Simple Attenuators - Design And Testing

[auflauf]

John, you are right. I imported the audio files from the recording at 0db and at -20db into REW and did a spectral analysis and smoothened rather drastically (1/6 oct). Then I superimposed the curves. Result: no discernible loss of heights. The 3-4db difference of the curves might arise from the sloppy and complicated workflow (mic sensitivity and output level set with non-calibrated sliders of the PC operating system, recording with a small program "Piezo", importing mp3 into Audacity, converting to wav, importing wav into REW).



Still, the perceived dullness remains.

So:
Attenuation level with the M2 attenuator apparently has no influence on frequency response if tested via
-frequency sweeps (post of 4. Oct 21)
-white noise (above)
-clean bright chords (above)

The frequency response tested included the chain of amplifier (input to end stage), speaker, cabinet and room.

What has not been tested was the influence of the attenuator on the guitar (highly unlikely) and the influence of level/attenuator/attenuation on human perception.
I guess it's psychoacustics, all the more as the sound level at the recording was rather low and in no way comparable to a rehersal room or a gig. Referring to a crude estimate from iPhone apps, the level of the chord playback was in the range from 65 to 75db unattenuated, thus 45 to 55db attenuated, which is lower than normal conversation level.

 

[JohnH]

That all looks right, thanks for following through.

I cases where a perceived loss of treble at very low volume becomes something that a user wants to adjust for, I think the capacitor across R2A is an option and easy to add. As I noted above, a 3 to 10 uF bipolar cap could be tried, subject to testing.

 

[auflauf]

Let me know if you want to measure my M3. I have a gig that I need it for in 2 weeks but after that I could send it to you as long as you promise to send it back.

Hello dbidshopbliss! If you are asking me: thank you for the offer. That would be very interesting. But I guess detecting differences between the M2 and M3 design requires a/b listening comparisons as well as more reliable measurements than mine done so far. It’s iefes‘ M2, so maybe the att willl be back at the owner in the coming weeks. And if both designs don‘t hamper the frequency response, what to look for and how to measure that? Amp-speaker interaction, dynamics, response to transients, all that quantitatively...I am not negative here, but this may require more than just semi-automatically measuring FR. I am having a few days of now, but will be busy the coming weeks. let‘s discuss this when you did your gig, ok?
Are you located in Germany in order to avoid international shipping?

 

[JohnH]

If you are up for more tests, there's more that might be investigated with the M2. It depends if you can get loud and drive the the amp hard, at least for the duration if one loud strum! , without overloading your recording system.

The reason is that the output impedance of a valve amp changes dynamically when the power amp section drives hard. The ability of the attenuator to respond to this change is the main reason for a reactive design. My tests are done at high amp volume, and it would be very interesting to compare with the different amp that you are using.

 

[dbishopbliss]

Hello dbidshopbliss! If you are asking me: thank you for the offer...
Are you located in Germany in order to avoid international shipping?

I'm in the US so probably not cost efficient to send to DE. I built an M2 and an M3... it might be more effective to help me figure out how to make the measurements you did so that I can post them.

 

[JohnH]

I'm in the US so probably not cost efficient to send to DE. I built an M2 and an M3... it might be more effective to help me figure out how to make the measurements you did so that I can post them.

It would be great to get some readings on your two boxes. We can do useful science in a range of different ways depending on what facilities you have. @auflauf is using some special tools and producing very nice plots. i might see what i can do with that REW software, which looks to be free in is basic version.

But, you can also work with just general musician gear. The reason is, we don't really need to know about absolute values of frequency response and SPL's etc, everything we'd like to know is relative, looking for the changes at different settings and comparing apples to apples, rather than accurately measuring a specific apple.

Here's how I do it.

Set up the amp, and close-mic it as if you were recording a track. Set it loud, so the power amp stage is driving
Record a few seconds of playing. make sure the strings are ringing cleanly. Open chords are best with all the strings, or a sequence of notes at each semitone up the first few frets on the low strings to explore bass response.
I use a delay pedal in my loop, set as a looper. Its very important to have consistent input as you explore settings, even though its not too important exactly what that input is.
Then with this annoying loop playing, record into the computer. I use a dynamic mic into a small mixer, then into Audacity.
Start at full volume, no attenuator, recording as high as you can but no clipping
Keep everything in place and the loop running, put the amp to bypass and put in the attenuator
record, stepping down through the attenuation settings

At that point, you could post a wav file and Id be very happy to post process.

But, eg with Audacity, you can then select a part, and do a frequency response plot, and export that to a csv file and so into excel. These plots are very spiky, so I smooth using a moving maximum to find the greatest level within a few steps either side of the frequency

Then plot results

For your M2 and M3 comparison, it will be interesting to compare particularly the low bass notes around 100 hz.


If you have means of doing a frequency sweep or a white noise input, that is also very good and tests other aspects. Probably keep that one at a clean level of amp drive. You can download wav files of these and play them from a phone or mp3 player.

If you have different amps, interesting to compare them, particularly if they have different output stages.


If anyone else who has built one of these would like to try something like this, Id be very grateful and it would be a big help to this project.

 

[auflauf]

Hello!
My setup isn‘t really sophisticated. The only non-ordinary piece is the relatively cheap measurement microphone. I bought it for testing and adjusting room response for my Hi-Fi-chain (which I never did). As John said: as we don‘t require absolute, but relative measurements, any decent microphone will do.
The free REW-software offers all features you might desire: frequency response as posted, elegant possibilities to smoothen the curves, waterfall analysis and so on. Room analysis is only one of the features. It might seem overwhelming at first, but once you understood where to start and where the preferences/options (divided into genreal and measurement-related, called „control“), its really easy. You just have to play around a bit to get acquainted at first. If you want to analyze the spectrum of a guitar signal or white noise instead of a frequency sweep, you have to offer them to REW as wav files.

Edit 15.Oct:

• If you want to do frequency sweep measurements (this is recommended for room analysis, it probably won't harm for frequency response measurements as I did): get REW, its free.
  
  
• Mic, sound interface
• • Get a microphone. You may even try a headset's mic, but sensitivity for low frequencies may be low. For a start, you may even try the built-in mic of a laptop.
  • Plug it into the combined headset jack at your PC/laptop.
  • If you are using a dynamic mic or a measurement mic, you may need an USB interface audio to USB. Min costed less than 20$. You may need a phantom power supply for electret microphones. I think a laptop plug provides phantom voltage, but (IIRC) it's suited for headsets, not for standard mics/measurement mics. I needed a phantom voltage source, it costed me 20-30$
  • Position the microphone in front of the amps speaker. Standard for hifi speakers is 1m. Standard for room measurements is your listening position. Standard for miking an amp is very close. As long as you keep the distance constant and ≤1m, it probably doesn't matter
  • Connect the output of the usb interface (or your headphone output) to the guitar amp)
  • I did not have success with splitters which go from combined mic/headphone jack to separate mic and phones jacks. There are two different connection schemes for mobile phones and... it was a headache, so I prefer the cheap usb interface.
    
• REW, setup
  • Open REW (I now refer to the Mac version, the windows version may be similar).
  • Preferences in the upper right corner of the std. window: select the usb device or whatever you want as your input mic. Select your output to the amp.
  • Slide mci sensitivity and output volume up high in REW and in the operating system. They are synchronized in a Mac and probably in a Windows device too.
  • Preferences/lower part "Levels". Select "main speaker", not "subwoofer".
  • Click "Measurements" in the upper left corner of the main window.
  • Name your measurment session.
  • Start to end 20-20000Hz is fine.
  • Length of the sweep 512k or 1M will do. For increased precision, you may select repetitions. I didn't
  • No need for acoustic timing reference, but it doesn't hurt.
  • Check levels until you and REW are satisfied
• REW, measurements
  • You may do all your measurements at low bedroom levels as I did and additionally at higher levels as John did.
  • Start measurement. You can then repeat measurements at different attenuation levels, they appear as different curves in one session.
  • if you want to analyze guitar chords, white noise etc, which is not built into REW, you need to play them and record them via some other software. I used a small music chain as source and recorded via audacity. You then convert/store the recordings as wav-files. You open them in REW. Their analysis is the same as written below for the built in frequency sweeps in REW
• REW, analysis
  • Click "all spl" in the window showing the measurement results. You can deselect curves in the lower pane.
  • Apply smoothing (menu strip at the top: "graph"), 1/12 may be a compromise.
  • use "controls" in the upper right corner of the results window: play with trace arithmetics etc, there you can add or subtract db in order to move curves up or down in the graph. Smoothing via menus/graph affects all curves, using "control" affects only the selected curve.
  • play with all the options in the measurement results window, eg. distortion which shows you among others the noise floor.
  • I exported the result graphs via screenshots. This is not the most standardized way, but I didn't find pixel graph export.

That‘s it, feel free to ask. To quote Nike: Just do it.

 

[Graham G]

ok great. Once you've concluded, you might be interested in putting together a simple, smaller, single box to do the same. Without all the attenuation settings, it can be distilled down to:

Output from amp 16ohms
standard speaker cable
box with one jack in, one jack out and three components (two resistors and a coil)
plug 8 ohm speaker into that.

Power is reduced -3db to 25W, as with the 3-way cable system.

Hi, John if you have the time I want to build the simpler Attenuator that you've described above, the M2 I've built & rewired to a fixed output then used as a load box is an ideal gigging volume for me & it allows me to get my Amp(HF modded ORI50) in it's "sweet spot" & for me the Amp retains all it's "tone & feel" without providing everybody with ear plugs.
I'd like to thank you again,the M2 as i'm using it has helped(along with the Headfirst Mods), to give me the best gigging sound I've had for years, maybe ever(& ever is a long time for me) with the possible exception of the days when I used to use 2 amps.
Thanks, Graham.

 

[JohnH]

Hi, John if you have the time I want to build the simpler Attenuator that you've described above, the M2 I've built & rewired to a fixed output then used as a load box is an ideal gigging volume for me & it allows me to get my Amp(HF modded ORI50) in it's "sweet spot" & for me the Amp retains all it's "tone & feel" without providing everybody with ear plugs.
I'd like to thank you again,the M2 as i'm using it has helped(along with the Headfirst Mods), to give me the best gigging sound I've had for years, maybe ever(& ever is a long time for me) with the possible exception of the days when I used to use 2 amps.
Thanks, Graham.

Sure no problem! Since we discussed before, I think there's more than one option.

1. The simplest is, you use your 8 ohm cab, plug it into this new box with three components in it, and plug that into the 16 Ohm amp tap. Its equivalent to how you've been using the M2 with the series lead.

2. There's also one with 4 components, where you use an 8 ohm cab and the 8 ohm amp tap.

3. Finally, a bit of reconfiguring, you can convert the full M2 to make the first stage switchable between 3.5 and 7 db, then switch the other stages off. Then you have all your settings in one box.

I reckon option 1 is simplest, unless you prefer 2. Either way, given you've figured out what you need, it will be a very simple reliable fixed unit.

 

[Graham G]

Sure no problem! Since we discussed before, I think there's more than one option.

1. The simplest is, you use your 8 ohm cab, plug it into this new box with three components in it, and plug that into the 16 Ohm amp tap. Its equivalent to how you've been using the M2 with the series lead.

2. There's also one with 4 components, where you use an 8 ohm cab and the 8 ohm amp tap.

3. Finally, a bit of reconfiguring, you can convert the full M2 to make the first stage switchable between 3.5 and 7 db, then switch the other stages off. Then you have all your settings in one box.

I reckon option 1 is simplest, unless you prefer 2. Either way, given you've figured out what you need, it will be a very simple reliable fixed unit.

.

Thanks John, i will be happy with the Option 1 fixed box, i only ever use the Amp for gigs(or practice) & really don't need it any lower than the way i'm using it now, could i ask a question, a Guitarist asked me at last nights gig could the way i'm using the M2 work for cutting down a 100w Amp ? & if so would it be about half the volume(as in the 50w) or would it be a different calculation ?.

 

[JohnH]

Hi @Graham G , ok option 1 then

With the 100W amp, yes the same system could work, but would need to be rated for the higher power. 100W will go down to 50W. But, I think most people would perceive that as a much smaller reduction in volume than 50%. While all the electrical numbers have very well defined meanings, perceived volume is a much more subjective parameter. I have heard it said that a 1/10th factor on power is about a halving of volume., ie -10db. Personally, I find our -7db setting, which is 1/5th power, is a good downwatd step, taking 100W to 20W.

 

[JohnH]

Hi @Graham G , here is the stripped back design for the series load box.



It will absorb half the amp power (mostly in R2), so in theory your 50W amp will send 25W through to the speaker.

It's likely to heat up, so a good aluminium box, thermal paste, ventilation, feet, etc are needed. The usual watch-its about mounting the coil apply - no steel bolts, raise it a bit off the case.

So as with your recent use of the M2 in this way, the box is intended to be similar to an 8 Ohm speaker, in series with a real 8 ohm speaker, to add up to a 16 Ohm load and you use the 16 Ohm amp tap.

There's a nice bonus with such a simple design. Although the diagram shows the 8 Ohm resistor on the speaker side, its actually reversible. Either jack can be the input or the output, so long as its connecting to the 16 Ohm amp tap and the 8 ohm speaker.

 

[Graham G]

Hi @Graham G , here is the stripped back design for the series load box.

View attachment 96202

It will absorb half the amp power (mostly in R2), so in theory your 50W amp will send 25W through to the speaker.

It's likely to heat up, so a good aluminium box, thermal paste, ventilation, feet, etc are needed. The usual watch-its about mounting the coil apply - no steel bolts, raise it a bit off the case.

So as with your recent use of the M2 in this way, the box is intended to be similar to an 8 Ohm speaker, in series with a real 8 ohm speaker, to add up to a 16 Ohm load and you use the 16 Ohm amp tap.

There's a nice bonus with such a simple design. Although the diagram shows the 8 Ohm resistor on the speaker side, its actually reversible. Either jack can be the input or the output, so long as its connecting to the 16 Ohm amp tap and the 8 ohm speaker.

John this is great , i'll order the components & build it this week .
Any progress on that Button we've discussed
Regards, Graham.

 

[fperezroig]

Hello! First of all thank you for such a great design and comprehensive thread. This is a really big contribution to the diy community. I'm about to order the components and build the M2 version of the attenuator but am still a little confused about the way I could connect a 4ohm cab to it.

Could I just halve the resistors and inductor values of the 8ohm version, and then connect the 4ohm amp tap through the attenuator to the 4ohm cab? What should be the values for R10 and R11 to get a third output for 8ohm cab with the proposed 4ohm version ?

Kind regards.
Fran

 

[diego_cl]

I'm currently building the M2 (16 ohms version) attenuator.

I can't understand how the 2nd speaker output works on the 16 ohms version.

With the 8 ohms version I suppose it works like this:

Input: 8 ohms amp tap
Output Speaker 1: 8 ohms speaker
Output Speaker 2: disconnected
Output Speaker 3: disconnected

Input: 8 ohms amp tap
Output Speaker 1: 16 ohms speaker
Output Speaker 2: 16 ohms speaker
Output Speaker 3: disconnected

Input: 8 ohms amp tap
Output Speaker 1: disconnected
Output Speaker 2: disconnected
Output Speaker 3: 16 ohms speaker

- Is the 2nd output in parallel?
- What's the use of the 2nd speaker output on the M2 16 ohms version?

With the 16 ohms version I suppose it works like this:

Input: 16 ohms amp tap
Output Speaker 1: 16 ohm speaker
Output Speaker 2: disconnected

Input: 16 ohms amp tap
Output Speaker 1: 32 ohms speakers ???
Output Speaker 2: 32 ohms speakers ???

I suppose that the 16 ohms version will work with a 8 ohms speaker as long as the reactive load is running, but what is the best use of its 2nd output?

 

[JohnH]

Hello! First of all thank you for such a great design and comprehensive thread. This is a really big contribution to the diy community. I'm about to order the components and build the M2 version of the attenuator but am still a little confused about the way I could connect a 4ohm cab to it.

Could I just halve the resistors and inductor values of the 8ohm version, and then connect the 4ohm amp tap through the attenuator to the 4ohm cab? What should be the values for R10 and R11 to get a third output for 8ohm cab with the proposed 4ohm version ?

Kind regards.
Fran

Hi Fran, thanks for your interest in this design.

The attenuator is safe to use with different cab Ohms, so long as the amp tap Ohms matches the design.

With a 4 ohm cab you can either build an 8 Ohm M2 and use it from an 8 Ohm amp tap and plug the 4 Ohm cab into it. Or, halve all the values and then you have a 4 Ohm M2 to use with a 4 Ohm amp and cab.

If you use the 4 cab with 8 M2, you lose a little more volume, about 2db more reduction at each setting, also the tone is very slightly brighter, which if you notice it at all, can be adjusted for at the amp.

But I think that if there is one main intended use for the unit, then its often best to build for that, and the flexibility for other cabs is a bonus. When halving all the resistor values and the inductor, sometimes there is not an exact value available in a common range. Try o keep the ratios between the resistors at each stage close to the design. eg, for the first stage, using standard values, R1=8.2, R2A=12, R2B=10. These are all a bit higher than 1/2x, but in proportion.

 

[JohnH]

hi @diego_cl , thanks for your message.

See above to Fran. The two outputs Out 1 and Out 2 are wired in parallel. With a 16 Ohm M2, they let you use it for two 16 Ohm cabs to make an 8 Ohm load if you want to. Still run it from a 16 Ohm amp tap though.

 

[dbishopbliss]

Hello!
My setup isn‘t really sophisticated. The only non-ordinary piece is the relatively cheap measurement microphone. I bought it for testing and adjusting room response for my Hi-Fi-chain (which I never did). As John said: as we don‘t require absolute, but relative measurements, any decent microphone will do.
The free REW-software offers all features you might desire: frequency response as posted, elegant possibilities to smoothen the curves, waterfall analysis and so on. Room analysis is only one of the features. It might seem overwhelming at first, but once you understood where to start and where the preferences/options (divided into genreal and measurement-related, called „control“), its really easy. You just have to play around a bit to get acquainted at first. If you want to analyze the spectrum of a guitar signal or white noise instead of a frequency sweep, you have to offer them to REW as wav files.

Edit 15.Oct:

• If you want to do frequency sweep measurements (this is recommended for room analysis, it probably won't harm for frequency response measurements as I did): get REW, its free.
  
  
• Mic, sound interface
• • Get a microphone. You may even try a headset's mic, but sensitivity for low frequencies may be low. For a start, you may even try the built-in mic of a laptop.
  • Plug it into the combined headset jack at your PC/laptop.
  • If you are using a dynamic mic or a measurement mic, you may need an USB interface audio to USB. Min costed less than 20$. You may need a phantom power supply for electret microphones. I think a laptop plug provides phantom voltage, but (IIRC) it's suited for headsets, not for standard mics/measurement mics. I needed a phantom voltage source, it costed me 20-30$
  • Position the microphone in front of the amps speaker. Standard for hifi speakers is 1m. Standard for room measurements is your listening position. Standard for miking an amp is very close. As long as you keep the distance constant and ≤1m, it probably doesn't matter
  • Connect the output of the usb interface (or your headphone output) to the guitar amp)
  • I did not have success with splitters which go from combined mic/headphone jack to separate mic and phones jacks. There are two different connection schemes for mobile phones and... it was a headache, so I prefer the cheap usb interface.
• REW, setup
  • Open REW (I now refer to the Mac version, the windows version may be similar).
  • Preferences in the upper right corner of the std. window: select the usb device or whatever you want as your input mic. Select your output to the amp.
  • Slide mci sensitivity and output volume up high in REW and in the operating system. They are synchronized in a Mac and probably in a Windows device too.
  • Preferences/lower part "Levels". Select "main speaker", not "subwoofer".
  • Click "Measurements" in the upper left corner of the main window.
  • Name your measurment session.
  • Start to end 20-20000Hz is fine.
  • Length of the sweep 512k or 1M will do. For increased precision, you may select repetitions. I didn't
  • No need for acoustic timing reference, but it doesn't hurt.
  • Check levels until you and REW are satisfied
• REW, measurements
  • You may do all your measurements at low bedroom levels as I did and additionally at higher levels as John did.
  • Start measurement. You can then repeat measurements at different attenuation levels, they appear as different curves in one session.
  • if you want to analyze guitar chords, white noise etc, which is not built into REW, you need to play them and record them via some other software. I used a small music chain as source and recorded via audacity. You then convert/store the recordings as wav-files. You open them in REW. Their analysis is the same as written below for the built in frequency sweeps in REW
• REW, analysis
  • Click "all spl" in the window showing the measurement results. You can deselect curves in the lower pane.
  • Apply smoothing (menu strip at the top: "graph"), 1/12 may be a compromise.
  • use "controls" in the upper right corner of the results window: play with trace arithmetics etc, there you can add or subtract db in order to move curves up or down in the graph. Smoothing via menus/graph affects all curves, using "control" affects only the selected curve.
  • play with all the options in the measurement results window, eg. distortion which shows you among others the noise floor.
  • I exported the result graphs via screenshots. This is not the most standardized way, but I didn't find pixel graph export.

That‘s it, feel free to ask. To quote Nike: Just do it.

My M3 is at the rehearsal space so I will need to pick it up on Saturday. I have an audio interface and one of those electet microphones from Parts Express somewhere. I will see what I can figure out.

 

[dbishopbliss]

Thats great news!
In principle, the footswitch stage can be any amount of dB's in the range. It's easy to work them out if you want an in between one.

Also, It would be possible to add another switch to the footswitch box to change the footswitch dB's.

But I'd suggest to experiment with the main box switches, particularly the -7 db switch, to help decide on values.

I haven't been keeping up with the entire thread much since I finished mine, but I was just thinking it would be nice to "boost" the volume of my amp as opposed to boosting the signal going into my amp and the footpedal idea seems like it will meet my needs.

My amp is already pretty dirty (Supro Coronado Clone) and adding a boost pedal really just makes things more dirty and compressed as opposed to bringing up the volume.

Could there be different footpedals that just get plugged in? That is... could I make a 3.5dB pedal and a 7dB pedal and then just use the one I feel like using depending upon the gig?

 

[JohnH]

I haven't been keeping up with the entire thread much since I finished mine, but I was just thinking it would be nice to "boost" the volume of my amp as opposed to boosting the signal going into my amp and the footpedal idea seems like it will meet my needs.

My amp is already pretty dirty (Supro Coronado Clone) and adding a boost pedal really just makes things more dirty and compressed as opposed to bringing up the volume.

Could there be different footpedals that just get plugged in? That is... could I make a 3.5dB pedal and a 7dB pedal and then just use the one I feel like using depending upon the gig?

Yes no problem. You can use the resistor values for the 3.5 or 7 db stages in the footswitches.

Or, you can also make an in-between value, like -5db (if anyone wants that I'll work it out)

Or, in one footswitch, you could add another -3.5 stage and another switch (could be a second footswitch? or maybe a toggle or slide?) so the overall footswitch box is selectable to go 0db to -3.5db, or 0db to -7db.