70's italian amp restauration/modification

Discussion in 'Amps/Cabs Tech Corner: Amplifier, Cab & Speakers' started by HH1978, Jun 22, 2019.

  1. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    Yes - but it looks like you figured that all out already - nice job!

    Regarding the PI:
    Cathodyne PI circuits actually have a positive feedback loop from the cathode to the grid - depending on the design, you can risk creating a positive feedback loop with a "too strong" tube. That's why you might be getting that effect with the 12AX7. At 12AT7 (or the 12AU7) are actually better in that slot typically as they can handle higher plate current than the 12AX7.
     
  2. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Thanks!

    That's good news. I'll change the PI to a paraphase anyway. I'm now studying various early Fender Deluxe schematics to understand how it works and what the values could be. If I understand the evolution of the line correctly, adjustments were made in the circuit to improve balance and gain, so I suppose the 5d3 would be a good starting point?
     
  3. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    If it were me, with the assumption it's not like I'm trying to reserve any "original vintage circuit" integrity, seeing the compliment/number of tubes and inputs, I would be shooting for more of a blackface era circuit - such as a Deluxe Amp (non-reverb) or a Pro. Here are links to some schematics to review:

    https://cdn.shopify.com/s/files/1/2325/2729/files/Deluxe-AB763-schematic.pdf?5926397210715640943

    https://cdn.shopify.com/s/files/1/2325/2729/files/Pro-6G5-A-Schematic.pdf?5926397210715640943
     
  4. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    So using a long tail inverter?

    I'm relatively new to working on amps, or to be precise, I usually make straightforward maintenance/repairs tasks on my own amps, that I know fairly well. Working on a gain stage design is a bit harder, but I really appreciate the help you offer, as there's nothing more exciting than a project through which you learn.

    I tried to figure the values using both 12AX7 and 12AY7 load lines and the explanations on Merlin's website, and here is what I ended up with :

    The HT supply is about 340 at the C node feeding the PI. So using roughly 30% of that for the tail, that would leave 230V for the triodes. That's the reference I used, drawing the load lines for a 100k plate resistor.

    For 12AX7, choosing a close to center/slightly hot bias (0,8ma), the cathode resistor would be around 625ohms, and the tail resistor would be 68k. For the 12AY7 (1ma), the cathode resistor would be 1,2k and the tail resistor 56k. Does that seem to make sense?

    If so, using this calculator, I would end up with a voltage gain of 26 for the 12ax7 and 16 for the 12ay7. Would that be suited for the amp? I didn't draw the load line for the 12au7, I would expect it to be lower.

    Or am I completely misunderstanding the whole process?
     
  5. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
  6. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    The answer would also depend on the voltage gain from preamp, I guess. But the stages look pretty standard I think.
     
  7. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    Yep - long-tailed pair is the way I would go. In a way, I think you might be OVER-understanding the process! ;)

    When all is said and done with following load lines and calculating resistances, the practical application is always a compromise - but a compromise that works (particularly in guitar tube amps). It technically is impossible to perfectly balance a long-tailed (or most other) phase inverters because at one point in the calculations, you will find you will require "infinite" resistances to perfectly balance output.

    The long-tail and the "asymmetry" of the plate loads (100k/82K0 are compromises to get "as close as good" for a well-functioning (if not technically "perfect") phase inversion. Your C node will provide good appropriate plate voltages for a 12AT7 implementation.

    Applying a phase-inverter circuit - pretty much exactly as you see it in the Deluxe, fed by your C node voltage, using a 12AT7 will get you a great result.
     
  8. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Yes, I know what you mean, and I think you're right, but that's part of the fun, learning as much as I can, so I'm not feeling like I'm just swapping parts without understanding what I'm doing :D

    I'll give the Deluxe version a try, I may always tweak it later to see how it affects the sound.

    I made some other tests with the amp before modification. I said the signal was weak yesterday, and that was actually the case. But I found out the inputs are reversed compared to the usual Fender arrangement, so the input 1 is the LO input and input 2 is the HI. Testing that input, the signal was much stronger, and the clean tone was more than acceptable. But the breakup is kind of ugly and fizzy, typical enough of only preamp tubes saturation. I guess that's one more clue the PI should be redesigned.

    The second channel (which physical positionning is also reversed, so at the left side of the control panel) is really dark. My bandmate only plays guitar, not bass, so I might also tweak that section later to get him two usable channels.
     
  9. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    Cool - great work in checking and testing and trying things out, by the way!

    Tough to say what could be leading to the ugly/fizzy breakup - that could simply be the the cathodyne phase inverter itself beginning to distort. A lot of people dislike the sound of a cathodyne PI going non-linear. (And could possibly be the reason for the somewhat larger grid stoppers on the output tubes - that can sometimes minimize that issue.)

    Or it could somehow be a cross-over distortion situation with the signals going into the output tubes being low. Sort of a faux-too-cold bias effect.

    But yeah, the amp has a PI that is designed for tubes that do not require much to drive them. I'm eager to hear about the results when you beef that PI up.
     
  10. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Hopefully I could get the time to get around it tonight :)
     
  11. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    BTW - you'll have to re-bias the amp when you change the PI. Right now, you very likely have a bias level that will be too hot when you introduce a PI with a stronger signal.

    Think about it this way, you have a smaller swing signal driving the output tube grids. To bias that properly, you're effectively moving the top of the upper swing of each push-pull cycle* close to 100% max plate dissipation. Another way of looking at it, the negative bias moves the cross-over point of the push-pull down below the 0V level far enough to fit the full swing of each signal under the 0V line.

    When you overlay a larger signal at that same bias point, those signals will not longer "fit" below the line.

    *better said - half-cycle - since during each full cycle, both tubes alternately swing 360 degrees
     
    Last edited: Jun 26, 2019
  12. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Thanks, I'll be careful with that. For now, I'll have to figure out the new layout. Once this is done, building it should not be too long.
     
  13. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Sorry, I missed your post, just noticed it now. You are correct, the actual voltages are 250V on the plate and 97V on the cathode, 347V total. By the way, the grid leak on that triode is actually 100k instead of 470k on the schematic, and there's a 2,2k from cathode to the junction with the 33k to ground, which is not drawed. There are a couple of other resistors and caps that are marked with different values than stated on the schematics. I don't know wether they have been replaced in the past, or if the schematic is not accurate. The schematic is from a FBT bass organ, the actual amp is a Manhattan 6024 bass. It was made at the same factory and is virtually the same amp, but further than that, I'm not sure.

    I just finished drawing the layout of the current cathodyne, now I'll work on modifying that layout to implement a long tail PI.
     
  14. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Here's my layout for the long tail PI. Does that seem correct? "In" is from the preamp coupling cap, "Out" 1 and 2 are to each power tube grid and 220k grid leak resistor.

    [​IMG]
     
  15. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    Yep - that's the basics of it. Looks like that wiring would work. I'm assuming that 82K resistor is NOT touching/connected to the out2 point. Keep in mind you can change the "strength" of the NFB signal by connecting to different OT secondary taps. The 8Ω tap will have twice the signal strength as the 4Ω tap, and the 16Ω tap (if there is one) has twice the strength as the 8Ω tap. In general, more negative feedback injected will provide a "cleaner" output, but will be attenuated further overall.

    Not to confuse things, I should have mentioned that you could also look at some of the Marshall LTP PIs (they use 12AX7s, but either tube would work). The "classic" Marshall set up is here. The difference is you could look at those to see how to implement a presence control which varies the NFB loop capacitance and resistance. Not sure if you have the real estate for another pot or not - just something to think about since you seem to enjoy thinking through the possibilities.

    Also check out the "presonance" circuit for kicks. (You might need to be a member of that board to check it out.)
     
  16. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    Thanks again! I'll take a closer look at the options later, and I certainly could add a presence pot, as I have 3 speakers output but only one connected. The others are old 3 pins connectors that I have no use for.

    I wired the PI, but failed somewhere. The sound goes through it, but is either weak (original 12AU7) or goes into ugly cross-over like distortion, and tremolo like oscillation (12at7 and 12ax7).

    Two things I haven't done is change the coupling cap from preamp (the original one is a .1uF, and I see Fender uses a .001.) and put a small cap between the two plates of the inverter (saw this just now on the schematic of someone that converted a amp similar to this one). Would one or both of these changes help with the issue?

    I took voltages on the PI socket : 160V at pin 1, 198V at pin 6, and 66V at the cathodes junction.
     
  17. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    Changing those caps could help - part of the issue might be bass frequency oscillation. But those voltages are a little weak for typical LTP PI circuits - you'll usually see something in the vicinity of 240V. What is the voltage at the C node?

    Looking back on the discussion - and I hate saying this, but you might not have enough available voltage to drive the output tubes in this configuration. Did you re-bias the output tubes? The PI needs to drive enough to work with/against the bias voltage. This might require looking at the signal voltages through the amp to see what is being presented to the inverter, then what gain is possible through the inverter. It may even require reducing the grid leak resistors on the output tubes from the bias.

    (At the end of the day, a cathodyne inverter may be the right way to go after all as it requires less voltage to deliver a sufficient voltage... yikes.)

    @HPB would probably have some great insight to the situation.
     
  18. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    He suggested a paraphase, I could always try that.

    I changed the coupling cap from preamp, and rebias, but that did not solve the issue. I tried with 5881's and 6v6's (rebiasing of course). Except than one is a bit more powerful than the other, that did not make much of a change.

    The voltage at C node is 345V. I'm not sure how to look at the the signal voltages, would that imply using a signal generator?

    I recorded a small audio clip, so you can hear what it sounds like exactly.
     
  19. Tron Pesto

    Tron Pesto Member

    Messages:
    746
    Joined:
    Feb 18, 2016
    Location:
    Mahwah, NJ
    Yeah - sounds like the output is too low and the output tubes are operating into cutoff.
     
  20. HH1978

    HH1978 Member

    Messages:
    85
    Joined:
    Jun 9, 2019
    I'll try the paraphase then. If that doesn't work, I will have to consider replacing the power transformer, if my bandmate agrees to.
     

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice