Simple Attenuators - Design And Testing

[jejj21]

Thanks @JohnH it was a fun project and has definitely given me new hope for other things. My 4x10 Bassman build may now come to fruition since I'm confident I can tame the beast should I ever want. :O)

 

[Freddy G]

Hey Guys,

So, I decided to take the extra COVID-19 tinkering time and build the M2 version of this attenuator. It works amazingly well and I am very happy with the results so I thought I would try and pay it forward some.

Attached are my drilling templates and part layout in case anyone else wants to give it a shot. I laid everything out in AutoCAD so all parts are to scale and precise. I also created a wiring layout in DIYLC to help me keep it all straight. You can print the PDF out (Actual Size) and cut out the pages and tape them directly on the enclosure for center punching, which is how I did it.

I did change the switch/resistor configuration to be in ascending order for my own sanity. I only included one output, since all of my amps can work on 8ohms. Other than that, all part values match with John's recommendation.

I used a Hammond 1590D enclosure and I bought it as well as all the resistors from Mouser. I could have probably found cheaper ones on eBay, but shipping to the USA was not in my favor. All wires are 18ga and the switches are all DPDT 6A ON/ON. The inductor is a 0.9mH 16ga Sidewinder Air Core from Madisound, held in place with a #10 stainless steel chassis bolt. I used #4 bolts for the resistors with the exception of #8 on the big 100W.

I'm attaching a couple pics of the final product so you can see how it looked once I completed it.

Thanks to John and Gene for the inspiration and knowledge.

Joey

View attachment 71701 View attachment 71702 View attachment 71703 View attachment 71704 View attachment 71705 View attachment 71706

Joey, that is beautiful work! and your layout drawings....could not be easier to follow. Excellent!

 

[Gene Ballzz]

Hey Guys,

So, I decided to take the extra COVID-19 tinkering time and build the M2 version of this attenuator. It works amazingly well and I am very happy with the results so I thought I would try and pay it forward some.

Attached are my drilling templates and part layout in case anyone else wants to give it a shot. I laid everything out in AutoCAD so all parts are to scale and precise. I also created a wiring layout in DIYLC to help me keep it all straight. You can print the PDF out (Actual Size) and cut out the pages and tape them directly on the enclosure for center punching, which is how I did it.

I did change the switch/resistor configuration to be in ascending order for my own sanity. I only included one output, since all of my amps can work on 8ohms. Other than that, all part values match with John's recommendation.

I used a Hammond 1590D enclosure and I bought it as well as all the resistors from Mouser. I could have probably found cheaper ones on eBay, but shipping to the USA was not in my favor. All wires are 18ga and the switches are all DPDT 6A ON/ON. The inductor is a 0.9mH 16ga Sidewinder Air Core from Madisound, held in place with a #10 stainless steel chassis bolt. I used #4 bolts for the resistors with the exception of #8 on the big 100W.

I'm attaching a couple pics of the final product so you can see how it looked once I completed it.

Thanks to John and Gene for the inspiration and knowledge.

Joey

View attachment 71701 View attachment 71702 View attachment 71703 View attachment 71704 View attachment 71705 View attachment 71706

Great work, effort and results! The round "speaker grill" pattern for vent holes in the bottom looks very cool! My only concern is that there are no holes drilled in the case, between the componets to provide flow through for the convection ventilation. Adding some feet may also be advisable.

Thanks For The Mention & Enjoy The Freedom!
Gene

 

[jejj21]

Thanks @Freddy G and @Gene Ballzz

I do have six small clear feet on the bottom to give it a little separation, but the gap is not huge. My amps are all 18W or less, so I haven't come close to warming this unit up yet. I will have to wait until this chaos is over to see how it does over a 1hr long jam session, but in my practice sessions so far, the case is not even the least bit warm using my Blues Jr. in the basement. I figure I can always drill more holes later if need be, but so far so good. I will continue to monitor and update the drill template if things start getting warm.

Joey

 

[JohnH]

Sounds fine! If you did need to add more cooling, then big holes in the base would do it. It lets air flow in below and out the top, without messing yr nice hole pattern.

 

[jejj21]

Hey @JohnH & @Gene Ballzz

After thinking more about this last night, I thought I would run some scenarios and see where in my design I could additional holes to give me a convection ventilation opportunity.

Attached is a V2 version of my drill pattern showing 9 additional 10mm holes. This pattern fits the layout and hopefully is more in line with what you are suggesting.

I think I can add the holes without removing all the components, but we shall see. Would've been a lot easier to add them when I drilled the others, so hopefully if we get this right it will help out the next guy.

Let me know what you think and I will try it out.

Thanks,
Joey

 

[JohnH]

hi @jejj21

Those holes look fine to me , so long as they don't clash with the case feet?

I think heat and power and volume, although all related, follow their own trajectory's in terms of significance. Like, fully cranked 20W, 40W and 100W are all loud, and are just a fairly small step of 3 and 4 db of volume between each. But in terms of how hot things get and how much that might hurt, temperatures seems to take off as you get to about 30-40W continuous. Seeing your build, the case looks to be pretty much identical to the one I used. I have a pair of 40W Marshalls (a DSL401 and a VM2266c). They sound best a maybe 20W,but even setting up the nastiest continuous cranked distorted looped riff, and then leaving it to cook, I just cant get the attenuator past 'fairly warm'.

These thick cases take a lot of energy to get them warmed up, and the temperature that they get to is effectively based on an average of power in over maybe 15-20 minutes. So any pause in the blasted riffs, or short break or clean playing will reduce average power heating it all up.

I've also run dc through it and left it for an hour. 7W dc in raised the temperature by 12 C but by 15 minutes it had only raised half that. I do some thermodynamics for work, so back figuring convection, radiation, planetary alignment etc, at about 60C starting from 20C, there is about 27W coming out and going in steady-state average. 60C is a limit for touchablity for things like hot pipes etc, and nowhere near having any effect on our electrical parts. I reckon that power is enough for most practical uses of say a 40W amp, if built well ventilated using such a case. But, at 100C, which is way too hot to touch, there is about 60W steady state and so for powerful amps like 100W, if to be fully cranked, I think we start needing fans.



I saw a difference in your switching compared to mine, with how the switches are wired. On my version, I did the stage bypass switches just with one pole, which disconnects the parallel resistor and shunts the series one, whereas yours uses the more normal two-pole bypass method. Not a big issue at all, given most switches have two poles. But I did it that way so there are less contacts being switched in the chain, and also it is possible to gang both sides of a switch into one for more reliability, halving the current through each contact.

Also, the reason I put the -3.5db stage last in the chain is for if it can be used independently such as in my M diagram on post 1. But physically, I also have them in logical 3.5, 7, 14 order.

Last thing to point out, your R1 is 12 Ohm which matches my first diagram for M2, though now I show 15 Ohm for that. No issue at all the difference in the results is fractions of an ohm and db. I think the theoretically best value is in between, like 13.5 ohm but that's not an available value, But for a 16 ohm version, x2 at 27 Ohm is a good standard option.

 

[jejj21]

Hey @JohnH

The circle grid pattern I made is on the base plate of the enclosure, along with the clear feet. The new holes would go on the top of the box, where I have all the parts suspended.

All of the wiring I've done in the past used DPDT switches, so I just kept that same logic as it makes sense in my head.

I did pull the 3.5 only option out because I figured if I'm going to use the M2, I'm going to want at least a 7db cut and I can go from there. It also simplified the wiring and resistor sizing, so I went with it.

When I started the process of laying it out and figuring out parts weeks ago, 12ohm was the value I saw so I ordered it. I'll leave it for now and if there is compelling evidence to change it, the swap would be simple enough. I'm certainly going to build another one, so I will make that size change and maybe the switch change and A/B the two to see if I can tell a difference.

Thanks,
Joey

 

[jejj21]

What I get for trying to do this on my phone, lol.

 

[JohnH]

I see, that all makes sense. I also never use the -3.5db-only option, but I built it just to prove it up in case it was a key feature for someone else.

 

[jejj21]

Totally understand, being able to add the option into the mix is cool for sure.

I will get out into the garage tomorrow and add some more holes and post a picture. I may try to do something cool around the inductor outline, just to give it some flare and extra venting near the 100W'r.

 

[jejj21]

Hey @JohnH

Thanks to a rainy Saturday morning, I got some time to piddle around on designing and drilling the extra vent holes. Decided to add some smaller holes vs. just the 9 bigger ones. I figured around the 100W/50W/Inductor side of the box would be my biggest heat concerns, so that's where I focused most of the holes.

Check it out and let me know if you think that is sufficient.

Thanks,
Joey

 

[JohnH]

I reckon that will run cooler if cranked now!

 

[What?]

JohnH, taking into consideration what you said previously about the Weber attenuators being 16 ohm with added resistive attenuation for lower impedances, I have been jamming my DSL 50 yesterday and today with a 16 ohm 412 cab, and it's rocking better than ever. I think I said on the forum somewhere yesterday that I wasn't digging my DSL, but after being able to crank the amp up to 7 again (without making my neighbors lose their minds) it makes such a big difference in the sound of this amp. I'm able to use the attenuator at down to around 2 out 10, and that gives me a manageable volume without any significant tone suck. Going lower than that starts getting flat and ugly. It's still on the loud side in a bedroom sized space with the attenuator set at 2 (more than loud enough to irritate people in the next room), but it is way below cop calling loud.

And I think I definitely will be building that last attenuator design for my Super Reverb. But on a related note to that point, with the DSL cranked up and boosted and shaped a bit with a graphic eq pedal, I'm able to get the same sort of cranked up tone (very much SRV tone ) that I get out of my Super Reverb out of my DSL on the green channel, which is unbelievably cool. I could never get that sort of sound with the amp down low. I always had the feeling that it was lurking in there, but not having a manageable volume level to be able to spend alot of time tweaking didn't allow me to find it. And with just a small amount of tweaking and switching to a humbucker loaded guitar, I'm able to get a nice fat and rich JCM 800 sort of tone from this same amp. Even the clean sounds so much better with the DSL cranked up, and being attenuated allows for a volume level for practically using it at home.

Any way, just saying thanks for dropping the knowledge here that allowed me to get this amp cranked up again and find some nice sounds. I had been jonesing for another Marshall variant lately, but I think the one I have had all along is doing what I want and then some with the attenuator. If I could only get some lows back into the red channel (lead channel design issue on these JCM 2000 amps) this would be the most versatile all around amp that does everything well for me.

 

[Gene Ballzz]

@What?
Are you experienced enough to safely perform amplifier mods? If so, I'll share my approach to correcting the tone (especially the bottom end) disparity between the two channels of many/most DSL models.
Let Me Know?
Gene

 

[jejj21]

Here is the latest templates for my build in case someone wants to try it.

 

[What?]

@What?
Are you experienced enough to safely perform amplifier mods? If so, I'll share my approach to correcting the tone (especially the bottom end) disparity between the two channels of many/most DSL models.
Let Me Know?
Gene

Yes, and it would be nice just to see if the red channel can get some oomph behind it. Looking at the signal flow diagram at http://www.lydian.ca/Marshall_DSL.html it looks like to me the problem is just before V1A. Maybe over here so as not to add noise to JohnH's thread about attenuators? http://www.marshallforum.com/thread...-green-channel-most-like.113262/#post-1966417

 

[jchjch]

@JohnH
Thanks for the great ideas. I haven't built it yet, but the idea occurred to me that two L-Pads connected in series, with the speaker side connected to each other, approximates a T-Pad. The impedance will be slightly below the L-Pad impedance at some point on the attenuation curve, but it will be minimal.

With that, I came up with the modified M2 design, using a stereo connected L-Pad for the 2nd stage of variable attenuation. Spice simulation seems OK, at both the Amp end and the Speaker end. The speaker never sees a load that is less than 8 ohms (going back toward the amp).

I'm attaching a drawing of the circuit, as well as a table showing how the L-Pad works. I'd appreciate some feedback.


In the table, I highlighted the lines with minimum impedance.

I also looked at what happens when the speaker is both open and shorted. At open, the amp sees 12 Ohms from R1, and it goes up with frequency. At short, the speaker sees 5.4 Ohms (Parallel combination of R1, R2a, R2b), and then it goes up with frequency.

Cheers!
Jeff

NOTES:
Rser and Rpar are calculated for an 8 Ohm L-Pad at the stated attenuation.
Zattn is the DC impedance of the L-Pad network and speaker, where it connects to R2a and R2b.
Zamp is the DC impedance of the entire attenuator network.

 

[JohnH]

hi @jchjch Thanks for your ideas. But to really know if it works and preserves tone you have to analyse it with all reactive components including an analysis model of the speaker, and take account and keep reasonably constant the impedance seen by the speaker. Have you checked it that way?

 

[jchjch]

hi @jchjch Thanks for your ideas. But to really know if it works and preserves tone you have to analyse it with all reactive components including an analysis model of the speaker, and take account and keep reasonably constant the impedance seen by the speaker. Have you checked it that way?

Absolutely - I use an electrical model with a resistor in series with an inductor which are in series with a parallel L-C. Actually tweaked the resonance frequency, inductance, and resistance.