Simple Attenuators - Design And Testing

[Gene Ballzz]

@crashcarlson
to the forum!

All +/Positive/Hot AND -/Negative/Ground connections must be connected by wire and isolated from the chassis/box/enclosure.

Just Sayin'
Gene

 

[JohnH]

hi @crashcarlson
Yes as Gene says, and you are right about the jack connections.

There's no chassis grounding wanted here. Should use plastic jacks. The two wires that come out of the amp are usually from just the transformer windings and are not grounded either. so we dont need or want either of them being live to the attenuator box.

That design C is a couple of years ago, but its still a good version if you just want resistive design as simple as possible. Get resistors ated at least 3x the power dissipations shown, assuming a 50W amp.

 

[crashcarlson]

@Gene Ballzz @JohnH - Thanks for the tips! I did order 3x or 4x above the dissipation shown for the resistors (I do have a 50W amp). Good to know about separating hot/gnd from chassis, I don't have any plastic TS jacks so I'll have to order those as well. I am going with the simpler one first to see if I like how it works with my amp (never used an attenuator before)...I'll probably end up doing the 'final version' you posted with the inductors somewhere down the road as well.

Thanks again!

 

[crashcarlson]

@JohnH

I decided instead to build the M2 version with the single inductor, that you have laid out in this post (only mod would be a single output jack instead of two, since I have a 16 ohm amp driving a 16 ohm cab)
https://www.marshallforum.com/threa...design-and-testing.98285/page-56#post-2000746

I’m not the best at reading schematics, but after reading through the posts, and looking at all the build pictures, I think I understand it. I hope you don't mind answering some quick questions just so I know I've got this right. The one part that I am not 100% sure on, is how the switches connect the resistors for the different stages. I took this from your M2 diagram and made my annotations:




(I have SPDT On-On switches, so 3 poles)

This is how I’m understanding the schematic - could you let me know if this is right?

• Wire comes off of pole 3 on the switch by R3/R4, connects to right lug on R4
• Wire comes off right lug of R4, connects to pole 2 (common) on switch by R5/R6
• Wire comes off left lug of R6, connects to pole 2 (common) on switch by R5/R6 (same as above)
• Wire comes off of “top” lug of R5, connects to pole 1 of switch by R5/R6
• Wire comes off of pole 3 of switch by R5/R6, connects to right lug of R6

Lastly - in this diagram, when R5 is connected via poles 1 and 2 on the switch - is this attenuation stage on or off?

Thanks again!

 

[Oliver Gardiner]

Yes, that is correct. When the switch is connected between pins 2 and 3, the wire connection simply bypasses the series resistor (e.g. R4) and the shunt resistor (e.g. R3) is disconnected - hence the stage becomes just a piece of wire and there is no attenuation. In the alternate position, both resistors are in circuit and the attenuation is created by the series resistor (e.g. R4) in series with the "speaker" (i.e. the load presented by the next stage which, simplistically, can be taken to be around 8 ohms). The shunt resistor is in parallel with the two series loads (e.g. R4 + 8) and serves to bring the impedance seen by the *preceding* stage back to around 8 ohms. Because the stages are cascaded, yes, pin 3 of each switch is also connected to pin 2 of the next switch as well as the output of the series resistor - it shouldn't really matter how you wire it as long as everything is robustly connected with wire rated for enough current (like your speaker wire).

 

[Grey]

JohnH, I haven't read through the entire thread but I think I did see simulations. Have you posted those for download? I'm a spice junkie and wouldn't mind playing around with them.

I like the versatility of the M2v. Have you thought about removing 2 of the jacks from each end and replacing them with option switches for the various input and output impedances?

 

[crashcarlson]

Yes, that is correct. When the switch is connected between pins 2 and 3, the wire connection simply bypasses the series resistor (e.g. R4) and the shunt resistor (e.g. R3) is disconnected - hence the stage becomes just a piece of wire and there is no attenuation. In the alternate position, both resistors are in circuit and the attenuation is created by the series resistor (e.g. R4) in series with the "speaker" (i.e. the load presented by the next stage which, simplistically, can be taken to be around 8 ohms). The shunt resistor is in parallel with the two series loads (e.g. R4 + 8) and serves to bring the impedance seen by the *preceding* stage back to around 8 ohms. Because the stages are cascaded, yes, pin 3 of each switch is also connected to pin 2 of the next switch as well as the output of the series resistor - it shouldn't really matter how you wire it as long as everything is robustly connected with wire rated for enough current (like your speaker wire).

This is perfect, thank you for the explanation! Got all the parts ordered, can't wait to build it and try it out!

 

[JohnH]

hi @crashcarlson ,yes I think you have it all right. Im glad your building the M2!, it's a bit better than C, andbits worth it' for very little added complexity or cost

 

[JohnH]

Hi @Grey
I haven't posted the simulation files, but there are quite a few charts of simulation results Most of them are from a dedicated spreadsheet which has all the theory but i dont post that But these circuits are pretty simple so easy to set up in any Spice program that you use.

 

[Grey]

Hi @Grey
I haven't posted the simulation files, but there are quite a few charts of simulation results Most of them are from a dedicated spreadsheet which has all the theory but i dont post that But these circuits are pretty simple so easy to set up in any Spice program that you use.

I was just curious because it's easier to go from a starting point than scratch. I'll probably do an M2 16ohm version myself. The weird part is I'd be doing it for a 15 watt amp but will have to over build it so I can also use it on a friends Bugera head with a 4x12. I'll have to check his impedances and wattage Sunday as I'm not sure what they are. This is my first tube amp so I'm not familiar with it all yet, I'm more into design power supplies.

 

[RoadShow]

Where can I get the 900uH inductor? I found some at Mouser like
https://www.mouser.com/ProductDetai...=sGAEpiMZZMv126LJFLh8y2AtSXNc0tXos1ZVI4wA1hY=
but the wire gauge is 20 and not 18 as prescribed.

 

[Oliver Gardiner]

Where can I get the 900uH inductor? I found some at Mouser like
https://www.mouser.com/ProductDetail/Hammond-Manufacturing/1540M22?qs=sGAEpiMZZMv126LJFLh8y2AtSXNc0tXos1ZVI4wA1hY=
but the wire gauge is 20 and not 18 as prescribed.

https://www.ebay.co.uk/itm/0-1mH-3-...var=611258617282&_trksid=p2057872.m2749.l2648

1mm is close enough to 18awg

 

[JohnH]

On the inductor, apart from the gage, that mouser link looks to have a core of some kind, but what we need is an air-core ie no core at all!

The ebay one linked by Oliver above looks fine to me. Ive not bought inductors from ebay myself because the ones I see are never clear about the wire gage, but that one does state it and 1mm diameter is 18 gage (tables quote 18 gage = 1.02mm, same thing)

In general, these coils are sold at places where components for speakers are sold, and they are used in speaker cross-over networks.

Ive bought mine from a couple of places down here in Oz. But in the US, we have had success buying them from Madisound. They do a range at 19 gage, which is OK. id be happy with that up to about a 50W amp and Gene has built a few using them. Good value too.

https://www.madisoundspeakerstore.com/air-core-19-awg-page-2

They seem to list a 0.85mH and a 1mH. To be honest, either of those would be fine for the 0.9mH inductor in the M2 design. Ive sometimes bought a bit bigger then unwound it a little to reduce the value (easy if you have a meter that measures inductance), but thats being picky.

For a use with a more powerful amp , I think they do a 16 gage range too.

 

[RoadShow]

Thx guys,

I appreciate the guidance!!

 

[JohnH]

hi @Grey

Here's a picture of the kind of model I use, I this case its in 5Spice. Im sure you would use a more powerful sim program but this works for me.


What I do is:

A frequency sweep analysis, feeding through an assumed output impedance of the amp (I based it on 20 Ohms measured from my VM, but also test over a range lower and higher).

I run two models in parallel, one directly into a circuit model of a speaker (based on Aikens model) and another via the attenuator.

In the above plot, you can see the cap and inductor C9 and L20 from design M3, which you can negate to test without for M2.

All attenuation stages are there but the last two are effectively switched off, using very low and very high R values. Then I look at the two output plots to see how they are the same shape.

 

[Grey]

hi @Grey

Here's a picture of the kind of model I use, I this case its in 5Spice. Im sure you would use a more powerful sim program but this works for me.

View attachment 79829
What I do is:

A frequency sweep analysis, feeding through an assumed output impedance of the amp (I based it on 20 Ohms measured from my VM, but also test over a range lower and higher).

I run two models in parallel, one directly into a circuit model of a speaker (based on Aikens model) and another via the attenuator.

In the above plot, you can see the cap and inductor C9 and L20 from design M3, which you can negate to test without for M2.

All attenuation stages are there but the last two are effectively switched off, using very low and very high R values. Then I look at the two output plots to see how they are the same shape.

Thanks. I'm going to see if I can knock this up in LTSpice. I've not done any work involving circuits for audio so I'm curious to see how the speaker comes out. I've done enough work on custom models that I think I can work something out that could be fairly accurate. Not to re-invent the wheel, I may do a bit of research first to see if it's been done somewhere before, which I'm sure it has. I'll post my work if it bears fruit so that others can have a go.

I do have one question though. On your schematics you show a representation of an audio jack in a format I'm not familiar with. What would be the pinouts for those? I get the sleeve part but the other two could be interchangeable in my mind. Are you using 3 terminal jacks for everything and just leaving one pin un-soldered or recommending others do so? The reason I ask is because the connections seem rather odd for the 4ohm input and the 16ohm and 8/4ohm outputs on m2v considering that speaker plugs are two terminal devices as far as I know. I have an Egnater tweaker 15 combo and the speaker plug is a two terminal arrangement. I guess what i'm trying to say is on the 16ohm output one terminal connects to R10 and the other to R11 and the sleeve to ground and this is confusing from my point of view.

 

[JohnH]

Hi @Grey , For most of the jacks on my diagrams, simple 2-wire jack sockets are all that are realy needed. But I still like to adopt the TRS stereo ones because the extra ring connection provides added physical grip.

But there are a few cases where I do use the extra connection. In the M2V its part of connecting in the added RL network that adusts the input impedance, and also on the output, the adjustment for use of 16ohm cabs. For this, Im using the basic barrel of a standard mono plug being pushed into the TRS socket to connect the ring terminal to ground via the plug barrel, and so make the added connections wanted. An alternative is to add more switches instead.

 

[crashcarlson]

For the "M2" design would I be correct in assuming that either of the outputs (Out 1 or Out 2) can be used with a 16 Ohm cab (I'm building the 16 Ohm version and the amp output will be at 16 ohm)?

 

[JohnH]

Yes Out1 and Out2 are shown as two sockets in parallel, so they are the same.

Having two Outs could be useful if you want to plug in a second 16 Ohm cab as well, which works fine with this design, so you'd then use your 16ohm amp, 16 ohm attenuator driving two 16 cabs which is an 8 Ohm load.

 

[Yamariv]

Hi John, thank you for all your hard work on this attenuator!

I just received all the parts to make a couple M2's and wanted to see if anyone has done a wiring layout for it by chance? I am getting a lot better with schematics but I'm a bit worried about messing something up. If there's a layout available that would be easier for my brain to compute. If not, I guess I'll be doing it schematic style, just thought I'd check first..

Cheers