Simple Attenuators - Design And Testing

Discussion in 'The Workbench' started by JohnH, Oct 21, 2017.

  1. fperezroig

    fperezroig New Member

    Joined:
    Oct 19, 2021
    Messages:
    4
    Likes Received:
    8
    Thank you John, I will halve all values and build a 4ohm version then. In order to use the M2 4 ohm using the 4ohm ot and an 8ohm cab, could I add R10=34ohm and R11=5ohm? Can’t wait to order and star building this! Best regards
     
  2. JohnH

    JohnH Well-Known Member

    Joined:
    May 31, 2008
    Messages:
    4,695
    Likes Received:
    2,777
    Location:
    Wilton NSW
    That's the right idea, but you can go to the nearest standard values eg 33 and 4.7 if they dont have a 5


    Here's a suggestion for a full set for a 4 Ohm M2, assuming its from a brand make like Arcol:

    R1 8 or 8.2
    R2A 12
    R2B 10
    R3 8 or 8.2
    R4 5.6
    R5 2.7
    R6 8 or 8.2
    R7 18
    R8 2.7
    R9 not used
    R10 33
    R11 4.7

    It also works fine shipping from China, much cheaper but much longer wait and a different range of values. Im happy to review if you like
     
  3. fperezroig

    fperezroig New Member

    Joined:
    Oct 19, 2021
    Messages:
    4
    Likes Received:
    8
    Thanks, I’m ordering the resistors from a big supplier, and they offer all these values. Hope to get this finished ASAP!
     
  4. auflauf

    auflauf New Member

    Joined:
    Sep 27, 2021
    Messages:
    11
    Likes Received:
    10
    I just measured treble caps in the M2 attenuator.

    My copy of an attenuator was lend to me by iefes (thanks again) and he allowed me to open it and to clamp some caps on. His attenuator is extremely well built, hats off, really!

    I tested caps at three positions:
    A. parallel to all resistors from input signal to output signal
    B. parallel to the voltage divisor resistor R2A
    C. parallel to the chain of resistors after the voltage divisor, that is after R2a to output signal, thus bridging R4 to R8

    Levels tested were 0db, -10db and -20db with cap values of 10µF and 4.7µF

    As I was a bit nervous about my provisorily clamped-on caps while plugged in a running tube amp, I hurried and only measured at 0db, -10db, -20db, but not at -30db and did only a very short listening test at -20db of the 4.7µF cap parallel to R2A (B.) versus no cap.

    Results
    Listening test
    The cap made the sound markedly brighter, the effect was even a bit too much for my taste and I preferred the non-cap-version. The sound characteristic was a bit harsh, edgy as the distorted harmonics were risen. As I played a bit louder, at living-room-level instead of bedroom level as before, I guess it really depends on the listening level more than anything else. I could imagine to like that effect if playing at lower levels or at -30db.

    Frequency response measurmenents
    First a confirmation that the frequency response is really independent of the attenuation level
    Aligned freq resp 23_10_21.jpg

    The effect of different cap positions:
    A. The cap parallel to input-output causes an extreme treble boost up to 17db. As one would expect, the effect depends on the attenuation level as low resistance values just bypass the cap. The effect is not useful for our intentions.
    10µF parallel EA.jpg


    B. The cap parallel to the voltage divisor resistor R2A gives a marked treble boost independent of the attenuation level, again this is as I would expect because the cap is parallel to a constant resistive value.
    10µF parallel R2A.jpg

    C. The cap parallel to the chain of resistors after the voltage divisor, from the end of R2A to output signal, gives less effect than pos. B above, but it increases with attenuation and produces some kind of level-dependent loudness which might be desirable.
    10µF parallel to R4-R8.jpg

    The 4.7µF starts boosting at higher frequencies than the 10µF cap. The latter can even be regarded (and might sound like) as a bass attenuator. Judging by the frequency response graph, 4.7 µF may probably the highest suitable capacity, less may be even better.
    4.7µF and 10µF parallel to R2A.jpg

    Bottom line: I like the M2 att. and I think I may build one. This would allow to play my Harley Benton Tube 15 at home just at the edge of distorsion. That means Gain 3-4, Vol 2. This may require -10db during the day and -20db in the evening and gain 6-7 and Vol 5 would require -30 db. Switchable attenuation levels should be 0db/bypass, -10db and -20db and a slight (adjustable?) treble boost with a switchable 4.7µF or 2µF cap in pos. B or C might be just the icing on the cake for late evening playing.

    Thanks John and iefes
     
    CO_Hoya, iefes and JohnH like this.
  5. JohnH

    JohnH Well-Known Member

    Joined:
    May 31, 2008
    Messages:
    4,695
    Likes Received:
    2,777
    Location:
    Wilton NSW
    Hi @auflauf , A great study as usual!

    Those measured plots are quite close to the equivalent in my SPICE analysis within about a db, once you allow that yours are measured off the speaker and so include the speaker response, while mine are just the electrical signal.

    If you do this, definitely across R2A is the right place.

    You can further adjust the effcct by adding a resistor in series with the cap. The lower traces on this graph are at -14db, a 4.7uF, 4.7uF in series with 12 Ohm (red trace) and no cap (lowest trace):

    M2 4-7uF 12 ohm.gif

    This one is similar using a 2.7uF cap and 22 Ohm:
    M2 2-7uF 22 ohm.gif

    BTW the upper curves are the signal at the amp, Full speaker aed also with the attenuator and added cap. You can see how the rise as seen by the amp drops at high frequency when the cap is switched in.
     
    Last edited: Oct 24, 2021
  6. CO_Hoya

    CO_Hoya New Member

    Joined:
    Jun 29, 2021
    Messages:
    7
    Likes Received:
    9
    I finally got around to wiring up (and troubleshooting) my version of John’s attenuator (I’m calling mine the M3V version).

    Glamour shot:
    [​IMG]

    With all of the bells and whistles, it comes in at a svelte 5.5 lbs (2.5 kg), and would make a good blunt weapon in a pinch.

    I’ve only tried it for a few minutes strumming some chords to confirm basic function, so I can’t contribute much to this thread in terms of the technical discussion above. I can say that the tone stays fairly well-preserved by my crap ears down until max atten (-31.5 dB). And at that level, my semi-hollow body was probably louder than the amp itself.

    I haven’t had a chance to try the line out, or taste test with the bass resonance circuit in and out (which ended up using 10mH / 300 uF for L and C). I expect at some point the bass LC will be hard-wired and I’ll repurpose that toggle or hole.

    Cheers to John and others for this thread. It was a fun project.
     
    Stephen H, donwagar, diego_cl and 3 others like this.
  7. JohnH

    JohnH Well-Known Member

    Joined:
    May 31, 2008
    Messages:
    4,695
    Likes Received:
    2,777
    Location:
    Wilton NSW
    Hi @CO_Hoya , thanks for posting that. It looks [email protected]$$!
    Looking forward to hearing more about it in due course.
     
  8. dbishopbliss

    dbishopbliss New Member

    Joined:
    May 24, 2021
    Messages:
    24
    Likes Received:
    8
    Being able to add a little sparkle for the late night heavy attenuation is a great feature. This may have to be included in my next build. I have the parts for another M2 and I'm pretty sure I have some 2uF caps in my drawer. I have a neighbor that wants one so I will give him my first M2 and keep the M2+ for me.
     
  9. CO_Hoya

    CO_Hoya New Member

    Joined:
    Jun 29, 2021
    Messages:
    7
    Likes Received:
    9
    Thanks, John.

    I have no update other than to post a few more pics showing off my work.

    Here are the guts as I was trying to figure out wire routing - you can see how I engineered myself into a mess which made circuit-tracing during troubleshooting a chore. I'd build into a bigger enclosure if I did it again.
    [​IMG]

    I milled some vents into the bottom of the box, not very well. There was a bit too much flex in the extrusion for that, and so I tried to clean up as well as I could. Thankfully, normally these are hidden from view. You can also just see the rear heatsink.
    [​IMG]

    Here's the rear - not very interesting. Note the failed attempt at routing a slot, which caused me to switch to a row of holes instead. Sigh.
    [​IMG]

    And here it is sitting on an amp to give some sense of scale.
    [​IMG]
     
    diego_cl likes this.
  10. JohnH

    JohnH Well-Known Member

    Joined:
    May 31, 2008
    Messages:
    4,695
    Likes Received:
    2,777
    Location:
    Wilton NSW
    Hi @CO_Hoya , I reckon you did an outstanding job getting that all together, especially given all the added features you got in there, and the fact that everyone's build is basically a prototype with respect to the layout and construction.

    I've also learned to wish I used a bigger case , even though they can be built compactly.

    Yours is the second M3 that we know of, after @dbishopbliss 's build. There's another may coming sometime soon on the TDPRI forum , where Bitsleftover is converting an M2 to an M3.

    The big question that we'd be very interested in your feedback on is, to what extent does the resonant circuit in M3 make a difference, as compared to M2?

    If it makes a difference, it is most likely to be noticeable when using the line-out to record, or feed a PA, or re-amp back into a cab. The latter is where Bitsleftover was finding a drop in bass since the solid-state re-amping doesnt let the speaker develop its own resonance, as M2 does when used passively.

    https://www.tdpri.com/threads/johndh-attenuator-build.1060083/page-6#post-11002872

    Also, with M3 used passively as an attenuator, you may get a few more harmonics on bass notes when driving the amp hard.
     
  11. diego_cl

    diego_cl New Member

    Joined:
    Jan 10, 2021
    Messages:
    20
    Likes Received:
    3
    I'll ruin the glamor of that beautiful M3V with my next question...

    Is there any problem placing the inductor pressed against a foam? (like static electricity)

    I didn't have nylon bolts, so I used some Old Spice deodorant parts :lol:
    Tried it just as a test, but it worked way better than I expected ... I'm just worried about that black foam


    01.jpg 02.jpg 03.jpg 04.jpg
     
  12. JohnH

    JohnH Well-Known Member

    Joined:
    May 31, 2008
    Messages:
    4,695
    Likes Received:
    2,777
    Location:
    Wilton NSW
    I'd never have thought of that! seems like a cool (and dry!) idea. I don't see any electrical issues with the foam. If the coil got very hot it could be an issue maybe, but usually the coil may get a bit warm at most, even if the resistors heat up more
     
    diego_cl likes this.
  13. malibu91

    malibu91 New Member

    Joined:
    Oct 23, 2021
    Messages:
    12
    Likes Received:
    9
    Hi All,

    I have spent the last week going through every page of this thread and am exhausted but excited to start my build.

    Firstly a great thanks to JohnH. My friend you are worth your weight in gold.
    I was close to pulling the trigger on a Tone King Ironman mini but couldn't justify the crazy price.

    I am going to build an M2 with bypass to tame my massive Princeton Reverb '65 Reissue.

    I am in Sydney and have ordered a Hammond 1590DBK case, resistors, and coil and am getting my shopping list for Jaycar as we speak.

    I just had one or two questions around a couple of items if you could be so kind as to offer some advice.


    1/ Toggle switches
    All DPDT. Which would you would recommend for the 4 x switches?

    Option 1:
    https://www.jaycar.com.au/dpdt-stan...2609d7c5ba833535597056cb270436&sort=relevance
    [​IMG]


    Option 2:
    https://www.jaycar.com.au/dpdt-mini...2609d7c5ba833535597056cb270436&sort=relevance
    [​IMG]



    Option 3:
    https://www.jaycar.com.au/dpdt-flat...2609d7c5ba833535597056cb270436&sort=relevance
    [​IMG]



    2/ Thermal compound

    How many grams is enough?
    Link to Jaycar options is below
    https://www.jaycar.com.au/search?te...SRFToken=a6cd3d46-5e81-4aa4-bff5-f8c85dfed4b5


    3/ Hookup wire

    Will this gauge work OK. I am intending to buy a metre or so in a few colours
    https://www.jaycar.com.au/red-heavy...os=5&queryId=db9880586b6243de1df02a7ff407bf3b

    I think that's about it. Again I really appreciate all the solid work that has gone into this attenuator!

    Cheers
    Jules
     
  14. JohnH

    JohnH Well-Known Member

    Joined:
    May 31, 2008
    Messages:
    4,695
    Likes Received:
    2,777
    Location:
    Wilton NSW
    Hi @malibu91

    I also based in Sydney (I live in Wilton south of Cambelltown). So Ive been through Jaycar to get everything that they have applicable to our project

    Ive tried all those switches. My own build uses the Option 1 standard toggles. Its OK but I dont like them much, they feel clunky and Ive had one fail internally for no good reason. My next build will use the dpdt mini-toggles with the flat levers, option 3 on your list. Although these are a small units and mount in just a 6.5mm hole, they have a nice solid but well engineered feel. They have surprisingly high current rating, 5A at 125V. The lugs are small though, and you probably cant get the 18 gage wire through them, so have to do a good close contact and well-flowed solder joint. I used them in a earlier prototype.

    On the heatsink paste, just one of the small 1.5g or 3g tubes will be fine. Its just a very thin smear that is needed, across the face contact. Actually you wont get much heat with the 15W amp.

    That hookup wire is fine.

    Get the plastic stereo jacks. Jaycar also has the Aluminium diecast cases. Also 10mm M3 nuts and bolts to mount the resistors

    Where did you order your coil?

    Good luck with your build and I hope you like it!
     
    Last edited: Oct 27, 2021
  15. malibu91

    malibu91 New Member

    Joined:
    Oct 23, 2021
    Messages:
    12
    Likes Received:
    9
    Many thanks for the advice John.

    Looks like it’s gonna be some warm days coming up for us in Sydney so I’m looking forward to that!

    I’m in the Hills District myself

    I’m pretty competent with a soldering iron, so I’m comfortable soldering onto small terminals on the suggested toggle switch.

    I managed to pick up a coil from Wagner locally on eBay. Hopefully it comes on some sort of spool, otherwise I’ll knock something up and use a non steel bolt.


    3E1F8D29-D01D-4EA1-9FA5-C8E4783DCC6D.jpeg

    The box and resistors I ordered from RS Australia, everything else I’ll get from Jaycar or Bunnings including the plastic stereo jacks.

    Now I have all the information I need I can’t wait to get cracking and finally crank up my Princeton.

    Thanks again
    Jules
     
    JohnH likes this.
  16. Gene Ballzz

    Gene Ballzz Well-Known Member

    Joined:
    Feb 19, 2013
    Messages:
    2,986
    Likes Received:
    2,962
    Location:
    Las Vegas, NV
    @malibu91 & @JohnH ,
    Ya know, at face value it seems almost comical to think about and hear phrases like "I need I can’t wait to get cracking and finally crank up my Princeton." What makes it truly funny is how well these units actually work for dialing in/taming even little amps like this into their sweet spots without ending up in jail or divorce court, as well as finding the perfect volume blend for use with a band! These attenuators are the greatest thing since sliced bread! :cheers:

    Just My :2c: & Thanks Again To JohnH,
    Gene
     
    Stephen H and JohnH like this.
  17. malibu91

    malibu91 New Member

    Joined:
    Oct 23, 2021
    Messages:
    12
    Likes Received:
    9
    I was hoping somebody would enjoy that comment!

    Don’t worry Princetons can get plenty loud when pushed just ask Mike Campbell.

    Mine’s a lounge lizard and never gets past 3. Thus this attenuator should be very, very good fun for me.

    My Mesa/Boogie 295 Stereo Simulclass has been banished to the garage lol. It has little in the way of manners.
     
  18. Gene Ballzz

    Gene Ballzz Well-Known Member

    Joined:
    Feb 19, 2013
    Messages:
    2,986
    Likes Received:
    2,962
    Location:
    Las Vegas, NV
    And speaking of manners, where are mine? Forgetting to :welcome: you to this forum and this fantastic thread is not typical of folks around here! Kudos for reading the whole thread. While there's a little chaff here and there, the thread contains some fabulous info, along with some gorgeous builds, like that recent great looking one by @CO_Hoya ! Hats off to him on that one!

    This is a really fun place with a lot of truly knowledgeable folks. There's even a really nice gentleman who was involved in the building of the very first Marshalls and some that currently work and design there that stop by occasionally! :D Once you get this attenuator under your belt, some of us will likely get you to build your amp, maybe even from scratch! :p
    Just Sayin'
    Gene
     
    Last edited: Oct 30, 2021
    CO_Hoya likes this.
  19. telesto

    telesto Member

    Joined:
    Apr 29, 2020
    Messages:
    65
    Likes Received:
    8
    Hi John and all,
    I haven't looked at this thread in a year or two, good to see it alive and kickin :) And nice to see someone else using REW. Cool tool. I just discovered it recently and was playing around with it testing some attenuator designs I'm working on, and figured I'd test it on some others as well. I did impedance plot comparisons of John's staged-pad style design vs. a rotary Lpad (*note: I didn't use any inductors, to keep it purely resistive) using the freq sweep generator and default -6dBFS setting in REW. I measured the voltage output to the speaker of John's, and then dialed in the same for the Lpad, so they both had exactly the same output to the speaker. Speaker was a 12" Randall, but that doesn't matter.

    At -14dB John's has a higher and better defined bass-peak, and there is a noticeable increase in the high-freqs over 4kHz. I would guess the higher-highs and lower-lows is the reason it has a little more "rounded" sound, and the Lpad having the lower bass-peak and lower highs, has a more mid-range focus (or a little more "flat"). Altho to be fair, the plots aren't too far from each other.

    14dB-JohnH-vs-Lpad.jpg

    Testing at -21dB, they both start getting flattened, altho John's keeps a better bass-peak and slightly higher response over 10kHz. But from other testings, I see most attenuators start going flat around mid-20dBs and below. Once the signal get's so heavily attenuated, I don't think any design is going to save it, it just starts going to a flat line in terms of impedance plots, and sound to an extent as well ...well ok, not too bad, I play mine at bedroom levels and am fine with it, but at that point, it might be a good idea to look into Aiken's reactive-load box and play into headphones if you're a golden-ear kinda guy ;)
    21dB-JohnH-vs-Lpad.jpg

    Looking back at the very first page of John's SPICE simulations, the plots all seem to look exactly the same down to -30dB, which makes me wonder is that really accurate after looking at real-world measurements (?). I ran my measurements twice to make sure I didn't make a mistake, there were some small variations, but generally they looked similar. Smoothing and zooming in could also change plot appearances a little, but the general trend was always the same. If anyone else wants to run some impedance plots in REW for the full spectrum of 7dB+7dB+etc down to -30dB, it would be interesting to see and compare with my measurements. Also with the inductor in and out to compare the effects. I didn't use the inductor and just used a series of two and three 7dB stages.

    For the Lpad vs stepped-bridge attenuator discussion, I think there is always a trade-off. The bridged design gives a slight advantage with the better freq response, but the you loose the flexability and simplicity of the Lpad design. And vice-versa, Lpad lets you dial in exact level of attenuation, but then you lose a little in the high/low ends. You can pick one, but not both :)

    As for the filter cap thing, I think it's generally a bad idea in any design. It boosts the highs artificially, but not the bass-hump, and just makes things worse IMO.
     
  20. malibu91

    malibu91 New Member

    Joined:
    Oct 23, 2021
    Messages:
    12
    Likes Received:
    9
    Thanks for that Gene,

    I'll admit I have been wanting to build a 5e3 for some time, but haven't been able to find a readily available kit locally.

    Cheers
    Jules
     

Share This Page