nomenclature, but i think you know what I meanses! I like to record as if going to a 2" from habit, so naturally I don't like to see the vocal going over @0dB on my monitor vu's on the SSL. Helioses are 0dB, NOT +4dB out., hence, I like to see not much more than -4dB on my vu's. this gives me @-10dB peaks in my Logic/ProTools, @ 4dB lower than I would normally leave with higher rated mic-pres. I prefer to leave @ 6dB in hand normally to allow for "divas" and screamers. I totally accept that exceeding 12v input will result in clipping( of the mic pre if nothing else,) but this must have been the case before substituting a transformer, since mic pre gain is the same, and there are no ic's(with their much reduced headroom over transformer) in the input anymore! Sowter is rated at +34dB . . . . My comparisons are at the same (lowish) levels as before, but this represents the(well, My!) real world of Digital.

my so called "mythical Threshold" IS what Jung was talking about, ie 0.5v/us for each Volt out! - still doesn't explain why API stuff sounds amazing even at rediculous output(will handle +30dB!) I know that actually the electronics are handling 6dB lower than that due to step-up 2503 transformer, but stll, shouldn't they require a slew-rate MUCH greater than 3v/us? I guess there is other black magic at play here! Or perhaps I just like the particular slew-rate distortion of the 2520, and not that of the 4136?

most interested to note that you don't think I'll find so much improvement in the output stage. This would make my life much easier! :green:

KInd regards,


ANdyP :green:

 

[dabo]

To strange and etc

I too find the LA-2 and LA-3 slow release to be very musical. My pair of LA-3's are my most prized compressors for that very reason. My LA-2 and every LA-2 in the studios I work in have that same glory. That is precisely why I was disappointed with the LA-4s until I fixed them.
The photoconductor cell in the stock LA-4 is nothing at all like the LA-2/LA-3 in release apart from being slow. It is driven by a red LED inside a plastic cup rather than the T4 arrangement in the earlier LA series compressors. The photoluminescent panel in the T4 has a fast on, slow dim-down curve that help lend the LA-2/LA-3 their wonderfully musical compression ballistic.
The red LED in the LA-4 has no luminescence latency. It is a very steep on/off light curve. That's why LEDs can be used in photocouplers for fairly high data transmission rates without edge rounding and signal loss (MIDI is a very slow example of this). The stroboscopic nature of the LED required the selection of a photoconductor with release retention to EMULATE the LA-2/LA-3 release. The technology of specific doping of the photoconductor element surface at the time was limited. The best version of the day is, to my ears, a compromise. Modern materials science has made available to us better choices. You can choose the release time constant.
As for the mention of FET based compressors:
If you examine the controls on the LA-4, you'll see that it was a deliberate attempt on the part of UREI to bring the compression ratio selection available on the 1176 (an FET based design) that was licensed to them from Teletronics into their optically driven series. They even went over to input drive/output level control ala the 1176 rather than the side chain drive/recovery gain topology in the LA-2/LA-3.
By careful selection of a modern photocoupler the LA-4 achieves the design intention the was not within reach when it was conceived. In the lower ratio settings my LA-4s have a compression ballistic like my LA-3s (with a different sonic signature) and in the higher ratios they behave like my 1176.
The presence of an input transformer does not remove the hole in the front panel bearing the legend "overload". I was merely offering an idea for what to do with it.
If to you the LA-4 is just like the LA-2 and LA-3 in compression ballistic and you think they sound the same then good-on-ya.
I know that I can really dig into a signal with the earlier UREI compressors and it still sounds smoothe. If I try to get anywhere near this with a stock LA-4, the release is too slow (or is not logarithmic) and it pumps very noticably. When I relax the compression to where the pumping goes away, the bad sonics of the 4136 make the negligible compression happening not worth the degradation of the signal.
If you have the time to try a mod, I think you'd be pleased.
If you love your LA-4 as it is, then forget I said anything.

David

 

[Kev]

[quote author="dabo"]If you have the time to try a mod, I think you'd be pleased.
If you love your LA-4 as it is, then forget I said anything.[/quote]There is no doubt that changing the opamps in the signal path is very worthwhile
.. or the above method of bypassing the input arrangement.

Changing the opamp in the side chain is a choice.
I chose to keep the sluggish 4136, as part of the basic sound of the LA4.

An LA4 is not a replacement for an LA2 or LA3.
Compared to an LA4 the 2 and 3 do sound similar BUT I still wouldn't say the 2 and 3 sound the same. They do share a compression envelope which is directly attributed to the T4B.

Topology doesn't spell the sound ... but does help to understand why and how they do sound the way they do.

 

[dabo]

I left the 4136s in the side chain/rectifier circuit for that very reason.

 

[Kev]

yes .. that was clear
my comments were for the wider audience

this can get confusing as all these things have been discussed here extensively and some of us make reference to that, such that we don't have to re-type it over and over

 

[strangeandbouncy]

Dabo,

i have actually removed the op-amps ic1a-c from ther circuit, replacing them with transformer, so nothing is connected to the led on the front . . .

:green:


ANdyP

 

[PRR]

> I hardly ever adjust it.... My particular combination of Neumann U48, Consant-impedance attenuator, Helios Mic-pre and La4a is about as universal as it gets. ...One size fits all. I can recreate every single vocal session everytime...

Ah, this is a "permanent" dedicated voice mike channel?

The Helios is internally single-ended, the LA4's gain-cell is single-ended: if you don't need to patch-in between, you could go right from one to the other, floating grounds as needed to suppress buzz in the unbalanced connection. No diff-amp or transformer(s). I'm a tin-ear but I've often heard the difference from getting unnecessary parts out of the chain.

The obvious place to feed the LA4 is the top of the Input control, but if you normally run it always the same level, you could even bypass the pot and use a fixed divider (or no divider if you run the Input pot all the way up). Pot-wiper distortion can be small, but is measurable.

> some mythical threshold of slew-rate

In this case, if the slew-rate of the LA4 input 4136 significantly exceeded the slew-rate of the Helios (possibly scaled by the level control), then the 4136 isn't the limit, the Helios is. Conversely, if the Helios passes signals that strain the 4136, ditching the 4136 could sound less strained. For any specific part, the "mythical threshold" is the point where it is no longer the system bottleneck.

I'm not too familiar with the Helios. If it is not compensated to run at low gains (old mike amps would not be run at low gain) then its slew-rate could be quite ample for its output voltage (and far higher than the old-IC unity-gain 4136). Or perhaps "just" ample for voice: even a spitty screech only has so much HF power, less than a complex mix of many instruments or some percussion sounds. Still, if the Helios is built with 741/4136-era transistors, but compensated for gain over 20, it should slew 5V/uS-15V/uS, which for its apparently ~5V peak output, comfortably beats Jung's criteria for full-range audio.

 

[strangeandbouncy]

Dear PRR,

I am not quite sure why I deserve so much wonderful attention from you, but please believe that I am very, very greatful indeed!

My Helioses all date from between 1971 and 1974 according to dates on both Beyer transformers and caps. I have a copy of the output driver schematic, but no scanner to scan it with, I am hoping to make it available to the "Bretheren" this week! I am certain that they don't work at low gain, and I will try to measure it "When I get around to it"!

I also figured that I could connect the two modules as single-ended, missing out the transformer, but it suffered from very serious HF roll-off - from@3kHz or so, which disappeared with any of the various transformers I tried.

Kindest regards,

ANdyP