FX loops and tonal effects

[Koreth]

Okay, so my understanding that the purpose of an FX loop is to offer you an insert point between the preamp and power amp, so that any effects that sound best post distortion can be run after the preamp adds the dirt, instead of getting dirtied up themselves by running them in front of the amp. If that's the case, would you want any tonal shaping in the FX loop itself (rolling off or boosting highs, lows or mids) or would you want the frequency response through the FX recovery and drive stages to be as flat as you can reasonably get it?

I ask this because of the FX loop setup in my Mesa Nomad. While I'm generally happy with the tones I'm getting out of it, a muddy sound is a common complaint with the Nomad, and nerd that I am, I had to ask why. So I grabbed the schematic off the Internets and played with LTSpice for a bit. The proposed fix for the muddy sound is to cut the 120pF caps out of the phase inverter, but SPICE says the only difference that makes is the highs being down -.1dB at 10kHz, instead of -.9dB. Frequency response is flat from lows to mids to highs regardless of the plate bypass capacitors. So unless I'm doing something very wrong here, I don't think the mud is coming from them.

going back one stage in the schematic revealed that the FX loop is always on, whether or not it is in use. All the signal from the preamp goes through the FX driver/buffer stage, and the recovery stage before it hits the phase inverter and power tubes.

Here's the schematic that I put into SPICE and the frequency response graph it returned.




The Green line is the frequency response as it comes off the plate, the blue line is from the junction of R5 and C3, from which it goes to the phase inverter. It seems to me that the lowpass filter formed by C3 and R5 is sucking all frequencies higher than the low mids. That would be a good way to make a muddy sound, yes?

What I'm trying to figure out is why Mesa designed it like that, and what to do about it. I can imagine maybe the preamp is voiced to bright and fizzy and and the FX loop was configured like this to balance it out. Seems like a weird place to do it, wouldn't the preamp or tonestack be a better place to do it? Maybe the amp has too much gain and it's necessary to roll off as much highs to keep the amp stable? But if that were the case, is rolling off everything past 100hz Necessary? Couldn't frequency response be kept flat until 3kHz, 5khz or even 10kHz?

[Chuck H]

Trouble is that circuit may be incomplete. Is this feeding directly into the PI input cap or is there another load at the following stage input? Also, are you sure that input resistor is 475 ohms? That value would hardly matter considering the impedance of the circuit.

Chuck

[Koreth]

Per the schematic, the grid stopper resistor on the FX recovery stage is 475Ω. I know, I thought it was weird too. That'd give the FX recovery stage a HF rolloff point of -3dB at 2.7MHz. I dunno, maybe the amp starts to oscillate when the rolloff point moves up to 2.8MHz with a 470Ω grid resistor. *shrug*

I hadn't thought of how the FX recovery stage might behave with additional loads between it and the phase inverter. I initally drew up just that stage because I was mainly interested in it's behavior, since that would be the next step back from the PI. I totally forgot about the master volume and how the FX recovery would interact with that load.

I modified the schematic in SPICE to include the master volume, a 100KΩ linear pot with a 220kΩ resistor strapped across it, and the .02µF coupling cap between the MV and the PI input grid. Running the simulation, it's not as bad as my above simulation would seem to imply.



Instead of peaking at 80Hz, being down -3dB at 300Hz and down -8dB by 10kHz, it peaks at 150Hz, is down -3dB by 725Hz, and is only down -4.4dB by 10kHz. I played with MV being set a various points between full off and full on. There doesn't seem to be any change in freq response, just attenuation.

That said, the relative boost to low mids still 'causes me to suspect the FX loop as a potential source for the muddiness that is often complained about. By shrinking the value of C3, I can move the rolloff point higher up, widening the low mid peak and flattening the frequency response of the stage. Or is there something about FX loops that I'm missing that makes a low-mid boost a good thing. Assuming I want to bring up the high frequency response up by shrinking C3, how high should I be aiming to shift the -3dB roll of point? Into the mids? upper mids? highs? Or just get frequency response flat up to 10kHz and call it a day?

[Chuck H]

I might try removing the C3/R6 circuit in your schematic. The cap is bleeding highs and how much is dependant on the secondary load value parallel with R6. Just lift one end of that circuit in the amp and see if it sounds better and behaves. I've pulled multiple top end bleeder circuits from Messy Boogers with good results and no instability problems. If you do get instability you can put the circuit back in with the lowest value cap/ highest value resistor that stabilizes the amp and still get an improvement. Be sure to try multiple effects in the loop to be sure it's performing to your satisfaction before buttoning it up

Chuck

[Koreth]

I have to wonder how so many various treble bleeder caps get scattered into Mesa/Boogie circuits. If people are removing them without detriment to the tone, and the consistency of layout on the PCBs makes them unnecessary for stability, why are they even there? Maybe they're left over from the design prototypes? A "better safe than sorry?" approach?" Maybe Mr Smith has an ear for darker amp voicings? *shrug*

[Chuck H]

Mesa's are notorious for having both low pass and high pass filters ad nausium. I'm pretty sure it's a voicing thing. But to my ears all the peaking and then padding just adds a bunch of hissy series resistance and makes the amps sound dull, flat and cold. I'm sure some are there to stabilize high gain circuits too. You find out how far you can take it as you remove them from these amps in handfulls.

But to be fair most players want a fat (dull can be mistaken for fat by many players) mid heavy (flat and cold can be mistaken for mid heavy by many players) tone. Mesa voices their amps for what the guitar shop demographic wants. And if you like your tone, you play better. Many good recordings have been done with Mesa products, which surprises me as much as anything since I've never had what I would consider a studio quality tone from a stock one.

You can always put the circuits back. Start pulling some of the top end bleeders and little top end feedback loops out and see if you like what happens. While your at it you may also find it a good idea to remove a couple of the treble peaking circuits too so the amp doesn't get too thin sounding. In the end you'll get to a more natural and lively tone.

JMHE
Chuck

[Pedro Vecino]

While I'm generally happy with the tones I'm getting out of it, a muddy sound is a common complaint with the Nomad, and nerd that I am, I had to ask why

The problem in Nomad channels 2/3 is not in the treble bleeders (you can check this by opening the presence and treble). The problem is in the following:
- The amount of lows in overdrive stages (lack of a good profile)
- Premature and not linear compression in tubes that Mesa used (12AX7 russian 2, russian 6L6... in 55 and 100 versions)
- Lack of internal adjustment in transitions of gain settings on channels 2 and 3. Channel 2 sounds very flabby in medium/low gain settings and channel 3 sounds very thin in low settings. With full gain the lows sound scattered and pulsation control is practically impossible.
- Poor bias adjust.

There are more details but these are the basic.
Regards

[Koreth]

The problem in Nomad channels 2/3 is not in the treble bleeders (you can check this by opening the presence and treble).
(snip)
There are more details but these are the basic.
Regards

I would love to hear the what how and why of said more details.

The reason I'm looking at the FX loop is because when the idea of modding the Nomad to combat a muddy sound came up on the Boogie Board, the guy who first started modding reported running the Nomad's Preamp send into the FX return of another amp and vice versa, the Nomad's preamp sounded lively going through the other amp's PA, and the other amp sounded muddy going through the Nomad's. That would certainly implicate the tone stack, phase inverter and power tubes as a source of mud, and not the preamp.

The problem is in the following:
- The amount of lows in overdrive stages (lack of a good profile)

As for the amount of lows in the preamp, I agree they could stand to be done better. Move that 15µF cathode bypass cap to the final gain stage, not the 2nd, to start. But changing the voicing of the preamp is going to be a more involved mod.

- Premature and not linear compression in tubes that Mesa used (12AX7 russian 2, russian 6L6... in 55 and 100 versions)

Tubes can be swapped out. Mesa may re-badge only a few types of Russian tubes, but there's also the Chinese and Slovakian tubes as well, and more than one type coming out of the Russian plants. If the factory tubes were the cause, I think someone would have reported how such and such non-mesa tube magically cured all tonal woes in the Nomad by now.

- Lack of internal adjustment in transitions of gain settings on channels 2 and 3. Channel 2 sounds very flabby in medium/low gain settings and channel 3 sounds very thin in low settings. With full gain the lows sound scattered and pulsation control is practically impossible.

I'm not sure what you mean by 'lack of internal adjustment in transitions'. Do you mean the transitioning is poor from higher gain to lower gain and vice versa? Or do you mean the way the circuit is laid out, the way the gain stages in the Nomad's preamp cascade into each other is done wrong? Either way, I suspect the 1st stage being parallel triodes with a 150k plate resistor might be part of it. That's equivalent to having a single triode with a 300k plate resistor. That's a lot of gain for a 1st stage.

- Poor bias adjust.

I'll assume you mean the biasing of the power tubes, as the preamp stages are biased in the same manner as almost every other tube amp (cathode bias), and the fact that bias adjustment I see most often used in the context of power tubes. The Nomad's bias-adjust isn't poor, it is non-existent. The same holds true for every Mesa amp. Mud doesn't seem to be a universal complaint with Mesa amps to me. Only on Rectos where someone has dimed the gain, bass and treble, scooped the mids, and then wondered why the sound is mushy and fizzy with no crunch. :P I won't disagree that the Nomad might benefit from making the power tube bias adjustable, so they could be run in their sweet spot for better sound instead of colder for better life. However, I'm not convinced it is a major contributor to a muddy sound.

[Pedro Vecino]

The details are numerous. It all depends on where one wants to go. But the first step is always the diagnosis, and this is done playing and valuing each channel (contours, harmonic content, texture, transitions in drive settings, agility, speed of response, attack, etc.) and balances between them (there are many sounds in amps which are not able to compose a musical set). Draw the line between balance and versatility is also subjective, of course.
This preamp step that has shown involves an evident attenuation in the treble and a profile in lows through the cathode, but should consider it as a part of the whole amp. 120pf capacitors in the driver are almost irrelevant compared with the previous.
If you feel that textures, transitions and balances are good does not need to attend to the preamp. But my idea is that all these imbalances (bad balance between strings, poor relation gain-sustain, "brumps" in the mutes, lack of agility and articulation, etc..) occur basically in the preamp. If you place a diagnosis like this over a cold bias adjustment (original bias adjustment is usually 23/25 mA through the plates) and a set of power tubes with excessive compression, even worse. These things tend to overlap. They are like layers of an onion when one wants to get to the center.
Tubes do not cure but can be used as ingredients in cooking a dish. Knowing each model and associating to each function within a design becomes a irreplaceable tool within the design and basic architecture stablishes. Same with capacitors.
Not all Nomads have the first two triodes in parallel. In many, you will find the plates and grids connected but one cathode disconected.
Modifications can be good, bad, irrelevant or without musical application. Altering the filter you will find sharper and colder tone. The texture and all that I mentioned earlier will remain the same.
Sorry for my poor english. I'm writing very slow in addition :-/

[Koreth]

I'll admit My relationship with the Nomad has been a strange one. When I bought it, it was a sound I found in channel 2 that sold me on the amp. I hadn't found a sound I liked in channel 3 yet. But oh man, channel 2. When I got the amp home I found out that the tone stack knobs were busted and would either short the circuit out, or short full on or full off. So the magic sound I found was found with a broken EQ circuit. After I got the amp fixed, and the EQ knobs were working as they should, I've had little use for the 2nd channel and been using channel 3 for my dirty sounds. Furthermore, the sounds that I loved as a bedroom rocker and the ones I didn't like swapped places once I started playing with a band. Go figure.

I'm hesitant to play with the voicing of the preamp in my Nomad. One it seems that will require more complex modifications. Also I'm getting sounds I consider useful out it. But you think changing the FX recovery circuit to even out the frequency response between the lows, mids and highs is looking in the wrong place and will make things worse?

[Chuck H]

It's all well and good to try to idealize each stage to interact with the next, and I do this when I design, but for a factory PCB amp it's a pita. I did do it to a Mesa Subway Blues with good results but the most profound improvements were a result of pullung out, bridging or otherwise disabling various superfluous hard wired tone shaping, local NFB and top end bleeder circuits. I'd start there. Do mods one at a time so you can decide if you like the change.

JM2C

Chuck