Simple Attenuators - Design And Testing

[JohnH]

Hi @TomBallarino , thanks for posting that. It looks really great and you have pushed the design forward in a few areas. So Id love to ask a few questions, which will help others:

How are you finding the results from it so far?

I'd really like hear about your impression of the difference between having the bass resonance circuit in or out, (ie keeping L1 in but shorting L2 and C1)

Whats your line-out sub circuit?

What values did you work out for the -1.75 stages?

What fan did you use and what's the circuit for it?

thanks agsin for posting!

 

[TomBallarino]

Hi @TomBallarino , thanks for posting that. It looks really great and you have pushed the design forward in a few areas. So Id love to ask a few questions, which will help others:

How are you finding the results from it so far?

I'd really like hear about your impression of the difference between having the bass resonance circuit in or out, (ie keeping L1 in but shorting L2 and C1)

Whats your line-out sub circuit?

What values did you work out for the -1.75 stages?

What fan did you use and what's the circuit for it?

thanks agsin for posting!

Hi,
of course, here are (some) answers:
I did not use the device in attenuation mode yet (at least not with proper volume levels, just "bedroom"-tests), so I cannot say anything about the efficiency of the resonance circuit. I am using it as a load box for recording most of the time, and I have to say that even in this mode the type of cabinet simulation may change the sound more than switching the resonance circuit, but as soon as I did some more tests I will report my findings.
The interesting part of the schematic is this:


I used a 40x20mm low noise fan (Noctua NF-A4x20 FLX), and took the circuit from a drawing of the Suhr reactive load if I remember correctly. Both fan supply and lineout are tapped directly off the amp output as seen in schematic.

I made some measurements with the Arta LIMP software to confirm that everything works as expected, here's a chart with everything set to full attenuation, and the resonance circuit not shorted (=active):

I was a little suprised by the height of the resonance peak, in my simulations (which included any real parasitic resistances, capacitances that I could get my hands on) the peak was higher.

I simulated the whole circuit in ngspice to get values for max power, max currents etc, which, of course, just confirmed your calculations!
As seen in the schematic I used 68/3 ohm resistors for the -1.75 stage. I am not sure how much of a boost is needed, I will report that too when I test in practice

Tom

 

[JohnH]

Thanks @TomBallarino

That's very helpful indeed.

I agree with the values for the -1.75db stage

The fan system looks nice and simple. You have a set of components there that you have found to work, with a specific fan that they work with. We have never had that info here before so now you have made a reference that others might try (at their own risk of course!). I've not tried a fan system myself, so have never been able to advise on specific components. If you found a need to spin the fan faster, a bridge rectifier of 4 diodes might do it instead of a single one. But there's no need to drive it any harder than needed to start it spinning when at very high power. Your box is nice and heavy, and as you found, it takes a long time to heat up. That means that where the temperature ends up in practice is based on an average power over about say 30 minutes.

Your line-out looks fine too. Quite a lot of voltage reduction with 47k and 5k. So those finding a need for more signal, maybe when using a smaller amp, might use a smaller value than 47k.

Here's my plots. representing what you have, set at -14db:

This one is the impedance. Red is based on a fit to a real measured 4x12 cab, and blue is the calculated impedance with that speaker after the attenuator, set to -14db:



The real speaker has a huge peak, but the attenuator tracks it very closely except at that very specific frequency. I think that's fine and is close over a broad range and matching the exact peak is neither necessary nor desirable.

I calculated with a low loss film cap (dissipation factor 0.001 at 100hz), an L2 coil of 0.4 Ohm (15 or 16 gage). I used a 9mH, which shifts the peak frequency a tad but not the height. My calculated impedances are just a bit higher than your measured ones, maybe more like your calculated values? This is apparent both at resonance, and at say 5kHz I wonder if the fan system, taking a bit of power, is making that small difference? Not a big deal though.

Here's the calculated signals, at a real speaker (red), at the amp driving the attenuator (green - equivalent to the lineout position), and at the attenuated speaker.

 

[Gene Ballzz]

@JohnH
I find it so hard to believe that folks actually come here and then end up buying a Bugera PS1 or Weber MiniMass! Of course, than they stae as a supposed fact that attenuators simply suck tone! Oh well, you can lead a horse to water, but.......
Great Stuff!
Gene

 

[TomBallarino]

I agree @JohnH , the line out resistor value 47k is a little bit on the high side. It was calculated for my DI box input (max +10dbU) and a 100W amp (so calculating with 130W spikes), and the pot fully open. This is a little bit of overkill, as you can always turn down the pot a little bit when really driving the load box with such a big amp.

With my component values I simulated the resonance peak to max out at around 24 Ohm:


As everything else (e.g. output levels at various settings) was nearly exactly as predicted through simulations this puzzles me a little, but if it sounds good?

I am thinking of building a M2 version without line out for live purposes as the current version is a little bit on the heavy side, and as you and some others have noticed, the bass resonance does not do very much soundwise when used as an attenuator. But I will do some tests first with the full M3 version.

Tom

 

[JohnH]

My calcs are very close to yours. But, i reckon the fan circuit could make the difference. The fan resistor is 82, so I'm guessing maybe about 100 ohm total for the whole fan circuit?. 24 in parallel with 100 is 19, close to the measurement. Could that be right? I think LCR circuits are just resistive at resonance, so maybe that simple math could be about right.

 

[TomBallarino]

My calcs are very close to yours. But, i reckon the fan circuit could make the difference. The fan resistor is 82, so I'm guessing maybe about 100 ohm total for the whole fan circuit?. 24 in parallel with 100 is 19, close to the measurement. Could that be right? I think LCR circuits are just resistive at resonance, so maybe that simple math could be about right.

I did the same impedance measurement with (green) and without (yellow) the fan connected:

It shows some difference, as expected. The measurement was done with about 1.5W@8Ohm, and the fan was therefore not turning. I expect slightly different absolute values with the fan connected and actually working, but I do not have the equipment to test that.

 

[Gene Ballzz]

Hey @JohnH & All Others,
Having built two of these already and at least two or three more in the planning stages, it has become evident that the most time consuming and pain in the butt parts of the process is the drilling for ventilation! Does anyone know of a good source for "pre-ventilated" project boxes for our purposes? In various sizes/shapes, etc. One like this seems pretty close, but is kinda pricey, not really the best size (for me at least) and possibly not availble globally:

https://www.ebay.co.uk/itm/273942089907?hash=item3fc83890b3:g:pFgAAOSwBGJeykJs



Shipping is also outrageous and the IEC connector is not necessary, although may be handy for fan power, but......

Watcha Think?
Gene

 

[Tone_Chaser]

I built the M1 for a Trainwreck Rocket clone I built because most of the time I couldn't get it loud enough to get it to edge of breakup. The attenuator worked great and it really works great with my drive pedals.

I'm building a high gain amp now that has a good master volume so I don't need an attenuator for it, but I'm getting involved with a project mainly as the vocalist but will be playing some guitar. They use processors instead of amps on stage. Could I use the simple 3dB circuit as a reactive load and not use a cabinet without damaging the amp? I'm building a line out into the amp so I could use an IR loader for speaker simulation.

 

[JohnH]

hi @Gene Ballzz
That looks to be a nice case. I found it also on eBay here, for the same equivalent price, which is a lot once its doubled by shipping. What I find tricky with the drilling is getting all the resistor screw holes right. Despite my best efforts or marking and centre punching there's always a couple that get ugly. For venting I use use a couple of lines of big holes. According to the physics of fluid flow, one 10mm hole is worth more than 4x a 5mm. I recently inherited a big floor-mounted drill press, so I'm hoping that will help reduce my workshop issues.

On that case it looks like it could all come apart for assembly. It would be great if all the resistors could mount to the side heatsinks. This would be good for heat control, and easy to wire up on the flat, before sssembling the case panels.

For other cases, I can buy Hammond-type diecast aluminium ones from a local shop (Jaycar), and that's the best option for me although theres no pre-venting. Id always want aluminium for conductivity.

 

[JohnH]

I built the M1 for a Trainwreck Rocket clone I built because most of the time I couldn't get it loud enough to get it to edge of breakup. The attenuator worked great and it really works great with my drive pedals.

I'm building a high gain amp now that has a good master volume so I don't need an attenuator for it, but I'm getting involved with a project mainly as the vocalist but will be playing some guitar. They use processors instead of amps on stage. Could I use the simple 3dB circuit as a reactive load and not use a cabinet without damaging the amp? I'm building a line out into the amp so I could use an IR loader for speaker simulation.

If you already have an M or an M2, you can use that as a load box if its the right ohms for your new amp. Just set it to max attenuation and no speaker. The amp doesn't know. You can add a line out circuit at the front, or you may well find that the signal from the attenuator output is at a good level to go to a mixer or IR. If you want a simpler dedicated box to do that with no switching, you could build that, I can help.

What all of those load-box scenarios don't reproduce is the bass resonance peak. So you may want to turn up bass somewhere. Its not yet concluded if this is really important or not (it only affects a few low notes), but in any case its all safe for the amp.

 

[Tone_Chaser]

If you already have an M or an M2, you can use that as a load box if its the right ohms for your new amp. Just set it to max attenuation and no speaker. The amp doesn't know. You can add a line out circuit at the front, or you may well find that the signal from the attenuator output is at a good level to go to a mixer or IR. If you want a simpler dedicated box to do that with no switching, you could build that, I can help.

What all of those load-box scenarios don't reproduce is the bass resonance peak. So you may want to turn up bass somewhere. Its not yet concluded if this is really important or not (it only affects a few low notes), but in any case its all safe for the amp.

I'm going to mount the one I built in the combo box so I'm going to build a dedicated unit for this. I'm building a line out in the amp, so just need something to take the load off the amp. Could build it into the amp as well but a stand alone unit is fine too.

 

[Gene Ballzz]

I'm going to mount the one I built in the combo box so I'm going to build a dedicated unit for this. I'm building a line out in the amp, so just need something to take the load off the amp. Could build it into the amp as well but a stand alone unit is fine too.

Building that into the amp (as a "dummy" load) might get a little dicey, due to the generated heat!
Just Something To Think About?
Gene

 

[JohnH]

I'm going to mount the one I built in the combo box so I'm going to build a dedicated unit for this. I'm building a line out in the amp, so just need something to take the load off the amp. Could build it into the amp as well but a stand alone unit is fine too.

This is about the simplest possible reactive load:



This one is for 8 Ohms. If you want 16 Ohms, double the values and pick the nearest standard values (ie 47 Ohm, 15 Ohm and 1.2mH). The resistors need to mount to a heatsink or aluminium chassis. The power ratings shown are probably good for about a 40W amp max. If you want to crank a more powerful amp, then Id go for a higher rating. A good way to get 200W for R2 would be two 15 Ohm 100W in parallel.

This circuit will do a very good job of tracking a real speaker from low mids up, and see if you want to add a bit of bass EQ to compensate for a bass peak at 100 to 120 Hz, which this does not do.

 

[Mcentee2]

Hello all, been away for a while enjoying my M2 build, can't believe how long it's been!!

Am building another amp now so am also pairing it with a new M2 - this one will have the full 100w R1, a complete bypass toggle as well as the other 3 toggles.

I have a larger Hammond box, painted grey this time, left over from another project so already has some air holes, lol!

Main question for posting now, is to ask if there had been any particular advice or gotchas learnt about layout?

Last time I just crammed all the resistors in more or less in circuit order.

Has anyone come up with any sort of "best practice" layout ?

 

[JohnH]

hi @Mcentee2 , great to see you again. I don't think there's one definitive layout yet, But I like the idea of using a larger box, for an easier build. How big is your Hammond box?

I think with those, its good if the box lid becomes the base, and most parts end up on the inside of what is then the upper surface. The coil too, if you have a neat way of fixing it, spaced a few mm from the aluminium, without a steel bolt involved. For those reasons, I think another good option is to let the coil be on the base, zipp-tied or siliconed down with a spacer.

Im planning to build a new one and write it up. But there are seversl great builds on recent pages. eg, by @malibu91 , on page 109.

 

[Mcentee2]

hi @Mcentee2 , great to see you again. I don't think there's one definitive layout yet, But I like the idea of using a larger box, for an easier build. How big is your Hammond box?

I think with those, its good if the box lid becomes the base, and most parts end up on the inside of what is then the upper surface. The coil too, if you have a neat way of fixing it, spaced a few mm from the aluminium, without a steel bolt involved. For those reasons, I think another good option is to let the coil be on the base, zipp-tied or siliconed down with a spacer.

Im planning to build a new one and write it up. But there are seversl great builds on recent pages. eg, by @malibu91 , on page 109.

Thanks John, I have read back a few pages and looks like these builds are going from strength to strength with your constant help and advice, so thankyou again!!

My box is pictured below, 22cm X 14.5cm X 6cm

So it should be large enough for good spacing, although I have yet to get all my resistors together to place mark them.

And yes, the intention is to fasten everything to the "top" of the enclosure, the flat lid currently has 4 rubber feet so will stay as the bottom

@malibu91 that's a lovely looking internal layout. I see you kept all the ground sides together in one place, and also you moved the stages around for a better spread of heat from your most used resistor combinations

My first thought was to somehow split up the R1 and R2 a/b group around the box still somehow minimising the wire runs.

 

[Tone_Chaser]

This is about the simplest possible reactive load:

View attachment 99292

This one is for 8 Ohms. If you want 16 Ohms, double the values and pick the nearest standard values (ie 47 Ohm, 15 Ohm and 1.2mH). The resistors need to mount to a heatsink or aluminium chassis. The power ratings shown are probably good for about a 40W amp max. If you want to crank a more powerful amp, then Id go for a higher rating. A good way to get 200W for R2 would be two 15 Ohm 100W in parallel.

This circuit will do a very good job of tracking a real speaker from low mids up, and see if you want to add a bit of bass EQ to compensate for a bass peak at 100 to 120 Hz, which this does not do.

Thanks! That is perfect. Building a 20W amp so that is well within the parameters. 8 or 16 would be fine too.

 

[Gene Ballzz]

@Mcentee2 & @Tone_Chaser
The layout by @malibu91 in post #2161/page #109 is absolutely stellar and well addresses the length of wire runs. I think the longest runs are created by including the BYPASS switch, so without a BYPASS, it would be even more clean and neat. IMHO, it's one of the better layouts for an M2 I've seen yet! He did add the BYPASS and delete the OUT 3 jack. Be forewarned though that in his first pic, there appears to be two #3 resistors and that his wired pic shows that the resistors got moved around a bit. To anyone going this route, see my comments after that #2161 for improved cooling.

For those with 16 ohm amp output capability, its my considered opinion/experience that the 16 ohm version offers a bit more flexibility in the use of these units, over an 8 ohm version.

> First, it most easily allows the use of an 8 ohm speaker without messing with anything, tonally or impedance wise. and while the 8 ohm unit is still "safe" for the amp, with a 16 ohm speaker, it is not as sonically happy as a 16 ohm with and 8 ohm speaker.
> Next, it facilitates "parallel" use at 8 ohm amp output, if your amp also has that 8 ohm option. If your amp has no 4 ohm option, you can't use an 8 ohm unit in parallel. What the "parallel" use provides is a stand alone -3.5 db attenuation, while still incorporating the reactive function.
> Most Marshall amps offer at least 8 & 16 ohm operation andonly "some offer an additional 4 ohm output. There are also a few that have only 16 ohm output!
​

@JohnH has graciously designed work around options for multiple input and output impedances. The only downsides are that these "work arounds" increase the number of components involved, complexity of the build and the margin of error in both the build process and use of the unit.

Bottom line is my suggestion to be realistic about what useage this unit will actually see and build it that way. If you REALLY need it to be a "Swiss Army Knife" so be it, but the builder/user had better know his stuff and be able to pay close attention to details. On the other hand, the simple units are cheap and easy enough to build, that extra options may be better accomplished by simply building a couple or several different units.

Just My $.02,
Gene

 

[Gene Ballzz]

And oh, FWIW, I'm working on a layout/wiring diagram for an M2, very similar to the one by @malibu91 , but with the BYPASS switch deleted. I will submit it directly to @JohnH to allow him to choose whether it gets posted or not.
Cheers!
Gene