Attenuator pad question

[mimmo]

Hi,

I have made pair of rca interconnect cables with -20.8 dB attenuation. I used 10k in-line and 1k0 as grounding resistor. The amount was just spot on to get more room to move with my volume pot. Now I am just thinking if I lost some of the low frequencies by using a L-pad if it is affecting the impedances too much.

So to speak, if I would like to have -20db of attenuation while keeping original output and input impedance ratio of my pre/power combo appropriate, is L-pad the right choice. Or should it be T-pad or maybe something else? This is a bit new area for me to explore...

My pre/power combo is following:

Prima Luna Prologue 3 preamp, output impedance 2.8 kOhms
Hypex Ucd180Hg poweramp, input impedance 100kOhms.

One option to try of-course, would be just removing the gain resistor from Hypex amp boards. But I am afraid the attenuation wouldn'tbe enough.

I would be very happy if I could get some help with my issue.

Best,
Mikko


Ps.
This is the original post I got my L-pad values from:

Hi,
fit the two resistors inside the RCA plug that goes onto the NAD poweramp input. No cable to help avoid the hf rolloff that would ensue due to capacitance between attenuator and input stage.

use a 10k in line and then choose a grounding resistor from the list.
Interconnect core to 10k to RCA centre pin.
and
RCA centre pin to grounding resistor to ground.

grounding resistor
4k7 => -9.9db these low atten give more risk of HF roll off
3k9=>-11db
3k0=> -12.7db
2k0=>-15.6db
1k5=>-17.7db
1k0=>-20.8db
750r=>-23db at this atten HF roll off almost zero @ 20kHz.

should be enough options to get you going.

BTW attenuation = 20 times [inv log (grnd res/10k+grnd res)]
This formula is not accurate because it does not account for the source impedance nor the poweramp input impedance but again it is close enough.

 

[DF96]

2.8K is a high value for output impedance. However, the main issue is the value of the output coupling capacitor. This, together with your 11K attenuator and 2.8K output, forms a high pass filter. I would be surprised if it can't drive 11K, but the high output impedance may be a sign of poor design.

 

[mimmo]

Thanks DF96!

Something I haven´t thought, really. Good points there.
For my life I have been dealing more with visual problems so some electrical things are somewhat new things to me.

I know the output impedance of my pre is really on the high side. On the other hand the Hypex has 100k put impedance that should work with the pre. But as I mentioned my problem seems to be the overall gain in my system which leads me to use the attenuator pads. And obviously this leads to another issue...

To be clear enough with my options that could somehow solve my problem I will tell more of my rig:

Phono Cartridge: Heed Zene MC, 0.4 mV;
Phono Stage: Hagerman Bugle 2, 60dB (DIY);
DAC: Beresford Caiman SE;
Preamp: Prima Luna Prologue 3, 2.8 kOhm out, Gain +/-12dB
Power Amp: 2x Hypex Ucd180Hg + SMPS400, 100 kOhm in, Gain 23dB (DIY);
Sonar OA-14 Omnidirectional speakers, 8 Ohm, around 86dB sensitivity.


And to cut it short, what would be the viable option to get things better? Maybe swap the pre, adding buffer in-between, skip the pre by making Hypex integrated, building passive pre, modifying gains in pre and power amp, maybe building the Grounded Grid Pre (Edit: this seems to have similar gain then the Prologue 3)?

Thanks!
M

 

[DF96]

Why do you have a 12dB gain preamp followed by a 21dB loss attenuator? The best this can do is add a little noise and distortion. Unless you have very long cables the first improvement would come from ditching the preamp. If you do have long cables then you need a better preamp: less gain, lower output impedance.

Anyway, you need to find the value of the output coupling cap.

 

[mimmo]

Hey,

The attenuator is just purely to tame the overall gain and to get the preamp volume pot into better position. Without it I am only barely able to move the pot before volume is too loud.

Actually, I think I might be thinking something similar you suggested. All of my interconnects are less than 1 meter, usually closer to 50 cm. So ditching the pre should not be problem at all. I was already thinking of building a unity gain buffer / passive with input selector for at least four sources.

This morning I managed to find some time to try the normal interconnect in place of attenuated one. This gave the low frequencies back but the room to move with volume pot is now restricted again. I really cannot move the pot past 8'o clock if 7'o is zero volume.

I will try to open the pre tonight to check the output cap value. I will post an image of the internals and/or to find a schematics.

Thanks for the patience,
M

 

[DF96]

A bit of Googling found a picture of the inside, but no circuit diagram. I am not sure which cap is used for output coupling. As you have the real thing in front of you, you are in a better position than me!

 

[mimmo]

Here are few images from the internals.
Sorry for the cell phone quality..


First one is overall layout. First capacitor close-up is from the front end, second one from the middle, last one close to the rca jacks.

In the last image the green, blue, black and white insulated wiring goes to pre-outputs. Btw, vol pot is Alps Blue 250k.

Best,
M

 

[mimmo]

The small black Solen caps in the first images are 0.22uFJ 630V, second ones below 0.47uFJ 630V.

 

[DF96]

What is the PCB in the middle? Output buffer, DC servo? It seems to be connected to the outputs, but so is the valve section (via the white twin wire running underneath the PCB).

 

[mimmo]

Hi,

I checked and at least, if trusting the text on the pcb it should be the power supply board. This is clearly printed on the lower left corner when looking at the last image. Of course it´s impossible to see this from the small file attached here.

When looking at the first image two left jacks are pre-out, third is tape out and rest are inputs

 

[adason]

mimmo, have you though about moving the attenuator to the output?

 

[mimmo]

Hi Adason,

As simple your suggestion is I must admit that I really haven´t thought of that. I think I have been fixed to some previous suggestion that the pad should be at input end..

But I will of course try it! You think this would make some difference?

I´ll try and report.

 

[GoatGuy]

Dunno…

I thought that the design goal was “do a pad, offer same downwind impedance, KISS”.

Your Prima Luna output is 2.8 kΩ
Your power amp is 100k input
The ratio doesn't much matter (in this case), but you want the same 2.8K facing into amp.

Using the Thevenin equivalent resistance identities:

Z_out = R_// = R₁R₂ / (R₁ + R₂)

Z_out = 2800

AND using the algebra-derived

R₁ = R_// / k
R₂ = R_// / (1 - k) … where k = 10^(dB/20)

You rather quickly find

R₁ = 28 kΩ
R₂ = 3.11 kΩ

You can check, but the output impedance now equals 2.8 kΩ. That old Thevenin equivalent resistance for series voltage dividers is a lovely thing.

However the loading ON the preamp is heavier: it becomes R₁ + (R₂ || R_input) or 28 kΩ + (3.11 kΩ || 100 kΩ) or about 30.7 kΩ. It was 100 kΩ. Not a material difference.

However I might be reading your original design all wrong.
The real point though is, your choice of resistors wasn't appropriate to simulate the condition you were trying to sim.

GoatGuy

 

[GoatGuy]

(and no, I still haven't cottoned to the modern fad of 2k8 instead of 2.8 kΩ … its the secret handshake to get into the Eternal Home for Ancient EE Goats. LOL)

 

[mimmo]

GoatGuy, thanks for your reply.

To clarify my thoughts the following is what I would like to achieve.

As I can barely use my volume pot before things get too loud I was thinking of making one of these inline attenuators. I decided to go with -20db and found that old thread with some resistor values presented. So I sticked with those. But on the other hand I was also worrying that my preamp output impedance vs poweramp input impedance figures would not be so favourable anymore

So I decided to try anyway as I had the material around. The amount of attenuation was good enough to get me to move the volume pot further. But I realized that some of the lower frequencies got lost somewhere. Clearly didn't realize that there is a possibility for high pass filter...

But if If I understood right, the resistor values you calculated would keep the amp impedances close to originals? Meaning those R1= 28k & R2= 3.11k. But on the other hand I would still have the high pass filter thing happening, right?

I appreciate your comments,
M

 

[DF96]

We still haven't established which is the output coupling cap, so we can't calculate the LF point. However, raising the attenuator resistor values should lower the LF point.

 

[mimmo]

I will keep studying the layout later on to find the right caps. I might drop an email to the manufacturer and see if they are willing to give more details. I will get back.

In the meantime I will also see if I have some appropriate resistors around so I could try raising the values.

 

[GoatGuy]

DF96 is way smarter about the filter-side being causal to the LF response different than I am. He is indicating that the 28 kΩ + 3.11 kΩ approach may lower the LF response to where you're getting back what you perceive is lost.

Give it a try!
Cheap, easy, et.c.

PS: ONLY DO ONE CABLE at a time!! Why pass by the opportunity to hook one to the LEFT channel, and the other to the RIGHT channel. A-B testing on the super-cheap.

GoatGuy

 

[mimmo]

Evening,

Just got my solder station cooled down.
Though, good point trying a/b channel by channel. Will do so.

I am learning every day, thanks for this!

M

 

[mimmo]

Hi,

It's been a while since writing on diyaudio. So, just small update and one new quick question regarding attenuator pads.

After trying few integrated and pre+power amp combos I have come to a conclusion that my problem really is related to having excess gain. So, no need for added pre-amp gain here. Best diy amp project so far has been Hypex UCD180 with Alps Blue. Still the pot does not allow too much rotation until things get too loud. But this can do now for a while, it's manageable.

Then the question:
If I would like to solder attenuator pads directly to the rca sockets inside the amp should I expect problems as all of the audio grounds are tied together? Or can I simply make individual pads on each input sockets and just connect all of the grounding resistors to audio ground?

If I wasn't clear enough with my question, just ask me and I will clarify.

Thanks,
M