| jh6you |
Hi, everybody.
Since I have a chance to get Elma 4-section rotary switch of 11 steps (4 x 11), I am going to make the balanced output volume attenuator for my SOBOZ. By the way, all always say that output impedance of the preamp is preferably to be low. Therefore, I am thinking to make a volume attenuator of variable impedance--i.e. the volume higher, the impedance lower. The sketch is shown in the attatchment. Is this worth trying?
Your advice or opinion would much appreciate. :) |
|
|
| wuffwaff |
Hi,
this way your output impedance will always be below 10 Ohms. At higher volumes your Bosoz will see a load of less than 100 ohms though. This will reduce gain quite a lot.
I don´t think this will work.
11 steps is also not really much (some people even find that 24 steps are not enough)
William |
|
|
| jh6you |
Thanks, William. I like your pessimistic view.
I have actually used the 11-step output volume for long year without inconvenience. Daily, I usually touch the volume knob within 3-5 steps. Meanwhile, if I want to shift up or down the global output level, I could do it by adjusting the input level or adjusting amp gain level. This however seldom happens.
By the way, I want to scratch my head further before I abandon my thinking. Could you give me a bit more specific reason why it will reduce gain quite a lot? |
|
|
| wuffwaff |
Hi,
just look at the original Balanced Line Stage article by Nelson Pass. Somewhere near the end it is mentioned that loading the amp with 600 Ohms or so won´t worsen the distortion but will reduce the gain.
Loading it with 10 ohms will reduce this even further.
William (realist, not pessimist;) ) |
|
|
| jh6you |
Yes, I clearly read: "Lower values of output potentiometers will reduce the gain of the circuit, but not otherwise degrade the quality."
And, if I open my eyes big towards Aleph P1.7 service manual, it sepcifies the output impedance as 0-800ohms. What does this explain?
I'm still scratching my head, now with both hands... |
|
|
| stefanobilliani |
| quote: | Originally posted by jh6you
And, if I open my eyes big towards Aleph P1.7 service manual, it sepcifies the output impedance as 0-800ohms. What does this explain?
|
It explains that you can't go up over 800ohm with the output impedance, since it is setted by the volume control circuit. |
|
|
| wuffwaff |
Jh6you,
if the gain is reduced at higher volume settings you won`t get higher volume at a higher setting! This may not reduce the quality but it won´t be of much use either.
If a volume pot is not buffered the output "sees" the pot and the output impedance changes with the pot setting. Depending on the output impedance of the gain stage your output impedance will go from 0 (pot turned to the left) to a max value (somewhere in the middle) and finally to the out put impedance of the gain stage at the max. setting.
William |
|
|
| jh6you |
William
I admit that I have not thought about your point. I will study more.
By the way, I have thought about the potentiometer (voltage divider) having almost constantly low R-thevenin at the output node, to make the potentiometer as a kind of "stiff" voltage source. Of course, the V-thevenin at the input node will be variable. |
|
|
| jh6you |
Hi, William
I am thinking to have one additional register at the input node of the potentiometer, e.g., selectively inbetween 30-220ohms.
My understanding regarding the amp gain reduction:
The original SOBOZ has the current source tails formed by registers. In this case, the value of the output potentiometer will affect the amp gain. In contrast, mine has the CCS tails formed by FETs so that the size of potentiometer could be at far low value. Do you think that now the blind is facing to the right direction? |
|
|
| wuffwaff |
Hi,
don´t know. I thought you had a normal BOSOZ. Could you show us a schematic?
William |
|
|
| stefanobilliani |
| quote: | Originally posted by jh6you
Hi, William
I am thinking to have one additional register at the input node of the potentiometer, e.g., selectively inbetween 30-220ohms.
My understanding regarding the amp gain reduction:
The original SOBOZ has the current source tails formed by registers. In this case, the value of the output potentiometer will affect the amp gain. In contrast, mine has the CCS tails formed by FETs so that the size of potentiometer could be at far low value. Do you think that now the blind is facing to the right direction? |
The gain of the BOSOZ , and every differential pair IMHO , does not depend from the quality of the "tail current sources" . That current sources can be passive (resistors ) or active (Constant CS) and the gain of the circuit still depends by the load resistors (or CCS or Variable CCS) and by the resistor value that join the "tails". |
|
|
| jh6you |
| quote: | Originally posted by stefanobilliani
... the gain of the circuit ... depends by the load resistors (or CCS or Variable CCS) and by the resistor value that join the "tails". |
You are 100% right and I agree with you.
In addition, let me have a physical experiment with two pipes each having length of 20cm, diameter of 1 cm and a 0.5cm-dia hole in the middle.
Pipe 1: The upper end of the pipe is open. The lower end is closed with a resistance material (tail registers). The middle hole is also closed with another resistance material (output potentiometer).
Pipe 2: The upper end of the pipe is open. The lower end is also open (CCS). The middle hole is closed with resistance material (output potentiometer).
Now, I put my mouth to the upper end and blow the pipes.
Pipe 1: Part of the air will escape through the middle hole. The amount will be depending on the relative resistances between the output potentiometer and the tail registers.
Pipe 2: Once the middle hole has resistance, high or low, the most air will escape through the lower end opening.
If I replace the air with electrons, I might have to suck the pipe, not blow.
I see the difference in this way between register tails and CCS tails, with the output potentiometer. Don't you think so?
Merry X-mas !!! |
|
|
| stefanobilliani |
| quote: | Originally posted by jh6you
I see the difference in this way between register tails and CCS tails, with the output potentiometer. Don't you think so?
Merry X-mas !!! |
I don't know exactly ... I was standing at sex and rock&roll not considering drugs...
:joker:
seriously your description apply better for a CMRR ( common mode rejection ratio ) explanation IMHO in a case of a single ended input .
But how with R5? ( see previous attachment)
If you short it, the tails will be in parallel ; and how if you use a resistor(R3) and a CCS(R4) at the same time?
Cheers and marry X-mas to all! |
|
|
| jh6you |
| quote: | Originally posted by stefanobilliani
sex and rock&roll |
Here is raining and cold. How is there?
Your circuit shows local part only. Anyhow, your amp gain is decided with the drain error voltage and the gate error voltage. By a mathematical approximation, it is expressed as (R1+R2)/R5.
The drain error voltage depends on the amount of current flowing through R1 and R2.
Now, let’s see the global picture. Some of AC current flowing across R1 and R2 would leak through the output caps and again through the potentiometers (let’s say, R6 and R7, probably of low values). Then, the above mathematical approximation will be getting more erroneous. This could be prevented by replacing the R3 and R4 with active CCS.
As mentioned before, the original article of SOBOZ having resister current sources says: “Lower values of output potentiometers will reduce the gain of the circuit …” Do you think the writer is D&D?
Let’s think about a little bit more. The sex, rock&roll would become more exciting… :)
Seasons Greeting !!! Cheers !!! |
|
|
| stefanobilliani |
| quote: |
Here is raining and cold. How is there?
|
Here there is beautiful weather perfect for holidays in snow fields... skying ... cold of course . :Olympic: :frosty:
| quote: |
This could be prevented by replacing the R3 and R4 with active CCS.
As mentioned before, the original article of SOBOZ having resister current sources says: “Lower values of output potentiometers will reduce the gain of the circuit …” Do you think the writer is D&D?
|
I agree with you .- BTW the NP articles and in particular SOZ ,BOSOZ has been great company and source of inspiration and in times not always happy (personal meditation) .
-I know well that if I use a pot of 10k the gain will be higher that
with one of 5k .
I am just thinking that the gain will stay the same using or not a CCS in the lower part ( sources ) of the circuit, taper included . ( someone often said that the differential pair is happy also driving a short ).
If you have to have a load ( the pot , say 470ohm) , from my approximated point of view, the total circuit gain at the output will remain the same and independent from the use (or not ) of the CCS ( at the sources of the signal FET)
I know that the CCS provide a better conversion from a single ended to a balanced signal where one of the gates of the differential pair is at ground .
I experienced no differences in gain with the D1 cascode stage with or without CCS:
the signal from the DAC is connected directly to the source of the "pipe" , its gate is at a virtual ground ; Drain is loaded with a resistor .
:tilt:
:xmastree: |
|
|
| jh6you |
Wow. Snow field? I'm missing skiing.
Alright, thanks, discussion seems to be enough. I will make the potentiometer as I think about, when I return back from the trip to S'pore. By the way, guys from S'pore, could you direct me where in S'gapore I can do window shopping of the elecronic components?
:deerman: :santa2: :deerman: |
|
|
| jh6you |
I’m unsure whether the title of the thread was correct or not in technical term. It might, however, don’t matter if the content has something.
I received the 4 x 11 steps rotary switches from a Norwegian friend and built up the potentiometer as shown in post#1. Some of resister values were changed due to availability in the shops nearby where I stay. Particularly, the shunt resister value was changed from 10 to 22 ohms. They were Taiwan-made metallic 1/4W 1% class. During fabrication, I found my hands were getting tough in soldering the small dots. The finish was ended up in rather ugly shape.
My BOSOZ (with CCS tails) is for unbalanced input and balanced output. The + input side was used to receive the input signal and had the volume potentiometer while the – input side was connected to the ground. I thought that this arrangement was the noise source to the speakers. I wanted to remove the noise so that I wanted to replace the unbalanced input potentiometer with a new balanced output potentiometer, wishing improvement in CMRR.
At the same time, I wanted to taste a new exploration, by making what I never done before. And, this guided me to the thinking of the post#1.
I replaced the new potentiometer last night. I was excited about the result. With no signal input, I turned up the volume knob step by step, bringing my curious ears to the 1cm distance from my 4ohms 90dB speakers. Until I reached to the top volume, the speakers were virtually dead quite. I wondered something must have been wrong. I decided to test music, and put one CD wandering around like a homeless dog. No music, wait a moment, it was because the knob position was at the start point. I stepped up the knob one click. Yeah! The music started at low level. Finally, I listened to the whole music loudly at the various volume positions. Wow! It worked.
I felt that the overall sound got better, particularly in high frequencies. I was wondering whether this could be due to the reduced noise level. And, it was right that the amp gain was resulted in reduction (I have no calculation). Anyhow, I got full satisfaction with the result. I felt that the overall music sounded more relaxed with better distinguished small details.
I will see how long I will live with this new potentiometer. If I see any room for further improvement with my BOSOZ, I will not hesitate to make it. Somewhat late though, Happy New Year!!! :) |
|
|
| jh6you |
Here is my final setting.
Works very fine for me.
In addition to this,
the input volume and amp gain level coordinate together.
Hope this will be a useful info. |
|
|
| tonywong |
Hi,
you can shop for electronic components in Sim Lim Tower and Sim Lim Square in Singapore.
regards,
tony
| quote: | Originally posted by jh6you
Wow. Snow field? I'm missing skiing.
Alright, thanks, discussion seems to be enough. I will make the potentiometer as I think about, when I return back from the trip to S'pore. By the way, guys from S'pore, could you direct me where in S'gapore I can do window shopping of the elecronic components?
:deerman: :santa2: :deerman: |
|
|
|
| jh6you |
| Thank you very much, Tonywong !!! |
|
|
| jh6you |
Hi, Wuffwaff and Stefanobilliano,
I was talking about sh*t of non-sense in this thread. Both of you (and other readers) will know what I mean. I feel full shame on my ignorance. I was wrong and you were right.
By the way, my BOSOZ has the drain resistor of 500 ohms and the gain control resistor of 200 ohms between the source pins. Therefore, the amp gain is approximated as about 14dB. Meanwhile, the above volume attenuator (the final one) provides a resistor between the drain and the ground of about 104 ohms with repect to AC. So, the amp gain is reduced by about -15 dB. Therefore, at maximum volume, my amp's total gain is about unity. This unity gain is considered enough for me.
By doing this, one thing I earned is that the variation of output impedance is managed low and narrow. I wonder whether this will work positive or not, though.
If you have any comment, would be appreciated. |
|
|
| dw8083 |
| quote: | Originally posted by tonywong
Hi,
you can shop for electronic components in Sim Lim Tower and Sim Lim Square in Singapore.
regards,
tony
|
Hi Tony,
I spent 4 years in Singapore. I used to love shopping in Sim Lim Tower and Square. I also liked Audio shopping in Adelphi shopping center by the court house. I never saw much Pass gear, but a lot of McIntosh, and Tube stuff. Pretty fun though.
I remember awesome grey market deals on Rotel and other gear made in Asia.
Take care,
-David |
|
|
| jh6you |
Phew¡¦
I have been polluting this thread with non-sense and misprint.
Okay, here is overall view of my attenuator.
The indicated gain values are actually total gain at output,
and they are calculated by Excel (carefully).
The maximum gain is slightly less than unity gain.
This one is working well for my system.
It is getting warmer here. It seems that spring is arriving soon. |
|
|
|
