No, your statement is a misconception. Sound perception involves hearing and brain and those are subjective. But the actual audio wave entering your ears can well be measured in far more detail than what you can perceive.
With a fully differential signal path +/-18Vdc supply is enough.
0.5mV@1Khz, Gain 1000@1Khz, overload 30dB gives 15.8Vrms or 22.4Vpk, divided over two differential outputs of 11.2Vpk each.
And yes before you ask, my preamp following the diff. phono preamp has diff in and outputs and so has my main amp diff inputs.
A reader that innocent should not be reading these sort of threads.
It means of course, 30dB overload margin over the whole range from 20Hz to 20Khz.
All information is fully described in Linear Audio Vol. 12.
You might have hoped getting to see (parts of) an illegal copy, but as I mentioned to you before, contact Jan Didden.
Hans
1.5Vrms
equals to 4.243Vpp
!
at 1kHz it is 424.3mVpp
approx 11Vp
need +/- 13 till 15Vdc
right!
#
we go back to 1kHz
212.15mVp * 100 = 21.215Vp
now we need +/- 24Vdc supply rails.
right!
Consensus!
As I said, without any additional information, a single figure is not very meaningful (to put it bluntly, “worthless”).
thx Andrew
greetings,
HBt.
So, how does this degrade the statement that correlation between measuring and hearing perception is very weak.
What you say could almost be interpreted as: what can't be measured can't be heard.
If so, we can simply conclude that we have radically different opinions in this respect, let it be.
Hans
I know, because I can look at a photograph of the assembled circuit board, and I can also read in the free Gerber files and thus completely reveal your mysterious circuit.
I have long since completed this process for myself.
This is one of the biggest mistakes of this month.
I'm not interested in your circuit, it's boring (for me!).
HBt.
You said that the claim that "everything is measurable" is a misconception. I just noted that this statement of yours is a misconception.
This statement is rather in contradiction with your questions about supply voltage and overload margins.
Without references and further information (!),
information of this particularly brief nature should be viewed very critically.
On top of it your posting is stated in a very unfriendly and almost offensive way.
Hans
Last edited:
Let's implement the whole scenario as follows for the sake of immediate comparability:
5mVrms and G now 100
corresponds to 0.707Vp as nominal output voltage
You will clip at 11.2Vp as you wrote !
11.2Vp/100
=112mVp or 79.2Vrms
leads precisely to an overlaod reserve expressed as a margin about 23.4dB. No more and no less.
That's some honest plain speaking, isn't it?
HBt.
Psst
Unbalanced, you often only need +/-12Vdc supply rails for this STANDARD.
When seen out of the context, right.
But here it was in the context that hbtaudio's remark "everything is measurable" was his reaction to perceiving sound differences between topologies.
Hans
I am always friendly and clear, I hope. Perhaps we could stick to the numbers ...
HBt.
What you calculate is not 23.4dB but 23.99dB. (20*log(11.2/0.707)
Now add 6dB for the differential output and you have 30dB
Hans
Last edited:
And only I know where the difference of 0.57dB comes from. A big mystery ..!
Logical, 24 + 6 = 30
Do you seriously believe that I don't know all this ???
And with 11.2Vpk and +/-18V supply it still isn’t clipping, so overload margin is even a few dB beyond 30dB.
Hans
Then we can dispense with the question of definition and the reference to this huge reserve. Then all that remains is the question of processing clicks and pops.
For this we now need meaningful examples:
what does the typical time curve Y(t) look like for such an interfering input signal? How does our transducer react
Only now does the question of the absolutely necessary reserve arise in the context, provided that these disturbances lead to an overload or some other form of saturation in comparison to the usfull signal (composed as a series of sin(omega t) or cosine according to Mr. Fourier) - "dem Nutzsignal", or information, the music.
Now the brainstorming and debate can begin.
HBt.
For this we now need meaningful examples:
what does the typical time curve Y(t) look like for such an interfering input signal? How does our transducer react
- "im Leerlauf" with R_load = infinitive (and C_load = zero) that means Rp= infinitive [V/A] and Cp = zero [As/V] equals |Z| infinitiv
- "im nominellen Lastfall" with correct termination, for example Rp=47kOhm and Cp=200pF
Only now does the question of the absolutely necessary reserve arise in the context, provided that these disturbances lead to an overload or some other form of saturation in comparison to the usfull signal (composed as a series of sin(omega t) or cosine according to Mr. Fourier) - "dem Nutzsignal", or information, the music.
Now the brainstorming and debate can begin.
HBt.
The self-builder can already be reassured:
with supply voltages of +/-12Vdc we are already completely on the safe side if we can control /swing symmetrically between the rails.
with supply voltages of +/-12Vdc we are already completely on the safe side if we can control /swing symmetrically between the rails.
Hi Gents, (I think only gents are here, any ladies?),
I see discussion on clipping is heating up with emotions 🙂, good but lets keep it in gentle manner.
What I know for sure, is that Hans can calculate, design, make and measure electronics, particularly vinyl reproduction. He has proven himself so many times!
Nick (OP) had something else to present, obviously he will need to wait a bit for us to actually read his paper (even he could translate it first do) 🤣
So I invite anyone to turn the heat on my old design that was already shown in 2 of my threads, about diy TT rebuilt and MC pream thoughts, many have seen it, but most did not I guess.
So, cart is Benz Ruby 2 , MC in this case, specified output 0.3mV @1kHz @ 3.54 cm/s. When measured with "standard" 5 cm/s @ 1 kHz it gives expected about 0.45 mV output @1kHz.
So, I amplify it with this:
PS rails are +-18V, minus say 3 V for opams it is 30 Vp-p achievable, but given it is fully balanced Vp-p is actually double, 60V. Lets say we can get 20V RMS out of it easily.
Now interesting, gain of each mic amp is set to 1000, since they are 2 and opposite polarity gain of first stage before RIAA is 66db, or x2000
Nominal output ruby in LF = 0.45 mV -20db (/10) * 2000 = 90 mV
Nominal output @ 1kHz = 0.45 - 0db x 2000 = 900 mV (about what scope says, this is measured and correct)
Nominal output @ 20kHz = 0.45 + 20 db x 2000 = 9 VRMS!
Signal after RIAA is so cool as passive riaa normalizes it to bass level (90mV) that is amplified with second stage. no issue.
But HF has 9VRMS nominal before RIAA and clipping comes at 20V RMS.... So what is the margin in db? Not very much isn't it? About 7db.
Still, in 20 years of using this set up me, neither anyone else, noticed any clipping what so ever....
If the record is messy and has clicks, pops and other nastiest, I'm not sure if I hear clipping or them nastiest, who cares, bad LP is bad....
I see discussion on clipping is heating up with emotions 🙂, good but lets keep it in gentle manner.
What I know for sure, is that Hans can calculate, design, make and measure electronics, particularly vinyl reproduction. He has proven himself so many times!
Nick (OP) had something else to present, obviously he will need to wait a bit for us to actually read his paper (even he could translate it first do) 🤣
So I invite anyone to turn the heat on my old design that was already shown in 2 of my threads, about diy TT rebuilt and MC pream thoughts, many have seen it, but most did not I guess.
So, cart is Benz Ruby 2 , MC in this case, specified output 0.3mV @1kHz @ 3.54 cm/s. When measured with "standard" 5 cm/s @ 1 kHz it gives expected about 0.45 mV output @1kHz.
So, I amplify it with this:
PS rails are +-18V, minus say 3 V for opams it is 30 Vp-p achievable, but given it is fully balanced Vp-p is actually double, 60V. Lets say we can get 20V RMS out of it easily.
Now interesting, gain of each mic amp is set to 1000, since they are 2 and opposite polarity gain of first stage before RIAA is 66db, or x2000
Nominal output ruby in LF = 0.45 mV -20db (/10) * 2000 = 90 mV
Nominal output @ 1kHz = 0.45 - 0db x 2000 = 900 mV (about what scope says, this is measured and correct)
Nominal output @ 20kHz = 0.45 + 20 db x 2000 = 9 VRMS!
Signal after RIAA is so cool as passive riaa normalizes it to bass level (90mV) that is amplified with second stage. no issue.
But HF has 9VRMS nominal before RIAA and clipping comes at 20V RMS.... So what is the margin in db? Not very much isn't it? About 7db.
Still, in 20 years of using this set up me, neither anyone else, noticed any clipping what so ever....
If the record is messy and has clicks, pops and other nastiest, I'm not sure if I hear clipping or them nastiest, who cares, bad LP is bad....
Cool stuff 😎
It is true that it is better not to play defective records with our favorite expensive system (often an MC).
Perhaps we should stick with the sphere and the inexpensive MM systems.
HBt.
It is true that it is better not to play defective records with our favorite expensive system (often an MC).
Perhaps we should stick with the sphere and the inexpensive MM systems.
HBt.