Simple Attenuators - Design And Testing

[Dwayne Eash]

@Dwayne Eash

Those of us who've been fighting the cranked tube amp being too loud battle for years (I've been a steadily gigging singer/gutarist for +50 years) have been through most of the same trials, tribulations, "new" ideas and schemes that you're cooking up. We've found what works and what certainly doesn't and are simply sharing our combined experiences with you. This has been an ongoing war since guitars first got plugged into amplifiers and got "THAT SOUND!"

One of the best additions to your rig involves how you use your guitar and in the long run boils down to developing very specific and controlled playing techniques and skills. One of those skills is hard earned playing/picking dynamics. I can plug a guitar with no volume and/or tone controls into a massively powerful amplifier cranked into its sweet spot and be "almost" whisper quiet and clean, with amazing clarity, tone and blissful harmonic overtones and tonal complexity by simply softly and judiciously "controlling" what my hands do on said guitar! I can then dig up a little dirt by "digging in" a tiny bit harder and then I can kill small pets and vaporize Rosemary's baby (she should know better than to bring a baby to a rock show, but that's another story) by really hammering on my guitar! Unfortunately, that blazing volume that gets produced is way more than most sane people (including venue operators) are willing to endure!

A bit more about dynamics. If we make a scale where "0" is silence and "10" is as blazingly hard as you can slam, many (MOST) folks keep their dynamic "average" (rhythms, etc.) somewhere between "6" and "8" and are then disappointed that slamming to "10" doesn't put their solo or other part "out front" in the mix. There's really not a very big, audible difference in "volume/sound pressure level" from "7" to "10"!

Now let's look at a more effective and beneficial way of controlling our volume. If a player develops the technique of keepin their dynamic average somewhere in the "3" to "4" range, dropping to "1" or "0" is still pretty dramatic and allows for nice dynamically apparent "stabs" into that "6" to "7" range, while still allowing blazing stings and/or solos at the "9" to "10" range. This is how many of the really great players have attacked this for years!

Now, I can certainly understand the attraction of stepping on some switch to make up for not wanting to put in the long hard hours of developing great techniques. The main purpose/intent of these great attenuator designs that @JohnH has worked so diligently and hard to develop and generously share with the world at large (for free) is to allow us crazy guitarsts and tone hounds to use larger than practical amps at sane and acceptable volume levels, without lossing our precious tone, nuances and dynamic response to playing techniques! It also helps to tame our volume levels when our exuberance in the heat of performance gets the better of us! It does not seem that the intent was ever to turn these attenuators into some new stomp/foot/pedal/thingie/toy to try to make up for the hard, diligent practice required to become a truly stellar artist! Might I politely suggest that if stomping on a pedal to control your sound is what you're really looking for, maybe you should look into high end modeling amps and/or boutique stomp boxes!

For me, these designs have been the most liberating pieces of gear I've ever owned! I will never be without one and plan to build enough of them to end up having one permanently installed and custom tailored for each in every amp I own!

Just My $.02 & Likely Worth Much less!
Gene

LOL, best post ever. You can paint it any fancy color you want, but you better play it well or it will not sound so good. hehe Might sound simple and trite, but it's nice to know others agree, the magic comes from within, and from your fingers, more than from a fancy stomp box or new digital effect.

I am for just me and the amp for the best in organic sonic ear candy. That's why a clean boost provided by the power attenuator, just releasing one attenuation stage, is transparent brilliance. You double the value of the project, with world class boost transparency, because "it's just your amp" the entire time.

It's just using, what is already naturally available, in a more convenient way is all. I LOVE foot-switch operation because I'm busy making magic with the hands. Sorta frees me up to do more, and stomp boxes for clean boosts is a match made in heaven.

Nice to meet another tone hound guitar purist.

 

[Gene Ballzz]

Oh & FWIW, at 65 years old, I play in a hard rock & rockin' blues band and still practice at 3-4 hours for every hour I perform! Playing guitar WELL is a life long commitment to a lot of hard work and not an act of instant gratification! Also remember that an audience will never notice the nuances of tone that we all obsess over, but finding that great sound is simply helpful to make us more comfortable on stage to in turn play our best and be inspired to do so. THIS is what an audiance really notices!
Just sayin'
Gene

 

[Dwayne Eash]

Oh & FWIW, at 65 years old, I play in a hard rock & rockin' blues band and still practice at 3-4 hours for every hour I perform! Playing guitar WELL is a life long commitment to a lot of hard work and not an act of instant gratification! Also remember that an audience will never notice the nuances of tone that we all obsess over, but finding that great sound is simply helpful to make us more comfortable on stage to in turn play our best and be inspired to do so. THIS is what an audiance really notices!
Just sayin'
Gene

Dig it man. No worries, I jumped a link to a new thread where I await John's answer to my question over my version of his design. We just got our signals crossed and pow, it's not linked up yet, so,,, it's all good. I can't wait to fire that thing up and let my greenbacks work again.

Until then, it's just my attenuation speaker from eminence. Not bad, but not a sweet pair of guitar and bass greenbacks either. So I totally agree, I need a power attenuator, to pull out the best in those old Marshall amps.

Holy cr@p, I turned it up to MV 6, and it was like, beam me up Scotty, feels like I have the ultimate power of the universe, at my fingertips, but more importantly, the lush singing overtones, and that's a nice,, way to treat your inner guitarist.

You get in a zone, and the music, sorta transforms the moment, it's like your in a timeless space, where purity and life meet.

I feel more welcome at my place. So here's a link to see more about my project goings ons. What becomes of the first stage, determines what stays or does not stay in the second and third stages. So I respectfully wait for John's answer, over there to (once again) demonstrate respect for this topic.

http://www.marshallforum.com/threads/johnh-reactive-power-attenuator-replication.114015/

 

[Martin56]

This is a great design. I ran across it after sourcing parts for an airbrake, and made sort of a hybrid with this idea, the airbrake, and what I had. Very handy.

I wondered if one could add some flexibility to the key first stage, if one added a resistor. Here are the optimal values for an 8 ohm load, treating the amp as 20 ohms. While this confirms John H's good choice of 15 ohms for R1 and 10 ohms for R2, with standard value resistors, it also gives some other options.

Martin

 

[JohnH]

hi @Martin56 , thanks for posting. Understanding how Airbrakes work and how they don't get dull at low volume was a stepping stone for my designs.

Your sketch shows variations on a 'T' pad attenuator, which is a standard module in electronics (guessing you know that!). The variations do approximately -8, -12 and -16 db, and you keep it consistent so the amp always sees 8 ohm and the speaker always sees 20 Ohm. I have a spreadsheet that generates these too. I found that by accepting a variation of an ohm or so, the lower limit in db's reduction for R0= 0 can be brought down from -8 db to -7db or a bit less.

Switching between your three versions needs switches on three legs, so that's why I started going for adding new switched stages after the first, so only one switch pole is needed per stage. Also, its hard to get the reactive coil or coils incorporated and properly balanced unless the first stage remains fixed.

 

[Martin56]

Sure thanks. I was only thinking of the first stage actually. The T as the first stage lets one keep the desired load at both ends for different values of dB reduction (but, yes, always more dB reduction in the first stage than 7 or 8dB), for the price of one more resistor. But your idea works great as is. Very nice. Martin

 

[Dwayne Eash]

Where are people getting the parts here in the USA? Seems like china has the supply for cheap, but that's a long shipping time..

 

[Boysenman]

Hi JohnH, thank you a lot for the schematic. I already tried your design in one of my amps with great success.

I wish to create an attenuator, where the edcor 10W transformer in my 5f2a clone reacts the same way, as it does to the Weber 10A100 speaker installed.

I've used LTspice, to recreate the impedance plot of your M3 design, along with series-parallel RLC circuit, simulating the weber speaker at the end of the attenuator. I've gathered the impedance plots of similar 10" alnico speakers, with a resistance of 5.6ohm, visually replicated the plot. It's no trouble to replicate the plot, but it's impossible also get a total impedance of 8 ohm. However i put the components, summing the two phasors will never yield 8 ohm and a good looking impedance plot.

Also, your m3 design (and Aiken) has a treble slope reaching it's peak amplitude much earlier than my circuit, and similar impedance plots on the internet. What's the reasoning behind designen the circuit like that - It is better sonically?

Hope you understand my questions.

Kind regards
Christian Boysen

 

[JohnH]

hi @Boysenman Thanks for your message. I'd be happy to investigate adapting the design for 10" speakers.

May I ask what did you build before? ie which design?

Also, what network of LRC do you find best captures the 10" speaker? is that the one in the plot above?

The basic designs adapt quite well to different speakers and different amps, even without changing values. But if I remember right, typical 10" speakers have a lower coil inductance than similar 12" ones, so the L1 inductor might be smaller. Can you offer an impedance plot of a 10"?

I havnt built M3 myself and I don't know of anyone who has. But M and M2 have been built. The treble rise matches measurements from a 12" Greenback cab and the impedance hits around 8 ohms at about 400- to 500 hz, as with a normal speaker. It varies by part of an Ohm dependent on settings.

I see you are modelling the turns ratio of the OT. In order to capture the amp properly, at least at low volumes, it's output impedance also needs to be included, which I use as 20ohm for an 8 ohm tap, based on my own amp. Do you have any parameter included for that?

Inteteresting, I'm happy to explore further with you

 

[Boysenman]

Hi JohnH.

I posted similar pictures earlier in this thread, on a different account(that I forgot the password for). Heres the back panel of the previous build. A fixed bias 2xEl84 Marshall plexi/jcm800 style 3 channel amp with fx loop and your M2 attenuator.

The 10" speaker impedance plot, is the plot you see in green. It is the circuit you see down to the right in the schematic. Your impedance plot is shown in blue.
The RL series circuit simulates voice coil parameters, where the RLC parallel circuit simulates suspension resistance, suspension inductance, and mass/air lod capacitance, respectively. As for now I've only graphically tried to replicate the 10" speaker plots found on the internet. Next step would be to solve the equavalent 2nd order differential.

Please school me on the 20ohm impedance? The turns-ratio is 8K:8, and afaik I know, it should be reflected back to the primary by your M3 circuit? Primary resistance is 147ohm, and secondary is 0.5, as per datasheet specs by the way.

It should be possible to make a spreadsheet that computes all necessary component values, given basic speaker and amp parameters, for any type of amp and speaker combination.

I look forward to working with you.

Kind regards
Christian Boysen

 

[JohnH]

The real output impedance varies per amp, dependent on its design and how much NFB it has, and also how hard it is being driven. My calcs aim at getting the small signal tone right, and then showing the amp a reasonably close impedance so that when it gets driven, it responds as normal. its too hard to model the real over-driven performance (unless you build the output circuit into Spice, its possible, haven't gone there though). So we focus on the small signals. If a valve amp has no NFB, its small signal output tends to be like a constant current source. Double the load ohms, current stays constant and signal voltage is doubled. This is equivalent to virtually infinite output impedance. Compare to an SS amp, which has extremely low output impedance almost zero. If you double the ohms of the load, current halves, voltage stays the same. A real tube amp with some negative feedback is in between. The more NFB, the lower the output impedance (which has little to do with whether its designed for 8 or 16 ohm speakers.

This effective output impedance is vital to the tone and its ignored in most (dull sounding) attenuators. If the speaker sees a much lower output impedance than the real amp, its bass peak and treble rise cant develop, they are squashed flat.

You can work it out by feeding a small pure signal through the amp and measuring output voltage using two different resistive loads. Then you do maths. I have a DSL401 with no NFB and I work out very high output impedance with this, 50ohms or more. My VM gave me 20 Ohms, which has turned out to be a good value to work with. But I suspect other amps with EL34 outputs and normal NFB are less. But luckily the circuit adapts to them well.

im not sure about the plots of yours. The speaker model has a 1mH coil as its main coil inductance, and this seems very high compared to typical parameters for a nominally 8 ohm 10" speaker. Im also not sure about how LTSpice is plotting impedance. From first principles, at say 10kHz, this coil inductance dominates and its impedance should be around 0.001 x 10000 x 2Pi = 63 ohm, which is what I have for it using my spreadsheet with your model of the speaker.

I have a spreadsheet that analyse the whole circuit, and switches stages on and off cyclically with a macro. The maths is too hard (for me) to 'solve' to directly get component values, but it lets you quickly test and home in on a good design.

 

[MAD64]

Hi JohnH,
first of all - great work and thank you very much for sharing and supporting this!
I've built the M version and found, that 7 + 3,5dB ist best for my needs (I don't want to bring down my amps to bedroom level, half volume is great for my needs) and is also very transparent with my NFB and non-NFB amps. If I remember correctly, somewhere in this thread is a table with different values for the the first (L+R) stage to get different dampings but it ends with 9dB. Can you give me the values for the 2x L and 3x R (for 8 ohms tap and speaker) from your magic Excel sheet when I want that 10.5dB from one stage? And for 16 ohms tap and speaker (all values x2, if I understand it correctly), is it useful to add R11 and R12 from M2(v)?

 

[JohnH]

Hi @MAD64 Thanks for posting.

I think for a -10 5db fixed version, might as well just hard-wire a -3.5 stage after the-7db Stage 1 from version M or M2.

Those added R10 and R11 parts are great if you want to run a 16 Ohm cab from an 8 Ohm attenuator. It just fixes input and output impedance so the amp keeps seeing not too much more than 8 ohms and the 16 ohm speaker sees the right amount to keep its treble and bass correct. Without them, a 16 ohm cab gets noticeably less high treble It all safe, and fine if you want a softer tone, But I like the clarity. It works with either M or M2. I use this feature to run a 16 ohm V30 cab from my VM combo which doesn't have a 16ohm tap.

For a general 16ohm attenuator, yes all resistances and inductances double. But in practice, you have to adjust slightly to use available values.

 

[MAD64]

OK, will do as suggested, I just wanted to save on parts ;-)

 

[JohnH]

OK, will do as suggested, I just wanted to save on parts ;-)

When I figured out the single-stage versions
(Ie no switching and designed to have a speaker straight after), I found that the configuration worked ok in the range -3 to -9db, by adjusting values. After -9 db, the output impedance as seen by the speaker was drifting outside of the target range, which would tend to start changing the tone. So it would need more parts anyway.

 

[Boysenman]

Hi Again JohnH, sorry for the late response. The way LTspice works, is simply plotting the step response of the circuit, with a sine wave sweep from 50 to 20khz (in my case). The Green picture you have seen, is simply the voltage across the speaker sim circuit at the end, divided by the current through the input resistor R6 to speaker sim

Again, I have not yet taken the time to actually sit down and meassure the actual values for the components, all I have done, is figuring out what component values looks most like the picture I've found on the web. This turned out to be with a coil inductance of 1mH, albeit that is very big. Smaller coil values, even when compensating the rest of the circuit, lowers the bass resonance, and gives a feeling of the bass-resonance being equally wider, due to the lower R11 values.

Kind regards
Christian Boysen

 

[JohnH]

The plot is labelling itself in some kind of db scale, which suggests a logarithmic scale. is there a way to plot on a linear scale to represent Ohms?

 

[Boysenman]

Hi JohnH. With a little investigating, I figured out how to make it linear and plot ohms.

Blue curve is with 0.25 voice coil inductance, and no leak resistance
Green curve is with 0.7 voice coil inductance, and 33 ohm leak resistance

I have also included an original plot from a 10" alnico P10R style speaker.
Green plot seems to capture the visual representation a little better, but I don't know how much I should read into that.

kind regards
Christian Boysen

 

[Baby Thomas]

Hello, I've been reading this thread for months and I learned a lot. Thanks to all contributing with JohnH ahead of everyone.

I built my version maybe two months ago and it's working great! My idea was to have simpler reactive only device with only 8ohm capabilities and 3 stages, but with the ability for one of the stages to be relay-foot-switchable for solo boost. For now this is not implemented yet (as well as the active cooling part) until I decide if I will use the device permanently. I never play with amps at home so I will only use the attenuator for rehearsals/live where I need to take off a few dBs and/or bump the MV 10-20% up without pealing the paint. I altered the stages so I have -3.5; -3.5; -7 dB and the first (bypass) stage is -3.5 which is fine for my needs.

I used it with a few amps (SLOClone and H&K Puretone for example) and working great. It's definitely better than Bad Cat The Leash (which I had recently but it's gone now) and much better THD Power Plate (which I had years ago). I'm not familiar with other reactive attenuators firsthand so I can't compare it with these devices. My device is based on Design M and honestly I can't hear ANY difference EQ-wise (which is not the case with the other aforementioned attenuators), however there's some... stiffness or lack of dynamics after the device is engaged. I'm not sure it's the attenuator though, maybe it's just the amp reacting different at higher MV volume levels. The change (if there's any) is minimal but I can't decide if I prefer the lower MV setting without attenuation. My "goal in life" is the purest signal path possible with maximum clarity and dynamics.

 

[JohnH]

hi @Baby Thomas thanks for posting and im glad that's working for you.
And that is one bad-ass mutha of a heat sink! Does it heat up?