image description


Pres coming round..
  Forum Index » StudioLive General Discussion 
Author
Message
roblof
Presonoid
[Avatar]

Joined: 12/04/2011 21:33:08
Messages: 1297
Location: Sweden
Offline

MikeRivers wrote:Interesting. Did you open it up and find the 33079s?

The official version is that I've never opened my sl - It was someone else's unit. Yes, they are 33079's.


And how do you explain the interaction when you plug in a mic with a line source connected, and plug in a line source with a mic connected? It doesn't appear to act like a 1/4" jack contact is acting like a switch because the mic input doesn't go away entirely when you insert a plug into a Line In jack. And vice versa. I didn't look closely, but it wasn't obvious that there was a switch in the XLR. I've seen some like that, but there's an extra pin in the center that's spring-loaded. When the plug is inserted, it pushes that pin which activates a switch.

I had to double check since you got me confused for a moment here. I tested various cases...

The TRS cuts the XLR input. I measured crosstalk from the xlr into the audio path but it was lower than -80dB of the original signal.

Also, you noted the input impedance of the line as 10 k on each leg. Did you measure that? I did, and they were unequal, which surprised me.

I made a quick measurement of the trs to verify, but the two resistors going into the diff.opamp was marked 10k so it's a no-brainer. Actually, the diff.stage is based on R1=R2=Rg=Rf=10k.

For the xlr they use the standard 6.8-configuration for phantom power and can be verified at the xlr. The only thing that got me somewhat confused was an extra 100k shunted to gnd.

The Z=1k resistors are after some dc-removal capacitors. The circuitry was very easy to follow since ch23/24 are mounted on an extra pcb and therefor quite easy to 'decompile'.

Given your statements above, it seems like you may have reviewed an bad apple or that the sl16 is a somewhat different beast compared to the sl24.

PreSonus is pretty closed about their R&D, but that's typical of this level of technology. I don't expect to see a real schematic unless it's reverse-engineered. It's not hard to make a functional block diagram from the way something functions. I've done that myself for reviews where I think a discussion of signal flow is useful and the manufacturer didn't supply a block diagram, but a schematic is, of course, different, and can reveal how something works, not just what it does at the functional level.

I like reverse-engineering as relaxation/therapeutic challange - It jogs my memory .

Pretty much of what presonus has made is more or less well-known-tested implementations and doesn't deviate too much. This makes it much more easy to verify ones findings.

The only thing that I've not bothered with at this time (saves it for a rainy day ) is the XMAX-implementation. It is made of old-fashioned transistors and like - I.e. the old way of doing things

This message was edited 1 time. Last update was at 26/09/2011 20:11:41

Behringer x32 with Dante, x32-rack, s16 stagebox, p16 monitor system, StudioLive 24, FirePod 10, Alesis HD24, M-Audio FastTrack Pro, Studio One Pro v2, NI Komplete 5-8 Ultimate, Cubase 4, Ableton Live 8 Suite, Reaper 4, misc Line6 gear/software, True Temperament Frets on basses and guitars, Korg Kronos, Roland vk-7, Yamaha 02r.....
MikeRivers
Presonoid
[Avatar]

Joined: 28/07/2010 15:44:52
Messages: 2247
Offline

roblof wrote:
The official version is that I've never opened my sl - It was someone else's unit. Yes, they are 33079's.


Nothing wrong with those chips. I was just curious since either you needed to look under the hood or have access to some secret documentation.


I had to double check since you got me confused for a moment here. I tested various cases...

The TRS cuts the XLR input. I measured crosstalk from the xlr into the audio path but it was lower than -80dB of the original signal.


Thanks for checking. I wonder how they do that. There could be an extra switch contact on the jack that (probably through an analog switch chip) that switches in a different circuit connected to the Line In jack. Leakage at -80 dB could be attributed to ground current. My notes only are that the Mic input gain goes down to a totally unusable level with a line input connected.

But I wonder why they didn't do it the conventional way (for this level of product) - with an attenuator between the mic and line inputs. Not only is it the cheapest way, but it works. If there's really an independent line input buffer, it probably doesn't really offer any better performance than the mic amplifier. But without really getting into the inner workings, there's no way to evaluate that. It does have a marketing advantage, however, as evidenced by all the folks who think it's silly to attenuate and then amplify the signal. Frankly, it doesn't bother me (I've had enough experience to have that opinion. It only bothers me that I don't know for sure how it works, or why they made the choice they did when it's pretty rare in mixers like this.


I made a quick measurement of the trs to verify, but the two resistors going into the diff.opamp was marked 10k so it's a no-brainer. Actually, the diff.stage is based on R1=R2=Rg=Rf=10k.


That pretty much explains it then. Remember that the output impedance of a differential op amp is zero. What the source going into the inverting input sees is the input resistor plus the feedback resistor, in this case, pretty close to 20k. However, the feedback current through that resistor back to the non-inverting input serves to raise the apparent impedance that the source sees. The non-inverting input, however, doesn't have feedback and the source sees the series resistor plus the resistor to ground reduced by the (relatively high) impedance of the op amp input in parallel. This is a perfectly good design for a differential input if what you're after is the voltage difference between the two inputs. But when you're looking for common mode rejection, which is pretty important for a mic preamp (less so for a line input since you're dealing with pretty high levels) you have to consider the resistance in the complete circuit which includes the source (mic) as well as the input impedance to the op amp.

If the resistance around the two loops is different, there will be a different current induced in each from an external electromagnetic field. Now, if the current was the same regardless of the loop resistance, it would be clear that you'd have more voltage from the induced noise dropped across the non-inverting input than the inverting input, hence poor CMRR. Since the currents are different due to the different loop resistance, it's not that simple. In the best world, you'd have half the current induced in the "non-inverting" loop than in the inverting loop, it's being dropped across double the resistance so you'd end up with the same voltage at both inputs, the differential amp would do its thing, and the induced noise would be canceled. But current induced by magnetic fields don't cooperate and that's not the case.

There are designs (actually one, that's been used over and over for many years by different manufacturers) that present the same impedance to both legs. This is what Mackie uses.

As for my measurements, I have an NTI Minirator Pro that I just plug in and it reads the current in each leg, converts it to resistance since it knows the voltage, and I read it on the display. I thought this was kind of odd, and measured it a different way and came up with the same numbers.


Given your statements above, it seems like you may have reviewed an bad apple or that the sl16 is a somewhat different beast compared to the sl24.


That could be. I believe they had some different designers on board before the 24 got off the table, but I've heard different versions of that story. I don't think it was a bad apple, though, as I checked several inputs (thinking maybe the first one I checked had a problem) and they all measure the same. It's not out of the question that there was a change since June 2009, but Serial number PG19040428 definitely had a different impedance from each line input side to ground.

The only thing that I've not bothered with at this time (saves it for a rainy day is the XMAX-implementation. It is made of old-fashioned transistors and like - I.e. the old way of doing things


I suspect that it's very much like the classic circuit I mentioned above. Here are some links that will get you going:

Cohen's 1984 paper updated
THAT Corp integrated mic amplifier

Here's how a famous manufacturer does it (slightly obscured out of respect)




This message was edited 2 times. Last update was at 27/09/2011 02:56:47

Visit http://mikeriversaudio.wordpress.com for some useful audio info
"It's much easier to look for a magic solution than it is to adapt to reality." - Allan Sloan
[WWW]
roblof
Presonoid
[Avatar]

Joined: 12/04/2011 21:33:08
Messages: 1297
Location: Sweden
Offline

The only thing that I've not bothered with at this time (saves it for a rainy day is the XMAX-implementation. It is made of old-fashioned transistors and like - I.e. the old way of doing things


Your fault Mike - You got me started . Aside from some very minor things this is pretty much the xmax:

http://sound.westhost.com/project66.htm

CMRR on the xlr is 65dB according to Presonus tech specs. The line input hasn't cmrr listed.

If the cmrr is skewed then couldn't one measure the delta by connect a unbalanced source to the input pins, making sure that a proper gnd is used, and then switch the wires.

In a daw it should now be possible to see any dB-diff and then calculate cmrr, or am I far fetching here?

Thanks for checking. I wonder how they do that. There could be an extra switch contact on the jack that (probably through an analog switch chip) that switches in a different circuit connected to the Line In jack.

Found no such thing. Looks like the TRS-connector is using insert-switches on the actual connector to break xlr(+) and xlr(-) directly at the source.

But I wonder why they didn't do it the conventional way (for this level of product) - with an attenuator between the mic and line inputs. Not only is it the cheapest way, but it works. If there's really an independent line input buffer, it probably doesn't really offer any better performance than the mic amplifier.

I can only guess - The xmax is maybe not a linear transform of the signal, but rather adds a characteristic to the sound to enhance the sound from connected mics.

When using a line source you probably want as natural sound as possible to be able to utilize external preamps to their fullest and possibly presonus had that in mind when constructing the input stages.

But, hey. It's only a guess...

As for my measurements, I have an NTI Minirator Pro that I just plug in and it reads the current in each leg, converts it to resistance since it knows the voltage, and I read it on the display. I thought this was kind of odd, and measured it a different way and came up with the same numbers.

You should borrow a sl24 and verify if the old readings still stands.

Here's how a famous manufacturer does it (slightly obscured out of respect)

Slightly obscured? "The image you are requesting does not exist or is no longer availbale".

I wonder how it looks when you obscure to its fullest

Behringer x32 with Dante, x32-rack, s16 stagebox, p16 monitor system, StudioLive 24, FirePod 10, Alesis HD24, M-Audio FastTrack Pro, Studio One Pro v2, NI Komplete 5-8 Ultimate, Cubase 4, Ableton Live 8 Suite, Reaper 4, misc Line6 gear/software, True Temperament Frets on basses and guitars, Korg Kronos, Roland vk-7, Yamaha 02r.....
skip jones
Presonoid
[Avatar]

Joined: 28/07/2010 15:38:49
Messages: 1744
Location: Corpus Christi, Texas
Offline

You know, this is like accidently stumbling onto the German forum . Some of it makes sense to me, but the bulk is way over my head .

Gentlemen, a lively discussion.

I'm going to have to re-read this a few times to even get a grasp, but as they say "you made me think!".

Skip
http://support.presonus.com/home

Windows 7 X 32 (SP1) : AMD Phenom II X 4 945
ATI Radeon 5450 / 512 RAM
8GB RAM/1T SATA 7200 RPM

PreSonus StudioLive 16.4.2, PreSonus FireStudio Project, PreSonus ABVSL 44, PreSonus ABVSL 1818, PreSonus Faderport

Yamaha S-08
Studio partner - Jack Russell named "Skooter"
PreSonus Studio One Professional 2.6.2
Fishman Triple Play / Logidy pedal
MikeRivers
Presonoid
[Avatar]

Joined: 28/07/2010 15:44:52
Messages: 2247
Offline

roblof wrote:
Your fault Mike - You got me started . Aside from some very minor things this is pretty much the xmax:

http://sound.westhost.com/project66.htm


The classic circuit. Nothing wrong there, and as I measured, at the mic input, the impedance of the two sides is pretty close. Not close enough for Bill Whitlock (Jensen Transformers) though.


CMRR on the xlr is 65dB according to Presonus tech specs. The line input hasn't cmrr listed.


That's not too far off the 62 dB that I measured. My CMRR setup is a bit crude but repeatable. It involves a couple of 150 ohm resistors matched as close as I could with what was in my junk box. But with the same setup, I measured 71 dB for a Mackie Onyx. That's a pretty significant difference if you're in a noisy environment, not much if you aren't. But haven't you ever set up a band on a stage backed up to the window with the big neon BEER sign?


If the cmrr is skewed then couldn't one measure the delta by connect a unbalanced source to the input pins, making sure that a proper gnd is used, and then switch the wires.


I have a test cable with two 150? resistors. The resistors are tied together at one end. The other endss are connected to pins 2 and 3. A generator is connected between the junction of the two resistors and pin 1. The common mode voltage not rejected is what comes out of the preamp.

Looks like the TRS-connector is using insert-switches on the actual connector to break xlr(+) and xlr(-) directly at the source.

That's so simple it's cheating. I wonder why they found it necessary to disconnect the mic input? Maybe just so its noise wouldn't be added when using the line input. Or just to be sure that if you have a mic connected, whatever it was picking up wasn't inadvertently added to your line input.

The xmax is maybe not a linear transform of the signal, but rather adds a characteristic to the sound to enhance the sound from connected mics.


Good guess, but most everyone, including me, characterizes the mic preamps as pretty much transparent. Certainly not substantially colored, don't you think?


You should borrow a sl24 and verify if the old readings still stands.


Hey, I told Rick that if he'd send me a couple of mixers, I could verify a lot of this questionable stuff as an impartial bystander. He thought it wasn't a bad idea, but I'm still waiting.

I wonder what was wrong with that link. Try this one: http://i.imgur.com/7nVBC.jpg

This message was edited 1 time. Last update was at 27/09/2011 02:56:03

Visit http://mikeriversaudio.wordpress.com for some useful audio info
"It's much easier to look for a magic solution than it is to adapt to reality." - Allan Sloan
[WWW]
MikeRivers
Presonoid
[Avatar]

Joined: 28/07/2010 15:44:52
Messages: 2247
Offline

skip jones wrote:
I'm going to have to re-read this a few times to even get a grasp, but as they say "you made me think!".


I guess my work here is done.

But I think it's good to get into discussions like this now and then. Some of it, for lack of accurate documentation, is useless for sure, but some of the principles behind how this stuff is supposed to work is bound to rub off.
Visit http://mikeriversaudio.wordpress.com for some useful audio info
"It's much easier to look for a magic solution than it is to adapt to reality." - Allan Sloan
[WWW]
roblof
Presonoid
[Avatar]

Joined: 12/04/2011 21:33:08
Messages: 1297
Location: Sweden
Offline

Found no such thing. Looks like the TRS-connector is using insert-switches on the actual connector to break xlr(+) and xlr(-) directly at the source.

Sorry, I was wrong about this one I shouldn't do stuff like this when I'm tired

It is actually pretty evident that the xmax preamp is out of the loop when the trs is inserted. If you look at my drawing the selector switch must be where I join the two line/mic-stages, right before the insert.

If you push gain to maximum you can clearly hear the noice from the xmax. That goes away when the trs is inserted.

Also, if my first statement were true then phantom power should not exist when trs was inserted, which it does.


However, due to my efforts I found out one other thing. If I use maximum gain without anything connected I find the physical inputs to become fairly microphonic . With that amount of gain it probably makes sense...
Behringer x32 with Dante, x32-rack, s16 stagebox, p16 monitor system, StudioLive 24, FirePod 10, Alesis HD24, M-Audio FastTrack Pro, Studio One Pro v2, NI Komplete 5-8 Ultimate, Cubase 4, Ableton Live 8 Suite, Reaper 4, misc Line6 gear/software, True Temperament Frets on basses and guitars, Korg Kronos, Roland vk-7, Yamaha 02r.....
SwitchBack
Presonoid
[Avatar]

Joined: 17/02/2011 01:10:03
Messages: 1428
Offline

Hi guys,

Looking at the diagram in the westhost link I'd say it would make sense to let the line trs cut in after the discrete preamp, right before the C2/C3 caps. That way you would skip an amplifier stage (including its noise) for the line input. Looking at the SL's gain dial there's definitely something non-linear going on between mic in and line in (bottom of the scale the difference is 20dB, top of the scale the difference is 40dB), which is also suggesting that it's not simple padding. Just my 2cts
RDUB
Presonic
[Avatar]

Joined: 01/08/2010 00:34:15
Messages: 169
Offline

Guys, thanks for the effort and the info! I'm almost keeping up with the tech facts...

I've done some experimenting with the XMAX compared to a UAD LA-610. As expected, a little more warmth, but also a bit more air with the UAD. I'm not noticing any difference using the various available inputs.
HP dv7 (i7, 8gb RAM, 2X500gb HD, NVIDIA Geforce 1gb video) running Win 7 Pro 64 bit
Studio Live 16.4.2
Studio One Pro
Capture
tysonviolin
Presonic
[Avatar]

Joined: 13/02/2011 06:07:21
Messages: 343
Offline

The xmax pres sound to like sytek or other driven solid state. Pre
TROPO.BANDCAMP.COM
SL24
SL16
Apogee
Studio One Pro
Logic
Ableton
Mac Book Pro
UA
Vintech

[WWW]
 
Forum Index » StudioLive General Discussion
Go to: