Simple Attenuators - Design And Testing

[Gene Ballzz]

Thanks for that @JohnH ,
I'm kinda planning my next couple builds and generally intending to do 16Ω versions of your simplified M2, so I need a different choke. All I have on hand is a 1.25 & a .8. Although I may just do one more M-Lite with the parts I have on hand! Please remind me (without me needing to re-read the whole thread) if there is any downside to the two choke M-Lite design over the single choke M or vice-versa? I think we already determined that the 1.25 & .8 (that I already have) substituted for the 1.1 & .7 made little discernible difference in your testing?

In all honesty, I have all the parts for three more of these (except only one pair of chokes), but have been playing around with a few different cofigurations & layouts, as all of mine are destined for semi-permanent installations and somewhat custom tailored to their applications. For example, I could never see a need for any stage #3 for my 5 watt amps!

Thanks Again John,
Gene

 

[JohnH]

hi Gene, this post....

http://www.marshallforum.com/thread...design-and-testing.98285/page-46#post-1973252

...was where I tried to tease out any likely differences between M and M2 types. The charts show that they are not quite identical but any differences are likely to be within a fraction of a db over a range of assumptions about amps etc. So really no difference. But we haven't had an M and and M2 in the same place at the same time yet, maybe until you or I get there.

But I like the M2 aesthetically. Understanding how to relate the output response to the triangle formed by the coil and the two resistors R2A and R2B was a nice bit of math (Star-Delta transform), and its front end although looking simple, has a nice WTF? aspect about it.

 

[LordoftheLivingRoom]

Hi @LordoftheLivingRoom those look awesome!
What is your technique with the faceplate writing? (and I appreciate the name check)
Also, do you find any difference in how they are working on your amps? Two classic amp circuits there!

I'm nearly there with new cleaned-up diagram. No changes though, its basically what you built.

Faceplates... my old nemesis. With these I just did up a quick thing in Gimp. They are simply printed out on paper stuck to the plastic front panel of the enclosure. And credit where credit is due.

I never really got much opportunity to crank either amp before. I can get both amps turned up to half or a bit more now. Before I was barely at 1. Totally different feel.

 

[Gene Ballzz]

@LordoftheLivingRoom
Fantastic looking builds! I'd love to see a pic of the jack panel end. Did you wire two output jacks in parallel, or ????

It also looks like there is no venting on the top? Have you noticed any heat issues during extended use, particularly with the JTM45?

Thanks & Nice Work!
Gene

 

[MAD64]

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.

In the meantime I've built the 7+3.5 dB 16Ohms version for a friend's vintage AC30 and he is happy. And it turned out, that 10.5dB was a bit too much so in the end I've added a switch to bypass the 3.5dB stage ;-)

 

[MAD64]

@JohnH I found some measurements of tube amplifier output impedances in a German book that might be of interest (all for 8 ohms and 1kHz):
AC30 - 70 ohms
Marshall 18W - 90 ohms
Tweed Deluxe - 80 ohms
Super Reverb (w/o NFB) - 180 ohms
Super Reverb (w NFB) - 17 ohms

 

[MAD64]

@JohnH What are the limitations of your model - will it work to just divide the components values for 8 ohms by 4 to build a 7+3.5 dB version for my Tweed Bassman's 2 ohm tap? I haven't read each and every post in this thread, so I apologize if this has already been discussed...

 

[Filipe Soares]

what's the latest version of the attenuator? I think I'll have to build one for 100w...

 

[JohnH]

Impedance values

hi @MAD64 , thanks very much for those impedance values. Its very valuable to see them. Real output impedance is such a significant parameter that it is amazing that there is almost no info about it online. Your post above has just at least doubled the total English-language data pool!

Does the book have any such values for a classic EL34 Marshall circuit? We are finding that in some cases, the values for those may be much lower.

As a side comment, despite electric guitars being invented by Americans, I'm always very impressed by the very interesting engineering analysis of guitars and their systems that has been done in German.

But anyway, it confirms that the base assumptions used in these designs are right on the money. Also highlighting the value of the reactive circuit to allow the output signal to adapt to such a wide range.

2 Ohm Design

Yes in principle, values should just scale down, divide inductor and resistors by 4 based on values for 8 Ohms. But then it may be necessary to adjust values to suit the available range and for that, the aim should be to keep the ratios of resistors within each stage about the same. eg, take Stage 1 which is the most important one. R1, R2A and R2B would go from 15, 22 and 18 to 3.75, 5.5 and 4.5 which would round up to standard values 3.9, 5.6 and 4.7. But power resistors can come in different series values so you might end up with 4, 6, 5 for example.

Power ratings would be the same, scaled according to the amp power.

But current through the attenuator coming from a 2 Ohm tap for a given power, is x2 that for an 8 Ohm version. If its a 50W amp at 2 Ohms then the full current is nominally 5A so that puts pressure on jacks and switches, and Id suggest 16 awg wire for the coil. If you were running it all with a full bypass switch, you'd be needing at least a 10A rated switch. That's a big advantage of not having full bypass (see page 56 diagram, Stage 1 is fixed), since only 20% of the power is being switched. So we are only switching 10W and for 8 Ohm the switch current is nominally 1.1A, so 2.2A if for a 2 Ohm build (so use at least a 5A switch).

Also, you could have a look at page 44 where there is a version of M2 with variable input Ohms. This was done by scaling to 4 Ohms then adding another 4 Ohm dummy load at the front. But if its to be dedicated to a 2 Ohm amp, id just go straight to 2 Ohm though.

But, although there have been quite a number of successful builds at 8 and 16, I don't know of any at 4 or 2 although they have been discussed a few times. I cant think of any other factors that would prevent this though.

 

[JohnH]

Hi @Filipe Soares , see page 56 for latest, also linked from page 1.

If it works for your amps, id suggest building at 16 Ohm for a 100W amp, and all power rating values get doubled, and use 16awg for the coil.

R1 would then be rated at 200W, but id suggest to use 2 at 100W in series or parallel, they are much more compact and also more commonly available, and you can spread them apart to mount in the case.

At 100W, cooling is critical and my starting point would be to mount resistors onto a big thick aluminium case (much better than steel for heat), with lots of wide holes in the top and the base. And/or, depending on your use and how its built, it may benefit from a fan.

 

[Filipe Soares]

Hi @Filipe Soares , see page 56 for latest, also linked from page 1.

If it works for your amps, id suggest building at 16 Ohm for a 100W amp, and all power rating values get doubled, and use 16awg for the coil.

R1 would then be rated at 200W, but id suggest to use 2 at 100W in series or a parallel, they are much more compact and also commonly available, and you can spread them apart to mount in the case.

At 100W, cooling is critical and my starting point would be to mount resistors onto a big thick aluminium case (much better than steel for heat), with lots of wide holes in the top and the base. And/or, depending on your use and how its built, it may benefit from a fan.

Thanks john, your attenuator is in my personal projects pipeline for a while. I think I'm going to do it next week.

 

[MAD64]

Thanks for your comprehensive answer. Gauge and switches/jacks is not an issue, e.g. my bypass switch is specified for 125V/15A. I'll give it a try with my tweed bassman and will report the results The impedance values I've quoted are from this book (see last page of this chapter, diagram 10.5.20). I'm not aware of measured values for an EL34 powered amp (neither with nor w/o nfb) but maybe I find some if I check my books more thoroughly...
https://gitec-forum.de/wp-content/uploads/2017/08/pde-10-5-7-ausgangs-widerstand-impedanz.pdf

 

[LordoftheLivingRoom]

@LordoftheLivingRoom
Fantastic looking builds! I'd love to see a pic of the jack panel end. Did you wire two output jacks in parallel, or ????

It also looks like there is no venting on the top? Have you noticed any heat issues during extended use, particularly with the JTM45?

Thanks & Nice Work!
Gene

Thanks Mr. Ballzz. The rear panel is just 3 jacks, one input and two output. Pretty boring hence no pics. The two output jacks are in parallel. I will one day run my Deluxe reverb clone 1x12 combo cab in parallel with the 2x12 I usually use with my JTM45 clone... It will be legendary. Parallel output jacks give me a ton of speaker combo options!

The enclosures do have vents in the top. Compressing the images must have hidden them. The top vents are manufactured slot vents, the bottom vents are my dodgy drilled holes.

I did get the JTM45 attenuator to begin getting quite warm after about 15 minutes with the amp cranked (everything on 11!!!), but I like the tone better with things backed off and the amp sitting just into overdrive for simple blues noodling. Then again, there are some early AC/DC tones in there to discover... It's amazing that I can attenuate down to living room levels with a JTM45 turned up to just over half (with some efficient WGS speakers too) and have some extra attenuation on tap.

Extended use... I don't get any of that these days. And I only know half of about 3 or 4 songs.... Too much building amps and stomboxes, not enough guitar practice. That'll begin to change soon.

 

[Gene Ballzz]

Thanks Mr. Ballzz. The rear panel is just 3 jacks, one input and two output. Pretty boring hence no pics. The two output jacks are in parallel. I will one day run my Deluxe reverb clone 1x12 combo cab in parallel with the 2x12 I usually use with my JTM45 clone... It will be legendary. Parallel output jacks give me a ton of speaker combo options!

The enclosures do have vents in the top. Compressing the images must have hidden them. The top vents are manufactured slot vents, the bottom vents are my dodgy drilled holes.

I did get the JTM45 attenuator to begin getting quite warm after about 15 minutes with the amp cranked (everything on 11!!!), but I like the tone better with things backed off and the amp sitting just into overdrive for simple blues noodling. Then again, there are some early AC/DC tones in there to discover... It's amazing that I can attenuate down to living room levels with a JTM45 turned up to just over half (with some efficient WGS speakers too) and have some extra attenuation on tap.

Extended use... I don't get any of that these days. And I only know half of about 3 or 4 songs.... Too much building amps and stomboxes, not enough guitar practice. That'll begin to change soon.

Yeah, well I'm headed out the door to a 3-4 hour jam that'll let even my Tweed Deluxe warm mine up a bit. Not hot, mind you, just a tad warm after a time! Now if I were using my JTM30, that'd get it mostly toastly with it's at least 40-ish watts!
Enjoy!
Gene

 

[JohnH]

Thanks for your comprehensive answer. Gauge and switches/jacks is not an issue, e.g. my bypass switch is specified for 125V/15A. I'll give it a try with my tweed bassman and will report the results The impedance values I've quoted are from this book (see last page of this chapter, diagram 10.5.20). I'm not aware of measured values for an EL34 powered amp (neither with nor w/o nfb) but maybe I find some if I check my books more thoroughly...
https://gitec-forum.de/wp-content/uploads/2017/08/pde-10-5-7-ausgangs-widerstand-impedanz.pdf

That paper is interesting. I read some of it by pasting into Google-translate, which did quite a good readable job.

It does have some plots for JTM-45:



The lower plots show output impedance dependent on presence setting, very low at low frequencies and maybe about 5ohms at 3khz if presence set at 5. Upper plots are if NFB is disconnected

Its much lower than I found on my VM, even though JTM-45 is its grandfather with similar tubes. But looking at the circuits, I think the VM may be using much less NFB at most settings. The JTM and VM feedback about the same fraction from the output tap, but JTM uses a 16ohm tap while VM uses the 8 ohm tap, so less signal. Also, at the settings I use, around 6-7, the PPIMV is probably reducing level by about another 50%

 

[em07189]

Well i arrived very late to this post so i have some questions, what components should i use to make a attenuator for 100 W amp?

Is it possible to make a attenuation level been activated by a jack footswitch?, im asking this because im thinking also as a boost for solos, for amps that dont have 2 master volumes.

 

[JohnH]

Hi @em07189

The latest diagram is on p56 and also see the first post and response to Filipe on this page above. The first thing is to decide which amp and cab Ohms to use, ideally 16, then use values for that, with x2 power rating.

For a boost switch, you'd build the full design as it is then add one more switched stage to give the variation you want, maybe another-7db stage? - with a footswitch. I don't think anyone has built this yet but there's a few ways it could work, It could be all in the main box and relay controlled, or a separate single stage in a floor box daisy chained after the main circuit just before the speaker, or via an insert jack.

 

[What?]

I still intend to build one of these for my Super Reverb, but I'm trying to make sure I have my 'Marshall' squared away first.

Seeing this thread brought something to mind though. As transparent as this attenuator design seemingly is, would this sort of resistor and inductor network be suitable for speaker level matching? For example, if you have a speaker with a 95db efficiency and another with a 100db efficiency, the 100db might drown out the 95db speaker. Could the 100db speaker be transparently tamed to better match the 95db speaker level?

 

[MAD64]

@JohnH The book is really useful, in some areas maybe a bit overloaded with math. Some parts of the German version are officially available for free on the web and there is also a nearly complete english version freely available. Re your thoughts on output impedance: As Manfred Z writes it's not only depending on the tubes and OT but also screen voltage (more than on anode voltage) and the operating point. And it's depending on input signal size (especially for class AB, so for nearly all guitar amps), but I assume for our usage here, the measured values for small signal input are good enough...

 

[em07189]

Hi @em07189

The latest diagram is on p56 and also see the first post and response to Filipe on this page above. The first thing is to decide which amp and cab Ohms to use, ideally 16, then use values for that, with x2 power rating.

For a boost switch, you'd build the full design as it is then add one more switched stage to give the variation you want, maybe another-7db stage? - with a footswitch. I don't think anyone has built this yet but there's a few ways it could work, It could be all in the main box and relay controlled, or a separate single stage in a floor box daisy chained after the main circuit just before the speaker, or via an insert jack.

Hi John! A volume boost of -7db or plus 7b from the selected level of attenuation, this option will give more flexibility to the power attenuator.