FX53 Classic Tube tone stack questions

[Mark Hammer]

First off, it sure as shooting does not sound like it amplifies 5700x at max gain.  It's dirty enough, but really seems to fall more in SD-1 territory, although the Tone control does not yield the sort of tonal extremes found in any of the TS-9 derivatives, and tends to be subtler.

The dual op-amps are a 1458 for the input and output buffer stages, and an LM353 for the clipping and tone stages, and NOT a pair of TL072 chips.  Quite possible the change may have been done for later or prior issues, but this is what I see on the board.

"Subtle" is not what I'd call the bass end of the tone control but the treble side certainly doesn't do much. It's far less shrill than, say, a Bluesbreaker.

I have the same opamps; I'm not sure where the TL072s came from. I may have put them in as a placeholder. I've experimented with a few different ones and haven't noticed much change. It would be interesting to put in some deliberately limited ones (i.e., OP07) for some testing, since I socketed my board.

The (2nd pair of) clipping diodes to ground (D3/D4), over on the right, are not there.  It does appear that something might have been soldered to those pads in past, but there is much about the board that has the sloppiness of construction in that era.

Mine has the hard clippers. Seems to be a popular mod to remove them.

The diodes in the feedback loop of the 2nd op-amp (D1/D2 on the left side) actually are in series with a 4k7 resistor, to "soften" the clipping further.

This sounds like a fun thing to try; I'll add it to my experiment list.

The feedback cap in that stage is 82pf, and not 120pf.  And although the 680R/1uf ground leg is there, there is NO 3M9 feedback resistor.  Instead, it is 1M5.  In tandem with 82pf, that yields a high-end rolloff beginning around 1.3kz.  Yes, one of the outside lugs of the 500k Drive pot goes to nowhere.  Well, it goes to a pad that is neatly soldered over, and no further.  You see that empty slot just below the 1M5 resistor.  The wiper of the Drive pot goes to pin 2, as per the drawing.  That yields a max gain of 2200x, which is still pretty dang high, but I gather the 4k7 series resistor and the high-end rolloff, takes much of the aggressive quality away.

I've attached mine for reference. It seems to match the schematic I posted. I didn't look very hard though before work.

The various Youtube demos don't sound like a gain godzilla either, FWIW. It might be worthwhile to put the hard clippers on a switch, or to experiment with LEDs in that position, John Greene Glasspak style, so the hard clipping doesn't kick in until the gain stage is cooking.

Yeah, that's on the list for experimentation! As well as different LED types. I don't have many hard clippers in my collection yet.

My serial number in https://serial-number-decoder.com/dod-pedals/dod-serial.htm says it was made in 1983 but I think I may have mixed it up either either the backplate for the flanger or the delay pedal, since the classic tube started in 1989. Actually no, that one says 1984. Who knows. I wish the model number was on the plate as well. I got these all off eBay and the delay and classic tube were both listed for parts. The delay pedal backside looks hand-drawn, very curvy and sparkly, while the classic tube has the usual hard angles.

The classic tube shows signs of repair, mostly in the form of a new battery connector and LED repair. Nothing looks disturbed on the board.

I wanted to thank you all for the wonderful conversation around this project. I'm an electronics newbie and the pedal building is a side hobby to relax and learn. And to get some cool new sounds on the board of course. (Shoutout to the Aion Vulcan octave fuzz, by far the coolest thing on my board so far.)

The marker scribblings at the bottom of my board say "FX53" as well, but say "35 H" instead of "7 H".  Does that mean anything?  I have no idea.

It IS quite possible that mine was modded by the 2nd-to-last owner (or even before then), but we have no way of knowing.  My puzzlement and urge to check the board on mine was really spurred by a seeming mismatch between the name of the pedal and the characteristics suggested by the components and values.  The pedal name suggests it was aiming for something capable of brute force, but also a little tamer than other pedals in the series.

Finally, I'll just note that stickers with the names of what I assume were either assemblers or QC testers line one side of the inside of the enclosure.  They are: Crystal Barlow (484), Donna Ramos (211), "Silver" (208) and Cesar Robles (576).  Using the dating-by-serial-number app you linked to, it says mine was made in or around January 1983.  But the site also indicates the FX53 was made between 1989 and 1997, that DOD recycled serial numbers and that the FX series were made in the USA and China.  So I don't know what to believe.

[ElectricDruid]

I think that looks very nice and much neater than DODs, as well as not taking up half an acre! 

[tootsMcgee]

Finally, I'll just note that stickers with the names of what I assume were either assemblers or QC testers line one side of the inside of the enclosure.  They are: Crystal Barlow (484), Donna Ramos (211), "Silver" (208) and Cesar Robles (576).

I see Crystal Barlow and Silver in mine!

I hear if you get a set of all 10 you get a free Happy Meal (tm).

As for the huge gain, I've noticed that on the stock unit in particular, the gain increases massively at the very end of the drive knob, rather like a fuzz pedal. My assumption is that because the opamp gain is given by 1 + 3.9M / (P1+680), that as the resistance of the gain pot approaches 0, the rate of change of the opamp gain increases (geometrically?)

e.g., going from 500K resistance to 450K changes the gain from around 7.8 to around 8.6. but going from 50K to 0 on the gain pot changes the opamp gain factor from 78 to 5736ish. So the last bit of knob travel is doing quite a lot. The last 30 degrees or so of travel are labeled as a "Boost".

The manual from https://www.synthxl.com/dod-fx-53/ says "The FX53 Classic Tube duplicates the warm, smooth distortion and long, singing sustain and harmonics of the early 60s amp stacks made famous by superstars like Jimi Hendrix and supergroups like Cream and Led Zeppelin." I guess that explains what sounds they had in mind.

I think that looks very nice and much neater than DODs, as well as not taking up half an acre! 

Thanks! The whole reason I wanted to do this was to make it smaller (and to practice my Kicad skills)

[Mark Hammer]

I hear if you get a set of all 10 you get a free Happy Meal (tm).

   

[Mark Hammer]

You know, reassembling the pedal now, I think it may well have been me that removed the second pair of diodes.  When I saw and recalled that I had installed an additional mini phone jack to enable remote switching, it dawned on me that I may well have removed those diodes.  The timestamp I have on the different FX53 schematics in my DOD directory is well before I acquired the pedal, so I may well have thought "I don't really need thoseones there, because I have these other ones over here".

Would I have changed the 3M9 to 1M5 and the 120pf to 82pf?  Would I have stuck the 4k7 in series with the diodes?  Absolutely no recollection of doing so,but the diode-pair removal seems in keeping with my approach back then.

So, let me put a stop to any possible rumours right now, and just say that the use of a MC1458 and LM353 chip-set, instead of a pair of TL072 chips IS a change on the production, but I cannot vouch for anything else I might have noticed is different in mine.  Should it be the case that somebody else finds the same changes in theirs that I have on mine, well THEN we can start up the rumour mill.  But for now, I'll just make a 90% retraction.  Apologies for unintentionally misleading.

[duck_arse]

The dual op-amps are a 1458 for the input and output buffer stages, and an LM353 for the clipping and tone stages, and NOT a pair of TL072 chips.

LF353. as good as exactly the same as a TL072.

[Mark Hammer]

Yes, although the 1458 not so much.

[antonis]

LF353. as good as exactly the same as a TL072.

Ancient days industrial standard..

1458 actually is a dual 741..

[tootsMcgee]

Is there a particular reason they chose that combo of opamps? I put TL072 out of convenience in mine because, well, it's basically what you see in the dictionary when you search for dual op-amp. Then I switcherooed the clipping stage one and flirted with others to see how they saturated without any diodes at all. The rc4558 at least sounded nice. This might be something to slap on a breadboard instead of working on a "completed" pedal.

I need to get an IC puller...

But for now, I'll just make a 90% retraction.  Apologies for unintentionally misleading.

No worries at all, it was enlightening to see the inside of someone else's unit. I have some mod ideas now!

[Mark Hammer]

1458 chips can be useful in high-gain situations that are intended to produce clipping in the chip due to headroom limitations.  In this instance, however, it is being used for the input and output buffers, in unity-gain mode. FWIW, a TL072 would be okay in that slot,but maybe a 5532 might be even quieter.

[PRR]

LF353. as good as exactly the same as a TL072.

In the day, the LF353 was different from a TL072. Different circuitry and process. Overlapping specs. '072 quickly got a penny cheaper than '353, and everybody jumped. I understand now everything is an '072 with varying code-stamps.

Some "dual 741" were literally that but I *thought* the early 1458 were improved 741-like guts. But then, late 741s consistently ran to the high side of original '741 specs.

Get some very high quality "gold" nail-polish, that always sounds better.

[Ben N]

Finally, I'll just note that stickers with the names of what I assume were either assemblers or QC testers line one side of the inside of the enclosure.  They are: Crystal Barlow (484), Donna Ramos (211), "Silver" (208) and Cesar Robles (576).

So the Tadeo Gomez, Lydia Sanchez, Lupe Lopez & Maybelle Ortega of DOD?

[tootsMcgee]

I've been experimenting with different component values alongside an LTSpice sim. The two big positive changes so far have been a reduced cap in the feedback loop (10p) and a reduced cap in the tone stack (1n instead of 3n9). The pedal still has more high end rolloff than I want, but it is getting me in the right area.

It's hard to do this without messing up the charm of the original pedal. When I hear that kind of organic, ragged, almost fuzz-ish distortion, it makes me think of a gain or clipping stage getting slammed with lots of low end. Removing too much of that turns it into something else. I just want a little more note definition in the upper registers so that my guitar doesn't sound like a beeping synth.

One nice thing about this circuit is that the sustain is prone to subtle but pleasant feedback. I don't know if it's inherent to the circuit design, or a byproduct of playing with higher gain designs than usual.

I tried a few different clipping diodes in the hard clip section--1n4148 and BAT46 are fuzzy, 3mm LEDs are a little clearer but not much. I think a switch to toggle soft, hard, and soft+hard would be a more fruitful thing to explore next than swapping more parts right now.

Sound demos eventually...I did put the clone with the schematic values next to the original pedal and in an informal A/B test I doubt you could hear the difference in a mix, so I know the circuit is as designed at least!

[tootsMcgee]

I just found out that PedalPCB actually has a PCB already for this series 😅 https://www.pedalpcb.com/product/fx5x/

That would've saved me some time...but designing my own PCB from the schem was great for education.

I added a soft/hard+soft/hard clipping selector switch, which is a fun addition. I'm not sure which I like best. The tone control is more useful than I thought. On the original, it sounds a bit more treble-y near or slightly right of midpoint. The boost on the high side feels more centered on mid-highs. I really feel that the gobs of bass response is a key part of the original's sound. It's not a tight-sounding pedal and never will be without more significant value changes.

Time to build another one with different values to experiment again!

[tootsMcgee]

NECROPOST

Picking up this project again after a break. I'm pretty happy with it. My diode mod has soft/soft+hard/hard clipping toggleable by an on-on-on DPDT switch. I've found that either soft or soft+hard is the sound I like, hard-only is kinda meh.

The biggest problem I've found with this pedal is that it's just not loud enough. Unity gain is pretty high on the volume knob and turning up the gain after a point just shreds the signal and doesn't boost the volume much. What kind of shenanigans can I do to make it, well, louder?

There's a non-inverting buffer at the very end at unity gain. I'm thinking of modifying that to add a few dB of gain there by cutting the trace and adding a 10k resistor across the feedback loop, as well as another 10k resistor to ground off the inverting input. In theory I think this should go from unity gain to 2, which I think is a simple 3dB boost? From there I can tweak the values.

Any pitfalls I should watch out for? Whatcha think?

[GibsonGM]

Just for grins, to keep us on track, could you post an 'as-built' schematic showing that output buffer? 

[tootsMcgee]

I was being pretty cagey with the schem/layout before I found out that PedalPCB did it better so I don't have to worry anymore Still, please don't yoink this for a commercial project. You have better choices and I don't want anyone to think I'd offer support for this mess! I'm a hobbyist.

The board layout is...brute force to say the least. I had help from someone much more talented than me, so it's not a complete disaster, but still.

Dev board:


Layout:


Schem:


The current experimental changes, as wired, are:
120p in feedback loop is 47p to raise treble cut-off a bit.
Added switches to disconnect diodes

Questions:
Do I need the 2M2 pulldown at all? The original FX53 schem does not have it. I blindly added it and now I don't remember why.

Planned changes for future board revision:
1. Add jumper for the gain knob pins 1 and 2 connection so it can be wired like the other FX5x pedals
2. If the buffer changes work out, scoot some stuff around to add fixed or soldering-adjustable boost there.

Overall though, even the original FX53 I have just isn't a particularly loud pedal. It can get very very distorted towards the last travel on the gain knob, but that doesn't actually raise the level much. It gets lost in a mix easily.

[tootsMcgee]

For fun (I don't know what I'm doing) here's my sim. The buffer *appears* to do what I want with that fixed resistor config, by moving the graph up without modifying it.

The gain knob is fixed at "halfway" (250k/250k voltage divider)

The graphs represent full bass boost, flat, and full mid boost (left, middle, right position)

By far the biggest offender for the treble rolloff is the 120p that is usually in the feedback network. The 1khz point is as follows with the tone knob in the middle (mostly flat) position

120p: -10dB
47p:  -5dB
10p:   0dB

[tootsMcgee]

Weather update:

I put the 120pf back in place in the feedback network. 47pf did make it brighter but also removed some of the oomph that makes these pedals awful/fun/awfully-fun.

It's still a little bit quieter than the original FX53 I have at full blast and I can't figure out why. At this point I have two guesses:

1. Something inherent to the switching mechanism (clean signal bleeding in?)
2. TL072 in place of the LF353 is clipping or otherwise not able to push the signal as high. (Yes, I know the schem I posted says TL072 for both, but the unit in my hand has a LF353).

About switching. Part of the switch mechanism seems to double as a filter? Looking at the original schematic again:


I removed Q2, the upper J201, but as part of that I also removed R? (220k) and C13, which may or may not have been a good idea. The diagram you guys made for me on pg1 notes that that area around J201 is actually part of a filter. But in my head I've always treated it as basically a way to short the signal to ground on the upper side. Should I have left the 220k + 1uF C13 in?

I'm waiting on my 353s so we can find out definitively if that's the issue.

Oh! And I found my intermittent switching problem. Whoever had a go at this pedal before me had rotated the rectangular power jack housing so that the contacts were facing up instead of sideways. This made the board contact it and bent the board a bit (not good!)

Finally, I was experimenting with taking photos of the board that I took out for cleaning and resoldering the offboard wiring. This isn't very useful as a technique because you can't see the cap values very well, but it looks cool so I'm sharing it.

[tootsMcgee]

Found it. Turns out BAT46 has a lower VF than 914/1n4148. So that takes care of that.

Now to build some of the variants...