Simple Attenuators - Design And Testing

[JohnH]

@JohnH
I've got a short list of dumb questions, but a little info to share first. So far only one unit has actually been built, modified and re-modified throughout the evolution process. This unit has proven (to my ears and feel) to be absolutely stellar in every way, including minimal, if any tonal and/or response anomalies or degradations with a fairly wide variety of amps ranging from 5 watts to 40-ish watts, from minimal to maximum attenuation levels. Even using an 8Ω speaker with this 16Ω version has not exhibited any negative issues! Some amps have negative feedback and/or PRESENCE circuits and a couple do not. I'm still pondering the custom layout/packaging/configuration for each individual unit for mostly permanent mounting in several various amps. This first 16Ω, M-lite got built utilizing our initial inductor values of L1 = 0.8mH & L2 = 1.25mH, as opposed to the current optimums of 0.7mH & 1.1mH. I've also been too lazy to address the mounting with regular steel machine screws and steel washers, as shown in the pics of post #178. When I got the coils for the first build, I ordered two sets of the same values and it turns out that MADISOUND does not have a 1.1mH in the 19 gage configuration, only a choice of 1.0mH or 1.25mH for L2, in the mounting style I like. I've got enough other components to finish four builds of the 16Ω, M-lite, except only enough inductor coils (0.8mH & 1.25mH) for the first two,unless those values are just plain WRONG and may cause unforeseen issues. My goal here is for all four units to be standardized (component value wise) with the only differences being in the mounting and layout configuration.

Following that, here come the dumb questions:

A> What problems and/or issues are being created by using the inductor coil values listed above?

1> Do I need to find a source for a suitable 1.1mH and swap out what I've already used, or is the difference so minimal as to be of little consequence?
2> Should subsequent builds simply use what I've already been using?
3> If I need to swap out inductors and still want to use the products from MADISOUND, should I go with the 0.7mH for L1 and then choose either the 1.0mH or 1.25mH for L2?​

B> How much and what kinds of problems are actually being caused by my current steel screw and washer mounting arrangement?
C> Is it possible that the value discrepancies, combined with my current mounting is actually contributing to this "test mule" unit being a "one size fits all" that operates nearly identically in all of my tested applications?
D> Bottom line, do I need to step back and re-think everything, or should I just go with what seems to have worked well, so far?​

Dumb questions, I know, but I'd rather make any changes now, before gearing up and collecting the remaining required inductors!

It's also my hope that these dumb questions don't complicate or confuse this thread!

Thanks Again @JohnH & Happy Holidays,
Gene

First, to others reading this and enjoying the good attenuation, I want to put out a big THANKYOU to Gene. Gene was the first to trust, build and validate these designs in their evolved form, almost 2 years ago now. His experience as a professional musician, builder and general good bloke continues to be invaluable.

OK Gene....

You have the advantage of having built a working unit that you've tried across many amps and found to be good for your uses. So I can understand the advantages of keeping consistency.

On the M design (for others: this has two coils, see post 1), the differences between 1.25 and 0.8 vs 1.1 and 0.7 are minimal. It's the ratio between the coils that is more important, which is pretty close with both recipes.

I have a $50 multimeter that also measures inductance, so since I have that and wanted to test exactly what I worked out, I used it to unwind a pair of slightly higher coil values down to a target. Its pretty easy, I just clip the meter across the coil wires as supplied and start unwinding the outer turns, without cutting the wire (once the wire is off the coil and loose, it no longer contributes to inductance, even if it is still part of the circuit). Then, when I get to the value I want, I snip the wire, re-tin it and secure it.

The way in which the one design is found to work well with many different amps is a lucky bit of magic that is hard to explain (ok, I cant fully understand/explain it!), but it is apparent in the analysis results. I tested this with a very wide range of assumptions about the effective output resistance of the amp, and the attenuators (M or M2 ) track this from high bass and low mids up to very high treble. The bass resonant peak is more dependent on the attenuator and so in theory would vary more. But the basic values that we use give a very good result. In theory, the more complex version M3 with a bass resonant circuit should track the low bass better across different amps. But, using just design M (and M2 would be the same), I have tried to reveal any issues with inconsistency in the bass on my two very different amps, and cant find any audible issue at all.

Steel bolts etc - these are increasing the coil inductances in unknown ways. There's no risk to the amp, the impedance seen by the amp is not affected until the upper treble range. But what would it do? It might affect the tone of the attenuator if one coil is being changed by a greater % than the other. It might also affect something in the 'feel' or 'dynamics') of the response, as you dig in with a signal that is driven hard in the power stage and decays - subtle maybe?

How about two ways to check this out?:

1. Play a while as it is, then take the bolts out and see if you can notice any difference. Try this with an amp with no NFB to expect the greatest possible effect. A more scientific version of that would be to set up a looped riff and record with a close mic to get two otherwise identical clips.

2. I have a prototype M and I can go and measure a coil inductance, then slip an M3 bolt through it and remeasure. - Ill do this....

 

[adew1]

I had a quick look for EU suppliers, and found this in the Netherlands, but Im sure there are lots of places:

https://www.soundimports.eu/en/crossover-components/coils/air-core-coils/?min=€0&max=€50&filter[]=641991&filter[]=628296

@JohnH Thanks for the link, John. I also found this place, also in the Netherlands:
https://europe-audio.com/

Thanks for the info re wire gauge, etc. Very useful. Cheers!

 

[JohnH]

Steel bolts!

So I cracked open an early prototype, which is now disused. The coil in the picture is about 40mm diameter, and disconnected it measures 0.41mH.

I dropped the two bolts into it on its axis, that are shown laid on top.

The small bolt which is a 10mm M2.5, raised the inductance to about 0.44mH so about +7%. The larger bolt is 30mm M3, and it raised it to 0.59mH which is +44%.



So, if mounting can be by short steel bolts or screws, it doesnt have too much effect. But a full length bolt through the coil axis causes quite a significant increase.

Fixings without carbon steel wouldn't do this.

Then, I placed the coil over the steel blade of a wood saw, as if it was in a steel case. Inductance went up about 10%. Thats with steel just under one face, so maybe it would be 20% more in a full steel chassis.

Aluminium is better for this.

 

[Gene Ballzz]

@JohnH
Thanks for that! I don't have a meter that will measure inductance so, greatly appreciated. When I can sonically test it all, in all amps, I'll swap to some nylon or non-magnetic stainless bolts and see if there is any difference. If there are any sonic differences, I'll let you know. I will likey just stick with the 0.8mH & 1.25mH coils for the rest of my current builds. I've pretty much nailed down my external layouts/configurations for most convenient use. I guess its time to start "strippin', drillin' & screwin'" along with a bit of solderin'!
Thanks Again!
Gene

 

[JohnH]

@JohnH Thanks for the link, John. I also found this place, also in the Netherlands:
https://europe-audio.com/

Thanks for the info re wire gauge, etc. Very useful. Cheers!

That looks like a great site! very well organised and comprehensive. I found this one for an 18 gauge 1.8mH:
https://europe-audio.com/Product.asp?mfr=Intertechnik&part=LU55/1.8/100&Product_ID=5316

….and there's also the Jantzen 'not on a bobbin' style

 

[Gene Ballzz]

That looks like a great site! very well organised and comprehensive. I found this one for an 18 gauge 1.8mH:
https://europe-audio.com/Product.asp?mfr=Intertechnik&part=LU55/1.8/100&Product_ID=5316

….and there's also the Jantzen 'not on a bobbin' style

@JohnH

The terminology in that EU link led me to a state side supplier. Not understanding a lot of the jargon, or even how to convert "u" Henry to "m" Henry leaves me at a loss to determine if they are a suitable source or not. They certainly offere a very wide range of products!

https://www.coilcraft.com

Here also is a link to the goggle search that comes back from the search criteria "air core coil"

air core coil

Just Lookin'
Gene

 

[MikeWinterhalt]

John, Thanks so much for sharing your journey in designing this. I just finished my build of the M2v. I'm just checking things before I try this thing out. I put the circuit in Micro-Cap Spice and calculated Zin, Zout.

When doing the calculations, I assumed 20 ohm as Zout of the amp at 8 ohm tap. Should we assume 10 ohms at 4 ohm tap and 40 ohms at 16 ohm tap?

 

[JohnH]

Hi @MikeWinterhalt , those values in your table look about right.

For checking out the design and the expected tonal balance, I used 20 Ohms for the output impedance of an amp with an 8 Ohm tap, and x2 or x1/2 that for 16 and 4, as you noted. But that's just a median set of values and I also consider a lot higher and a lot lower because it varies according to the amp design. Also, those kind of values are based on what is seen at audio frequencies. If you are looking for ohms values to use as a dc resistance check before powering up, then the dc output resistance of output transformers would probably be a lot less.

Id say your circuit analysis is looking right, so if you run it without the assumptions about amps and speakers connected, it should give you some check values for dc resistance that you can measure on your completed build.

Or, if you want to check it out initially to make sure its safe for the amp, then if you measure the dc resistance at the input (no amp), with speakers connected, it should be around 4, 8 or 16 ohms (depending of which input setting), within about say 20%

 

[JohnH]

@JohnH

The terminology in that EU link led me to a state side supplier. Not understanding a lot of the jargon, or even how to convert "u" Henry to "m" Henry leaves me at a loss to determine if they are a suitable source or not. They certainly offere a very wide range of products!

https://www.coilcraft.com

Here also is a link to the goggle search that comes back from the search criteria "air core coil"

air core coil

Just Lookin'
Gene

Hi Gene,

I think in electronics generally there are applications for inductors that are several orders of magnitude different from those that we need. Often they are about tuning and shunting and blocking at Mhz frequencies. When you see uH etc, that's a shortcut for a greek 'mu' symbol, denoting micro henries, ie 1/1000th of a milli Henry (mH). There's also nano-henries (nH) which is 1/1000000th of a mH, or a billionth of a Henry. I had a look through that coilcraft site but they all seem intended for those kind of applications, I didn't see anything for us there.

The ones we need always seem to be designed for audio speaker applications. Its good that we can draw on that reasonably common product group to find the bits we need for our rather uncommon application.

Also, its worth talking about why we need air cored, instead of coils wound on an iron core. I got that point from various online threads about reactive loads. When a coil with a core gets high current, it can saturate and this momentarily reduces the inductance. We don't want that since the whole idea is to get the unit doing what we want with high amp power!

There's another place where inductance comes into our guitar tone, in the pickups themselves. Thousands of turns of very fine wire on a core leads to very high inductance values, of several Henries. eg a PAF humbucker might be around 4H which is 4000mH.

 

[Travis Knight]

Hi, looking forward to building this; I like the idea of reactive vs resistive attenuation - seems a more "natural" simulation of speaker load to me. Though, I wonder if anyone has thoughts on how a home theater volume control might work for this application. Being an attenuator, wonder if there's a play? Something like the OSD SVC100 (sorry, first post, no links)

 

[RoadShow]

Hi Folks,

So, over the holidays I got my attenuator put together built for 8 ohm with the 16 ohm option using a 1.0mH coil. Trouble is, using it with an 18W Marshall it's almost complete attenuation using it in both 8 and 16 ohm set ups. I've looked this over and over and am stuck. Maybe a fresh set of eyes will see where I've gone wrong?

Attached is a picture of the layout and wiring, my marked up schematic, and a drawing I made for wiring it up.

With the switch in the 1-2 position, at the input I read 8.8 ohms when the plug is in the 8 ohm outputs reading 18.1 ohms, and I read 8.1 ohms at the input when the plug is in the 16 ohm output where I read 28.1

With the switch in the 2-3 position, at the input I read 16 ohms when the plug is in the 8 ohm outputs reading 25.3 ohms, and I read 13.3 ohms at the input when the plug is in the 16 ohm output where I read 32.9

Please open my eyes!! I hope my mistake isn't too dumb...

 

[JohnH]

Hi @RoadShow

Mysterious! It looks to be well thought out and neatly built, and I'm not seeing the issue yet.

If you could run some more resistance tests, with an 8 Ohm speaker plugged into Out1/2, and measure the input resistance in all settings. It should stay around 8 ohms within about 20% or so. Also check the switches are working, and that the 1-2-3 lugs are operating the way you expect. I had thought that all toggles were similar, until I found out that they aren't.

The jack sockets look to be metal, not plastic bodied, hence the 'ground' side, which usually is not a ground but is just one end of the OT output winding, may be live to the case. Unlikely to cause the issue, unless something else is touching it too. (which is why plastic bodied is better)

Also, I'm suspecting something about the jack connections, only because its hard to see them in the pics.

I'm assuming that your amp at full volume is seriously loud? with minimum -7db attenuation on this M2 design, 18W is still delivering about 3.5W to the speaker, which is usually still too loud.

And just for others looking in, we have collectively built dozens of these now so we know that they work fine.

 

[JohnH]

Hi, looking forward to building this; I like the idea of reactive vs resistive attenuation - seems a more "natural" simulation of speaker load to me. Though, I wonder if anyone has thoughts on how a home theater volume control might work for this application. Being an attenuator, wonder if there's a play? Something like the OSD SVC100 (sorry, first post, no links)

Thanks for posting, I found the link to those units, designed for combining and controlling sets of ceiling mounted audio speakers.

These kind of units are often put into attenuators, and they tend to be responsible for the way the tone drifts off and gets dull as you turn down. They are designed to pay attention to the impedance seen by the amp, but not so much to that seen by the speaker. Controlling the output impedance seen by the speaker is the key to this design and the main reason it stays so consistent at all volumes. So although it can 'work' I'm not recommending it. The small switched steps that we have are actually audibly very small, and generally there's no need to fine tune any further. With the switches and carefully judged component values for each stage, we can maintain the required control.

I did work out a version with a variable pot, trying to mitigate and control some of the disadvantages, but its not tested and will depend on the specific pot used.

 

[RoadShow]

Hi @RoadShow

Mysterious! It looks to be well thought out and neatly built, and I'm not seeing the issue yet.

If you could run some more resistance tests, with an 8 Ohm speaker plugged into Out1/2, and measure the input resistance in all settings. It should stay around 8 ohms within about 20% or so. Also check the switches are working, and that the 1-2-3 lugs are operating the way you expect. I had thought that all toggles were similar, until I found out that they aren't.

The jack sockets look to be metal, not plastic bodied, hence the 'ground' side, which usually is not a ground but is just one end of the OT output winding, may be live to the case. Unlikely to cause the issue, unless something else is touching it too. (which is why plastic bodied is better)

Also, I'm suspecting something about the jack connections, only because its hard to see them in the pics.

I'm assuming that your amp at full volume is seriously loud? with minimum -7db attenuation on this M2 design, 18W is still delivering about 3.5W to the speaker, which is usually still too loud.

And just for others looking in, we have collectively built dozens of these now so we know that they work fine.

@JohnH THX John!!

The switches verify good.

With an 8 ohm speaker plugged in the readings vary from 7.9-8.8 ohms. With a 16 ohm speaker plugged in the readings vary from 7.3-9.6

The big news is that when I removed the jacks from the housing and let them lay loose like the switches it all started to work!! I have plastic jacks on had so I'll be swapping those later today and buttoning it up.

I will say that the minimum -7db puts it at a good quiet level with the 18W'r and guitar turned all the way up. I'll try it out with my JTM45 later.

Thanks for the tip, that was it!! Reading this thread a couple times over the last year and following along I knew it had to be something I was doing.

Bill

 

[JohnH]

Hi @RoadShow

That's great news, and I'll be looking forward to hearing about how it goes with the JTM45.

Those new resistances are spot-on as expected. Im still a bit curious though, about what was happening with the current metal jacks. All the sleeve connections on the jacks are all connected to each other anyway, so if they were all connecting to the case as well that shouldn't have really changed anything, unless something else was connecting to something?

 

[RoadShow]

Hi @RoadShow

That's great news, and I'll be looking forward to hearing about how it goes with the JTM45.

Those new resistances are spot-on as expected. Im still a bit curious though, about what was happening with the current metal jacks. All the sleeve connections on the jacks are all connected to each other anyway, so if they were all connecting to the case as well that shouldn't have really changed anything, unless something else was connecting to something?

Hi @JohnH ,

OK, I hooked it up to the JTM45 with a 16ohm load and it works fantastic. IMO it serves itself better on something like that or bigger. I think for use on an 18W Marshall the default -7db is sufficient. With gain and volume all the way up it's barely audible in the next room.

Hmm, I'm thinking I could build just the front end with smaller power resistors for use with the 18W, maybe 25W resistors? All I would need is R1, R2A&B, and L1 and I'd have a mini attenuator.

If I built the 8ohm version with the 16ohm option would I use the same values for R10 & R11 of 10 & 68 ohms?

Thoughts?

Thx, Bill

 

[Osman Qureshi]

Hi @JohnH ,

OK, I hooked it up to the JTM45 with a 16ohm load and it works fantastic. IMO it serves itself better on something like that or bigger. I think for use on an 18W Marshall the default -7db is sufficient. With gain and volume all the way up it's barely audible in the next room.

Hmm, I'm thinking I could build just the front end with smaller power resistors for use with the 18W, maybe 25W resistors? All I would need is R1, R2A&B, and L1 and I'd have a mini attenuator.

If I built the 8ohm version with the 16ohm option would I use the same values for R10 & R11 of 10 & 68 ohms?

Thoughts?

Thx, Bill

Hi all,

I'm new to the forum. What a great community you all have here!

Before I start building a JohnH attenuator myself, I would also be interested in any design tweaks that would make it even more suited for a 20W amp. And does the attenuator work as a dummy load without a speaker attached (the DSL20HR requires a speaker attached to use its emulated out, so could I just attach the attenuator without the speaker)?

By the way, I'm very impressed with the design of the attenuator--is there a name for it yet? I haven't seen a design with an inductor before--ingenious! With the MiniMASS now costing north of $170, I won't be surprised if clones start showing up on eBay, or even Alibaba.

Thanks, Osman

 

[JohnH]

Hi @JohnH ,

OK, I hooked it up to the JTM45 with a 16ohm load and it works fantastic. IMO it serves itself better on something like that or bigger. I think for use on an 18W Marshall the default -7db is sufficient. With gain and volume all the way up it's barely audible in the next room.

Hmm, I'm thinking I could build just the front end with smaller power resistors for use with the 18W, maybe 25W resistors? All I would need is R1, R2A&B, and L1 and I'd have a mini attenuator.

If I built the 8ohm version with the 16ohm option would I use the same values for R10 & R11 of 10 & 68 ohms?

Thoughts?

Thx, Bill

Its great that its working. You can definitely just build one with the front end and leave off other stages as you wish. Component values for the resistors stay the same in terms of Ohms, but you can scale back the power ratings if you want to.

R1 takes about 1/2 the total power, and I reckon on a factor of x3 to allow for overdriving the amp and imperfect cooling. On that basis with an 18W amp, R1 rating is 18 x 50% x 3 = 27W. If your 18W amp is heating up the current build so its reasonably warm, then maybe use 50W for R1 and 25W for R2 A and B. If its staying cool with your use, maybe you can use 25W far all. Or use all 25W and use two in parallel for R1 (eg two 30 Ohm in parallel to make a 15)

On the inductor, stick with 18 or 19 gage and the same inductance.

Just a comment on volumes though, and id be interested in any further comment from others particularly @Gene Ballzz . At home, Im running with my VM2266, which is basically an updated JTM45 and puts out about 35W-40W from its KT66 output stage. I often run it at full -31.5 db attenuation at home, and its like moderate TV volume late at night - still very clear and loud enough. That's cutting 35W down to about 25mW, and I'm at about 6 on the amp volume! Now its much more fun at about -21db or -24db, but then it is filling the house with sound. The equivalent to those settings for an 18W amp would be just 3db less. I reckon for a small gig I might be at about -14db with this amp. Obviously it depends on lots of things like how hard you drive the amp, what the speakers are etc

Im sure we all have different perceptions about how loud is loud, but your rig as described seems like it might be being quieter than Id have expected overall. That's just a thought, no prob if its all working at your end.

 

[Gene Ballzz]

Hi all,

I'm new to the forum. What a great community you all have here!

Before I start building a JohnH attenuator myself, I would also be interested in any design tweaks that would make it even more suited for a 20W amp. And does the attenuator work as a dummy load without a speaker attached (the DSL20HR requires a speaker attached to use its emulated out, so could I just attach the attenuator without the speaker)?

By the way, I'm very impressed with the design of the attenuator--is there a name for it yet? I haven't seen a design with an inductor before--ingenious! With the MiniMASS now costing north of $170, I won't be surprised if clones start showing up on eBay, or even Alibaba.

Thanks, Osman

@Osman Qureshi

I call it the JohnH Design, JohnH Attenuator or simply "The JohnH."

I'm not certain that these will be appropriate for use as a "dummy load" but it is not required for your DSL20 to use the "emulated out." All you need to do do is turn the amp on and leave the "LOW/STB/HIGH" switch in the center "standby" position. The speaker (while still connected as a load) will not produce any sound, but the "emulated out" should function just fine!

And To The Forum!
Gene

 

[JohnH]

Hi all,

I'm new to the forum. What a great community you all have here!

Before I start building a JohnH attenuator myself, I would also be interested in any design tweaks that would make it even more suited for a 20W amp. And does the attenuator work as a dummy load without a speaker attached (the DSL20HR requires a speaker attached to use its emulated out, so could I just attach the attenuator without the speaker)?

By the way, I'm very impressed with the design of the attenuator--is there a name for it yet? I haven't seen a design with an inductor before--ingenious! With the MiniMASS now costing north of $170, I won't be surprised if clones start showing up on eBay, or even Alibaba.

Thanks, Osman

Hi Osman and welcome to our thread. There's nothing particular that would make the design better for a 20W amp as compared to the base designs which target 50W amps. But if you want to, you could reduce the 100W and 50W resistors down to 50W and 25W respectively - with the same Ohm values.

Yes you can definitely use this as a dummy load with no speaker, you just set the attenuation to max. At this setting, the amp doesn't see much of the speaker anyway (about 0.1%!), its fully loaded by the attenuator. But, I see Gene has pointed out that you may not need to worry anyway, using your particular amp (but I know there's been a couple of versions of DSL20's, which may differ in this respect).

If silent/headphone playing/recording is a key thing that you need, you might consider adding a line-out. ie, its another output fed with a small signal via a resistor an a pot. From this you can go to an external cab-sim or IR box. Some guys have built this feature in - easy to add later if you save enough space. Note that a line out wont sound any good as a direct signal though, it needs the cab sim etc. Whether that's worth doing might depend on how good the amps own emulated out is.