nuvistor said:
mikeks said:
If it's 'musical', then i don't want it....!
...amps must never be 'musical'....
nuvistor said:
I've known about the 1st stage trick for almost 25 years
------
Hopefully others can enjoy this knowledge.
If you have noninverting gain you really should try the 1st stage trick,
with cascoded matched dual JFETs if you can,
or matched NPN with emitter resistors.
IME almost as good as an OPA627 and much better than the 5534 for clarity,
and the reduced OL gain allows unity gain without added Ccomp.
this lets you add a discrete 1st LTP stage without added phase shift,
yes!Like I have shown before and in Attachment example.
Shows my own custum JFET opamp input LTP.
(The weaknesses of NE5534 are not many. But in order to minimize bias current & noise the LTP pair in NE5534 is biased as low as ~75 uA per transistor.
I think I made a measurement long ago, indicated this level.
Of course this have also some negative effects = Lower slewrate, lower Bandwidth.
Mikeks, Michael said
If it's 'musical', then i don't want it....!
...amps must never be 'musical'....
I tend to agree with you
Krell and other REAL High prformance Amplifiers have not often been blamed to be musical.
On the contrary, real high standard amplifiers do not add much (2nd harmonics) to Music.
This is why, in compare, they are not as 'rosy' and warm colored.
Lineup said:
If you go buy Michelangelo 'Mona-Lisa' for your own wall.
Then Nobody can stop you from put on green / yellow sunglasses
just because you fancy Mona-Lisa better in that color.
I is up to you
Me, I belong to those, that would like to enjoy (or not enjoy ....) Mona-Lisa's smile in a 'natural' state & color.
Even if a painting is not = reality, I would like to experience her
through the eyes of this unique man.
Michelangelo, one of the Greatest artists of The Millenium #2
Attachments
Hi lineup -
Voltage between 5534 pin 7 and 8, and pin 7 and 1, is about 2.5V when the part is used normally. 2.5 to 3.5V is OK with a custom LTP such as yours. The 5534 chip contains a 13k load resistor between pin 7 and 8, and pin 7 and 1. With your component values, expect about 1.13V, since the drain current is 1mA and the effective drain resistance is 1k+235 ohms in parallel with the internal load resistor. I suggest an external load resistor value of 2.5k per side and a 1k trimpot, this should give an effective drain resistance of about 2.5k and 2.5V across the load. The LSK389A transconductance of 10mS at 1mA is low enough for unity gain stability without an external Ccomp.
Voltage between 5534 pin 7 and 8, and pin 7 and 1, is about 2.5V when the part is used normally. 2.5 to 3.5V is OK with a custom LTP such as yours. The 5534 chip contains a 13k load resistor between pin 7 and 8, and pin 7 and 1. With your component values, expect about 1.13V, since the drain current is 1mA and the effective drain resistance is 1k+235 ohms in parallel with the internal load resistor. I suggest an external load resistor value of 2.5k per side and a 1k trimpot, this should give an effective drain resistance of about 2.5k and 2.5V across the load. The LSK389A transconductance of 10mS at 1mA is low enough for unity gain stability without an external Ccomp.
nuvistor said:
When I measured my NE5534, there was less than 1 Volt across 11 kohm (nominal 12k).
This indicates something like 300-400 mV across second stage emitter resistors.
Which makes sense to me.
I do not know if different manufacturers bias the input stage differently.
I would not think so.
Of course, looking at NE5534 topology, there is not much
that would stop us from increasing current in SECOND Stage, too.
A bit of experimenting would give a good answer
lineup said:
Of course, looking at NE5534 topology, there is not much
that would stop us from increasing current in SECOND Stage, too.
A bit of experimenting would give a good answer
Increasing the drain resistor voltage drop too much will affect positive limit of input common mode range. I have tried as much as 5V with +/-15V supply without problems, using Signetics parts from 1980s, but then maximum positive input will be less than 10V. It's possible that different or newer parts have different voltage drop, experimentation as you suggest is probably best.
nuvistor said:
I had to measure again.
My NE5534 are Texas NE5534P.
You were right nuvistor.
I measured drop across input collector resistors.
I did this only at one examplar (the one I could find) and at
A) +- 5 Volt (10Volt). B) +-15 Volt (30Volt).
Result:
I have changed in my schematic. See attachment
LSK389 (2SK170 can of course be used)
input JFET pair runs at 1 mA each.
2k2 + 235 Ohm = 2k435
Parallell 2k435 // 12K, gives collector resistor: 2kOhm.
1mA x 2kOhm = 2.00 Volt
Which corresponds to my measuring.
regards
LineupAttachments
Just as a matter of historical interest, does anybody know who first released the NE5xxx - was it Signetics? Who was the designer? Maybe someone on the forum knows. This op-amp has been around for 30 years and still going strong - quite a feat and clearly tells th e story of a product ahead of its time and designed right for its intended application.
Keep on trucking old timer!
Keep on trucking old timer!
The LSK389 fets give lower noise than the internal bipolars in this case.
Usually, you hav e to look at your source source impedance to make the call. With JFET's fed from high source resistances, there is very little [noise current x Input impedance contribution], so you are left with jus t th e equivalanet input noise voltage, which at 0.9nV/rtHz is very low anyway. If you are using low source impedances (say <100 Ohm) the noise current on bipolars is not really a factor.
In the case of this proposal from Lineup, looks like the LSK389's will offer a lower noise floor across the board compared to the internal input devices. You can DC couple your input source as well with the FET's - nice touch.
Usually, you hav e to look at your source source impedance to make the call. With JFET's fed from high source resistances, there is very little [noise current x Input impedance contribution], so you are left with jus t th e equivalanet input noise voltage, which at 0.9nV/rtHz is very low anyway. If you are using low source impedances (say <100 Ohm) the noise current on bipolars is not really a factor.
In the case of this proposal from Lineup, looks like the LSK389's will offer a lower noise floor across the board compared to the internal input devices. You can DC couple your input source as well with the FET's - nice touch.
Bonsai said:
* At medium source resistance( say 15-30 kOhm ), the noise will be at same level.
* At lower ( say <=10 kOhm ) bipolar can have noise advantage. At least compared to more normal noise JFET.
* Higher input impedances (like 47 kOhm and further to MOhms) will benefit from JFET input.
As we have almost no bias current,
even at higher gains, we will not have any DC-offset problems
caused by resistor choices. (not compared with bipolars)
I like JFET for this thing
Not much headache cancel DC-offset, caused by bias current + High Gain.
--------------------------------------------
Now, if you have less good matched JFET pair (V-gs),
you will have to un-balance currents in input.
My Question
BJT needs good current balance in input .
( D. Self show a difference of only 2% increasing Distortion by a level ).
Anybody knows how sensitive JFETs are to such un-balanced currents?
.. my guess is, that they are less sensitive ... but we should strive for equal current ..
I first came across the fact that the NE5534 input stage could be by passed , in a thread on this forum , a few years ago. Unfortunately I couldn't find the folder where I saved that info. So I started experimenting on my own.Later when I asked for more info again on the forum, I hardly got any response.
It's also taken me a long time to get around to actually building my version of the modified input stage. I hope I get enough time to test it. I am expecting it to be worth all the trouble that I've taken. My input uses two SK170's as I couldn't get hold of the dual FET easily.
I don't care if there are better ( more expensive ) solutions . As long as this one sounds very good, it would have been worth it.
Great to read all the posts on this thread. It's heartening to know that there are several people out there who don't trash the 5534 . I find it great in many applications. And I do have exotic opamps to compare it with.
Cheers.
It's also taken me a long time to get around to actually building my version of the modified input stage. I hope I get enough time to test it. I am expecting it to be worth all the trouble that I've taken. My input uses two SK170's as I couldn't get hold of the dual FET easily.
I don't care if there are better ( more expensive ) solutions . As long as this one sounds very good, it would have been worth it.
Great to read all the posts on this thread. It's heartening to know that there are several people out there who don't trash the 5534 . I find it great in many applications. And I do have exotic opamps to compare it with.
Cheers.
Bonsai said:
Rupert Neve was the mastermind behind the 5534 and Signetics were to manufacture it. His consoles mid 80-90's had only them in and discretes were added around this ic to make it a solid balanced power stage for most studio applications. The work he did for the audio industry earned him the Grammy award. Check out his website.
richj (ex Neveite)
Here a link with 1st hand references which seem to back that:
http://recforums.prosoundweb.com/index.php/m/0/13715/64/0/
Posts #193227 ff., at about the middle of the page.
There seems to be some confusion about the actual manufacturer, three brand names "Signetics", "Philips" and "Mullard" are thrown in, which in the end may (or may not) have been linked to one single actual IC fab, at that time.
- Klaus
http://recforums.prosoundweb.com/index.php/m/0/13715/64/0/
Posts #193227 ff., at about the middle of the page.
There seems to be some confusion about the actual manufacturer, three brand names "Signetics", "Philips" and "Mullard" are thrown in, which in the end may (or may not) have been linked to one single actual IC fab, at that time.
- Klaus
richwalters said:
KSTR said:
Rupert Neve.
I did not know he designed op-amps, like NE5534.
Thanks for info.
But I knew about Neve Mic Amplifier / Mixer modules, as I have studied them some years ago.
A search for: Neve amplifier & similar gives great many hits.
Here, for example, is a collection of old Neve stuff & some datasheets:
Dan Alexander Audio - Neve Information, Module Layouts / Specs
/Lineup
----------------------
edit:
Websource:
http://amstudios.net/amstudios/htdoc/Pages/Studio_A_CONSOLE_tech.html
Rupert Neve - 80 Years in the Making
An externally hosted image should be here but it was not working when we last tested it.
The 5534 was the best op amp for audio for its time, which was early eighties onward, perhaps even late seventies. I used it all over the place in my THD analyzer. Performance-wise it was unbeatable and is still very good. It was the op amp de jur for studio consoles back then.
There are some, however, who later claimed that its sound was not as good as hoped. This is the issue of difference between subjective and objective measurements that haunts audio. When properly applied, there is nothing in measurements of it that I have been able to find that would seem to compromise its sound. Nevertheless, that is what some people say based on listening tests.
The 5534 did not employ a full complementary BJT process; i.e., it did not have decent PNP devices. As such, like most op amps of the time, it was at a disadvantage. For one, the output stage cannot be made symmetrical without complementary PNP devices. The 5534 can be improved somewhat by pulling down its output with a resistor or current source, forcing the output stage into class A emitter follower operation.
Modern op amps have surpassed the 5534, largely because they have excellent vertical complementary PNP devices. The very best ones are dielectrically isolated.
The 5534 was first marketed by Signetics.
They are cheap and are still a very good choice for audio instrumentation.
Cheers,
Bob
I read these comments, most are test in simple circuits, I believe the majority chooses the opamp Burr Brown because it has low DC offset, low TDH and output impedance.
I chose the opa 132 for a pre-amp.
NE 5534/2 is necessary DC servo it's hard to find in projects preamp, the vast most decouples the DC with a capacitor that is not the best option.