Simple Attenuators - Design And Testing

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
4,993
Reaction score
3,192
Location
Wilton NSW
Hi @Gene Ballzz , I think the case looks good for the basic versions, given some planning . Would it need some components on the top and on the bottom in order to fit? It's not huge in surface real-estate.

On the coils, I think both those values are within range of the sweet spot sych that there's not much difference. There'd be no real difference in basic clean tonal balance (which is set by the resistors), and maybe a tad difference in the reactive harmonics when you drive the output.

For me, I'd pick the 1.0mH, since my meter has an inductance range and I can use it to unwind them a few turns. Ive done this before, and if you keep the meter running as you unwind, (not cutting the wire yet), you can watch the Henry's reduce as each turn comes off and so is removed from the coil. Then trim and resecure the rest of the coil. But even if not, I think I'd still pick 1.0.
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
@JohnH ,
Thanks for that reply. Kinda what I figured. The current search for inductor coils, especially 0.9mH is pretty daunting. Most are either #20 gauge, #14 gauge (pricey) or aren't even secured wit zip ties! I really dislike the zip tied units anyway, for multiple picayune reasons, but one is that they simply look cheesy! I really like the mounting convenience and potential extra sturdiness of bobbin wound units.

On another note, the #20 gauge units "claim" to be good for 200 watts, but it seems that would defeat the whole purpose of wiring these units with #18 gauge wire? Although, I guess that since full power signal does not really pass through that coil, maybe #20 gauge is sufficient for a 50 watt build? Maybe not!

And given that I don't have an accurate means of testing inductance, might you suggest a "guesstimate" number of turns to remove from this Madisound #19 gauge, 1.0mH unit, to get it closer to the 0.9mH ballpark?

Thanks Aagin John,
Gene
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
4,993
Reaction score
3,192
Location
Wilton NSW
ok, I can have a go at working that out. Another reason not to go down to 20ga is to control resistance, which we don't really want in the coils.
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
I had no idea that the small value difference in resistance is critical, although I can see how the difference between 0.48 or .50 and 0.67 or 0.73 is in the neighborhood of 40% or more! Thanks for that reminder. Can I assume that the lower the DCR the better, or more inconsequential it becomes? For example, a #14 gauge, 0.9mH is like 0.23 DCR.

Of course, this would be "electrically" perfect,


but then there's the conundrum of securely mounting it to avoid damage or ugliness! Yeah, I know, zip ties and a couple disks to squeeze it between, but…….., I'm not really a fan of "squeezing" these coils, given the relative "fragility" of the varnish coating and the fact that these units might get bounced around in gig bags, while still hot from use, etc! I really want to enhance the cosmetics a bit, with flush, countersunk mounting screws and certainly no zip ties protruding through the outside of the enclosure! I'm guessing the Madisound #19 gauge, 1.0mH, on its bobbin, unwound a bit is the best solution? Of course, "engineering" a suitable "bobbin" might be an option, but rather tedious and time consuming, at best!

Gene
 
Last edited:

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
Hey @JohnH ,
I just accidentally sent you an email from my junk account and I'm going to send it again from my real email, just in case you think the first was SPAM!
Thanks Again,
Gene
 
Joined
Jan 9, 2021
Messages
10
Reaction score
5
I'm building a M2 design. Does it make sense to swap the order of the switches into stage 3, then stage 2 followed by stage 4. So you get -14db, -7db and -3,5 db?
 
Joined
Jan 9, 2021
Messages
10
Reaction score
5
Hey @JohnH ,
What do think of this enclosure?


s-l1600.jpg
I know, I know, I'm not John. Like the shape, but the joint between the front and the top cover has a gap. If it bothers you you could add some moulding. I haven't bothered to convert the imperial measurements. The 165 mm length doesn't allow for much play. I used a longer one I had for another project, but that one is a bit to long, it doesn't quite fit on the top of my JCM800 head. There isn't much room on the top.
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
I know, I know, I'm not John. Like the shape, but the joint between the front and the top cover has a gap. If it bothers you you could add some moulding. I haven't bothered to convert the imperial measurements. The 165 mm length doesn't allow for much play. I used a longer one I had for another project, but that one is a bit to long, it doesn't quite fit on the top of my JCM800 head. There isn't much room on the top.

I did a full mockup using those dimensions and everything fit wonderfully, even using the huge ARCOL/OHMITE 100 watt resistor for R1. The gap did not concern me, but I had hoped to get him to do two more identical rows of vents on the top. Unfortunately, these boxes are no longer available and will not ever be again, unless at a much higher price!

Oh Well,
Gene
 

BlueX

Well-Known Member
Silver Supporting Member
Joined
May 5, 2020
Messages
647
Reaction score
1,588
Location
Sweden
I'm building a M2 design. Does it make sense to swap the order of the switches into stage 3, then stage 2 followed by stage 4. So you get -14db, -7db and -3,5 db?
Maybe post # 2,207 on page 111 (!) answers your question:
 

stickyfinger

Well-Known Member
Joined
Nov 11, 2010
Messages
868
Reaction score
466
I did a full mockup using those dimensions and everything fit wonderfully, even using the huge ARCOL/OHMITE 100 watt resistor for R1. The gap did not concern me, but I had hoped to get him to do two more identical rows of vents on the top. Unfortunately, these boxes are no longer available and will not ever be again, unless at a much higher price!

Oh Well,
Gene
Crap that was actually the best looking enclosure i had found after searching for a few weeks even before you linked it. I originally wanted to make a tube preamp out of it.
 
Joined
Jan 9, 2021
Messages
10
Reaction score
5
Yeeey, it works!

IMG_5951.jpg
And it's guts:
IMG_5947.jpg
Rubbed the paint off where the resistors are mounted with thermal grease. I didn't want the paint to thermal isolate. I intended to make slots in the back, by "connecting" the holes, but sticked with the holes. 3D printed a little mounting thingy for the coil. I used the colors to keep track what I'm doing. It get's warm, but not hot after playing it on over 8 at volume level. Something like red = hot = in, orange is a bit cooler etc. Used different colors for each stage. I didn't connect ground to the chassis.

Edit: Thank you @JohnH
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
Yeeey, it works!

View attachment 110301
And it's guts:
View attachment 110302
Rubbed the paint off where the resistors are mounted with thermal grease. I didn't want the paint to thermal isolate. I intended to make slots in the back, by "connecting" the holes, but sticked with the holes. 3D printed a little mounting thingy for the coil. I used the colors to keep track what I'm doing. It get's warm, but not hot after playing it on over 8 at volume level. Something like red = hot = in, orange is a bit cooler etc. Used different colors for each stage. I didn't connect ground to the chassis.

Edit: Thank you @JohnH

Great use of space! Nice, neat and tidy layout. I'd expect the need for more/better ventilation, but if it doesn't get actually "HOT" with your amp suitably cranked to it's sweet spot, you're good to go! Ain't it liberating to be able to get great sounds that don't elicit screams of "TURN IT DOWN?"
Nice Job! :thumbs:
Gene
 
Joined
Jan 9, 2021
Messages
10
Reaction score
5
@Gene Ballzz thank you. I’ve heard “turn it down” quite a few times in my live 😂😂 Really nice to hear it roar again, but not that loud. Like a tiny tiger kitten haha. I like the sound. It’s awesome.
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
4,993
Reaction score
3,192
Location
Wilton NSW
Hi @Gert-Jan van der Heiden . That looks like a really nice build, clean and simple and I like the 3d coil-mount gadget.

Just on the order of stages, yes you can reorder them and it makes no significant difference. A little history....

At the start of the thread, you can see the previous design M, which had two coils, but otherwise works like M2. In both these, the -7db stages (the reactive Stage 1 and the switched -7db) are the most perfect ones that have the ideal input and output impedances. The -14 and -3.5db stages have the right output impedance but are allowed to have a somewhat different input impedance because the amp generally never sees them directly, only via Stage 1. This allows them to remain as nice simple 2-resistor stages that can be switched with a single pole.

But, with M I wanted to explore getting a dedicated -3.5db setting, to fill the gap between full power and -7db. The -3.5db stage works quite well on its own, so in M, I put that last and the switching allows that you can use it on its own. I works fine, you can see and hear results in Post 1. I put the -7 stage first in line, since it is more ideal than the -14 stage. So I had -7 (reactive), -7, -14 then -3.5. That was the electrical order, but the physical order can be different, and like with binary numbers, -14, -7, -3.5 is probably most logical, as you look at the switches on the panel.

Now we have other better ways to get a -3.5 or -3 db setting if needed, plus, its a setting that is rarely used since its only very slightly less volume than full power. So for the base M2 designs, its not shown. But I kept the order of stages just for consistency.

But with further analysis, I found it makes really no difference what order the stages are wired up after Stage 1, and of course in any case they all have to work after stage 1 without the other two. So for easier wiring and understanding, Im now showing them -7(reactive) -14, -7, -3.5 in diagrams that Im doing. Im keeping the R numbers though, so R3, R4 etc is consistent through the thread.
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
4,993
Reaction score
3,192
Location
Wilton NSW
I did a full mockup using those dimensions and everything fit wonderfully, even using the huge ARCOL/OHMITE 100 watt resistor for R1. The gap did not concern me, but I had hoped to get him to do two more identical rows of vents on the top. Unfortunately, these boxes are no longer available and will not ever be again, unless at a much higher price!

Oh Well,
Gene
That's a shame! I would have looked very nice

Im happy with the Hammond styles for mine, and the box 'lid' is the base which looks a bit better i think due to the slight taper of the sides. These boxes also give you continuous Aluminium for what is then the top and sides, to get max conduction and cooling to resistors mounted there. I put the coil on the lid (ie final base), so if I ended up with anything zip tied or with an unusual screw, its underneath so not seen. Also, its weight presses down, so on my prototypes, once Id figured out that steel screws are not good with the coils I actually just used dabs of silicone to stick the coils down. Perfectly safe for my use, except maybe not ideal for rattling around in a gig bag!
 
Joined
Jan 9, 2021
Messages
10
Reaction score
5
Hi @Gert-Jan van der Heiden . That looks like a really nice build, clean and simple and I like the 3d coil-mount gadget.

Just on the order of stages, yes you can reorder them and it makes no significant difference. A little history....

At the start of the thread, you can see the previous design M, which had two coils, but otherwise works like M2. In both these, the -7db stages (the reactive Stage 1 and the switched -7db) are the most perfect ones that have the ideal input and output impedances. The -14 and -3.5db stages have the right output impedance but are allowed to have a somewhat different input impedance because the amp generally never sees them directly, only via Stage 1. This allows them to remain as nice simple 2-resistor stages that can be switched with a single pole.

But, with M I wanted to explore getting a dedicated -3.5db setting, to fill the gap between full power and -7db. The -3.5db stage works quite well on its own, so in M, I put that last and the switching allows that you can use it on its own. I works fine, you can see and hear results in Post 1. I put the -7 stage first in line, since it is more ideal than the -14 stage. So I had -7 (reactive), -7, -14 then -3.5. That was the electrical order, but the physical order can be different, and like with binary numbers, -14, -7, -3.5 is probably most logical, as you look at the switches on the panel.

Now we have other better ways to get a -3.5 or -3 db setting if needed, plus, its a setting that is rarely used since its only very slightly less volume than full power. So for the base M2 designs, its not shown. But I kept the order of stages just for consistency.

But with further analysis, I found it makes really no difference what order the stages are wired up after Stage 1, and of course in any case they all have to work after stage 1 without the other two. So for easier wiring and understanding, Im now showing them -7(reactive) -14, -7, -3.5 in diagrams that Im doing. Im keeping the R numbers though, so R3, R4 etc is consistent through the thread.
Thank you. Like the idea of keeping the reference numbers the same.
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
@JohnH ,
Given your relative expertise in cooling, is the die cast material of the Hammond boxes as good of a heat sink/thermally conductive as extruded aluminum? Also, how much better or worse for cooling, is it to have all the resistors mounted to the top, vs to the bottom of the unit! I definitely prefer the Hammond boxes having the lid on the bottom, as the other way gives it an odd "keystone" look! Also the screws holding the lid are handy for also mounting the feet, if they can be found with a small enough diameter!
Just Pickin' Nits, As I'm Wont To Do,
Gene
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
4,993
Reaction score
3,192
Location
Wilton NSW
Aluminium is all good, whether its die-cast or extruded. Its about 5x as conductive as steel, and probably 2x as thick in a given application, so 10x better. So between those two AL cases, it all comes down to the thickness of the material and the surface area. My belief is that the conductivity across a junction say box to lid is not as good as just within the material, so on a Hammond, best to fix hot parts to the larger section. Also, the top of the attenuator will get better air cooling than the base. But you have to enjoy seeing all the screws etc. - I quite like the Steam-Punk vibe of the screw heads!
 

Gene Ballzz

Well-Known Member
Joined
Feb 19, 2013
Messages
3,539
Reaction score
3,994
Location
Las Vegas, NV
@JohnH
Another dumb question from the peanut gallery, that may have already been answered? For the R8 in an 8Ω M2, 5.6Ω is called for. If my only choices were 5Ω & 6Ω what would be the better choice and what would be the consequences, compared to the specced 5.6Ω? Would it also require a different value for R7?
Thanks For Helpin'
Gene
 

JohnH

Well-Known Member
Joined
May 31, 2008
Messages
4,993
Reaction score
3,192
Location
Wilton NSW
@JohnH
Another dumb question from the peanut gallery, that may have already been answered? For the R8 in an 8Ω M2, 5.6Ω is called for. If my only choices were 5Ω & 6Ω what would be the better choice and what would be the consequences, compared to the specced 5.6Ω? Would it also require a different value for R7?
Thanks For Helpin'
Gene

Hi Gene, I looked at that before, to try to work out a set of values to buy from typical Chinese sellers - see the first diagram. I went for 5 Ohm at R8 and the diagram shows R7 staying at 33 Ohm. Theres no tonal consequence (less than 0.1db at any frequency) but it shifts the attenuation down by about 0.15db. When you step through the range, then the series of increments available at each step, intended each to be -3.5db, then vary from about -3.3db to -3.7db. With the specced values the range is about -3.5 to -3.6 db, so a tad more consistent.

If you change R7 from 33 to 30, with R8 at 5, then its back to virtually the same as specced.
 
Last edited:


Top