The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
Neve "cloudiness"/Transformer Questions (lots)
Old 29th December 2003
  #1
Lives for gear
 

Neve "cloudiness"/Transformer Questions (lots)

Is the notorious Neve low/low mid cloud from the harmonics generated by the transformer, the low end phase ****, high end phase shift, all, or something else?

Also, when there is a current that is sent to the output transformer, does it create that singled ended style distortion, low frequency distortion, or is it just simply to null out DC?

Is there much amplitude nonlinearities in transformers across the entire bandwidth, or is it primarily in the low frequencies?

Do transformers slew limit the signal?

Does hysteresis occur in greater amounts in louder signals simply because of amplitude, or is it some kind of inductance property?

Sorry. I've been researching this stuff and I'm having a hard time finding definate answers.
Old 29th December 2003
  #2
Lives for gear
 
Geoff_T's Avatar
 

Hi

What a lot of questions!

I'll give a clue that you can probably use to work out the remaining answers.

The BA283 in many ways resembles a classic tube class A transformer output. The collector load of the 2N3055 is the primary of the transformer and around 70mA of current passes through the primary. Most transformers do not like dc through their windings and the 1166 has gapped laminations to allow this situation.

Think tape technology for a moment. Decent tape recorders have ac bias and the audio is superimposed on top of this magnetic bias. Even cheap cassette recorders have bias when recording albeit I've seen dc bias on the cheapest models.

Now think back to the Neve output stage... the dc current passing through the primary is biasing the transformer and the audio rides on top of that.

This is all part of the magic....

Old 1st January 2004
  #3
Gear Addict
 

Jbuntz, GET A ****ING LIFE.
Old 1st January 2004
  #4
Lives for gear
 

Here's a few, everybody please join in...

**Is the notorious Neve low/low mid cloud from the harmonics generated by the transformer, the low end phase ****, high end phase shift, all, or something else?**

All of the above, plus more!! It has to do with a moderately low open loop gain amplifier trying to correct a fairly non-linear transformer's response problems, which are largely caused by it being a gapped transformer with a significant current screwing up the magnetic properties of the core.

**Also, when there is a current that is sent to the output transformer, does it create that singled ended style distortion, low frequency distortion, or is it just simply to null out DC?**

That single ended style of distortion is the result of the single ended amplifier driving the transformer. The low frequency distortion is mostly the result of core saturation, which is mostly screwed up (in mostly a sonically pleasing way) by the large dc primary current flowing through the transformer allowing the single transistor to modulate said current. There is no nulling of the current in a circuit of this type.

**Is there much amplitude nonlinearities in transformers across the entire bandwidth, or is it primarily in the low frequencies?**


The amplitude non-linearities are actually pretty minor in the passband, which is roughly 5hz to 50Khz in most of these circuits +/-1db when properly loaded and driven.


**Do transformers slew limit the signal?**

You betchor ass. It's a fundimental property of bandpass filters.
Does it matter? Not to me.

**Does hysteresis occur in greater amounts in louder signals simply because of amplitude, or is it some kind of inductance property?**

Hysteresis actually is a greater problem in smaller signals. By definition, hysteresis is the reluctance of an object to change state, like a teenager, if you're trying to get him to do something, he won't really move with a small bat, it really takes a Louisville Slugger to get some action.
Old 2nd January 2004
  #5
Lives for gear
 

Quote:
Originally posted by mcsnare
Jbuntz, GET A ****ING LIFE.
As you can see, Dan Kennedy likes this stuff and if it wasn't for people who were interested in it, we wouldn't have quality products like the Great River stuff. I'm developing some products in which this is useful information.
Old 9th January 2004
  #6
Moderator
 
Tim Farrant's Avatar
 

Jbuntz,

I think (and hope) mcsnare was just joking with you. Good on you for taking an interest in this stuff, I think Dan has answered your questions and I can add nothing further. I would interested in hearing about the products you are working on too.

Tim.
Old 9th January 2004
  #7
Lives for gear
 

Part of the sound is the loading on the input transformer. It is partially loaded with a resistor, and partially by the resistor in the feedback network. Since the amplifier isn't totally linear and also has a rolled-off frequency response (via all the capacitors peppered in there) the transformer's real loading is not exactly the same as a purely resistive load would be.
Old 10th January 2004
  #8
Lives for gear
 
Geoff_T's Avatar
 

Quote:
Originally posted by dale116dot7
Part of the sound is the loading on the input transformer. It is partially loaded with a resistor, and partially by the resistor in the feedback network. Since the amplifier isn't totally linear and also has a rolled-off frequency response (via all the capacitors peppered in there) the transformer's real loading is not exactly the same as a purely resistive load would be.
Hi

If you look at the circuit of the 1073's sensitivity switch you'll find the resistive loading on the secondary of the mic transformer is all over the place. The thing expects to see a 4800 ohms load but has a 12K resistor (added when a lot of other anti-click resistors were added) and then sees a variety of impedances depending on which step the switch is on.

As the secondary impedance reflecting through the transformer contributes to the input impedance it follows that the input impedance alters depending on the position of the gain switch and this, in all probability, affects how the microphone reacts with the module.

I personally believe this is a design flaw, like a few others lurking in the design of the module, but it all adds to the "sound" that one associates with this module.



Old 10th January 2004
  #9
Lives for gear
 

That switched resistance network thing had me stumped too until I figured out that the gain of the first preamp stage reflected back through the 15k feedback resistor (the resistor between S and U) made up the missing resistance, and when you compensate for that, the loading is almost always about 5k.

That's also why there's a 12k then 18k then no load when you go from -40 to -45 to -50 - the impedance seen looking into that 15k resistor goes down to about 5k when the gain of the amplifier is raised to 28 dB from 18 dB.

I needed to get my electrical engineering textbooks out to figure that one out, haven't looked at them in about ten years. Also, I got confused another way when I was figuring it out - the gain from U to S is roughly 10 dB lower than the gain from U to P. This circuit configuration is not found often in most of today's circuitry, at least I haven't seen it anywhere other than in a Neve and in my EE textbook, and maybe in some RF stuff.
Old 10th January 2004
  #10
Lives for gear
 
Geoff_T's Avatar
 

Quote:
Originally posted by dale116dot7
That's also why there's a 12k then 18k then no load when you go from -40 to -45 to -50 - the impedance seen looking into that 15k resistor goes down to about 5k when the gain of the amplifier is raised to 28 dB from 18 dB.

Hi

The only problem with that is that from -20 to -50 it feeds the second amp with the 18K feedback either via a short, an 18K shunt, a 12K shunt, and then four attenuators whose total impedance isn't too consistant.

The load on the transformer has to jiggle about a bit, methinks!

Actually, my issue 3 circuit (dated 17 May 74) doesn't even show the 12K on the transformer.

Back around 1972/3 there were two mods to all the Neve channel amps that made a lot of work for the tracers in the drawing office. The EQ was originally totally by-passed by the 2 pole EQ switch and then an improved version was incorporated on every module to leave the EQ's input permanently connected to the pre stage and just switch the outputs.

Also, there was the mod to add around three 12K resistors to the transformer, and a couple of locations on the gain switch (especially that link at -50) to reduce the clicks when the switch was rotated.

For anyone not already aware, the "off" position on the switch between -50 and -55 is to prevent a God awful click when the extra gain stage is switched in.


Old 11th January 2004
  #11
Lives for gear
 

Don't know why there was a problem on the 1073, because there ain't no click on an MP-2NV.

Maybe a function of make before break vs break before make?
Old 11th January 2004
  #12
Lives for gear
 

That may be the case - I built up a copy using a non-shorting switch and the audio drops out between each setting but there's no serious bangs.

That 18k resistor used to load the amplifier has an effective impedance of 10.4k when the amplifier is set to 18dB gain, 7.8k when set to 23dB, and 5.42k when set to 28dB. The 15k resistor has an impedance of 3.5k when the gain is set to 30dB.

I also forgot something, the gain of the amplifier to the feedback point S is 20 dB lower than the gain of the module, not 10dB.

-55 and higher = 3.5k
-50 = 5.4k
-45 = 5.4k (7.8 parallel with 18k)
-40 = 5.6k (10.4 parallel with 12k)
-35 = 5.0k
-30 = 4.8k
-25 = 4.8k
-20 = 4.8k
Old 11th January 2004
  #13
Lives for gear
 
Geoff_T's Avatar
 

Quote:
Originally posted by dale116dot7
That may be the case - I built up a copy using a non-shorting switch and the audio drops out between each setting but there's no serious bangs.

That 18k resistor used to load the amplifier has an effective impedance of 10.4k when the amplifier is set to 18dB gain, 7.8k when set to 23dB, and 5.42k when set to 28dB. The 15k resistor has an impedance of 3.5k when the gain is set to 30dB.

I also forgot something, the gain of the amplifier to the feedback point S is 20 dB lower than the gain of the module, not 10dB.

-55 and higher = 3.5k
-50 = 5.4k
-45 = 5.4k (7.8 parallel with 18k)
-40 = 5.6k (10.4 parallel with 12k)
-35 = 5.0k
-30 = 4.8k
-25 = 4.8k
-20 = 4.8k
Hi

So, taking your figures, the load on the secondary varies between 5600 ohms and 3500 ohms.

If 4800 reflects 1200 to the primary then
5600 = 1400 ohms
and 3500 = 875 ohms

Are we not agreeing that the input impedance varies with position of the gain switch? That was my original observation.

Old 12th January 2004
  #14
Lives for gear
 

Yes, I can't argue with the math nor with your observation. It does change but not as much as I initially thought it did before I calculated it out. I was expecting the impedance in the -40 to -50 to be way out to lunch but it only shifted a little bit.

The low gain mode seems like it's fairly even (1200 to 1400 ohm load on the mic). I wouldn't expect a large change in sound from that sort of an impedance change, but down to 875 seems like a lot. Even then, I haven't noticed a big change in sound when switching from -50 to -55 but what do other people think?
Old 12th January 2004
  #15
Lives for gear
 

Yep, that was my initial impression when I actually took the circuit apart. But as you've seen, and many years of dealing with stuff that was technically way out of whack but still worked proved to me, +/-6db doesn't ****ing matter, if it sounds right.

And that's the point.
Old 13th January 2004
  #16
Lives for gear
 
Geoff_T's Avatar
 

Quote:
Originally posted by Dan Kennedy
Yep, that was my initial impression when I actually took the circuit apart. But as you've seen, and many years of dealing with stuff that was technically way out of whack but still worked proved to me, +/-6db doesn't ****ing matter, if it sounds right.

And that's the point.
Hi

Indeed, it all adds to the characteristic charm/sound of the module.

Another issue is the topology of the output where the negative feedback is taken from a point very close to, and perilously decoupled from, the +ve power rail input.

Old 13th January 2004
  #17
Lives for gear
 
ulysses's Avatar
Quote:
Originally posted by dale116dot7
I built up a copy using a non-shorting switch and the audio drops out between each setting but there's no serious bangs.
Compare that to the Telefunken 672 where a non-shorting gain switch causes the amplifier to go to maximum gain between steps. Decidedly less pleasant!
Old 14th January 2004
  #18
Gear Maniac
 

Perfect timing with this thread. I've had the chance to borrow 1073 and 1066 modules recently after a long period of not using them. When tracking bass, drums and other instruments where the gain was set to -35 and under everything sounded great. Things got interesting when tracking B3 (with R121s) and acoustic guitars (with C12 and ELA M251s). Initially the gain was at -50. Not a sound that would knock your socks off. When switched to -55 the difference was _not_ subtle. The sound became _so_ much fuller. There was actually bottom end and a great deal more depth.

I thought that the modules were in need of some major maintainance; there were a lot of switch problems with them. So interesting that this behavior is part of the design (flaws). Thanks everyone for all this great info.
Steven



Quote:
Originally posted by dale116dot7
Yes, I can't argue with the math nor with your observation. It does change but not as much as I initially thought it did before I calculated it out. I was expecting the impedance in the -40 to -50 to be way out to lunch but it only shifted a little bit.

The low gain mode seems like it's fairly even (1200 to 1400 ohm load on the mic). I wouldn't expect a large change in sound from that sort of an impedance change, but down to 875 seems like a lot. Even then, I haven't noticed a big change in sound when switching from -50 to -55 but what do other people think?
Old 19th January 2004
  #19
Lives for gear
 
Empire Prod's Avatar
 

The whole "Neve cloud" thing has always baffled me as I am yet to actually hear it. 1073's have always sounded thick, clear and sit in the pocket very well. Even with very high track counts I have not encountered any "cloud" or "haze". Is this issue a bunch of spec sheet bull that doesn't translate to real life? I'm just curious because every one of the top recording facilities I work in seems to have racks full of Neve 1073's not Great Rivers. Why are they throwing money away on "cloudy/hazey" pre's?
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Similar Threads
Thread
Thread Starter / Forum
Replies
ROCK777 / So Much Gear, So Little Time
38
gwrr / Gearslutz Secondhand Gear Classifieds
1
DONNX / Gearslutz Secondhand Gear Classifieds
19
thethrillfactor / Gearslutz Secondhand Gear Classifieds
2

Forum Jump
Forum Jump