Bas, I saw your attachment rather late. I have some remarks: The virtual earth of the op amp in question becomes a little bit inductive when frequency rises, so connecting a C across could give instabilities!
http://www.by-rutgers.nl/Pictures/AD826 IVconv Zin.JPG
As You could see on my web site I put a 2 ohm resistor in parallel with the LT1028. Then the input impedance only changes at the high end and never exceeds 2 ohm:
http://www.by-rutgers.nl/Pictures/AD826 IVconv Zin shunt2.JPG
without any instability problems.
Dear PA05SU,
I saw you had a huge topic about I-V conversions
I will read it all. It is a whole subject on it's own (I-V conversion). Lot to learn lot to read.. I never had problems with implementing the AD797 as advised by the datasheet. Because the AD797 is not used as unity gain, and compensated it remained stabile in my setup's. With kind regards,
Bas
in this months MJ -
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
Then download the NE5534 datasheet, study the internal layout of this chip and use your own imagination
Bas
Oh yes, sure. BUT this is one of the points in Self's book: NOBODY KNOWS THE INNER SCHEME! He also carefuly takes conclusions on the published diagram (WITHOUT VALUES) but he warns for the changes which could have been made in the last decenia/years. It is all secret. Today a solution works fine with a 15 years old sample from the junk box but if you replace the old 5532 with a just produced one (from a different supplier) heb je de poppen aan het dansen!
As a diyer, who only builds two or three critical circuits with them this is no problem, but in a (smal) production..... Never design on not specified data!
Glad to see that thoses venerable 5532/5534 are still
considered up to date..
One has to look at the internal schematic to see
that it was a breakthrough in high end audio.
3 stages of voltage amplification and still
stable at will with exceptionnal perfs and
sound clarity..
My favourite op amp to this day...
considered up to date..
One has to look at the internal schematic to see
that it was a breakthrough in high end audio.
3 stages of voltage amplification and still
stable at will with exceptionnal perfs and
sound clarity..
My favourite op amp to this day...
Last edited:
Dear PA0SU,
Well you don't need to know exact how the NE5534 is build inside, but it is to get the idea. Together with only a few other opamps this is the only one who gives you access to the the VAS stage. That is all you need to know basically. Then it is up to the designer what you going to do with this output.
With kind regards,
Bas
Mine too since a while. I love it, especially the exceptional low noise which is really a welcome on some applications.
With kind regards,
Bas
the only place I have seen the 5534 internal compensation discussed in any depth is Huijsing, "Operational Amplifiers, Theory & Design" 2001
it is interesting that it is still seen in current generation "Flagship" audio DAC I/V app circuits
do use the output "Class A bias" pull down trick - the difference is easily seen in I/V transimpedance apps with DAC I steps causing asymmetric feedthrough as the feedback C shorts the current to the output stage - the internal pull down is very much slower
it is interesting that it is still seen in current generation "Flagship" audio DAC I/V app circuits
do use the output "Class A bias" pull down trick - the difference is easily seen in I/V transimpedance apps with DAC I steps causing asymmetric feedthrough as the feedback C shorts the current to the output stage - the internal pull down is very much slower
Last edited:
Hi Bas - I'm relentlessly curious about I/V circuits. This one you've cited (looks to be from ADI by the drawing style, fonts) is shown with a very old DAC chip - the AD1862. I think I got samples of this part around 1990 - it looked good on paper but on the scope it showed a significant discontinuity around the zero crossing. Its weird that it would be coupled with the AD797 which is a much more modern opamp.
What's interesting is that for their more modern DACs, ADI don't show that 2nF capacitor even when the datasheet recommends the AD797 for I/V (AD1955 and AD1853 for example). So which DAC did you listen to with the 2nF cap? Was it an old fashioned multibit or a noise-shaped one?
Also I note in another post you say you trust Doug Self's ears as well as his engineering skills. Are you suggesting he listens to all of his designs? Having just read his latest (the 550 pages one) I've found no reason to believe that he really thinks circuits sound different. In the introduction he wastes quite a lot of time debunking the 'subjectivist' position but can't resist in so doing referencing Randi and having a dig against parapsychology at the same time.
Cliff, the Neve MD and its NE5534's are still in service. I can't remember the model (80's vintage) other than the 32 track plasma PPM meters, automated motorized faders and a large patch bay on the right. Any ideas on that ?
The studio does not have air conditioning, and the console ran quite warm. +/-19VDC rails and eventually the op-amps "aged", due to the heat. Their output offsets increased or even changed polarity which made the output electrolytics short and then the cermet potentiometers went noisy from the DC. So to overhaul a "strip", I change all the IC's, electrolytics and the potentiometers. A lot of work but worth it.
tricks with the NE5534
I've seen three tricks with pins 1, 5 and 8 on the NE5534:
-Take output from pin 5, bypassing the IC's output transistors (chip's weak point) and add your own discrete output stage.
-Bias pin 5, moving the IC into class A.
-Bias pins 1,8 to change the input diff amp operating point, great for optimizing noise by matching the op-amp's I/P to "different" source impedances. Analog Devices also mentioned this technique but I'd have to dig for it.
I've seen three tricks with pins 1, 5 and 8 on the NE5534:
-Take output from pin 5, bypassing the IC's output transistors (chip's weak point) and add your own discrete output stage.
-Bias pin 5, moving the IC into class A.
-Bias pins 1,8 to change the input diff amp operating point, great for optimizing noise by matching the op-amp's I/P to "different" source impedances. Analog Devices also mentioned this technique but I'd have to dig for it.
The output offsets changed polarity because the opamps aged? Do you have any circuit details to share, this could be interesting. On the surface sounds like poor design if the polarisation of an output electrolytic depended on the offset voltage of the NE5534. Its typically got very low offset but its worst case is nothing special.
<edit>
One of the things that makes the NE5534 such an interesting part to creative designers. The input stage can be bypassed completely using these two pins. So a JFET pair could be substituted, or some lower noise bipolars.
Last edited:
Regarding output from 5534 pin 5, it's connected to the current-sinking half of the output stage, and will sink as much current as pin 6, but cannot source current. Pin 6 is internally connected to pin 5 by a resistor and diode, which are for negative output current limiting, so pin 5 is not current limited. For bipolar output from pin 5 with a bipolar power supply, an external resistor or current source from V+ to pin 5 is required, and the output will always be in class A. 5534 is rated to drive 600 ohm loads to +10V so 15mA should be safe.
I think Class A bias by pull-up to V+ is better for the all-NPN output stages like 5534 and MC33078 if a large bias current (10-15mA) is acceptable, I've observed 40V/us with 5534 pin 5..
I think this is the best use of 5534 pin 5, eliminating thermal feedback between input and output stages and optimally driving the discrete output stage.
Last edited:
I think Walt Jung had an app using pin 1 & 8 as INPUTs from a distrete front end to bypass the 5534 input stage for even lower noise.
Probably around 1970.
See also http://www.diyaudio.com/forums/solid-state/29360-ne5534-misunderstood-10.html#post457320
jd
Last edited:
Hi Bas,
Thanks for your input. It is important since I am about to design a pre-amp.
Panson
I used a variant of this circuit with the AD1853 in a Benchmark DAC1 upgrade, and I tried it with a old TDA1543. The measurements and listening results where really promising. I remember I did changed some values and I simulated the circuit and optimized it, I need to look it up again.
Also I note in another post you say you trust Doug Self's ears as well as his engineering skills. Are you suggesting he listens to all of his designs? Having just read his latest (the 550 pages one) I've found no reason to believe that he really thinks circuits sound different. In the introduction he wastes quite a lot of time debunking the 'subjectivist' position but can't resist in so doing referencing Randi and having a dig against parapsychology at the same time.
Well let me nuance a bit. Personally I can't know if Mr. Self actually listen to his circuits or not. But My guess is that the man has a huge experience, and tried so many circuits already in the older days, that he can rely now in this experience. I owned a Soundcraft Ghost mixing console designed by him, and for the money it was an amazing device. It is a great fact that with so many routings, audio busses, filters he was able to create such a good sounding mixing console for the money. And he did "something" that a lot of others didn't and that made his mixing console so good for the money. (that "something" is explained on his website where he reveals several tricks for routings and audio busses).
Later on when I bought his book out of interest I build his blameless amplifier (just copy/paste), and I liked it a lot. In his 5th edition book he showed a very clever (but old school) topology with many paralleled NE5532 instrumentation circuits to reduce noise. I did those tricks already out of myself, but his book confirmed me I was on the good way. Also this circuit sounds awesome.
Last but not least, He have some very clever techniques for (pro audio) variable gain input amplifiers while retaining equal input impedances, and also those circuits sound really good and are clever solutions.
My conclusions. I would't be surprised if Mr. Self designs on "auto pilot" those days, but I do believe he investigated a significant of time to experiment and listen to circuits in the old days. His circuits are clever and do sound good.
With kind regards,
Bas
Having just read his latest (the 550 pages one) I've found no reason to believe that he really thinks circuits sound different. In the introduction he wastes quite a lot of time debunking the 'subjectivist' position but can't resist in so doing referencing Randi and having a dig against parapsychology at the same time.
In general I agree with Douglas Self, but in my opinion you cannot measure all parameters which are responsible for 'the sound'.
Henk ten Pierick (once again!) developped a method to rank amps on sound. He found out that de 5532 is not the winner! For LOW impedances (like in an I/V-converter) the very expensive LT1028 is better en in high-impedance applications the OPA(2)314 is the winner.
In D. Selfs last (2010!) book: Small Signal Audio Design, he sais that it should be impossible to judge better op-amps any more because the signals have been passed hundreds of 5534's before it has been put on CD.
Here he makes a huge mistake! Many of my test-signals with which I judge my progressions stem from vinyl, other CD's, my own recordings, radio!!! and many other 'imperfect' sources, somtimes havily compressed.
It is amasing and hardly beleavable how different the tracks sound when I'm talking about 'environment', depth, hight, definition, etc.
One example: the water, from a radio-recording from a lady who sat down a riverside, is realy WET on the one CD-player and not at all on another!
And for a good reason. You can try to build this circuit with FET opamps, the OPA2134's or the OPA627's. But for the nitpickers, the noise-floor is simply lower with the NE5534's. One can argue of this higher "noise-floor" would be a problem or not, but for a reference state of the art design, with such a high quality DAC, I can imagine TI wanted the highest resolution and lowest noise floor possible. The NE5534's with it's multiple nested feedback loops are hard to beat in that segment.
I heard the actual circuit with the PCM1794A, and it sounds nothing more then stunning, with pinpoint detail, 3d depth and great low end definition.
With kind regards,
Bas
Well, Douglas Self has a great record in the audio industry both on the hi-fi side, and even more so in the recording industry. I've just finished reading his 'Small Signal Audio Design' and in the front section he goes through the various op-amps available on the market and in his view there are only 3 winners: the NE553x, AD797 and the new LM4562. The NE553x is the outight winner on a cost performance basis. Regretably, he does not cover any of the current feedback op-amps, but I think noise would be his objection. PAOSU above highlights his comment about the hundreds of NE553x's that a typical recording goes through on its way from the microphne to the customers ears. Its highly unlikely that the signal is being irretreivably modified for the worse by that final NE553x. Trust your ears to be be able to judge for you what sounds good and what does not. But don't make the mistake of assuming that if it does sound worse, it must be this or that op-amp. Its just too complex and there are too many variables to be able to make that kind of leap without some serious investigations and double blind testing.
As an aside, I'd love to know who designed that first NE553x at Signetics. If he's around, maybe we could get him on the forum. Hats off to Scott Wurcer BTW, designer of the AD797 and good to have him on the forum.
As an aside, I'd love to know who designed that first NE553x at Signetics. If he's around, maybe we could get him on the forum. Hats off to Scott Wurcer BTW, designer of the AD797 and good to have him on the forum.