Simple Attenuators - Design And Testing

[JohnH]

Dwayne, id suggest to take one step at a time and get it working. But are you thinking of doing VVR first rather than attenuator?

I don't know anything about VVR's but if there is much more to discuss about them it might be better on another thread.

 

[Dwayne Eash]

Dwayne, id suggest to take one step at a time and get it working. But are you thinking of doing VVR first rather than attenuator?

I don't know anything about VVR's but if there is much more to discuss about them it might be better on another thread.

I was hoping you'd let me know and update on the diagram, so I can search out how much this should cost. I expect to spend a couple hundred on getting my amp going like it should. Here's hoping I can figure out which way is most effective, soon. Are you waiting on something from me?

I'd like to do both, but I also need to buy a set of tuners too as my set just failed, so no more playing until I fix the guitar.

 

[Dwayne Eash]

True, John has done alot of work to find a way to blaze a trail over or thru the metal wall of attenuation
A VVR is basically just a pot and MOSFET, but you have to know exactly where to put it, and make sure it is done well, as you are tapping into the highest voltage in the amp, 300-500 VDC range.
Someone on this forum installed one, you can read here: http://www.marshallforum.com/threads/18w-vvr-installation.71431/

I can do intense bible theology, even textural criticism, and that's some heady deep stuff. I do alternative energy, and alternative health and medicine, I've built a dozen computers and a network from scratch, but I don't know electronic schematics.

But I can probably install what's needed, if someone points me to the right parts. How about a variac and just apply it to that spot, instead of a VVR? I prefer buying a variac for inventor experimenting.

 

[telesto]

I was hoping you'd let me know and update on the diagram, so I can search out how much this should cost. I expect to spend a couple hundred on getting my amp going like it should. Here's hoping I can figure out which way is most effective, soon. Are you waiting on something from me?

John's M2 diagram is here. Resistor power ratings are the same as the M1 I guess. Total cost should be well under $100 depending on where you shop and shipping costs, etc.

I can do intense bible theology, even textural criticism, and that's some heady deep stuff. I do alternative energy, and alternative health and medicine, I've built a dozen computers and a network from scratch, but I don't know electronic schematics.

But I can probably install what's needed, if someone points me to the right parts. How about a variac and just apply it to that spot, instead of a VVR? I prefer buying a variac for inventor experimenting.

I think your best best it to build John's attenuator, and get a Variac. VVR is "invasive surgery". Variac is just a step-transformer that you plug your whole amp into, so very simple. Just make sure you get one that can handle your amps power. Not sure how much one costs that can handle a guitar amps power, but I guess around $200 or so (?)

In any case, while it's interesting to compare VVR, Variac and attenuators, I think we shouldn't derail John's thread and keep the focus here on attenuators, and discuss the pros and cons of different volume reduction methods in another thread



...coming back to the attenuator: I'm slowly re-reading this thread trying to understand the concept behind it. So far I'm up to Page 3, and understand that John is not just doing the traditional speaker to amp impedance matching, but also going the opposite way and matching the impedance of the amp to the speaker. Which I think is the crux of the design? Do I have it right so far John? I didn't get to the inductor part yet, but I suppose it will somehow enhance the dynamics of the impedance in a reactive way?

Oh, and I remember you mentioned somewhere about NFB increasing the impedance of the amp (?) Is this good or bad in terms of this attenuator? Or would it be best to adjust the Presence/NFB and find a "sweet spot" and leave it there? And does the NFB need to be coming from the same tap? Eg. NFB is usually taken from the 16ohm tap, but will it matter to the attenuator if the attenuator is connected to the 8 ohm output and NFB to the 16ohm tap?

 

[Dwayne Eash]

John's M2 diagram is here. Resistor power ratings are the same as the M1 I guess. Total cost should be well under $100 depending on where you shop and shipping costs, etc.


I think your best best it to build John's attenuator, and get a Variac. VVR is "invasive surgery". Variac is just a step-transformer that you plug your whole amp into, so very simple. Just make sure you get one that can handle your amps power. Not sure how much one costs that can handle a guitar amps power, but I guess around $200 or so (?)

In any case, while it's interesting to compare VVR, Variac and attenuators, I think we shouldn't derail John's thread and keep the focus here on attenuators, and discuss the pros and cons of different volume reduction methods in another thread

...

I'm not doing the M2 design. If I do this, might as well have two footswitchable clean vol boosts, so I added two extra of the lowest attenuation stage for footswitchable control. I submitted to him MY modification and am waiting on him to ok the design so I can verify the part count and expense. Until he gives my design his pointers, I can't do anything on the design.

I can see your one of the smartest around when it comes to electronics, you should be proud of being such a professional. But I am am not familiar with electronic schematics, and even if I did, I don't know what each part does in the circuit. How would I know unless I work with it in a meaningful sense.

Just because I am asking for pointers so I can proceed with my project, does not mean I wish to derail the focus on this thread. That's a bit insulting. Just because I am not on your level, that should make you happy to welcome others. You continue in a discussion that only the few understand and can participate in.

I can pick up a variac of 500 or a 1000 watts, for about 90 bucks. I suppose I could make the attenuator for $150-175, but I don't know how much the footswitch end of things costs. So I was hoping he would help me find a suitable footswitch and relay to use, at least something to start with, as his designs do not seem to use footswitches, so I assume that unusual request on my part, has put a snag in his ability to verify parts for my design.

I think I based mine off of M3. So M2 is not exactly correct.

I'm glad your asking him questions, the design seems a bit simple for such a good result, but if it works, then I'm all for it. I feel certain Ted Weber's MASS attenuators is likely going to be hurt by this development. But more importantly, if we guitarist are better off, then I'm all for it.

I'm not here to disrupt, just to take care of my amp that lacks singing sustain is all. No worries, after I get answers, I'm outta hear. I bought my amp through this forum, and that was about 10-11 years ago. I have not been back until now as I needed to get my amp going again. I don't push myself on others. Maybe after I leave, I wont be back for another decade.

 

[JohnH]

...coming back to the attenuator: I'm slowly re-reading this thread trying to understand the concept behind it. So far I'm up to Page 3, and understand that John is not just doing the traditional speaker to amp impedance matching, but also going the opposite way and matching the impedance of the amp to the speaker. Which I think is the crux of the design? Do I have it right so far John? I didn't get to the inductor part yet, but I suppose it will somehow enhance the dynamics of the impedance in a reactive way?

yes, that's right.

@Dwayne Eash , I can help with the schematic, to include the relay footswitch control. It will be based on the M2 design (EDIT: I just saw your last post, you can also use M which is what you drew, but M2 is easier to build).

Please could you review each of the following:

Amp tap: 8 Ohms (this should match you main speaker or speaker pair. alternative is 16 Ohms)

Stage 1: -7db reactive always on

Stages 2, 3, 4: -3.5, -7, -14 db each with a switch

Stage 5 and 6: -3.5, -7 db each controlled by a foot-switched relay, with footswitch plugged into a jack. when the footswitch is not plugged in, or if dc power is not provided for the relays, these stages will be off ie bypassed, not attenuating

Parts

1 input jack
2 output jacks wired in parallel (use one for an 8 ohm cab or both with two 16 ohm cabs)
input socket to plug in a 9V wall wart

2 jacks sockets for footswitch cable at main and footswitch box

All jack sockets to be plastic bodied TRS (ie stereo with ring contact)

3 toggle switches rated at 5A minimum

Aluminium clad case mounted power resistors

Air-cored inductor 16ga or 18ga wire

hookup wire, 18 awg stranded

solder

M3 nuts and bolts to fix resistors to the case

Thermal grease

Feet for case

Spray paint if you wish

diecast Aluminium case for main box, a smaller one for the footswitch

stereo jack plugs and 3 core flex for footswitch carrying dc (could be a lightweight mains flex or a stereo audio cable)

Questions:
Do you want the option of running a single 16 Ohm speaker (through attenuator still using 8 ohm amp tap) ? Its perfectly safe to do it with no additional parts, but the treble tone is not optimised (you lose a bit). One more output socket and two more resistors fixes this.

I recommend not to have an overall bypass switch, and not to have attenuation less than the always on -7 db stage. its safer, less current flows and its simpler wiring. If you want full volume, just don't use the attenuator. please could you confirm if you agree.

What would be your sources for parts? are you in the US? do you have any retail electronics stores where you could pick out most of the common components? probably need to order at least the resistors and inductor, probably separately.

 

[Dwayne Eash]

To clarify, I downloaded the M3 design to make my mods. But I also downloaded another drawing. Maybe I used more than one drawing.

 

[telesto]

Dwayne, I think we got some wires crossed. I'm definitely not that smart, and you are definitely welcome to participate in the conversation I wasn't aware you had asked John for a specific request and was waiting on that, I just assumed you were waiting for the M1 on the first page to be updated to the M2. But looks like John just got back to you on that

yes, that's right.

Whew. So then I don't need to read the next 30 or so pages?

Then just two questions:
1) Can you explain a bit on NFB affects on the attenuator?
2) How did you settle on the incuctor values? And don't say "complex mathimatical formula" At least say what you based it on, eg. compensates or simulates impedance changes in the amp? Or did you just play with values until it "sounded right"? And on page 11 or so, the inductor only gave "subtle" improvement to the resistive design. Is that still true today? Or did some new revelation happen between page 11-49 regarding the inductor?

 

[Dwayne Eash]

I guess I don't understand the differences between the three different models. I saw the three output taps and prefer'd that option, that's why I thought I used M3. Opps, I guess it was M4 that I used. I downloaded more than one diagram.

This part is confusing to me, especially now, as I thought I did the M4 version.

I feel a little bit like I need an explanation of the differences between each version. I read what you wrote about them, but I still don't understanding the differences.

I could use a wooden box, or better yet, a DIY metal plate, some something repurposed. I need to use anything I can re-purpose.

Yes for three output jacks.

Yes for gain stages as indicated.

As to attenuation on off bypass. If all it takes is adding a switch, then no, I do not agree, it should be installed so I don't have to uninstall the attenuator in order to bypass it. I don't mind a slight hit on highs. I play an SG on a JCM 800. Both are high pitched instruments, so some loss on the high end seems nearly desirable. But you know the tonal loss better than I do.

Yes, I agree with 7 dB as minimum constantly attenuating. No problem.

I'm in the US, I am very unfamiliar with parts sourcing, but I sorta figure mouser might come into play, IDK.

Thanks for your time and energy invested. I believe this is a great project, so if I don't do this project presently, it would be done next time I have the money to do it. So this is not a waist of time. I want to use a minimum of attenuation (to keep the sound more transparent) combined with a variac, both.

 

[Dwayne Eash]

Dwayne, I think we got some wires crossed. I'm definitely not that smart, and you are definitely welcome to participate in the conversation I wasn't aware you had asked John for a specific request and was waiting on that, I just assumed you were waiting for the M1 on the first page to be updated to the M2. But looks like John just got back to you on that


Whew. So then I don't need to read the next 30 or so pages?

Then just two questions:
1) Can you explain a bit on NFB affects on the attenuator?
2) How did you settle on the inductor values? And don't say "complex mathematical formula" At least say what you based it on, eg. compensates or simulates impedance changes in the amp? Or did you just play with values until it "sounded right"? ...

Thanks for clarifying and offering a welcome. Much appreciated. I might stick around. But I am an inventor, and into alternative energy and alternative healing, and humanitarian causes. So I stay extremely busy. We the people deserve so much better than what we get.

You are asking the kind of questions I'd be asking if I could, but because I don't understand what things do to a circuit and you do, so I just stay out. You bring a functional reality to the discussion, however, I should not underestimate JohnH's intelligence. He also has a good grasp of electronic circuits and electronic theory.

We guitarists have been needing something like this forever. So I am all for it. I just find it odd that more guitarist's don't want foot-switchable clean boost(s). Am I the only one??? Nah, that's can't be true, as it could be the best in transparent clean boost, indistinguishable between it, and the amp.

Because,, it is the amp.

 

[JohnH]

I started giving a letter to different designs once it looked like this would be a journey, starting with A. But when I got to M I stopped with letters, since I like M for Marshall.

M is the one most have built including me, it has two coils. It comes in a few versions depending on features included.

M2 was a newer idea last year and it performs the same but with one less coil. It's a nice idea and simpler. But what it does is the same as M.

M3 has additional components focussing on the low bass resonance, But it makes it quite a lot more expensive, and so far doesn't seem necessary. It's hasn't been built to my knowledge.

Any of then can have one, two or three output sockets

On boxes, attenuators heat up when an amp is driven hard, and a metal case (or a substantial heat sink) with plenty of vents is important for cooling.


NFB

Negative feedback reduces the output impedance of the amp at normal volume, reducing the high treble and bass peak. without NFB, amp output Z would be much higher, and I've measured that effect on my other amp with no NFB The attenuator adapts to this, see charts on page 46. It means it helps to make each amp and speaker combo sound like itself, at lower volume. You can hear it and calculate it, but I can't explain it.

The inductance is picked so that the signal at the amp going into the attenuator is very close to that of an amp driving a speaker in terms of how it rises with frequency. The amp then can react to that changing impedance, but that's not modelled in the analysis.

 

[telesto]

There is a question about the amp impedance that is bothering me. You said somewhere (in the last 49 pages) that you measured your amps 8ohm output and found it to be 20 ohms. How is this possible? We put 8 ohm speakers on the 8 ohm taps because we need to do impedance matching for optimal signal transfer. If the amp is showing 20 ohms and we connect an 8 ohm speaker it would stress the amp and not sound very good. (Actually a 16 ohm speaker would be better.) So in this sense, your measured 20 ohms makes no sense to me

You design the attenuator to still show the amp 8 ohms (why not 20?), but then to show the speaker an 18 ohm amp impedance. If there was no attenuator, the speaker would see the "natural" amp impedance, of what should (?) be 8 ohms. Some other attenuators show the speaker lower impedance, like Weber's L-Pad shows 2 ohms to the speaker. And my variable L-Pad shows probably even less, which makes for a very poor sound. Coming back to Aiken again:

In this case, the input impedance also stays constant, but the output impedance (*of a reversed variable L-Pad) gets higher as you increase the attenuation. This is great for guitar amps, up to a point. The increasing output impedance lowers the damping factor, which enhances the interaction between the amp and the speaker, giving a natural bass and treble boost as you increase the attenuation (sort of a "built-in" Fletcher-Munson effect compensation!). The problem is that it gets to be too much, and you end up with too much bass and treble boost and "hangover effects" as it was called in the old days of audio, and you get a flubby, fizzy tone at high levels of attenuation. This can be alleviated by limiting the maximum output impedance with an additional shunt resistor at the output.

So, as you design your attenuator, I think the amp impedance doesn't matter, or at least you show it the 8 ohms it expects for successful impedance matching, so that is fine. But then you show the speaker a higher impedance, which is what Aiken is saying is "good" in respect to attenuators.

My next question then, did you experiment with showing the speaker other levels of impedance? eg. what if you show the speaker 30 or 40 ohms? Would it sound better or worse? ...ah, wait, yes you did, you tried with 26 and said it was to shrill. Based on your plots and listening tests, you were happy with about 6 ohms showing to the speaker. But now you have Stage 1 configured to show 18 ohms to the speaker, so what happened there? You changed your mind and decided 18 was better than 6?

Coming back to the 20 ohms as the amp resistance: You calculate with 20 ohms amp impedance, and then calculate Stage 1 to show the speaker 18 ohms. Using 8 ohms as amp resistance in the same calculation would mean showing the speaker 15 ohms. Not really a big difference I guess. Main thing I guess is that it's higher than the circa 2 ohms that some attenuators give, which end in anemic attenuated sound. So I guess the the "sweet spot" of impedance to show to the speaker would be somewhere in the 2-20 ohm range, or actually more in the say 6-18 ohm range. Your original U-pad attenuator on Page 1 was showing 8 ohms to the speaker, which made sense to me.

*edit: sorry for editing this post several times as I worked some things out in my head in real-time

 

[JohnH]

lots of questions!

Actual Amp output imoedance has nothing to do with what speaker ohms it is designed to drive. Impedance matching, although part of some devices design, is not a thing for guitar amps. An amp designed for an 8 ohm load could have any output impedance from almost zero (A typical solid state amp) to almost infinite ( a valve amp with no NFB, running clean). You've seen how this affects speaker response.

You can't measure output impedance directly, but it's not too hard to work it out. You feed a small steady signal into the amp and feed it into say an 8 ohm load and measure the output voltage. Then repeat unchanged but with a different load resistance, say 16 ohms. If the signal remains constant, amp output resistance is very low. It it increases in proportion to the load, then it's very high In between you can take the ratio and work it out.

 

[telesto]

Hmmm...ok. True, DCR of an OT is near zero. So what is the difference between 4-8-16 ohm secondary taps? Since it's a coil, I guess something to do with the winding ratio, and how it reacts to different impedance loads?
Yes, sorry, lots of questions Well, I'm also kind of thinking out loud, and finding some answers to my own questions, but yea, feel free to jump in, waters nice
I found this somewhere, I guess you did the same or similar:

To measure the output impedance at the secondary of the transformer, as driven by the amplifier, do this:

Apply a 1 kHz signal to the input of the amplifier, so that the output is not too large, but easily measurable, perhaps 1/10 volt. Carefully measure the output voltage with no load connected. Next connect a variable load resistance to the output, and adjust its resistance until the output voltage is reduced by half. Measure the value of the load resistance. That value will be the output impedance.

Or you can use a fixed resistor. Measure the open circuit output voltage with sufficient drive to give you a small but easily measurable output voltage; call this voltage Vo. Now connect a resistor of suitable value; start with 10Ω. Measure the output voltage with the load resistor, RL, connected; call this voltage VL.

The output impedance is given by Zo = RL*(Vo/VL - 1).

 

[telesto]

Ok, so I was playing around with the fundamental concept of trying to show the speaker around 20 ohms, while still keeping the amp seeing 8. Keeping also the "simple" concept of just using 2 resistors in a reverse L-pad form. And it's a kind of tricky game to play! There are not that many possibilities. Using standard resistor values, I was only able to find 8 combinations that fit (chart below). I was focused more on keeping the amp seeing 8 ohms, and letting some "slack" with the values the speaker sees.

Do these look OK to you? I guess you also did some similar calculations over a wide range? You chose the -7 dB combo. I think you said to go with less attenuation would throw off the grand scheme when additional stages are added? So I guess the -9dB should be fine for Stage 1 as well? For your -3.5dB stage you went with 33 and 5.6 ohms, which shows the speaker a nice 18 ohms, but then shows the amp 9.6 ohms, which is maybe not the best option (?) I guess it's fine as additional stage, but in some early schematics, you bypassed Stage 1 and had the -3.5dB as a stand-alone stage. Or were you more focused on showing the speaker 18 ohms, and letting slack in what the amp sees?...Ok, yes, now I see you made a "single stage" chart like what I just did, but you did for 16 ohms, and were more focused on getting the speaker to see 18 ohms and giving slack in what the amp sees. What was the reason for that?

 

[JohnH]

You are reading my mind.

yes there has to be some slack in these numbers, given real resistor values and also not all may be available in high power types, and sometimes a different range is sold other than the standard series. I like my numbers to be as exact as possible, it helps me home in on how things are working, but then I use them to make the best available engineering compromise for a practical design.

People talk about mismatches between amp and speaker, and is it Ok or not, and its unclear enough to be a subject of discussion. But they are talking about x2 factors, 8 to 16 etc. In a tube amp, im pretty happy that a much smaller variation is definitely OK, so for an 8 Ohm load, if it happened to turn out between 7 and 10 then I don't see a problem for the amp, if it allows something to be achieved. So when I offer a -3.5db option, the amp sees about 10. Im not worried about that though Id rather it was nearer to 8. And when as is usually the case, that -3.5 stage is after the nicely balanced -7 stage, then it makes much less difference.

 

[telesto]

Ok, great
Yes, I know there is some wiggle room with impedances. But why did you target 18 ohms for the speaker to see? You were focused on keeping that, and allowing the slack on the amp end. I would think the other way, since you were OK earlier with the speaker seeing 6 ohms, that if it sees now somewhere around 15 or 16 it's still good.
In any case, yea, the 15 and 10 combo is a sweet spot of having both, 8 ohms to the amp, and 18 to the speaker, so I would agree that's the best choice.

 

[JohnH]

In the early posts, I was exploring (still am). I was happy that anything worked at all with just a few resistors, given all the BS and $ that get wrapped around this topic.

Once I'd learnt a few things, I wanted to push higher and I became more picky.

Other than the small compromise in the -3.5 only option, the design gives good control of both impedances at all settings

 

[Dwayne Eash]

I started giving a letter to different designs once it looked like this would be a journey, starting with A. But when I got to M I stopped with letters, since I like M for Marshall.

M is the one most have built including me, it has two coils. It comes in a few versions depending on features included.

M2 was a newer idea last year and it performs the same but with one less coil. It's a nice idea and simpler. But what it does is the same as M.

M3 has additional components focusing on the low bass resonance, But it makes it quite a lot more expensive, and so far doesn't seem necessary. It's hasn't been built to my knowledge.

Any of then can have one, two or three output sockets

On boxes, attenuators heat up when an amp is driven hard, and a metal case (or a substantial heat sink) with plenty of vents is important for cooling.


NFB

Negative feedback reduces the output impedance of the amp at normal volume, reducing the high treble and bass peak. without NFB, amp output Z would be much higher, and I've measured that effect on my other amp with no NFB The attenuator adapts to this, see charts on page 46. It means it helps to make each amp and speaker combo sound like itself, at lower volume. You can hear it and calculate it, but I can't explain it.

The inductance is picked so that the signal at the amp going into the attenuator is very close to that of an amp driving a speaker in terms of how it rises with frequency. The amp then can react to that changing impedance, but that's not modelled in the analysis.

That was very well put, awesome, and quite helpful. And I understood the vast majority. Thanks much.

As to the bass thing, uh oh, I'm kinda a bass nut. hehe I gotta have the warmth of my bass. I include a bass speaker in my rig, so my rig is probably warmer and has fuller mids than most. I feel lucky in my choice of speakers. So I don't know how transparent it already is.

So maybe I wouldn't mind one that includes lower frequencies, depending on the cost. How about a compromise, and do an in between version of M2 and M3?

I want sonic transparency, or the least change to my tone.

Do you have sound clips of the amp, using the attenuator, and not using it, and volume equalized after the amp, like reduce the mic level. Might have to make a couple recordings to get the recording equalized, or maybe that's easily done in the software, IDK.

That sort of sonic example, with and without the power attenuator, and also volume equalized somewhere after the amp, would help me examine how transparent it treats the signal after attenuation.

I'm guessing you might attenuate using mid 20's dB. Wherever the sonic difference might be noticeable. And have that, compared to without an attenuator, without changing the amp settings, and just adjust the volume after the recording. Maybe you can manually do this by adjusting the mic volume, to result in the same dB's.

If need be, I would give up a clean boost pedal, to gain a more transparent attenuator. And I'd give up the bypass switch too if that helps the tone. If less switches means less tone loss, then I'd be willing to make more of the stages, continues stages. Let me know.

I'll have to go see if you already posted such an audio demo. Thanks again! Great post. Your design is a very cool project. Very glad to see this.

 

[JohnH]

Hi Dwayne,sound clips as you request are already done and posted in post 1. Its attenuator M which will be the same as M2. Not sure if you saw these before.

Here they are:

Attenuator M: Max attenuation to non-attenuated:

https://vocaroo.com/i/s1QgVDnl1XQi


Attenuator M: Normalised:

https://vocaroo.com/i/s1cnuucLz8yK

Its a looped riff placed in my delay pedal working as a looper, pushing the amp quite hard (harder than I usually do). The first recording is with me then stepping through each setting on the attenuator from max attenuation through to zero attenuation, recorded from the speaker through a dynamic mic with no other settings changed. Then the second one is the same recording, boosted in software so each setting is the same volume. The recording level at the mixer was set for the loudest setting, so it s pretty quiet recording to start with.

Obviously your amp and playing, set your way, will sound different, but just listen for the consistency. The riff is played twice at each attenuation setting.

There's actually nothing wrong with the bass in M or M2, which is why I have never built M3. You still get the resonant bass peak which is created at the speaker. If there is a difference, it would be in the added distortion of the bass tone when you hit the very lowest notes very hard. it would only be the fundamental note as affected by the power amp, not the harmonics coming from the guitar, nor from the preamp or any distortion pedals.

The only difference between M2 and M3, is another coil and large cap. But to do this right with the right parts is about another $100. But, if we base your design on M2, these parts could be easily added later if you leave space for them.

There should be no tone loss with good switches. Can add it all as you wish. The reasons for suggesting not to have a bypass is that it is one switch that you need to be very careful with. Don't ever flick it while playing or else the amp can see transients. But that's not different in any such commercial attenuator of any type. I have literally never used my bypass switch on my 40 and 50W amps, nor has Gene who gigs regularly using smaller amps. Maybe if you have it, you could put it at the back while all the others are at the front? Other switches you can flick on and off at any time.