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.

Continuing saga of plexi clone build with harsh distortion

Discussion in 'Shock Brother's DIY Amps' started by itsGiusto, Oct 20, 2020.

  1. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    See https://www.tdpri.com/threads/plexi-clone-build-cannot-figure-out-why-harsh-distortion.1044918/ for my last thread on this.

    Following the last thread, I basically started from close to scratch, and rebuilt from the ground up to try to make my build much cleaner, and hopefully get rid of the harsh distortion. New PT, new OT, new choke, all new board components (but same board), (some) new filter caps, better grounding scheme, tried to do better wire routing and lead dress all around. Aaaaannnd....

    It sounds mostly the same. Still have those weird overtones...
    Furthermore, the tubes are still getting really really hot, and occasionally the HT fuse blows. This really makes me think there's parasitic oscillation happening somewhere, burning too much power and making the output tubes work too hard.

    Here are some images of the new build:

    [​IMG]

    Input section:
    [​IMG]

    Potentiometers and post-PI section:
    [​IMG]

    Power section:
    [​IMG]

    Filtering section:
    [​IMG]

    Output section:
    [​IMG]

    Preamp section:
    [​IMG]

    Side view:
    [​IMG]

    Other side view:
    [​IMG]

    Other side view:
    [​IMG]

    Final side view:
    [​IMG]

    LMK if there's anything you want to see in more detail, or if you have suggestions on anything that could be better.

    I met up with someone on the forum, and we started running some signals through the amp and seeing how it looks on the oscilloscope. A lot of weird results, which I'll post about in a bit.
     
    Last edited: Oct 20, 2020
  2. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    All oscilloscope experiments were done running a sine wave (800hz) through the high-bright channel.

    First strange oscilloscope thing:

    When probing here, after the mixing of the bright and normal channel, right at the input of v2 (see green arrow):
    [​IMG]

    If the channel volume is turned most or all of the way up, the sine wave looks asymmetrically clipped:
    [​IMG]

    This holds true even if I remove V3 and the two power tubes.
    However, if I remove V2, the wave is no longer clipped:
    [​IMG]
    I find this strange, because I'm not sure how V2 could have a hand in clipping a signal at its input, before V2 even has the signal passed through it.


    Second strange oscilloscope thing:
    When probing here, after V2, and the 250pf cap that leads to the treble control (see green arrow):
    [​IMG]

    The signal looks very weird, and like it's maybe ringing. Here you can see what happens as I turn the channel volume up and down from about 5 to 10:


    I don't know why the clipping from V1 and V2 would lead to a weird looking signal like that.
    Interestingly enough, if I disconnect this from the wire leading to the treble pot, but still probe in the same location, the weird signal goes away, and leaves a more normal looking wave. See this video where I'm going back and forth, connecting and disconnecting that location to the treble pot by means of alligator clips:
     
  3. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    Third strange oscilloscope thing:
    When probing the very output, on the speaker output jack after the OT (see green arrow):
    [​IMG]
    With the channel volume most or all of the way up, playing with the tone controls, mostly treble and middle, I can get weird looking notches in the output signal, and modulate the notch width and depth:


    Bass and presence have some effect on the notch, too.

    I'm unsure how those notches would come to be. Could this be the weird overtones I'm hearing? If the presence is really high, it almost has an octaving effect, the notch gets so deep that the frequency of the main wave looks like it's doubled.

    I guess, the big thing that I'm not sure about is, I built this amp to distort. The problem with using sine waves on oscilloscopes is that I am not expecting the sine wave to make it cleanly through the amp. I want it to distort. But it's difficult for me to know what is the "good" distortion, the distortion of how a plexi should sound, vs the "bad" distortion, other artifacts that are not wanted, and how a plexi should not sound. So if anyone here has experience with similar amps, especially bassmans and plexis, and running sine waves through them and can tell me if these look normal or not, it'd be very very valuable.
     
    Last edited: Oct 20, 2020
  4. elpico

    elpico Tele-Afflicted

    Posts:
    1,148
    Joined:
    Sep 14, 2011
    Location:
    Vancouver BC
    Those are both normal, expected things to see unfortunately.

    The first one is simply you clipping the grid of the second stage. When the input signal pulls the grid up near the same voltage as the cathode grid current starts to flow. That makes it impossible for the previous stage to raise the grid voltage any more and it simply flat lines at that point.

    On the second video you're looking at the output of the treble cap so the EQ is different than the input signal. The higher frequencies have been cranked up relative to the fundamental. Those higher frequencies were generated at the corners of your clipped sine wave.
     
  5. elpico

    elpico Tele-Afflicted

    Posts:
    1,148
    Joined:
    Sep 14, 2011
    Location:
    Vancouver BC
    Oops I was writing when you posted the last one. Same story though. I don't think we can put all the info you need to make sense of these into a forum post, but I guess the short story is the only time the output signal would look like the input signal is if the amp had absolutely no distortion and no tone stack.

    You can't change the sound of a signal without changing the shape of the signal. And guitar amps are very much designed to change the sound. Every tweak of the tone controls will change the shape of the output wave. Turning up the volume until a stage clips will likewise change the shape of the output wave. This is normal and expected. You're adding new sine waves with higher frequencies than the original when you clip the signal. These are the little squiggles you're seeing, higher frequency waves mixed into your original one.

    There's gotta be a good youtube video demonstrating this.
     
  6. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    Cool, thanks

    Awesome, makes sense. So this is supposed to be there? It happens in any normal plexi?

    Cool, that makes sense about the high frequencies causing the ringing-shape. But I wanted to clarify what I was doing in that second video. I was probing connected to the output lead of the treble cap, and I was connecting and disconnecting the wire that leads from that treble cap to the treble pot. Why would the signal look so much weirder when it's connected to the treble pot than when it's disconnected?

    Cool, that totally makes sense. I do understand that changing the sound and the shape are related. But I just don't really know if this is expected, the weird notching, for example. Is it changing the shape in the CORRECT way, as opposed to doing something it's not supposed to do.
     
  7. elpico

    elpico Tele-Afflicted

    Posts:
    1,148
    Joined:
    Sep 14, 2011
    Location:
    Vancouver BC
    Because the cap can't change the EQ of the signal by itself. The cap is only one half of a high pass filter, it needs the other half (the pot) to complete the circuit and make it work. Connecting/disconnecting the pot essentially turns the EQ effect on and off, so the wave shape changes.

    Now that is much harder to say, I don't know if anyone can tell you that. Honestly I don't know how practical it really is to try to locate certain elements of a complex sound by looking at the time domain. Do you have a recording of the weird sound?
     
    itsGiusto likes this.
  8. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    Thanks. I don't think I have a recording from my rebuild, but this video showcases the problem. Granted, a lot has changed, like for example, I removed the variable NFB and took it completely down to stock:


    But those overtones are still present. And then, of course, there's also the question of why the HT fuse blows occasionally.
     
    Last edited: Oct 20, 2020
  9. Tom Kamphuys

    Tom Kamphuys Tele-Holic

    Age:
    41
    Posts:
    613
    Joined:
    Sep 18, 2018
    Location:
    The Netherlands
    I've been probing my jcm800 quite extensively and more people did that around that time. As elpico said, it all looks ok. Except the notches, those I find strange. Some other guy on the forum had something that looked like that and it was solved by using a shielded cable for the output of the volume pot.
     
    itsGiusto likes this.
  10. Tom Kamphuys

    Tom Kamphuys Tele-Holic

    Age:
    41
    Posts:
    613
    Joined:
    Sep 18, 2018
    Location:
    The Netherlands
  11. Lowerleftcoast

    Lowerleftcoast Friend of Leo's

    Posts:
    2,258
    Joined:
    Dec 30, 2019
    Location:
    california
    You have described the symptoms:
    HT fuse blows
    Odd distorting overtone
    Tubes too hot

    I will assume you have the bias squared away.
    There may be a possibility of a bad tube or filter cap but, I agree, this is consistent with oscillation.

    Perhaps you can catch oscillation on your scope. Check out:
    *How to fix tube amp parasitic distortion Grid stop resistors D-lab electronics*
    on youtube.

    There are measures that can help.
    Anode to cathode snubber cap on V1 like the Marshall 800. 100pF to 220pF.
    Increase the PI snubber to more than 47pF.
    Add/or increase grid stoppers.

    So that is the easy stuff.
    ____________________________________________

    I know you don't want to hear the hard stuff.
    I get it... Marshall did not do it and others get by with it so why can't my build? Well at the factory the amp would be sent to the *did not pass inspection* pile. It would be scrapped or fixed. The ones that got away with it would be sold. Point is, not all the amps built to the poor design worked properly.

    Ceriatone shows a very specific ground scheme they claim has been screened to avoid grounding issues. Their scheme does not follow well know best practices for grounding. Does it work well? IDK.

    So have a read through Aiken amplification grounding. It mentions oscillation issues due to ground problems. Have a look at the Valve Wizard grounding as well.

    I think you truly want to understand this stuff. Most of us would like to know but don't understand why sometimes the electrons don't always behave the way we expect them to. We know more than we did back when these amps were built though. So, we know a single point ground scheme leads to the least trouble. Some of us here at TDPRI still hold onto the two point ground scheme. I use to. It is usually quiet.

    The best offense is to use *best practices* throughout the build. I noticed on your new re-build you still have a less than optimal heater wiring layout and the ground scheme has ground loops. The amp is not quiet. It is still quite noisy which can lead to the very oscillation that could be the root of all of this amps troubles.
     
    Last edited: Oct 20, 2020
    akuster777 and Nickfl like this.
  12. _Steve

    _Steve TDPRI Member

    Age:
    40
    Posts:
    80
    Joined:
    Jul 14, 2020
    Location:
    Los Angeles
    Just a note to say that its normal for the bright channel to be super piercing (especially at low settings) and really really noisy. Lots of people when they first try one think that its broken but that's the way they are.

    Im not as experienced as others on this forum but my suspicion is that it's not a lead dress issue as you've already re-housed everything. But to help diagnose if it is you can prod the connections and move the hookup wires using a chopstick or something else non-conductive while picking the low E string really hard to see if that 'overtone' goes away.

    I would suggest focusing on why the HT fuse is blowing. It's probably a bit easier to diagnose, and theres a high chance when you solve that you will also fix your overtone issue. My instinct would be to watch how much current is going through some of the key stages of your amp like the power tubes and the dropping resistors with the amp volume on full while turning up and down your input sine wave, or picking your low-E really hard if you're brave. Hopefully you will see some excessive currents and narrow down where the issue is.
     
  13. Tom Kamphuys

    Tom Kamphuys Tele-Holic

    Age:
    41
    Posts:
    613
    Joined:
    Sep 18, 2018
    Location:
    The Netherlands
  14. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    That thread is interesting, and it has me thinking. I have two 68k resistors on the board which are mix-resistors, coming from both the channel 1 and channel 2 volumes, then there's no grid resistor. I wonder if it would be worth me trying to disconnect the normal channel, and just use the bright channel, and put the 68k as a grid resistor directly on the V2a pin, since I know that grid resistors are supposed to go right there to minimize noise. I could also try shielding the wire.

    Edit: I meant to say 470k mix-resistors, not 68k
     
    Last edited: Oct 20, 2020
  15. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    I can try probing without having an input signal to see if there's any oscillation happening that I can pick up on. I'm not seeing anything take over, though, like d-lab shows, when I have a signal running through the amp.

    I can also try grid-stoppers on all my amplification stages to see if that'd help, though I'm not 100% sure of what values to use.

    I've tried varying the PI snubber a lot, (that's the one that goes between the PI plates, right?) and it's never affected the oscillation. It makes the signal sound very dull and dark, but the oscillation is still there.

    If I were to try to get to a single grounding point, I'd want to see how someone else has done it for a similar amp. I know that it's probably not as simple as just desoldering all the ground points and connecting them together - a good grounding scheme is more complicated than that. So I'd like to know what I should do to have the best chance of the ground scheme being good.
     
  16. Middleman

    Middleman Friend of Leo's

    Posts:
    4,243
    Joined:
    Aug 29, 2007
    Location:
    MV, CA
    Here's all I got and it will be controversial as some people think it's hooey.

    Any and all resistors in the EQ sections replace the metal transistors with carbon comps. That will change the harshness for the better. Just my own experience.
     
  17. Lowerleftcoast

    Lowerleftcoast Friend of Leo's

    Posts:
    2,258
    Joined:
    Dec 30, 2019
    Location:
    california
    Can you explain how the amp acts when the tubes seem hot and/or when the fuse is about to blow?
    You can go quite high without altering the voicing much. For testing to get rid of an oscillation, I would go as high as 470k. I might try 100k on the power tubes. If it is determined that an oscillation is averted with grid stoppers, I would then back off to more socially acceptable values. Maybe around 10k - 22k stoppers if they would still keep oscillations at bay.
    Yeah, the 49pF snubber cuts highs way above our hearing range. Once that pF value gets into the hearing range audible highs are lost. Usually 100pF is OK.
    You are right, it takes some thought. Let's see if the other measures will stabilize the amp. BTW I don't remember if new/different power tubes have been tried???
     
  18. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    I haven't really seen any distinguishing characteristics. Usually I'll just be playing, and all of a sudden, my sound starts to fade away, signifying that the fuse blew.

    Do the resistors need specific power-handling capabilities? 1/2 watt?

    Yeah, I've tried new power tubes, and new other tubes as well.
     
  19. Lowerleftcoast

    Lowerleftcoast Friend of Leo's

    Posts:
    2,258
    Joined:
    Dec 30, 2019
    Location:
    california
    1/2 Watt are fine.
     
  20. itsGiusto

    itsGiusto Tele-Holic

    Posts:
    521
    Joined:
    Dec 31, 2017
    Location:
    Massachusetts
    This is interesting. If I put this jack converter into the bright-hi input:
    [​IMG]
    and turn the volume to between 8 and 10, and probe the very output of the amp, I end up getting this signal:


    Could this be the parasitic oscillation I'm looking for? It seems to be a 50,000hz sine wave.

    Note: this only happens if a jack like that is in the input. If the jack is not in the input, then the input is grounded, so it seems to go away. If I use a cable instead of that jack, then the length of the cable ends up picking up lots of other noise and I get tons of buzzing which probably drowns out this 50,000hz signal on the scope, making it too hard to find.
    Update: Actually it is there if the cable is in instead of the jack, it's just harder to see. But I can see it now.

    Update: I've traced this as far back to the beginning of the amp as I could (still same conditions, with the jack adapter in, volume all the way up), and it seems that this oscillation does not appear on the input of V2, coming from the volume pots and mixer resistors, but it does appear on the output of V2, when the signal's on its way into the tone stack. It's also present on the small lead connecting V2a's plate to V2b's grid
     
    Last edited: Oct 20, 2020
    Lowerleftcoast likes this.
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.