John, The ADA4898 uses Nathan's spin on the multi-tanh input that we discussed at length. In essence distortion-wise it has the same performance as Barrie's simplest two area ratioed diff-pair circuit (which is public domain). The catch is that the low noise .9nV is preserved here but not in Barrie's circuit, unless the currents are unreasonable.
PMA, I followed up the question from, Stinius: 'can you be a bit more specific' with a specific request, and that I thought that Scott could best answer my question. Did I really do anything, but direct a question?
Everyone, I am trying very hard to be reasonably polite and get information exchanged here. Please give my the 'benefit of doubt' that I am contributing in a positive way, because I am doing the very best that I can.
By the way, PMA, could you please model the Nathan patented input stage for us? Scott can give you the patent number, if you need it. It would be most useful. Thanks in advance.
Everyone, I am trying very hard to be reasonably polite and get information exchanged here. Please give my the 'benefit of doubt' that I am contributing in a positive way, because I am doing the very best that I can.
By the way, PMA, could you please model the Nathan patented input stage for us? Scott can give you the patent number, if you need it. It would be most useful. Thanks in advance.
Last edited:
John
This has already been done by Edmond.
It is Patent US6963244
This is only one way to implement the multi-tanh.
stinius
For you, at this stage, no. For others, I may think about
What about posting your own understanding and concerns about the tanh principle, and I'll tell you if you are correct or wrong? That would be free
Have you read and understood Gilbert's paper published in IEEE? Here's a link for a free copy: http://ee.sharif.edu/~comcir/readings/mixers/mixer-gilbert.pdf
Read, try to understand the paper, then we may have an educated discussion about. Edmond can help you as well with this.
P.S. The link above is very slow. If you are having problems, I can send you a copy in private.
Last edited:
Thanks for your offer, but I took Barrie Gilbert's paper to Denver last week, and gave it to my associate to read. It was he that I was asking the questions for. How about that higher order infested transfer function? Good for audio? Looks like a multi-pole Chebyshev filter. That is my concern.
That’s cool, but not necessary.
I’m fine, let the show go on.
Would you like to continue on the multi-tanh discussion, or should it rest for a while?
Cheers
stinius
Last edited:
Well, the delay in my input has not brought forth any more info., but audio design marches on.
One factor that I might address about the Blowtorch, or at least repeat, is the importance of power supplies in both amps and preamps.
First, a little history of power supplies. In the very early days amps and radios were often driven by batteries, that were often recharged by gasoline engine-generators.
For example: "Another early Delco product was an electric power plant suitable for providing electricity for homes or farms not yet served by central power systems. The Delco power plants generally utilized one-cylinder internal combustion engines to drive an electric generator." Oct. 2009 'Proceedings of the IEEE' p.1737
Next, the electrolytic capacitor was developed to eliminate the batteries. Interestingly enough, the electrolytic caps might have been 2uf, BUT large chokes were also used in a π network (C-L-C) to get acceptable hum, and to remove the need for batteries. In the tube era, the C-L-C or C-R-C networks worked well enough, with low powered amplifiers. Finally, solid state diodes took over the job of tube rectifiers. This is both apparent in the Marantz and McIntosh amps and preamps, from the late '50's. In a way, this was a step backward, because the tube rectifiers actually behaved better in that they did not generate RFI due to the slowness of the switching transition of the solid state power diodes. Ultimately caps got much bigger, cheaper and more reliable, and chokes were removed to save weight and cost, in later designs, both tube and especially, solid state. By the late 1960's, computer grade electrolytic caps, of 5000uf or so were available, and that is what many of us stated with. Individual power diodes were replaced with solid thermal blocks of 4 diode bridges, and we thought everything was fine and dandy.
In those early years, virtually everything was analog, except main frame computers, that were remotely located, and not in every home. The power lines were much cleaner, I would presume, up to about 1975. Then fax machines, home computers, and a number of other sources, began to compete with our hi fi, for our attention and the power line got more garbage prone. Electric stoves, using SCR's, and light dimmers did a lot to the power line, as well. Up to 50MHz has been reported on the AC power line in easily measurable quantity in many cities, AND it depends on the time of day, as to how much is generated. Ever find that your hi fi sounds better after 10pm?
Well, we have found a number of solutions for these problems, and yet get good power supply regulation, and normally we attempt to revert to the equivalent of that early 'battery power supply, especially in preamps.
One factor that I might address about the Blowtorch, or at least repeat, is the importance of power supplies in both amps and preamps.
First, a little history of power supplies. In the very early days amps and radios were often driven by batteries, that were often recharged by gasoline engine-generators.
For example: "Another early Delco product was an electric power plant suitable for providing electricity for homes or farms not yet served by central power systems. The Delco power plants generally utilized one-cylinder internal combustion engines to drive an electric generator." Oct. 2009 'Proceedings of the IEEE' p.1737
Next, the electrolytic capacitor was developed to eliminate the batteries. Interestingly enough, the electrolytic caps might have been 2uf, BUT large chokes were also used in a π network (C-L-C) to get acceptable hum, and to remove the need for batteries. In the tube era, the C-L-C or C-R-C networks worked well enough, with low powered amplifiers. Finally, solid state diodes took over the job of tube rectifiers. This is both apparent in the Marantz and McIntosh amps and preamps, from the late '50's. In a way, this was a step backward, because the tube rectifiers actually behaved better in that they did not generate RFI due to the slowness of the switching transition of the solid state power diodes. Ultimately caps got much bigger, cheaper and more reliable, and chokes were removed to save weight and cost, in later designs, both tube and especially, solid state. By the late 1960's, computer grade electrolytic caps, of 5000uf or so were available, and that is what many of us stated with. Individual power diodes were replaced with solid thermal blocks of 4 diode bridges, and we thought everything was fine and dandy.
In those early years, virtually everything was analog, except main frame computers, that were remotely located, and not in every home. The power lines were much cleaner, I would presume, up to about 1975. Then fax machines, home computers, and a number of other sources, began to compete with our hi fi, for our attention and the power line got more garbage prone. Electric stoves, using SCR's, and light dimmers did a lot to the power line, as well. Up to 50MHz has been reported on the AC power line in easily measurable quantity in many cities, AND it depends on the time of day, as to how much is generated. Ever find that your hi fi sounds better after 10pm?
Well, we have found a number of solutions for these problems, and yet get good power supply regulation, and normally we attempt to revert to the equivalent of that early 'battery power supply, especially in preamps.
Actually, I was hoping that PMA would help me with Multi-tanh or perhaps Dimitri. I need an objective, (not negative or positive) person to analyze the ADA4898 by emulating its input stage, as close as possible, and look for 'glitches' in the transfer function. First, I had to find which multi-tanh input stage was used, then to carefully evaluate it for higher order distortion. Scott helped me with the first question, but no offers were put forth for the actual evaluation of the distortion. Therefore, I changed the subject. Did I miss anything?
I know I'm missing something obvious here but the question that keeps popping into my shallow mind is: why don't you build a bare bones circuit around the chip and listen to it (or have those whose ears you trust listen to it)?
ExpressPCB, a few hours layout time, $55. for the PCB's, a standardized power supply and layout to drop it into and you can compare virtually any topology/device you can imagine darn close to instantly (well maybe not instantly), measure them even. Prototyping technology has come a long way (cheap/quick/easy; heck, if I can do it anyone can!).
It really does distill down to what it sounds like and you can control the variables allot easier in hardware if you pay attention. Even if you don't understand all of the variables at least your comparing apples to apples.
The cool thing is the circuits are operating in hot-rod mode with no features to get in the sonic way, which makes it easy to sort out the real differences. I've been doing this for years and it works well.
- Status
- Not open for further replies.