Asher Guitars WD Music Products Amplified Parts Mod Kits DIY Nordstarnd Pickups darrenriley.com
Asher Guitars WD Music Products Amplified Parts Mod Kits DIY Nordstarnd Pickups Warmoth.com
Asher Guitars WD Music Products Amplified Parts Mod Kits DIY Nordstarnd Pickups Warmoth.com

Reverb/Vibrato standalone - schematic review request

Discussion in 'Shock Brother's DIY Amps' started by alathIN, Jan 8, 2019.

  1. clintj

    clintj Friend of Leo's

    Apr 4, 2015
    Idaho
    http://www.valvewizard.co.uk/smoothing.html

    As usual, Merlin's articles are quite good. As he notes, most guitar amp supplies filter down to around 1 Hz or so, and he gives the equations used. The LFO circuit normally bottoms out at 6 Hz, give or take, so standard values should be fine. If it's off the first node and audio is off the third as I think I read earlier on, that should be good.

    Sent from my Pixel 2 using Tapatalk
     
    alathIN and moosie like this.
  2. alathIN

    alathIN Tele-Meister

    Age:
    53
    393
    Oct 25, 2017
    Indiana
    Man, I wish I could just download Merlin straight into my head.
    That is super helpful. It does look like I may need to tweak these values a bit - but before I do this, I may need to make further modifications to the power supply (see below)

    Unfortunately, the isolation of the LFO is not quite as clean as you state.
    The first node drives the entire tremolo circuit, including the "summing" triodes that put the lowpass and highpass signals back together.
    The first node also powers the reverb driver.
    And the "middle" triode in the tremolo - which I take to be the one that turns one LFO signal into two out of phase LFO signals - is physically located within the same bottle as the reverb recovery triode (but uses a different power supply node).

    That doesn't look optimal to my untrained eye.
    Possible changes:
    1. Add a cap and a small resistor to the filter cap doghouse, and just make a new separate node - don't need a big resistor because I don't need a big voltage drop between A0 and A1. Power the whole harmonic trem, and nothing else off of the first node.
    2. I could add an extra tube socket and keep the trem phase inverter separate from the reverb. I would have two unused halves of two dual triodes, but other than frugality I'm not sure that's a real problem. I could use a lower power tube in the reverb recovery and use both halves of it which might not be a terrible idea anyway.

    One other Reverb related question:
    All these Fender-derived reverb circuits have a resistor and cap, in parallel, both of which are in parallel with the entire reverb circuit - ie, a resistor and cap in the dry signal path.
    I've seen wildly different values for this resistor. On RobRob's page (taken from the original Princeton reverb) this resistor is 3.3M. Rob notes this changes the tone of the amp.
    I've also seen values as low as 100k for this resistor.
    On balance I'd like to keep the dry signal tone as original as possible - but surely there must have been some reason Leo put a huge resistor there?
    What problem can ensue by lowering that value too much?
     
    Last edited: Jan 11, 2019 at 5:13 PM
  3. alathIN

    alathIN Tele-Meister

    Age:
    53
    393
    Oct 25, 2017
    Indiana
    So I redesigned the power supply to add an extra node just for the harmonic trem.
    I also replaced the reverb recovery triode with a dual triode, to keep the reverb and harmonic trem more separated.

    updated power supply.JPG

    I think this math is correct.
    For filter caps
    C = i/(2*f*v) f = AC frequency 60Hz, v = ripple voltage i = DC current (average)

    For Merlin's ideal less than 5% ripple voltage
    C = i/(2*60*20)
    C = 3.54 x 10e-6 F
    C = 3.54 uF
    which is really small compared to what I usually see for filter caps.
    I am guessing this is because I am drawing so little current (no power tubes?)

    With a 20uF first filter cap, I get 3.54 ripple volts at the first node. That would be what supplies all 5 triodes of the harmonic vibrato. The ripple volts of course drop dramatically from there.
    Note: the reason 3.54 seems to be the answer to everything is that 20 = target ripple voltage, and 20 = the filter cap I have laying on hand.

    So that all sounded good until I got to the frequency calculations.

    Using a 16uF filter cap as I originally had done, the frequency at the first node is 9.9 Hz which is quite above the 6 Hz target for the harmonic trem.
    A 20uF filter cap doesn't help much, only lowers the frequency to 8 Hz
    If I use a 32uF filter cap there, I get 5 Hz frequency at that node.
    So it looks like I need to increase that first filter cap.
    My idea was to go ahead and use a dual can cap, so the first post-rectifier cap and the first node filter cap are both 32uF.

    If I increase all the subsequent filter caps to 20uF, they're all below 1Hz.

    I included the choke because it was in the power supply I originally lifted from. I'm not sure I really need it.
    But the way I've drawn the power supply, I've got the choke between the first and subsequent power nodes - again, just to even further isolate everything else from the harmonic trem. This is probably overkill, but I've got the choke and I don't see a downside.

    I think this fixes my problems?
     
  4. Snfoilhat

    Snfoilhat Tele-Holic

    Age:
    37
    770
    Apr 8, 2016
    Oakland, CA
    I have some power supply design thoughts that I'd share here, and you can take or leave. They are general rather than specific to your circuit, but you may find them helpful.

    Chokes: If you are going to have a choke at all, the earlier the better. A choke is in inductor, L, and with a capacitor will form an LC filter, just like the combination of resistor and capacitor (RC) in the filter supply that are low-passing unwanted frequencies in the power supply to ground.

    Unlike a resistor, a choke can have a decent sized inductance without a large resistance, so you can get some decent filtering (look at the choke on any post-1955ish Fender -- it sits between B+1 which feeds the power tube plates and B+2 which feeds the power tube screens). Fender design philosophy is to get as high a voltage as possible on the screens for best headroom, but still have some kind of filtering because the screens are sensitive to noise+ripple. That's why from the Deluxe Reverb and up the product line there is a several dollars choke instead of a several cents 1k resistor.

    Caps: today's electrolytic filter caps are small. Even the slightly ridiculous boutique ones that are intentionally oversized (so gossip says) so that vintage amps don't look less cool are smaller than some of the originals, and the prices for 8 or 16 or 20 or 50 microfarads (@500V) are basically flat. I doubt this was the case in 1955. So unless there is a good reason (like you are trying to design in some kind of sag), why not use 50+ microfarad caps throughout (or up to whatever capacitance your rectifier of choice can handle). Then you can get the same filtering with smaller resistors, shed less energy as heat, and have higher voltages at all the power supply nodes? And more nodes total, since there's less risk of whittling away all your voltage before you get to V1.

    For a PT that is giving excessive voltages, I wouldn't use a choke at all and would then use larger resistors throughout.

    I also wouldn't use a tube rectifier for just about anything, but that usually comes down to personal taste.

    Again, I missed the nuances of this thread's first page and might have missed some decision or design element that makes what I've written irrelevant. Looking forward to hearing a new harmonic trem device!
     
    alathIN, moosie and clintj like this.
  5. clintj

    clintj Friend of Leo's

    Apr 4, 2015
    Idaho
    Are you planning on running a tube rectifier as drawn, or solid state? If tube, which one?

    These design questions are exactly what my holdup is in designing just a tube driven harmonic vibrato box, BTW. Just how much isolation and separation is really needed in real life for any odd side effects to not be noticeable? Will any side effects just be buried under the tremolo itself? Lastly, how best to implement grounding to minimize loops with this number of nodes and circuit overlap?

    Your changes may be overkill, or may be just what it needed. Only one way to find out, right?

    Sent from my Pixel 2 using Tapatalk
     
  6. tubeswell

    tubeswell Friend of Leo's

    Jul 1, 2008
    NZ
    You might want to consider an extra gain stage for the reverb recovery (wet) side. The gain structure of the unit in your schematic would leave a pretty strong dry signal, but only a weak/mild reverb recovery signal, which might be overwhelmed by the dry signal. The output transducer of a typical reverb pan is akin to the signal of low-impedance single coil guitar pickup where the strings are plucked softly. If you only have one (reverb recovery) stage to amplify this, it will be competing against a much stronger dry signal from the harmonic vibrato output. YMMV
     
  7. alathIN

    alathIN Tele-Meister

    Age:
    53
    393
    Oct 25, 2017
    Indiana
    Yes, tube rectifier - 5U4G. The data sheet says you can run up to 40uF filter caps without trouble, and fortunately the 32s appear to give me what I need.
    I think I might try the "star plus" grounding scheme described here by Moosie, where everything run off of each power node has it's own star, and all the stars connect to one spot (rather than all separetely to the chassis).
    Hopefully I've gotten rid of some of the circuit overlap by eliminating the double triode straddling two different nodes, one of which has an LFO.

    I'm glad to hear you say this, because to avoid having 1/2 of a dual triode in one circuit and the other 1/2 in a different circuit, I'm giving reverb recovery its own 9 pin socket.
    So the strange move I was making just for noise reduction turns out to have a good purpose all its own.
    I'm planning to use both sides of a 12ax7 there - but if that's not enough I could probably be talked into using a pentode there...?

    Snfoilhat, I'm 50-50 on the choke. According to Merlin's formulas it doesn't look like I need it.
    On the other hand I am not aware of any problems it is likely to cause, and maybe there's no such thing as a power supply that's too clean, or too much separation between the LFO and the other circuits?

    I am one of those 'tube rectifier or death' people. Not because I think it is objectively better, especially for this use, but just because I like tubes. The more the better. Especially when it's a great big strangely shaped tube like a 5U4C. I know this is whacked but I just love the Frankenstein's Lab vibe. Other than all the static noise and unsafe grounding he had going on in there ;-)

    Will keep everyone posted on the progress.

    Which realistically means I'll come back crying for help next time I get stymied.

    This must be the world's greatest concentration of people who have built harmonic trems. Very glad to have all this experience to draw from.
     
  8. alathIN

    alathIN Tele-Meister

    Age:
    53
    393
    Oct 25, 2017
    Indiana
    I have 81% convinced myself to try two triodes in parallel here, rather than series. A) could likely use more gain but may not need as much as 2 in series, B) a "thicker" wet reverb signal could be a good thing, C) lower noise could be a good thing.

    This is all gut and guesswork here.

    Thoughts?
     
  9. alathIN

    alathIN Tele-Meister

    Age:
    53
    393
    Oct 25, 2017
    Indiana
    OK, I think this is the final version or very close.

    Changes:
    Updates to power supply after reading Clintj and Merlin.
    Recalculated all the anode resistors to give each tube the voltage spec'd in the original circuits (or as close as I can get)

    Changed the single triode reverb recovery to two triodes in parallel

    Took out the input and output isolating transformers and using a ground lift instead.

    Chickened out on the EF86 in the high gain preamp
    Re-reading Rob Robinette's Marshall page brought me back home on how all the parts of the cold clipper preamp work together, and it seemed like replacing the two triodes between the cold clipper and the tone stack with a pentode could have unwanted effects on the clipping as well as how the tone controls work. So this goes back closer to a "direct import" circuit. A preamp with cold clipping and a TMB tone stack is something I don't have in my rig now; I'm more in the mood for it to work as advertised than get some kind of strange experimental outcome.
    I have gone back and forth on this about 200 times and may flip back again.

    Updated Review Copy copy.jpg
     
  10. tubeswell

    tubeswell Friend of Leo's

    Jul 1, 2008
    NZ
    Try it and see. It should be easy enough to convert to cascaded stages if a parallel stage isn't enough
     
    alathIN likes this.
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.