Low power inline attenuator

[Zozobra]

Afternoon folks,

I've just about finished up a 3-4W ecc99 amp for home recording and even at this low power it's probably still a bit too loud to crank without upsetting my neighbour, who is a lovely person so I want to avoid that. With this in mind I've been trying to understand the 0.1/1W inline attenuator found in the 1W 12AU7 based Marshall's, which I've posted a snippet of below.



The interaction with the OT taps is confusing me a little. My first thoughts is something like this and I'd appreciate a sanity check

16 ohm load

We have the 82R resistor in series with the 16R load, all in parallel with the 18R resistor to ground, thus

Rtot = 1/(1/(82+16) + 1/18) = 15.2 ohms

This looks like a sensible load for the power amp.

The 8 ohm tap doesn't do any thing here so we can ignore it.


8 ohm load

We have the 47R in series with the 8R load, all in parallel with the 18R resistor to ground on the 16 ohm tap.
This is where the interplay with the 16 ohm tap confuses me. So do I halve the value of R18, such that:

Rtot = 1/(1/(47+8) + 1/(0.5*18)) = 7.7 ohms

This looks like a sensible load for the tap. Have I missed a trick here?


Additionally, how should I calculate the dissipation in the resistors (I'm sure it's simple ohms law stuff but I'm having a brain fart today and I'm not entirely sure whats happening for the 8 ohm load...) and what is the purpose of the capacitor? It's omitted on the JTM1 for some reason.

Cheers
Ian

[Zozobra]

I think I understand the power dissipation now if my understanding of the circuit is right...

[Chuck H]

The load at HF in the above schematic is shunted to zero impedance. Where with an actual speaker load the HF would be much higher impedance than even the rated load. That's never caused a problem IME but I just thought I would note it. I'm sure the cap is there for tonal, rather than impedance corrections.

Players choice in purely resistive attenuators seems to lean toward those that use a higher resistance than the correct load match. This is likely, again because actual speakers have an AVERAGE impedance that is higher than their rated impedance and only match their rated impedance in a fairly narrow band. So by using a higher impedance, say, a 100% mismatch (like a 16R resistor to replace an 8R speaker load) the amp may actually feel and respond and operate more like it does with the speaker. Where the "correct" 8R load would represent a lower impedance at most frequencies to the amp than an actual speaker would. Just food for thought.

You seem to be reading the circuit correctly and I've never had any trouble or complaints just using the rated load value replacing speakers with resistors. YMMV and as noted, most players seem to prefer the higher resistance WRT the rated speaker impedance. No one gets hurt trying it and you can decide what you like for yourself

[Zozobra]

Thank for looking over it. Good point on the higher resistances being potentially preferable. I'll give that a go. My usual supplier doesn't stock 5W 18R resistors but they do have 22R so I'll try that Now to work out the best way to wire this to a switch.

[J M Fahey]

That is a cheesy attenuator but hey, that amp is practically a toy, so ....

DonŽt overthink it, zozobra got it right, just missed an extra step, focused on impedance (good) but forgot attenuation
a) this system attenuates by placing a 6X resistor in series with speaker .... a glorified headphone attenuator circuit .
So the 8 ohms tap now sees 47 ohms in series with 8 ohms or 56 ohms total; or 82 ohms in series with 16 ohms for 98 ohms total.
Attenuation is accomplished and if it were an SS amp weŽd stop here, but a tube amp wonŽt like 7X expected impedance on a tap so they load *one* of them with the proper resistor, in this case 18 ohms in the 16 ohm tap, which gets in parallel with all taps since they are connected by the transformer itself.

Doing things on different taps is no big deal because anyway the tube plates see whatever is connectd to any of them and "reflected" to them by transformer action.

In a nutshell, a simple and clever circuit, and although the effect of crude attenuating by simple series resistor and nothing else makes for buzzy sound (the plague of all simple headphone outs) ,in this case may help give the speaker bite.

[olddawg]

Honestly, I have never used a purely resistive output attenuator that I thought sounded good. You can get a little more control using L pads: MCM Electronics - Search Results for L pad
If you are serious about recording consider a Weber Mini Mass attenuator. They use an actual speaker motor. The 25 watt version cost less than $100. If you gig and use a variety of small(ish) amps I would recommend the 50 watt version. All have a built in line out and you can get a footswitch option: https://www.tedweber.com/
There are also always a few used ones available.

[Zozobra]

Thanks guys, I'm not expecting miracles from it but it's a nice bit of functionality for not much cost and it'll probably sound better than turning down :P

This doesn't sound bad IMO.