Simple Attenuators - Design And Testing

[JohnH]

Hi @stickyfinger

On rotary controls, I did figure something out in December 2019. See post 634 on page 32

https://www.marshallforum.com/threads/simple-attenuators-design-and-testing.98285/page-32 dated 21 December 2019.



A fixed -7db stage, then another switched stage and a rotary with 6 positions giving -3db steps. If anyone is interested, please read the post about it above.

Never tested but it follows all the design principles. The issues with it compared to a standard M2 are, there's more parts needed, and finding a good enough rotary switch with the right current rating. By limiting the rotary to 6 positions and using just 2 poles, it should open up the selection. It may be good for small amps. Lets say its a 5W amp . After Stage 1, there's about 1W, which into 16 ohms is 0.25A or 0.35A with 8 Ohms. The switch spec should be x2 that, at 125V.

Some might prefer the variable control of a pot or L-pad. But with these, its not feasible to control the resistance values properly to maintain tone as volume is reduced. This is evidenced by most commercial units using Lpads, which sound muffled as you turn down. Also, looking at the green resistors above, you can see that they are a chain of specific values designed for the relevant attenuation step, and that amounts to an overall taper across the range that isn't found in any standard pot

Overall, I put this up again for discussion but really, the standard M2 and its variants is a lot simpler and once you get a feel for what each switch does, it works intuitively. Plus, once the right volume is found, you just leave the attenuator set and get on with playing!

 

[JohnH]

hi again @stickyfinger
Interesting about the other attenuators and discussions that you referred to above. Just wanted to note that with the M series designs, even though only the first stage has reactive components, that's actually enough to show the amp the right impedance curve and they are balanced so that the following stages can just be resistive through to the speaker. What the speaker sees is then something like an amp output impedance,

That Faustine unit looks like a beast!

 

[Gene Ballzz]

@JohnH &All,

I just wired up my first 8 ohm, M2 attenuator. Actually it was a rewire for another member (I won't mention his name) whose soldering was poor, at best and had burned up a couple of the switches, etc.

While a 16 ohm unit still sounds fine on a 16 ohm amp tap, into an 8 ohm speaker, works well and is perfectly safe, the exact same amp and speaker sound noticeably better and more dynamically responsive with the dedicated 8 ohm, M2, out of the 8 ohm amp tap. I just ordered the parts for an 8 ohm build!

I thought you'd all like to know?

Simply Sharin'
Gene

 

[donwagar]

By now, you would think this thread would be a sticky.

 

[Gene Ballzz]

By now, you would think this thread would be a sticky.

If I recall correctly, some of the moderators, etc, here may not be real fans and/or subscribers to attenuator usage. There are many folks who believe no attenuator is really very good or useful!
Just My ,
Gene

 

[JohnH]

@JohnH &All,

I just wired up my first 8 ohm, M2 attenuator. Actually it was a rewire for another member (I won't mention his name) whose soldering was poor, at best and had burned up a couple of the switches, etc.

While a 16 ohm unit still sounds fine on a 16 ohm amp tap, into an 8 ohm speaker, works well and is perfectly safe, the exact same amp and speaker sound noticeably better and more dynamically responsive with the dedicated 8 ohm, M2, out of the 8 ohm amp tap. I just ordered the parts for an 8 ohm build!

I thought you'd all like to know?

Simply Sharin'
Gene

HI Gene, thanks and yes that is very interesting. Based on the numbers that I run, an 8 ohm speaker out of a 16 Ohm attenuator is in theory equivalent to about 1 to 1.5 db less mid volume as compared to presence and resonance frequencies, if set the same. So it could be that. Or, it could be slight difference between M and M2 designs, or just the amp sounding different on different taps.

It supports the idea that the best version to build is where amp, attenuator and speaker all match Ohms for the main rig in use.

A slight tone tweak on the 16 Ohm build to make the theoretical tone into an 8 speaker match better would be a 39 or 56 Ohm resistor switched in parallel across the speaker.

But, there's magic involved in these things and the theory only captures some if it!

 

[JohnH]

By now, you would think this thread would be a sticky.

I have asked for that before, but no reply was received

 

[Gene Ballzz]

HI Gene, thanks and yes that is very interesting. Based on the numbers that I run, an 8 ohm speaker out of a 16 Ohm attenuator is in theory equivalent to about 1 to 1.5 db less mid volume as compared to presence and resonance frequencies, if set the same. So it could be that. Or, it could be slight difference between M and M2 designs, or just the amp sounding different on different taps.

It supports the idea that the best version to build is where amp, attenuator and speaker all match Ohms for the main rig in use.

A slight tone tweak on the 16 Ohm build to make the theoretical tone into an 8 speaker match better would be a 39 or 56 Ohm resistor switched in parallel across the speaker.

But, there's magic involved in these things and the theory only captures some if it!

Yes indeed John, it appears that it is in the mids that a difference is heard. I'm surprised that in the 1db to 1.5db range it would be so audible, but you know us tone chasers! What is subtle to others is a whole universe of difference to some! That and the fact that the mid range is where we shape most of our tone, etc.

For example, there are several signature, feedback notes (from specific songs) that I seek to achieve with certain guitars and no pedals. Certain speaker setups and responses make them difficult to achieve, while others just tend to do them naturally. Certain tonal differences can make or break such nuances!

Thanks Again,
Gene

 

[Lancer X]

hi @Lancer X , it's ok to operate the main stage switches -14, -7.5 and -3.5db 'on the fly'. There is enough permanent continuity through the circuit to control this when any one switch is operated in a particular moment of time. The foot switch option is the same and it's also true of the -7/-3 5 switch on stage 1 shown recently.

Thd only switch I think should not be operated unless the amp is off or on bypass is the 'full bypass' switch, if you have one. It's a good reason not to have one IMO, so I don't draw it on the base diagrams. I don't think the bypass switches on commercial units work any differently really. The issue is fly-back voltages back into the amp. I know enough to be aware of a possible risk but not enough to say it isn't one!

Oh! Cool, thanks everyone. I must have been thinking of a message where someone cautioned against hot-switching the bypass.

Glad to hear this. I'm thinking about building this circuit right into my next amp build, so very cool that hot switching is okay.

 

[mountainhick]

Never mind... It took a while but I found the answer to my question.

 

[BuckTennington]

I love you for designing this and I want to make this however I have no idea where to start on this thread to build it. I’m having trouble following this thread.

 

[Gene Ballzz]

I love you for designing this and I want to make this however I have no idea where to start on this thread to build it. I’m having trouble following this thread.

FIRST: to the forum!
NEXT: If you share what amp(s) you want o use it with and what bells and/or whistles you want to add, we can help you. And yes, this thread is pretty lengthy and cumbersome, although well worth the read, once skimming past some of the inevitable chaff. Lots of great testing and technical info as well as modifactions and additions to the original design, which by the way works fabulously in it's bare baones basic form, with very economical aspects. Depending on parts/components choice and sourcing, the price ranges from $75 to $150 for a basic unit. Obviously, that prices goes up a little, depending on what feature you may want to add. The price to build a basic unit is
THEN: All of the well updated component value and electrical info is in post #1 of this thread, along with links and/or page numbers to lead you to ther related info. Actual layout diagrams start on about page #110 or #111.
Let Us Know?
Gene

 

[stickyfinger]

Any way I can make a M2 that would work with a 16ohm amp and use a 8 ohm speaker?
I have a fully broken in 8 ohm Greenback that I would like to install in my DSL 5C that only has a 16 ohm tap.
Thanks for the help!

 

[JohnH]

Any way I can make a M2 that would work with a 16ohm amp and use a 8 ohm speaker?
I have a fully broken in 8 ohm Greenback that I would like to install in my DSL 5C that only has a 16 ohm tap.
Thanks for the help!

Yes a 16ohm M2 can use an 8 Ohm speaker with it. This is because the first -7db stage is always engaged, so the amp doesn't see the speaker directly. But it can't convert 16 to 8 as an unattanuated setting.

 

[DeluxeReverb]

I love you for designing this and I want to make this however I have no idea where to start on this thread to build it. I’m having trouble following this thread.

I was in a similar situation as you. I heard about this attenuator while researching other products such as the Weber MiniMass, but I was intimidated by the length of the thread. However, the first posts by @JohnH have been updated to include the latest circuit design and parts specs.

You’ll need to balance your budget with your intended build timeline, and be prepared to order parts from multiple suppliers, but your efforts will be well rewarded. In my case, I ordered from 3 US suppliers that I’ve used before, and while not the cheapest, they got me the parts quickly. I could’ve saved money if my timeline was more relaxed and if I’d used eBay sellers.

The updated circuit diagrams at the very front of the thread, combined with the layout diagrams on pg 111 (thereabouts) are the gold here. The rest is interesting - at least to me - but not essential.

DR

 

[JuliusJ]

Hello folks, this is my first post. JohnH, congratulations on this super project ! And Hello Gene Ballzz, i will greet you first ! Briefly, i have been learning tube amp technology for three decades. I am familiar with LTSpice, a great program. Years ago i was an avid reader at AX84 and the Music Electronics Forum. I think i need to inform the folks at MEF of the JohnH design !

I have been following this thread for months and have some things to contribute. When i write "you" i am mostly referring to JohnH. Sorry it is long but not too technical. I will be posting tech oriented stuff and observations.

1) I went to bed the other night with the first stage of the M2 design firmly in mind and went through all the circuit paths in both directions. The triangle with the shunt looked like a yield roadway sign, new and strange to me. Flatten out out the inductor horizontally ! Shortly, i got it ! I thought, eureka ! This stage has the same structure as a Bridged T attenuator. That is what i think has been derived here with a lot of time and hard work. The inductor is the left side of the T. R2A is the bridge resistor. R2B is the right side of the T. R1 is the shunt. Simple and intriguing. Maybe you knew this, but i wanted everyone to see this now, firstly because i thought it would save you the work and trouble of coming up with custom formulas since you mentioned you were looking into doing Delta Y transforms. The properties and formulas for the bridged T are well known. The formulas can be adapted for use with the inductor i am sure. My first thought about this structure was to try swapping the inductor position with R2B and see what the result is if any different or usefull. I believe the inductor is currently in the best position as we want the input impedance to be low and the output impedance to be high and close or equal to the output impedance of the amp. Second though, try an inductor on both sides of the T, test, and then parallel a resistor across both of them, and test. This looks more balanced, and food for thought. There are other attenuators we can look at like the balanced designs and figure out where to place the inductor(s), but those look more complicated.

2) Four attenuator units can be built and wired in series parallel like a 4x12 cabinet for use and analysis. This might help soak up the power for higher wattage amps, at -3db each unit, set to full load, this will give a -12db starting point, please confirm. Put that in parallel with another attenuator and speaker for fine loudness control.

Values for 32 ohm and 64 ohm attenuator units can be derived easily. Please see the excellent calculator here:

k7mem "dot" com/Res_Attenuator.html

I used this calculator and easily derived all the resistor values listed for the M2. Since we want to match the amp output impedance, select Match Z1, very important ! For an 8 ohm tap, i used 18.6 ohms as the source impedance Z1 and derived the same resistor values listed for design M2. For a 16 ohm tap i used 37.2 ohms as the amp output impedance to get your numbers. The calculator also shows what standard values come close. Please try out to confirm i am doing it right. it's fun and addictive ! I think this calculator is a miracle, just what i was looking for last night, and i am so glad i did not give up searching. Just what we need for our purposes. Custom attenuation values can be found.

3) The type 2 L pad you discovered at the beginning is also quite a eureka. At first i did not think using the online calculator would give me the correct results, because it is the regular series + shunt voltage divider, in which the input impedance is the sum of the two resistors, and the output impedance is the parallel formula. I am looking to confirm the input and output impedance formulas for the shunt + series divider arrangement. I think it might be just the opposite of the series + shunt voltage divider.

Some notes:

4) Looking back at Design G post #111 we can derive the bridged T structure by combining L1 and L2, and eliminating R2.

5) On the Aiken design a noted error, R11 at 16 ohms was copied on post #92. i noticed that was changed to 30 ohms, on post #489, good.

Ok that's it for now, thanks and talk to you all later.

 

[JohnH]

Hi @JuliusJ , welcome to our thread and thanks for all your thoughts!. That attenuator calc page is very smart. the only other ones I've seen just seem to do matched impedances.

At this point I do it all on a huge excel sheet, everything is coded into it, including the speaker complex impedance vs frequency. To test a new design, i put in the values I want to try and hit a button, and then it whirs its macros for a minute or so and plots every frequency at every attenuation step with 4, 8 and 16 ohm speakers. The Star/delta transforms are built in too. So that's all pretty handy! I have smaller routines for working out T groups to use for each stage.

But really - the thinking and research went something like this:

Aiken shows us how a speaker impedance can be modelled out of analyzable components - and Mike Lind posted an accurate measured plot to work to
An amp has a higher output Z that I thought, and it is not too different to being as if its resistive at most levels (a discussion I had on TGP)
Amp needs to see the right reactive load to respond correctly, and speaker needs to see a highish resistive load for the same reason
But, if the attenuator shows the amp an inductance that is all in series, then that inductance develops most of the treble, leaving a dull sound to go through. But if its all in parallel then the attenuator sees too bright a tonal balance. So Stage 1is a mix of inductance in series and parallel, which started with two inductors in Design M, but then developed into one combined inductor, in a series/parallel position for Design M2. M2 creates a tricky circuit to calculate using voltage dividers, but the maths of the Star/Delta transform turns it back into a simpler to calculate arrangement
After Stage 1, the amp is seeing the right load but the signal through intermediate stages is almost at nil phase change, and can travel happily through the series of resistive stages before interacting with the complex speaker impedance

 

[stickyfinger]

Yes a 16ohm M2 can use an 8 Ohm speaker with it. This is because the first -7db stage is always engaged, so the amp doesn't see the speaker directly. But it can't convert 16 to 8 as an unattanuated setting.

Thanks John! I really appreciate all the time and effort you put into this design and helping all of us.

I have a Rockcrusher I use for my big Marshalls and every now and then I forget to switch the ohm setting on the unit. 8ohm head going into 16 ohm attenuator and then into a 8 ohm cab and the tone changes and vol drops and I usually notice it right away.

I assume I can expect the tone to change using a 16 ohm M2 into a 8 ohm speaker?

 

[JohnH]

Hi @stickyfinger , if you use an 8Ohm speaker in a 16Ohm M2, the amp is fine and sees almost no difference, the speaker tone will have about 1 to 1,5db more resonance and presence, in theory., very much like turning up those two controls , if you had them. A small change that you may or may not hear.

Going the other way, a 16 speaker into an 8 ohm M2, there's a bit more of a difference in reduction of those frequencies. I can hear it, though its still safe for the amp . So for the 8 Ohm design I added Out 3 which compensates for this small tone change.

 

[Browneyesound]

So for a 100W version of the M2, you should double the watt rating of the resistors?