• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Line level attenuator

OP
Davide

Davide

Senior Member
Joined
Jul 6, 2020
Messages
460
Likes
171
Location
Milan, Italy
Well, thanks everyone then. Your help is really useful!
 
OP
Davide

Davide

Senior Member
Joined
Jul 6, 2020
Messages
460
Likes
171
Location
Milan, Italy
I am again writing about the issue because the Rothwell attenuators I took turned out to be unsuitable. The volume was not turned down by 20db and there was also a noticeable channels imbalance, in addition to the impression of a "bad" sound compared to the digital attenuation alone.
Generic commercial products are probably not suitable for all combinations of devices, as I understand the relative output and input impedances matter.
Having said that, I saw that inside the Rothwells there were three 8.2k resistors in U scheme.
If I did it right, the calculations had to correspond to 10db of attenuation, so I also have the doubt that there was an error in the product I purchased.
Anyway, I decided to build the attenuator myself because it's really simple. What is less simple is to understand how to calculate the resistance values.
I target always to -20db, considering 100 ohm output impedance of Gustard A18 and 47k ohm of input impedance of HYPEX NC502MP.
What is not clear to me is whether the declared impedance is relative to the single pole or to the sum of them... but I think I understand that it is for single pole.
Leaving this aside, from the table of k factors it results that 20db correspond to k = 10, therefore a ratio of 9 between the series and shunt resistors.
So, since I read that the shunt resistor determines the impedance seen by the amplifier, I consider that they must be about 200 ohms (100 X 2), and consequently the series resistors must be 900 ohms each (200 X 9 / 2 because it's balanced). As a result, the DAC should see 2000 ohms of load.

Now, I understand little of these things, I simply applied to the letter a tutorial found online (uneeda-audio.con/pads/).

Could anyone kindly tell me if all this is correct?

Thanks
 

kchap

Addicted to Fun and Learning
Forum Donor
Joined
Jun 10, 2021
Messages
579
Likes
566
Location
Melbourne, Oz
I am again writing about the issue because the Rothwell attenuators I took turned out to be unsuitable. The volume was not turned down by 20db and there was also a noticeable channels imbalance, in addition to the impression of a "bad" sound compared to the digital attenuation alone.
Generic commercial products are probably not suitable for all combinations of devices, as I understand the relative output and input impedances matter.
Having said that, I saw that inside the Rothwells there were three 8.2k resistors in U scheme.
If I did it right, the calculations had to correspond to 10db of attenuation, so I also have the doubt that there was an error in the product I purchased.
Anyway, I decided to build the attenuator myself because it's really simple. What is less simple is to understand how to calculate the resistance values.
I target always to -20db, considering 100 ohm output impedance of Gustard A18 and 47k ohm of input impedance of HYPEX NC502MP.
What is not clear to me is whether the declared impedance is relative to the single pole or to the sum of them... but I think I understand that it is for single pole.
Leaving this aside, from the table of k factors it results that 20db correspond to k = 10, therefore a ratio of 9 between the series and shunt resistors.
So, since I read that the shunt resistor determines the impedance seen by the amplifier, I consider that they must be about 200 ohms (100 X 2), and consequently the series resistors must be 900 ohms each (200 X 9 / 2 because it's balanced). As a result, the DAC should see 2000 ohms of load.

Now, I understand little of these things, I simply applied to the letter a tutorial found online (uneeda-audio.con/pads/).

Could anyone kindly tell me if all this is correct?

Thanks
I was asking a similar question on placing a pad between the DAC and the amp. I proposed what you could call a balanced L pad it's a simpler design than the H pad. https://www.audiosciencereview.com/...d10-balanced-review-usb-dac.25094/post-853758
 

kchap

Addicted to Fun and Learning
Forum Donor
Joined
Jun 10, 2021
Messages
579
Likes
566
Location
Melbourne, Oz
I missed the part where you said "I saw that inside the Rothwells there were three 8.2k resistors in U scheme" You had already come to the same solution I suggested.
 

Matias

Master Contributor
Forum Donor
Joined
Jan 1, 2019
Messages
5,031
Likes
10,803
Location
São Paulo, Brazil
I also experienced channel imbalance several years ago with Rothwell RCA pads. It amazes me that they still did not fix this in their products after all this time. Bad quality. Buy something else.
 
OP
Davide

Davide

Senior Member
Joined
Jul 6, 2020
Messages
460
Likes
171
Location
Milan, Italy
Yes, in fact, the quality of the Rothwells is doubtful, although the internal resistances are 1% tolerance, but maybe they are counterfeit.
In any case, the output impedance of these attenuators are higher than the DAC one, so it's not very good for the amplifier (this impedance should be as low as possible, as confirmed by HYPEX). Maybe that's why they sound bad ...
However the choice of the pad resistance scheme should be dictated by the impedances involved.
To find out if the impedance created by the attenuator is good for the DAC just see the specification of the opamp of the output stage (typically LPF).
In my case the LME49860 works from 600 ohms upwards ... slightly varying the performance (better or worse due to parameters).
Anyway I bought some "audiophile" resistor to build the pad (Takman REY, they cost very little) so, in a little while I'll let you know ...

PS. Also AES suggest U-pad scheme with two 1k series resistors and variable shunt resistor to change the attenuation level (this work for DAC with <300 ohm output and amplifier with >20k ohm input). Just google it.
 
OP
Davide

Davide

Senior Member
Joined
Jul 6, 2020
Messages
460
Likes
171
Location
Milan, Italy
Dear Forumers,
I just wanted to update you on the matter.
Today I assembled my attenuator with Takman Rey resistors in U scheme (1k + 1k + 226R), for an attenuation of 20 db and I connected everything "flying" to the NC502MP board.
What can I say ... everything works correctly! The Hypex goes very well with my NHT C3s, and the attenuation level is perfect, in the sense that if I put my Gustard A18 at 0db I get exactly the volume I had calculated (the aim was not to risk destroying the speakers and ears) .
I specify that on the pc I also have Dirac Live which applies an attenuation of about 12db (obviously variable depending on the correction it applies).
The only thing I had to correct is the phase, via the reverse function on the Gustard A18.
I made some measurements with Audacity reproducing a tone of 1khz and I saw that with the "normal" phase of the Gustard, the wave was recorded in reverse (I created a track with 1 second of initial silence to be sure to capture exactly the start of the wave).
Applying the inversion on the Gustard the recorded waveform matches that of the track, at least according to Audacity.
Now, unless my Umik-1 is phase reversed, I believe I have operated correctly.
It seems strange to me having followed the wiring diagrams to the letter ...
However I have noticed that in the Gustard manual there is some confusion regarding the phase inversion function. They basically say the opposite of what is written online about the European rather than American phase in XLR connectors.
Anyway ... little changes ... it sounds good anyway!
I would like to thank the whole forum for my result ... without the information present here I would never have been able to do all these steps.
 

PurdueDeep

Member
Forum Donor
Joined
Aug 17, 2021
Messages
42
Likes
11
You're building exactly the same system as I am almost. Thanks for the really helpful thread!
 
OP
Davide

Davide

Senior Member
Joined
Jul 6, 2020
Messages
460
Likes
171
Location
Milan, Italy
You're building exactly the same system as I am almost. Thanks for the really helpful thread!

I am happy to share my experience, especially if it is helpful to someone. :)
 
Top Bottom