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

So this looks fun: Matchless Spitfire

Discussion in 'Shock Brother's DIY Amps' started by JuneauMike, Jan 4, 2019.

  1. peteb

    peteb Friend of Leo's

    Apr 25, 2003
    Cascadia
    Does reverb have phase?


    Does the phase of the reverb matter?



    Interesting questions.



    Looking into it. Reverb being delay is phase shift. It gets fuzzy and delayed but apparently there is phase and in phase will be louder than out of phase.


    I don’t know much about reverb but the phase of the signal is an are of interest.
     
  2. bwacke

    bwacke Tele-Meister Gold Supporter

    360
    Oct 28, 2007
    Toledo, OH
    Mike, you might want to read a little about long-tailed pair phase inverter operation vs. differential amplifier operation and paraphase phase inverter operation. If you don't ground the cap on the grid of the second triode of an LTP, you won't get two out-of-phase original signals at the plates of the circuit to drive your push-pull power amp. The grounded cap on the grid of triode 2 allows that triode to amplify the signal from triode 1 as a circuit called a "grounded grid amplifier" where the input from triode 1 is coupled to triode 2 by the cathodes of the tubes. For a grounded grid amp to work, the grid must be bypassed for AC signals so the incoming cathode signal is amplified in phase (signals on the grid will be amplified out of phase just like the signals on triode 1 and will not drive the power amp correctly).

    Vox uses a paraphase inverter on some amps, which feeds the output of triode 1 to the grid of triode 2 so that the outputs of the two triodes are 180 degrees out of phase.
     
    JuneauMike likes this.
  3. JuneauMike

    JuneauMike Friend of Leo's

    May 5, 2015
    Alaska
    Thanks, I've been slowly trying to understand the LTP, concertina PI, etc. What does Matchless do in their Chieftain?
     

    Attached Files:

  4. FenderLover

    FenderLover Friend of Leo's

    Jun 11, 2009
    Minnesota
    As you pointed out in your post #30, I can attest to the Chieftain as it is the Matchless that I have. Also notice the Vox AC50 and the Watkins Dominator. Like you said, all proven designs.
     
    JuneauMike likes this.
  5. JuneauMike

    JuneauMike Friend of Leo's

    May 5, 2015
    Alaska
    Thanks. I agree that it works. But the warning to AC ground the grid keeps coming up and I don't quite follow it. I get that the cathode needs AC ground in gain stage operation, but in PI it seems like the cathode and plate are working overtime to make gain to spare. So its possible to lose some and still have a strong working circuit. If it was critical for the PI, wouldn't Vox and others just add a small cap around the tail R? And in terms of the phase of the signal, I'm totally lost there beyond grasping the need for it.
     
  6. Ten Over

    Ten Over Tele-Meister

    304
    May 13, 2015
    Central California
    One of the things that may be confusing you is that the usual explanation of a LTP, always vague, is not really how it works. Of course being vague means that they can construe their language to mean whatever they want or need. The input from stage 1 is coupled to stage 2 via the tied cathodes? How so?
     
  7. JuneauMike

    JuneauMike Friend of Leo's

    May 5, 2015
    Alaska
    The bias resistor sets the current for both plates, having a huge effect on gain. The tail is the local negative feedback and balances the outputs for both plates. That's how I see them tied.

    My reading on LTPs tracks well with what you are describing, I have yet to have that "light bulb" moment that we get when something is explained at our level so well that we automatically understand all the constituent parts that were confusing previously. I'll keep pressing on though.

    I really appreciate your contributions.
     
  8. David Barnett

    David Barnett Doctor of Teleocity

    But Vox also uses the diff amp as a mixer in the AC30, where the TB channel and Normal channel feed into G1, and the Vib/Trem channel feeds G2. In hifi amps G2 is frequently used as the return for negative feedback. In these cases the diff amp still works as a phase inverter/driver, but the second grid is not bypassed to ground.
     
  9. Ten Over

    Ten Over Tele-Meister

    304
    May 13, 2015
    Central California
    This may take a few tries.
    LTP Constant Current Fig 1.png


    LTP Figure 1 PNG.png
     
  10. Lynxtrap

    Lynxtrap Tele-Holic

    546
    Feb 29, 2016
    EU
    Every signal has phase. But think about the simple wave form from a guitar, then imagine the complexity, time shifts and layering of reverb. Then think about how to make the dry and wet signals out of phase.
    Can't be done.
     
  11. danlad

    danlad Tele-Meister

    258
    Mar 24, 2015
    Here and there
    Stealing the simple one tube reverb and putting it into the other input of the LTP works well. I built something similar a few years ago. Ultimately I rebuilt it and reused the transformers but it wasn't the fault of the reverb circuit. That part was probably the bit that came out best.

    It really surprised me how well that reverb circuit worked. You'd think from all that driving with 6K6s or doubled up 12AT7s that Fender did that using one triode would be lacking, but it really wasn't at all.

    I don't have the schematic I sketched, but it was single channel and I split the reverb feed off after the first gain stage and fed it into the other PI input. I could turn down the dry signal and play just the pure boing of the reverb which was... weird.
     
  12. JuneauMike

    JuneauMike Friend of Leo's

    May 5, 2015
    Alaska

    Well that was really good guidance from @Snfoilhat and I appreciate it. Left to my own devices I would have plugged it into the main signal line after V1 and between a mixing resistor the way Fender does it. And after taking a second look, I really think he's right about the limitations of the first gain stage in this schematic. Now if I was doing more of a Lightning circuit I'd feel more confident mixing the reverb at the first PI input ala Fender.
    I found another example over at Ampgarage last night of a Lighting using the PI for their Reverb send (or maybe it was a Rocket, hmm). I'm sure it will work, I just wish I understood why it works. Ha. It makes sense, it's just a signal going into the LTP from the back door. You've got a signal, you've got the exchange, you've got a ground reference keeping all the rubber bands taut so to speak, you've got the output. The LTP shouldn't care which door the signal uses.

    I'll keep digging. I keep searching "Reverb + PI" and "LTP+Reverb" but not really getting much. Can anyone think of what I should be looking for. Does this type of circuit have its own name?
     
    Last edited: Jan 8, 2019
  13. Snfoilhat

    Snfoilhat Tele-Holic

    Age:
    37
    770
    Apr 8, 2016
    Oakland, CA
    I have been following this thread w/ interest, because it has raised some cool questions, and I simply don't know enough to contribute much more.

    The early Vox AC10, AC15, and AC30 all used the long-tailed pair's two grids to mix two channels. And of course later amps like some of the Matchless amps mentioned did something similar. But the negative points raised earlier in the thread are also important. I re-read the LTP chapter in Blencowe's book after seeing these responses, and he confirms there is decrease in LTP performance, but doesn't give any indication of how much. Ideally one grid is AC grounded.

    So just like in so many other gear questions, it comes down to values and trade-offs, and these are really hard to discuss without some standardization of terms and measures and even then it's going to come down to subjective stuff like tone and feel.

    Even a very straightforward question: which reduces gain more 1) inserting a reverb mixing network into the Spitfire (Princeton Reverb style) or 2) reducing gain at the LTP by adding resistance (the reverb circuit) to the second control grid which results in PI gain loss through common-mode rejection?

    I suppose that could be estimated on the back of a napkin. But what about the dynamic, playing effects, the different tones and player input and expectations?

    From a design point of view, how do you rank all the different factors? Number of expensive parts (like number of triodes), complication of layout and time to build, available space, possibility of oscillations, rate at which units are returned to retailers, how much the circuit performance is influenced by good or less good tubes, price point and other marketing stuff?

    The web is full to the brim of examples of circuits we know will work. But deciding what's better or worse is really complicated.

    Just to throw one more idea out there, the last amp I built I mixed the dry and the wet channels in a circuit block that's called a common-anode mixer, and it worked great (IMO), but it costs an extra 12AX7 (both halves).
     
  14. JuneauMike

    JuneauMike Friend of Leo's

    May 5, 2015
    Alaska
    If this ever gets built it would be a great candidate to answer that question since it's so easy to gin up a board that allows me to move a couple of wires and add a resistor in the circuit.

    To answer your overriding question, I think approaching it with an engineering ethos is the correct answer, apply the minimum force to the problem necessary to get an acceptable outcome. Strive for fewer components, maximize commonality in values and components, use the least amount of resources to create a serviceable product. That's the beauty of Leo's work. And that is, in my uninformed opinion, where Vox sometimes stumbles. I wonder what an AC30 would look like if Fender designed it. Assuming it sounds and performs exactly the same. I bet it'd be a whole lot lighter. And more simple.

    Schematic, or it didn't happen. ;)
     
    Last edited: Jan 8, 2019
    symbiotic likes this.
  15. Snfoilhat

    Snfoilhat Tele-Holic

    Age:
    37
    770
    Apr 8, 2016
    Oakland, CA
    Single-ended-tremolo-amp_II_cathode-bias_reverb_07.png
    Is this better than mixing resistors and then a recovery gain stage? I have no idea!
    IMG_2501.JPG
     
    JuneauMike likes this.
  16. Ten Over

    Ten Over Tele-Meister

    304
    May 13, 2015
    Central California
    The PI gain loss is not because of common-mode rejection because there is no common mode. The gain loss is due to the output impedance of the source that is applied to the stage 2 grid. The Matchless one has a 100K Reverb pot, so its output impedance is something less than 100K. 100K doesn't disrupt the PI very much, but the 1M that some amps use lowers the gain significantly and creates a greater imbalance between the PI outputs.

    LTP Figure 6.png
     
    JuneauMike likes this.
  17. JuneauMike

    JuneauMike Friend of Leo's

    May 5, 2015
    Alaska
    @Ten Over how would you calculate that impedance in example 6B and 6C? What is the formula?

    The question I'm getting at, as you probably understand, is if 100k is good, then would 50k be better. Is 250k too much? Etc.

    Also, does lowering the grid leaks do anything to the output impedance. The general rule is to use something in the range of 100k-1M but most every example I see uses 1M. I think it has some influence on the amount of input impedance, right? Bootstrapping a higher or lower amount of input impedance would affect the amount of output impedance, right?

    And actually, if you find the reverb to be weak couldn't increasing the LTP tail resistor help strengthen that signal. The tail balances the output signal. But then again, a slightly unbalanced signal adds second order harmonic content. .... So confused...
     
    Last edited: Jan 8, 2019
  18. Ten Over

    Ten Over Tele-Meister

    304
    May 13, 2015
    Central California
    The 100K Signal 2 Output Impedance in Figure 6B is the impedance that the PI sees from whatever the source is for signal 2. The Matchless Cheiftain has a 100K pot (VR7) for the reverb, so whatever that pot is set to relative to ground is the output impedance. The PI sees that 100K pot as a resistance between Grid 2 and ground when it is responding to a signal at Stage 1. (That 100K pot is actually in parallel with 220K which is connected to the plate of V4B which has an output impedance, but that is a whole 'nother thing).

    Figure 6C just shows what happens if the reverb happened to have a 1M output impedance instead of 100K.

    Another thing to note here is that the impedance of whatever is connected to Grid 1 is going to affect the signal applied to Grid 2.
     
  19. Ten Over

    Ten Over Tele-Meister

    304
    May 13, 2015
    Central California
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.