Simple Attenuators - Design And Testing

[matttornado]

I made this volume level chart in case anyone is interested. Please let me know if I missed a step or if you see any errors. Thanks

 

[aceofbones]

Finally finished my build, mostly. I do have to redo my input and outputs, totally forgot about the switched jacks so may have screwed that up. Anyways, I haven't tested it yet but this is where I'm at. Thanks John!

 

[JohnH]

hi @aceofbones , very nice indeed! I particularly like the neat dressing of the wiring, the 'bus' connections between jacks, and the elegant looking case (your design?), with the funky retro dymo labels.

I think the jacks are all wired fine, given that you connect across the switches, and we aren't using the switches in this design.

Speaker plugged in, and test input resistances as would be seen by the amp, 7 to 10 ohms for an 8 ohm build.

 

[JohnH]

hi @matttornado , table looks find to me. On building, I suggest to lay the switches out on the front panel 14, 7, 3.5, so its more like binary numbers. It doesn't have to be the same as the electrical order.

And thankyou for taking us onto page 100!

 

[aceofbones]

hi @aceofbones , very nice indeed! I particularly like the neat dressing of the wiring, the 'bus' connections between jacks, and the elegant looking case (your design?), with the funky retro dymo labels.

I think the jacks are all wired fine, given that you connect across the switches, and we aren't using the switches in this design.

Speaker plugged in, and test input resistances as would be seen by the amp, 7 to 10 ohms for an 8 ohm build.

Thanks very much John. Ok, glad the jacks are good to go. I thought they were ok like that, but then saw your warning about the switched jack wiring within the last couple of pages so thought I would double check. I found the case on AliExpress and thought it would provide good cooling having the air gap below and it came vented on either side. I just had to drill the holes in the top. It cost as much as all the other parts combined, but I figured it was worth it.
I’m just finishing up my amp build, but I will report back after I have that fired up and have the attenuator working with it. Cheers.

 

[JohnH]

hi @matttornado , thanks again for the table idea. I thought it would be interesting to add up the db's for each setting on an M2, and also show the actual nomjnal amp power that you end up with, based on 50W coming in. Its amazing how little power is delivered at max attenuation, and yet this setting can be still quite useful late at night.

 

[mike58]

Please, post a link for that case, it’s really ubercool. Sort of retro sci-fi with a touch of laboratory thrown in

 

[auflauf]

Hi!
I am deeply impressed by the expertise and experience gathered here. Maybe my questions have been answered in this thread, I apologize for that, but please forgive me for not reading through all 99 pages. My amp is only a Harley Benton Tube15 (Monoprice Stage Right in the US) with two EL 84 valves. It is said to be a descendant of the Laney CUB 12 which's first stage is said to be loosely based on a Marshall design - if that could serve as a ticket in this forum?

I want to build a simplistic reactive attenuator with a fixed value of eg 20db attenuation and one single switch to toggle attenuation/bypass. If I understand it correctly, the second coil as in M3 or the combination of parallel and serial inductors as in M-lite is not worth the effort, and M2 appears to be the sonically best (or the most cost effective?) iteration of John Hewitts design - is that the consensus here? I would then settle with schematic M2 as from 2. July 2020 and 18. March 2021, omitting all resistors except R1, R2a and R2b.

May someone help me out with these questions:
What is the formula to calculate R1, R2a, R2b and L1 for a given attenuation for an 8 Ohms speaker? I understand it is a loaded voltage divider, but my results don't add up.

Did someone try a "bright cap" of eg 10µF or 4.7µF parallel to R2a?

As shorting switches are hard to find, does a simple non-shorting switch which may leave the amp output open for milliseconds harm the amp?

Finally: db attenuation in this context refers to power, not to voltage? Half the power is -3db, half the voltage is -6 db, half the perceived loudness is -10db. Thus for half the perceived loudness how much calculated attenuation do I need? 31% of voltage or 10% power?

Any help is much appreciated, thanks!

 

[aceofbones]

@mike58 Here you go…
https://a.aliexpress.com/_mLadpLc

@JohnH I just tested resistance as you suggested and it ranged from 8.1-9.2 ohms depending on switch combinations. Looks like it’s good to go. Thanks again!

 

[JohnH]

Hi @auflauf and welcome to our thead. Thanks for your interest in this design.

Yes M2 is the best base design. The M and M-lite versions in the early pages work just as well, but not better. M2 simplifies the build with the same performance. M3 is more accurate in the very low bass, in theory, but it really makes minimal audible difference and makes no difference at all except when you hit the low frets on the low strings.

There's no simple formula for the component values, since with the resistive and reactive components, its too complex to solve directly. Instead, its a step by step iterative process using complex numbers. i do it on a spreadsheet which is about 100x100 cells, which also has to model the speaker. If you want to investigate yourself, Id suggests using a Spice program.

But, the first stage is highly optimised for several parameters, even if it looks simple. So it's not easy to vary it. If you want to build a fixed version, Id suggest to build stage 1 as it is, then add one or two other fixed stages to it.

But id also ask, do you really know what attenuation you want? It may not be obvious what you need until you try it. It depends on your own percepetions and also how you use your amp. Happy to discuss.

If you are building with an overall bypass switch, id recommend that you never operate the switch unless the amp is on bypass or off. Hence shorting or non-shorting doesn't matter. I think most, or maybe all toggles are non-shorting.

On other attenuator designs, the treble capacitors try to compensate for loss of treble due to the attenuator design, typically because an L pad is used that damps the speaker at low settings. We dont have that issue, so we don't need this feature as a work-around. If you want to try this just for tonal adjustment thoug, id suggest to do it on a later stage, such as R4. I dont think anyone has done this so far though.

 

[auflauf]

Thank you for the kind reply. Concerning the attenuation: I play the Tube 15 in a rented flat, gain at 4-5, volume at 1 to 2 or gain at 1.5, volume at 3. Even this is sometimes a little loud during the day and definitely too loud during the night. I guess I would like to attenuate the perceived loudness to a quarter to two thirds of the original. Desired sound should range from clean to a little crunchy and sometimes singing overdrive - classic late 60s blues and bluesrock, ... I liked a silverface fender champ and a modeled VOX AC 30. I'm just dipping my toe in the water with the 'modest' Tube 15, my first valve amp.

Concerning the schematics: got it and I am even more impressed. And you are right, the resistors are way, way cheaper than the coils, so they are not that much of a factor.

 

[JohnH]

Those Tube 15's sound like a great entry way to try tubes. its still hard to really relate db settings to perceived volume though. Some people say that -10db is like half volume audibly, but its a big loose variable. If your are not sure, the switched stages help you explore and select what you need. The individual increments of -3.5db are very small steps. If you go to post 1, there are some sound clips stepping up through the stages to full volume.

The coil that you need for M2 is about $10 to $15 usually.

 

[auflauf]

Just as a ballpark figure, I guess 0, -10, -20, -30 db would be fine. If the reactive stage would have -8 db, then one -8db and one -16db switchable resistive stage would allow -8, -16, -24, -32 db.

Does it make a difference if the reactive stage had eg -20db without a resistive stage or -8db reactive plus -12 db resistive?

What is your experience and opinion with loudness (the change in frequency sensitivity of the ear with the acoustic power)? Given that your design is sonically very transparent, is there any value in a "bright cap" parallel to R2a? I guess a switchable capacitor ruins/alters the carefully fine-tuned combination of resistive and inductive loads? The same question for a "bass boost coil": where could one put that in order to have an increasing effect with increasing attenuation? Serial or parallel to R2a? Elsewhere?

 

[JohnH]

To be honest I see no real value in adding treble bleed caps to this, and so far in 100 pages, I don't think anyone has wanted to add one either!. I think that the Fletcher-Munson hearing effects get blamed for a lot of issues with attenuators when it is mostly about the design of them. I play mine loud and also quiet. I find if I am playing loudly, and them have to go right down to very low volume, it sounds muted and dull. But if I start in a quiet environment, and begin at low volume, it sounds fine and clear and full. Our ears adapt all the time. On the attenuator, the real tone is virtually identical at all volumes.

On the bass cap and coil shown in the M3 design, it doesn't actually relate to the level of bass that you hear. There is still a bass resonance generated by the speaker interacting with the attenuator, in a similar way that it would usually interact with the amp. The only difference is, the amp doesn't see this so in theory doesn't respond in quite the same way. Actually, what you get is a very slightly cleaner more solid bass, which is not a bad thing at all.

Based on the above and what you describe, I'd suggest to build the M2 as it is, but leave out the last -3.5db stage. This will give you a good range of -7db steps. Ive had that arrangement in an earlier build, and its quite good when you aren't needing to match into a band mix. And everything is worked out ready to go. Even without that -3.5db stage it'll get you down to about the same volume with your 15W amp as I get to with my VM, which is about a 35-40W output really.

 

[Graham G]

Hi Graham, that's great news and congrats on getting it together.

(There's another work-around, though I think it doesn't work with your rig. It gives a -3db attenuation, ie half power. You plug the speaker direct to the amp, and plug the attenuator as a load-box without a speaker, set to max attenuation, in parallel. Then you set the amp to lower ohms, eg 4 ohms with an 8 ohm speaker and attenuator. I think that's the bit that doesn't work with your 8 ohm cab and the Origin.)

Hi John, could I ask a couple of questions about the above "option",
Could this work by using a 16ohm out(& Speaker) & then using the 8ohm out for the M2 ?.
Or does the Amp need an ohms switch ?, the ORI50 has a 16ohm out & 2 8ohm outs, but no adjustable ohms switch, I also believe it to be Class 2 wiring, I don't know if that makes any difference.
I still have the original 16ohm Speaker.
Thanks again.

 

[JohnH]

Hi Graham, yes I see how the Origin is configured looking at the manual, with three output jacks and no switch.

What you describe above would put 2x the load on the amp compared to what is intended, since it'd be getting a full 16 Ohm load and a full 8 Ohm load. The idea I posted above would work if you were using a 16 speaker and also had built the M2 for 16 ohms. Or with your gear, it would work if the amp had a pair of 4 Ohm taps.

But, if you want a half power, -3db set up there is a way, that requires a bit of wiring. Basically youd set the M2 at max attenuation with no speaker, and wire it in series (not parallel) with the 8 ohm speaker. Now you'd have two 8 Ohm things in series to make a correct 16 Ohm load, and plug that into the 16 Ohm amp output. That could be done as a switch addition to the attenuator, but Id suggest it may be more reliable in the form of a specially wired cable, with three jack plugs for amp, M2 and cab. The change to his arrangement would need to be while the amp was off or on standby.

The theory is all good, but the risk, in my case, would be my errors in plugging/switching it in.

 

[Graham G]

Hi Graham, yes I see how the Origin is configured looking at the manual, with three output jacks and no switch.

What you describe above would put 2x the load on the amp compared to what is intended, since it'd be getting a full 16 Ohm load and a full 8 Ohm load. The idea I posted above would work if you were using a 16 speaker and also had built the M2 for 16 ohms. Or with your gear, it would work if the amp had a pair of 4 Ohm taps.

But, if you want a half power, -3db set up there is a way, that requires a bit of wiring. Basically youd set the M2 at max attenuation with no speaker, and wire it in series (not parallel) with the 8 ohm speaker. Now you'd have two 8 Ohm things in series to make a correct 16 Ohm load, and plug that into the 16 Ohm amp output. That could be done as a switch addition to the attenuator, but Id suggest it may be more reliable in the form of a specially wired cable, with three jack plugs for amp, M2 and cab. The change to his arrangement would need to be while the amp was off or on standby.

The theory is all good, but the risk, in my case, would be my errors in plugging/switching it in.

Ok so as a "quick fix" solution, for my 8ohm M2 & 8ohm Speaker wired as in my sketch ?
Must all switches be left "ON" & cannot be altered ?, so the -3.5db stage could not be used ?

So the correct(long Term) solution is a 16ohm Speaker & reconfig my M2 for 16ohms ?.
I'm sorry to be a pain in the arse, but I hadn't realised that -7db would cut my Amp to 10W.
Regards, Graham.

 

[JohnH]

Ok so as a "quick fix" solution, for my 8ohm M2 & 8ohm Speaker wired as in my sketch ?
Must all switches be left "ON" & cannot be altered ?, so the -3.5db stage could not be used ?

So the correct(long Term) solution is a 16ohm Speaker & reconfig my M2 for 16ohms ?.
I'm sorry to be a pain in the arse, but I hadn't realised that -7db would cut my Amp to 10W.
Regards, Graham.

I'm happy to do you a sketch of this (its not quite as yours). Once its done, it should work fine, no need to rebuild anything and it should sound right too. But its a one step fix for getting a small -3db cut. The attenuator doesn't control the level, and has to be set at max. Then to go to lower settings, youd replug the wires back to the normal way.

If you try this and it turns out to be just right, then we could maybe consider other ways to implement this setting.

btw. Are you running a combo or a head? It affects whether the speaker end of this is a plug or an inline socket.

 

[Graham G]

I'm happy to do you a sketch of this (its not quite as yours). Once its done, it should work fine, no need to rebuild anything and it should sound right too. But its a one step fix for getting a small -3db cut. The attenuator doesn't control the level, and has to be set at max. Then to go to lower settings, youd replug the wires back to the normal way.

If you try this and it turns out to be just right, then we could maybe consider other ways to implement this setting.

btw. Are you running a combo or a head? It affects whether the speaker end of this is a plug or an inline socket.

Hi John, if you can do me a sketch that would be great & yes I have a Combo, so the Jack lead is hard wired to the speaker.

 

[matttornado]

I finished the footswitch today. @JohnH I'm measureing 75.5 ohms resistance. Is that correct?