Hi to check the THD i am using LTSpice and the models available on the basic configuration I am still struggling with adding models ... i am a little retarded
Anyway i am attaching the original buffer present in the old vintage Kenwood 700c
My curiosity was originated by the very positive opinions on its line stage sound ... very clean and detailed
My principle is that an amp chain that sounds good must be formed by stages that sound good by itself
I am usually look at those stages following the volume pot and called flat amp or buffer amp or output amp to understand a little about the schematic
However i have been convinced that when really great performance are the goal nothing but better opamps
I see some very high end manufacturers switching from discrete circuitations to opamps these days
And i am talking about price of 140 kUSD
Attachments
You could make a very low noise power supply
If you mean the self noise of the circuit should be minimal
I do not have a clue about its PSRR but it could be decent
Last edited:
In simulations the power supplies are noiseless, if I correctly assume? Hence by definition it is about 'self-noise' as you call it.
My answers on noise are off topic by definition
I mean that I did not learn how to make reliable thermal noise simulations or calculus except for one resistor or maybe for a resistive voltage divider.
I do not even know if the source noise of a coil is the R_dc or its impedance Z(f), or its ESR(f). I do not have a model for ESR for coil, except its R_dc, skin depth effect, proximity effect, self and surrounding Eddy current losses, etc.
If for coils it is R_dc+effects(f) then how this works for Capacitors?
If it is the Z(f), then why reactive energy would give any noise? It means a photon has noise (hence losses) which is not true.
If it is ESR alone, then why Rg of a coil is its impedance?
Then the propagation and amplification of noise (any kind) in active circuits: for solid state physics the small signal BJT amplification models were taught properly, also all noise mechanisms known, but maybe I skipped - somehow, by mistake, if any - the classes or seminars on applied electronics: for the noise models of BJTs, their circuits, and actually for anything but pure R's.
In so far, self learning did not worked well for me in this topic.
I have a lot to make up. For the moment unfortunately I must rely on external help.
Thanks a lot again So if i understand rightly you achieve regulation and noise suppression only with passive parts ? Very interesting
Nevertheless i would not excluded the use of a simulation SW like LTSpice i am trying to use
It could be very helpful to design also passive filters i understand I am very bad at math and the sim can help me a lor with the very heavy calculations needed for the selection of component
Like calculating the noise rejection vs Hx of a LC filter and see how it changes changing L and C values
I have fallen in love with LTSpice I cannot imagine how the work of audio designers was before simulation SWs became available The huge amount of calculations they had to do
I think that it can help also the beginners to understand how things work when changing working conditions
I am testing some circuits taken from the service manuals of very expensive equipment using discretes Even if i have a strong feeling that i will end with a buffer made with opamps ... i see them used everywhere these days There must be a reason
Nevertheless i would not excluded the use of a simulation SW like LTSpice i am trying to use
It could be very helpful to design also passive filters i understand I am very bad at math and the sim can help me a lor with the very heavy calculations needed for the selection of component
Like calculating the noise rejection vs Hx of a LC filter and see how it changes changing L and C values
I have fallen in love with LTSpice I cannot imagine how the work of audio designers was before simulation SWs became available The huge amount of calculations they had to do
I think that it can help also the beginners to understand how things work when changing working conditions
I am testing some circuits taken from the service manuals of very expensive equipment using discretes Even if i have a strong feeling that i will end with a buffer made with opamps ... i see them used everywhere these days There must be a reason
Sorry my bad
I am very poor at electronics
To be honest I haven't understood well your point
Maybe your point Is that simulation SW are imperfect?
I think anyone can agree with that
It's just a good tool for analysis of circuits
A very good tool
But the last word can only come from measurements
I am very poor at electronics
To be honest I haven't understood well your point
Maybe your point Is that simulation SW are imperfect?
I think anyone can agree with that
It's just a good tool for analysis of circuits
A very good tool
But the last word can only come from measurements
Don't bother. I am known to be often not enough clear.
I kindly asked for noise optimizations help from gurus, who know how to do this in the proper way: analytically + experience.
Why, because blind optimizations using only the Spice tool are a) not really optimizations, and b) not trustworthy enough in my hands
I kindly asked for noise optimizations help from gurus, who know how to do this in the proper way: analytically + experience.
Why, because blind optimizations using only the Spice tool are a) not really optimizations, and b) not trustworthy enough in my hands
I see now clearly your point Of course you are right At very least LTSpice is very useful to see what will never work
I have begun simulating commercial units that should work I see that very low THD+noise are very difficult to achieve even at simulation
I cannot imagine how hard is to get that in reality I think that LTSpice fixes the max performance obtainable in optimal conditions (perfect power supply)
I am trying to understand a little about theory using it I am afraid i have to study a little theory first
I have begun simulating commercial units that should work I see that very low THD+noise are very difficult to achieve even at simulation
I cannot imagine how hard is to get that in reality I think that LTSpice fixes the max performance obtainable in optimal conditions (perfect power supply)
I am trying to understand a little about theory using it I am afraid i have to study a little theory first
Hi and good morning This is quite worrying because the sim runs smoothly
If you will find the time could you kindly sim with TinaTi the attached circuit and see what THD comes out ?
just to run a check
With R1=22K load Tina gives this result: see below. It is slightly worse THD (0.003%) than the LM317 optimized by LV and Marcel, but let's see about the noise.
With 10KOhm loaded lines the THD will increase to 0.01% and for Headphones 600Ohms the THD will further increase to 1.8%, according to Tina. Maybe it is not optimized...
In contrast, the LM 317 performance for THD is stable also with strong loads.
Still, I will try to build it to verify and let you know.
With 10KOhm loaded lines the THD will increase to 0.01% and for Headphones 600Ohms the THD will further increase to 1.8%, according to Tina. Maybe it is not optimized...
In contrast, the LM 317 performance for THD is stable also with strong loads.
Still, I will try to build it to verify and let you know.
check the SMs for Revox A77&B77: Input amp and Rec amp. The manuals are free for download on internet and easy to find.
They also measure excellent (for a tape recorder) or pretty good (for other uses).
This is the Tina result (below): same H2 as above, but H3 is decreased slightly to just below 100dB. Beware of my changes to 'optimize' and make it work for 21V.
For 70V and Biased accordingly, the THD will be significantly better. Not sure about the noise.
For LM317 (using TI proprietary model) the Tina predicted almost identical as measured results.
I would be happy to spend some money for a professional version of Tina, but I would probably not do this for LTSpice. No offense to anyone.
Thank you very much again for your kind and valuable advice
I missed that you want to drive headphones Sorry my mistake
Instead I am looking for something driving only a power amp
The lowest Zin i have has a 5kohm/ ?pF input I have read to add 1nF of capacitance in parallel to the resistance I have to do that
Again i am sorry I missed your requirements
I am a beginner and i think that is wise to start from a line stage Then if i get something we say appetite comes with eating
The headphones have Prio-2. I can drop this requirement.
What I am concerned in your circuit above, is that good THD requires 22Kload. The standard 10K load is already worse than LM317.
If good THD is stable until 5K, it would be even nicer for me.
Your Kenwood proposition will also be built.
As you see, the results above are for 5K load and almost as good as LM317.
The 10K or 22K simulations are slightly better. Many thanks!
It will worth my time to try to build it for measurements (without headphones), but at the end of the day it will be hard to implement it in my PCB without cutting at least one trace.
As you see, the results above are for 5K load and almost as good as LM317.
The 10K or 22K simulations are slightly better. Many thanks!
It will worth my time to try to build it for measurements (without headphones), but at the end of the day it will be hard to implement it in my PCB without cutting at least one trace.