1. Win a Broadcaster or one of 3 Teles! The annual Supporting Member Giveaway is on. To enter Click Here. To see all the prizes and full details Click Here. To view the thread about the giveaway Click Here.

5881 in a 5f2a

Discussion in 'Shock Brother's DIY Amps' started by jsnwhite619, Jul 11, 2020.

  1. cobaltu

    cobaltu Tele-Meister

    Posts:
    164
    Joined:
    Feb 19, 2010
    Location:
    New York, NY
    Is the power tube cathode bypass cap still 25uF?

    If you wanna add some heft, try bumping that cap to 100uF! I tried this on my Allen Classic 10 and if got rid of some of the boxiness these amps are famous for, gave it an 'Oomph' I always though it lacked. It's subtle, but appreciable. A word of warning though, in my amp I cut bass earlier in the circuit first before I did this. The first cathode bypass is 10uF and the first coupling is 10nF.

    I think it's best to trim bass earlier and let it out at the end. it"s funny but the most bass attenuation in these Tweed circuits is at the power tube cathode bypass cap.
     
    tubegeek and Kevin Wolfe like this.
  2. jsnwhite619

    jsnwhite619 Friend of Leo's

    Age:
    36
    Posts:
    3,567
    Joined:
    Sep 10, 2013
    Location:
    Georgia
    I have not because I don't have one. Honestly, though, the 12ay7 gets it done if you have room to turn it up a bit. Plan to order the speaker today, so I'll report back once it's here. But, I've tried the C-Rex in enough amps that I'm 99% sure I'm gonna like it with this one.

    I have 50uf on the bias resistor, and 25uf on both sides of V1. This is getting a full-sized 5e3 cabinet and a Cannabis Rex, so I may cut back in the preamp once it's all together. I generally love the 1058 & Ragin Cajun with everything, but this just really seems to need a 12" speaker. The Cajun is in a full-sized cab, but it just doesn't work in this situation.

    I tried it back & forth with a 5y3, 5u4, and GZ34 rectifier, and settled on the 5u4. 5y3 was really gnarly, but the bass didn't hold up and it was really lacking. GZ34 was too clean. Funny enough, I played it over the past weekend and forgot the GZ34 was in it. I thought I was losing my mind because I couldn't figure out why I wasn't getting the overdrive I expected...wrong tube.
     
    Kevin Wolfe likes this.
  3. Kevin Wolfe

    Kevin Wolfe Tele-Holic

    Age:
    63
    Posts:
    540
    Joined:
    Jun 13, 2017
    Location:
    Lakeland FL
    That’s funny right there!
     
    jsnwhite619 likes this.
  4. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    Do any of you guys ever optimize the plate load and cathode resistor values for 12AY7 or 12AU7 when you make these swaps? The 12AX7 values aren't really ideal for AY and AU, you might get EVEN BETTER results. Take a look at the data sheets "Typical operation" section. (The 12AY7 has a typo: "heater supply voltage: 150V" is actually PLATE supply voltage.) In particular the 12AY7 looks like it would best have a much lower plate load and a larger cathode R to be in its wheelhouse.)

    The 12AU7 actually seems like it's pretty happy with typical 12AX7 values. So never mind that part. But see for yourself what I'm talking about..
     

    Attached Files:

    jsnwhite619 likes this.
  5. cobaltu

    cobaltu Tele-Meister

    Posts:
    164
    Joined:
    Feb 19, 2010
    Location:
    New York, NY
    DISCLAIMER: I am no expert at drawing tune load lines but I've done it before a few times and haven't killed any tube (yet). I too am interested in 'idealized biases' for 12ay7s.

    If you want the easy fix, leave the 100K plate resistor and change the cathode resistor to 2.7K. With a standard 12ax7 bias, 12ay7s run pretty warm - though it still sounds pretty darn good.

    Digging into this further, I think a more ideal bias for a 12ay7 is with a ~2K cathode and a ~80K plate resistor. This creates a more 'technically' ideal bias for a 12ay7.
     
    Kevin Wolfe and tubegeek like this.
  6. jsnwhite619

    jsnwhite619 Friend of Leo's

    Age:
    36
    Posts:
    3,567
    Joined:
    Sep 10, 2013
    Location:
    Georgia
    Yall really have me interested here because I most always prefer a 12ay7 to a 12ax7 unless the former just does not provide enough volume on smaller amps.
     
    tubegeek likes this.
  7. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    There are a couple suggestions in the data sheet. What do you get typically at the supply node above the plate R in your amps?

    It'll be interactive: the different current through the AY will result in a different drop through the power supply dropping resistor, and the choice of cathode and plate Rs will also affect the current draw. Bit of a juggle - maybe choose a plate R, look at a cathode R target value on a load line, tweak the dropper based on the current and supply V that you see working well. It makes sense that folks don't bother - close enough is close enough - but I wonder if there is anything to be gained? Maybe even better headroom in the 12AY7?
     
  8. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    Definitely a good start to the process.

    OK here's a swipe at it:

    200V supply at the top of the plate R. 40K load line (will intersect the current axis at 5ma/0V.)

    That line looks like it crosses the 2V grid curve at about 1.75mA. 2V/1.75mA gives a 1.14K cathode R.

    So a starting point would be, 40K plate R. I'll be back in a minute after photoshop.
     
    Last edited: Jul 30, 2020
    Kevin Wolfe likes this.
  9. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    OK here's a 2nd iteration.

    12AY7load line.jpg

    47K load line from a 200V supply: 200/47= 4.25 mA. So the load line will intersect at 4.25 mA on the left and 200V on the right. (Red.)

    Picking a bias voltage of 2V, we see the operating point looking like 2V, 1.6mA. 2/1.6=1.25K. (Yellow.) Closest we're going to get to that is a 1.2K resistor in standard values. That ought to get us a voltage at the plate of 125V and we can measure that to see if we've got it working right.

    Now: the 12AX7 that was in there already was probably running about .8mA, not 1.6. So the dropper on its power supply node will need to be adjusted to drop to 200V with 1.6mA drawn through it. It depends on what the voltage at the previous filter node was and also whether we really care about hitting the 200V supply voltage exactly or not. There is NO reason to care about that: I took it off the top of my head. But this gives you some ballpark ideas at the very least.

    What I would do if I wanted to mess around with this:

    If we're taking the stock 100K/1K5 12AX7 values as a starting point, I'd parallel another 100K with the existing 100K plate load for a 50K plate load, and I'd parallel a 6800 ohm with the 1500 ohm cathode R for 1228 cathode R. Then I'd look at the (22K, say) dropper and parallel that with another (?) if I felt I was working at too low a plate voltage compared to the calculated 200V supply/125V at the plate values.

    Then I'd plug in a guitar and let 'er rip.


    3rd Iteration (blue):

    Suppose we start with 250V and the same 47K plate load. If we wanted to keep 2VBias, we'd have about 2.35 mA, which conveniently looks like 820 ohms, a standard Fender Tweed amp value. (Usually shared for a 1640 ohm equivalent, but NOT TODAY!)

    That gives us an amazingly clean operating point. Look at the difference between the voltage swing on the red op point and the blue one as we travel along the load line towards 0V grid curve and 4V grid curve: the red version gets all bunched up near 200V and the blue stays very symmetrical. The prediction here is that this would be a very clean, pedal-friendly (because of the large available 3 or 4 V pk to peak swing) operating point. Gain would be about 25.75 (neglecting the load of the next stage because it won't be much) as follows:

    Output swing from approx 93V to 196 - that's 103 V pk-pk. Input approx 4V pk-pk (0 to -4.) 103/4=25.75.

    12AY7 2 plate curves.jpg
     
    Last edited: Jul 31, 2020
    D'tar likes this.
  10. elpico

    elpico Tele-Afflicted

    Posts:
    1,190
    Joined:
    Sep 14, 2011
    Location:
    Vancouver BC
    First thing you have to decide is what "ideal" means to you. Most power out? Most voltage gain? More headroom/harder to overdrive? Easier to overdrive? Lowest THD cleans? Lowest noise?

    They're conflicting ideals found at different operating points so whether a change makes things "better" or worse really depends on what you're trying to optimize.

    Fenders 12ax7 values aren't ideal to achieve any of those things with a 12ax7 even, but they achieve another goal - reasonable compromise
     
  11. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    Agreed. This is mostly just an exercise to see what might happen if you took a second look at the operating point as you swap tubes. We already know what Leo would use for a 12AY7 because of the Tweed Deluxe and Bassman: 100K plate, 1640 equivalent (820 shared) cathode. But that 250V blue line operating point above, man, that's a pretty nice optimization for a FEW parameters, better than I really expected. And I guess what I was coming from was a gut feeling that 100K was NOT the perfect plate load for a tube with rp as low as 12AY7's. (Note that 3rd iteration was an edit, reload the post.)
     
  12. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    EDITED for correction. SEE POST #55


    here's what that looks like for a 240V supply (picked to make the math easy with 80K.) In orange:

    12AY7 80K 2K corrected.jpg

    Using the 80K/2K/240V conditions:

    Operating point looks like 1.3 mA, VBias -2.6V, 137V on the plate. This is an excellent operating point.

    See post #54 below for an explanation of the two tangents I drew.

    Obviously I took a preference for a hotter bias which gives a LOT of headroom but this operating point is also good from that standpoint. This is probably not even close to necessary for the 1st input tube but if you wanted to go nuts with pedals driving it, the larger bias/hotter current/lower plate load R might be less susceptible to clipping the input in that case.
     
    Last edited: Jul 30, 2020
  13. elpico

    elpico Tele-Afflicted

    Posts:
    1,190
    Joined:
    Sep 14, 2011
    Location:
    Vancouver BC
    Yeah that 100k should be made far higher right?

    Not what you had in mind I know :)

    But that's the direction it should be changed if the goal was either "cleanest cleans" or "highest gain". Just trying to point out that stages don't have "ideal" values of their own. It all depends what you're trying to achieve. You're making the load resistance lower which is in the direction of optimizing power output at the trade-off of slightly lower gain and higher THD.
     
  14. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    EDIT for correction to analysis of the 80K load line.


    While a larger plate load almost always does do what you've said, I don't see a whole lot of room for improvement, looking at those plate curves, in terms of gain or THD. And there's another factor that impacts gain and - in my firmly-held **OPINION** tone, which is drive/power/low output impedance.

    If you take a look at the markup (I started doing this before I saw your comment BTW, I'm doing this out of curiosity rather than for the sake of argument, honestly) there is a comparison to be made along these lines.


    the graph below should be used for the 47K operating point's rp calculations. IGNORE THE ORANGE LINE in this first graph. The 2nd graph below it has a corrected diagram.

    12AY7 stage Zs.jpg

    The rp at an operating point can be estimated by drawing a tangent to the grid curve at the operating point and using delta V/delta R using two points on the estimated tangent. This is far from an exact science but it can be revealing. (Please treat the precision in these calculations as highly unlikely to say the least. The overall trend is very clear though.)


    The 47K/1.25K/250V example:

    116V/4.3mA = 27K ohms rp. Add 1.25K for the cathode R, you get 28.3K, 28.3K||47K = 17.7K.

    This will be less susceptible to loading effects than a 12AX7 stage and even a little less than the other operating point for the 12AY7 (see below: its output Z is 25.2K.) In my ***OPINION*** I usually like the sound of lower-impedance drive better - even (especially!) in early stages of an amplifier. This tube will drive the tone stack's complex impedance more effectively than a 12AX7. As you say, @elpico : POWER.

    It may well be that THIS is what people like about the 12AY7's sound, and not the different gain. But who knows?

    The next graph is a repeat from the post above where I promised to explain the tangent lines. It is correct (I hope) for an analysis of @cobaltu 's suggestion of 80K plate/2K cathode. This is an excellent suggestion.

    12AY7 80K 2K corrected.jpg

    Using the 80K/2K/240V conditions:

    Operating point looks like 1.3 mA, VBias -2.6V, 137V on the plate with a 240V supply.

    The two tangents I drew (using the closest two grid lines) have slopes of 90V/2.5mA = 36K and 120V/3.6mA = 33.3K. Average them together you get 34.7K and add 2K for the cathode R to get an estimated rp = 36.7K. 36.7K in parallel with the 80K plate load: output Z of 25.2K.

    BTW: no disrespect or criticism intended to you, @cobaltu - your selection still whomps the piss out of a typical 12AX7 stage, if you ask me.

    Apologies to anybody who doesn't care about this stuff or doesn't find drawing lines on graphs to be fun!
     
    Last edited: Jul 30, 2020
    Kevin Wolfe likes this.
  15. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    APOLOGIES TO @cobaltu - I drew the 80K load line as if it were 160K and drew some conclusions based on that. It's all fixed now. I have gone back and fixed those posts. The overall results are not terribly different EXCEPT that @cobaltu 's suggested operating point looks much more interesting to me than it did at first, because of its nice linearity and good headroom.
     
    Last edited: Jul 30, 2020
    jsnwhite619 likes this.
  16. jsnwhite619

    jsnwhite619 Friend of Leo's

    Age:
    36
    Posts:
    3,567
    Joined:
    Sep 10, 2013
    Location:
    Georgia
    Slow day at the office? :lol:
     
    tubegeek likes this.
  17. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    What office? ;)
     
    jsnwhite619 likes this.
  18. elpico

    elpico Tele-Afflicted

    Posts:
    1,190
    Joined:
    Sep 14, 2011
    Location:
    Vancouver BC
    I'm not saying you're doing anything wrong or that I think you should make the plate load larger. By all means, be curious! Everyone should investigate what different operating points do.

    I might be doing a bad job of expressing it but all I'm trying to say is articles sometimes give the impression that there's a "correct" operating point for a stage and all you have to do is slap a line here and there and boom you've made the amp more "ideal", but that's really not how it is. Its trade offs in every direction.
     
  19. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,882
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    Try and stop me!
    :D

    I did NOT take your comments as criticism, more suggestions from an alternative perspective. There are approximately 139,127 (number of TDPRI members) perspectives about any issue that comes up here,* so that ought to tell you (me) something.

    Agreed. Just trying to illustrate how some of those tradeoffs play out and how there is the possibility of getting a better? different? disastrous? result by exploring alternatives to What Leo Did.

    The cool thing about guitar electronics is that there are NO solved problems. Wrong is right, trash is gnarly, grunge is class. Whatever sounds good is good.

    Sometimes you can reconstruct from an engineering perspective what may have made it sound good. Sometimes you just get more confused. Anyway I like drawing load lines!

    * except "should I buy another guitar?" which seems to have really only one answer on TDPRI.
     
  20. jsnwhite619

    jsnwhite619 Friend of Leo's

    Age:
    36
    Posts:
    3,567
    Joined:
    Sep 10, 2013
    Location:
    Georgia
    Back to the build - though appreciating the discussion - here is the cabinet after the second coat of shella Still clear, shoplight gives the color.
     

    Attached Files:

    Tele Slacker and tubegeek like this.
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.