Edit: August 12 2018: This thread has developed to include a range of resistive and now reactive attenuator designs - I've changed the title accordingly. A Fixed Attenuator This is about a simple fixed-level attenuator that I've been using with my Vintage Modern VM2266c. I built this mainly for low volume practice at home, in order to be able to get the volume up a bit but without disturbing others too much. Also, there's a lot written about attenuating a valve amp and how it does or does not change the tone, and I wanted to understand this better myself. The following is still in a testing form. Schematic I've tried a few recipes, but the current one that I've been using these past few months has three fixed resistances, combined to in theory take -7db off the input power, which is a x0.2 power reduction, so 40W becomes 8W. Each resistance is made up of several 5W power resistors. With 40W from the amp and an 8Ohm load, each resistor is dissipating between 2W and 4W. The idea of the three resistances is to not only match the Ohms as seen by the amp, but also to control the impedance as seen by the speaker. This gives the speaker a bit of damping but not too much, as when directly connected to a tube amp. This is different to an SS amp which has super-low output impedance, which creates high damping and suppresses some of the speakers natural response. Construction It's currently built on tag-board, inside an aluminium box 120mmx90mm. Plastic jacks make sure the box is not connected at all. Sounds in use I like the sound of my VM2266c best at about 5 on the dial, at which point it is way too loud for home, but this box makes it feasible. At much lower settings, the attenuator squashes the residual hiss and hum and lets me get up to 2 or 3 for a reasonably quiet practice. At anywhere up to half-way volume, the resistors are not noticeably warming up at all. Although, based on the nominal 5W ratings of the resistor, it could theoretically absorb the full amp output, as currently built it's not really set up for that. The resistors would need to be 10W types to give a margin, and it needs to get some better ventilation happening too. So I don't use it for rehearsal or gigs, and there really is no need with this amp since it doesn't need to be cranked to 10 to sound great. Tests Here are some tests from the VM2266c: I miced up the amp with a Rode M1 and recorded via a neutral mixer into Audacity. I put a strum from a bridge Hb into a loop pedal, then recorded at Volume 5.5. I used LDR mode with Body at 6 and Detail at 8. Then with no change of settings and the loop still running, I closed down the amp, inserted the attenuator, switched on and re-recorded. Normalizing both traces showed that Audacity was finding a -9.7db difference in the attenuated, rather than -7db as calculated. This is the sound: There are three of the full strum, then three of the attenuated strum, to show what this approx -10db attenuation sounds like. Then there are three of each, full then attenuated but normalised to equal total volume. With those you can listen for tonal differences: https://soundclick.com/share.cfm?id=13646224 This is the first part of the waveform, full and then attenuated and nomalised: THey are very similar but you can see a few differences. Then I exported the frequency spectrums into a spreadsheet to compare them, by subtracting one from another. This showed that, after adjusting for overall level, the attenuated sounds have a slight boost in the range up to about 1500hz and a slight dip above that. The differences are about 1 to 2 db. Its not much but you can just hear it. It's not enough to really change the character of the tone or turn a good sound bad, but it is a small noticeable difference, within range of EQ if needed. This is the two spectrum plots. the green line is the difference in db, showing how the attenuated sounds has a db or so more mids and less treble: I hope that is interesting to some. I'm finding this useful, and it was a good learning experience to test it.