Magnatone Custom M10 - REVERB

[trobbins]

TelRay, can you elaborate on your spectrum software ?

I am used to REW and some other scope/spectrum software, that I must admit makes it very easy to do a spectrum measurement, and hence make changes (such as tone pot settings or part value tweaks) and visually review the change in spectrum. I must admit I haven't used that on an amp with reverb, or looked up how that could be done, although I know REW (and possibly others) has the software smarts to make measurements on reverberant rooms and just time window the 'direct' signal (and hence remove any room reflections) which may be akin to just measuring the response of the first pass through the reverb, and remove any reverb spring reflections.

[TelRay]

hi trobbins!
i am not using a spectrum analyzer per se but a parametric dynamic EQ plugin from WAVES (F6) that includes a spectrum analyzer interface
i use that because even if i can guess that what i see in the graph matches with what i think is missing i can actually boost that frequency and confirm i like it better that way. in this case, i’d be happy if i could bost the reverb a couple of dB in the 4-5 KHz area
i was not trying to analyze the reverb in terms of density, number of reflections, time, etc but simply tone. that’s why i have chosen this “simple” tool
the way i did the analysis was by recording a stacatto single guitar note on a delay / looper stompbox (the sound had no sustain so i could hear just 100% reverb after the initial sound)
i had the looper playing that guitar note and recorded it (with a standard SM57 mic) playing it through the amp at the different REVERB settings (0, 5 and 10). those are the files attached to the post above.
feel free to download them and play with your tools and post what you think

[TelRay]

btw, what do you guys think about placing a modern transistor to substitute the current original reverb recovery 2N2614?
I have no way to test a transistor and just assume it's working correctly because the REVERB works

[trobbins]

TelRay, yes I thought you were using some audio production software.

I was just thinking through an alternative test setup that could allow the base frequency spectrum response to be measured, without the addition of reverb tank 'echo' signals. That type of benchmark could identify if it is the tank driver and recovery circuits that provide the main frequency shaping, and not the remainder of the amp.

[TelRay]

thank you for the proposal
I’ll be happy to try if you can provide some guidance
I took a look at REW and it seems to be relaying on calibrated microphones to take the measurements. I do not have any. Is this absolutely necessary or a calibration I could perform myself?

[Helmholtz]

btw, what do you guys think about placing a modern transistor to substitute the current original reverb recovery 2N2614?
I have no way to test a transistor and just assume it's working correctly because the REVERB works

As reverb and dry signals are mixed before the transistor, it is not a simple reverb recovery stage. Rather it recovers both dry signal and reverb.
A modern transistor won't improve anything.

[trobbins]

I took a look at REW and it seems to be relaying on calibrated microphones to take the measurements. I do not have any. Is this absolutely necessary or a calibration I could perform myself?

REW is best used with a USB type audio interface (such as a Focusrite - I have a vintage but excellent EMU 0404 USB), which provides an audio 'output socket' to inject a test signal in the the amplifier input, and an 'input socket' to interface to the speaker output connections of the amplifier. REW can then be set up to inject a variety of test signals to the amp, and display the response at the output of the amp (either as a scope waveform, or a frequency spectrum plot, or compile a run of tests such as for distortion sweeps or waterfall plots). Like all tools, REW has lots of options and setups, so can be a bit bewildering at first, and yes its original and still main use is with audio systems using microphones and speakers (but that is not what I use it for). And adapting an audio interface to use for amplifier testing may also not be straightforward, as like an oscilloscope there is a need to include a probe to allow speaker level voltages, and like an oscillator signal source there may be a need to match test signal levels and impedances. It's all a great test tool system, but requires a learning curve and effort to set everything up, but once it is set up it takes just a minute to put it on the bench and connect it to an amp and then get pretty much instant test results.

But the type of test setup that may be needed to window a direct reverb signal, without subsequent echos, is unknown territory so perhaps I shouldn't have offered it as an alternative to your existing equaliser software technique. Do you have a DI to acquire a signal from the amp's speaker output, as an alternative to using a mic ? You may also be able to generate a test signal on a PC and apply that to an amp input. If the test signal was a short burst, then a recording via a DI may show up the direct burst, and then echo bursts. You could compare direct burst magnitudes at different frequencies to form a spectrum response 'curve'.

[TelRay]

thx trobbins,
i might contact you offline to discuss the details of the method, sounds interesting.
and... Happy 2021 for everyone!

[TelRay]

Is the dry signal from the vibrato channel ok for you? Do you notice any difference between channels 1 & 2 when vibrato and reverb are disabled on channel 1, and both channels have the same settings ?

hi all, hoping everyone is doing fine after all this time

I finally (FINALLY!) decided to keep working on this amp and tried what trobbins suggested but at a (LOUD) volume that allowed me to hear not only the losses in HIGHs but also in the LOW end when comparing channels 1 (VIB/TREM off) and 2 (NORMAL).

What I hear is what's represented in the graph below: measuring the output with the scope (speaker still connected) i see that both channels have the same output in the 600-1,000 Hz range but the losses in the highs and lows are very noticeable on CHANNEL 1


(all voltages P-P)

The next step was to follow the signal path throught the circuit.

Firstly I set both channels volume so that with an input of 140 mV p-p at 800 Hz (a frequency at which both channels showed the same output) they would show the same voltage at the input (PIN 7) of the 12DW7 tube (where CH1 and 2 meet) marked in YELLOW



Secondly I set the input signal at 80 Hz as it is a frequency at which there is a big gap between the 2 channels output and followed the voltages through CH 1's signal path.

All VOLTAGES are mV Peak to Peak



NOTE A: if the REVERB TANK is disconnected this value increases to 4,700 mV pp
NOTE B: at this point CH2 shows 600 mv pp at 80 Hz and CH1 shows about the same value at 800 Hz.

Thank you in advance for your ideas!

PS: the reverb tank is an accutronics 4FB2A1C suggested in a previous post as the original one was missing

[Delta362]

Were all the tone controls set to the same settings between the two channels? Do you have the same level signals at the outputs of the 2nd half of the first 7025 for each channel (highlighted in yellow)? Your description of the high and low frequency levels sounds like there are some heavier cutting filters in channel 1 at those frequencies in the second and third preamp tubes (12AU7) for channel 1.

[TelRay]

thank you Delta362 for the suggestions, they have revealed something odd (that I am still far from understanding).

To your questions, in order:

- CHANNEL CONTROLS: both channels had the TONE fully open and CONTOUR SWITCH at NORMAL. The GAIN for both channels was set so that they would show the same Voltage at the INPUT PIN of the 12DW7 (highlighted in yellow) and the position of both pots turned to be about the same (a bit below half way)
- SIGNAL 2nd half of the first 7025: surprisingly NO! They do not show the same values. See schematic below (highlighted in red)

Input signal for both 80 Hz and 800 Hz about 140 mVpp




What I did a couple of days ago was to physically swap CH1 and 2

That is:



Then CH2 (now going through VIBRATO and REVERB, even if they are OFF) was loosing HIGHs and LOWs and CH1 going to PIN 7 12DW7 (through the 0.022 uF cap) was sounding great. That's what I thought the issue should be in the VIBRATO / REVERB area (maybe i did it wrong???). But what you asked to measure shows otherwise.

[Delta362]

My suggestion was trying to see if there was a different at the yellow point since the input stages for both channels seem to be identical. I was initially looking to see if something was getting lost since I would expect the signals to be practically identical (minus tolerances of R, C and tube gain). Then we could start chasing left or right in the circuit from there.

Your test of swapping channel 1 and 2 there seems to indicate that both front ends are working the same. That leaves the hi and low cuts stemming from the next few stages of the vibrato/reverb circuit.

I don't recognize this symbol in the schematic.

[x-pro]

I don't recognize this symbol in the schematic.

VARISTOR is drawn and labeled in the diagram below the table.

[Delta362]

VARISTOR is drawn and labeled in the diagram below the table.

That was a complete miss on my end. Thank you!

[Helmholtz]

I think the band-limiting is a consequence of the signal passing through the tremolo section.
This contains several filters and uses (modulated) phase-antiphase cancellation at different frequencies.
So it's not simple amplitude modulation.

BTW, I can't open the full schematic from post #1 anymore.