Help diagnosing Marshall Class 5 failure

Discussion in 'The Squawk Box' started by bwest, Oct 6, 2017.

  1. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Hello there! I was inspired by Lyle Caldwell on this forum and, having read about his mods I made some basic mods to the Class 5 myself (which worked). These sounded fine, but I was inspired to go all the way...and essentially turn the preamp section into that of an honest-to-goodness classic Plexi (1959 super lead was the inspiration, but they are all about the same). This required rewiring of the amp, which (having checked and double-checked), I believe I have done properly.

    Now for the odd part which I would love some help diagnosing. When the gain is low the sound is clean and undistorted, and roughly the volume I'd expect out of the amp - all good. Once I turn the gain up though, when I play power chords the volume cuts out, never to return. I've changed all three of the tubes independently and, unless >1 tube is blown, swapping tubes doesn't fix the issue. I've read Blencowe's "Designing Tube Preamps for Guitar and Bass" and Kuehnel's "Guitar Amplifier Power Amps" and nothing I've encountered in either book (or at least recall encountering) explains this behavior. If it was blocking distortion the sound should come back. There is still some slight hiss out of the amp after the sound cuts out, and if I let it sit for a while and discharge the caps I can turn it on again, after which it works, at least until played loud and then it quits. The fuses all look fine.

    Since a sound clip is worth 1000 words, I've posted a link to soundcloud below with a recording of the failure (apologies for the playing - also I say I will try swapping the tubes in the audio - I tried it after to no effect).
    https://soundcloud.com/user-619860691/marshall-class-5-amp-failure-diagnostic

    Have any of you seen/heard anything like this? What could be failing? Note that I didn't change the power supply and I didn't change the power amp section. Thank you in advance for any help you can offer!

    In return for the help I receive here, assuming this all works, I'll post my LTSpice model of the circuit and instructions on how to make the changes.

    Sidebar: This is my first post so if I'm messing something up here please have mercy

    Thanks again,

    Brent West
     
  2. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Quick follow-on. Attached is the circuit as it stands today (with rewire, assuming that I did the wiring correctly). Note that there are a few floating elements - these are potential future mods that are in the schematic if I choose to make those mods in the future, but they are not currently "active" (e.g. the 5H power supply inductor instead of the stock resistor). Note also that these are not physically in the head as it is wired today - they are just in this printout from LTSpice
     

    Attached Files:

  3. Cjsinla

    Cjsinla Premium Member

    Messages:
    2,625
    Likes Received:
    1,477
    Joined:
    Jul 7, 2014
    You're using an attenuator on a 5 watt amp? At any rate, try it without the Mini Mass and see what happens.
     
    Last edited: Oct 6, 2017
    michaelinokc likes this.
  4. Standard 64

    Standard 64 Senior Member

    Messages:
    2,523
    Likes Received:
    387
    Joined:
    Jul 2, 2008
    seems like one of the worms escaped from the can of worms you opened
     
    Cjsinla and thatF-INGflyingVguy like this.
  5. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    @Cjsinla I took the mini-mass out of the loop and can confirm the amp still fails in the same manner. BTW I really like the minimass - I use it as a master volume
     
  6. Cjsinla

    Cjsinla Premium Member

    Messages:
    2,625
    Likes Received:
    1,477
    Joined:
    Jul 7, 2014
    Right, more crunch at lower volume. I would check all your resistors and capacitors to make sure they are within spec. It sounds like something is drifting when you drive it. Also re-check your wiring layout and solder joints.
     
  7. D'tar

    D'tar Senior Member

    Messages:
    302
    Likes Received:
    94
    Joined:
    Sep 23, 2014
    Always suspect tubes first! Mods as well. Check all solder joints. tube sockets. When it is in fail mode, do a voltage chart. Compare it to when it is working fine (you should have a chart already). In fail mode apply a steady signal to the input and trace it through the signal path until it disappears. You can use your phone as a signal generator.
     
    bwest likes this.
  8. Slater529

    Slater529 Premium Member

    Messages:
    1,151
    Likes Received:
    2,154
    Joined:
    Jul 18, 2011
    Sounds tube related to me. Have you tried swapping tubes? If you don't have spare tubes, you should get some. Until then, you could at least try swapping the positions of the V1 and V2.

    Any progress to report?
     
  9. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    @Slater529 I've tried swapping each tube individually - that leads me to believe (unless multiple tubes are failing) it isn't the tubes. I will likely try swapping all tubes soon
     
  10. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    @D'tar that is a really good idea. Because I have a spice model of the circuit I know what each voltage should be at idle. I can idle it pre and post failure and see what is off. Also to your point I can feed it a steady state signal and track the failure. I should have some time to do this over the weekend and will report back
     
  11. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Ok quick update: Swapped both tubes at once - not that. Went through and tested all DC voltages while in "fail" mode - all read fine. Tested AC voltages in fail mode - seem to have isolated to R7 (originally 220K but I changed to 470K as per original plexi). I'm guessing a bad solder joint is at play (though what is weird is I would think it would conduct better at higher voltages...but TBD). I'll probably desolder and resolder, or just hit both sides with the iron and see if it fixes it. FWIW on visual inspection both look good and it measures as complete...TBD
     
  12. D'tar

    D'tar Senior Member

    Messages:
    302
    Likes Received:
    94
    Joined:
    Sep 23, 2014
    What I failed fo mention earlier is probably the easiest, quickest way to isolate a problem area... The "pop" test. When checking for voltage, volume at about half, start at the output tube(s). When you probe the plate you should hear a pop noise through the speaker. Probe the grid and you should hear it again. On to the next plate, then grid etc, until you get to v1. As you get closer to v1 the pops should become louder as they are amplified through the tubes. On the cathode follower the pop will sound on the cathode.
     
  13. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    @D'tar thanks that makes a lot of sense - I sort of accidentally fell into that method which seems to work
     
  14. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Update here: I desoldered R7, threw it out (thought it metered fine - only 30 cents) and soldered in a new one. Didn't work - same issue.

    Now here's the weird thing - when I touch the left (signal) side of R7 with the probe, even really lightly, the sound comes back and then fades out. I'm guessing now this might actually be an oscillation. I'm looking for someone with a scope I can borrow locally, but until I find one, I'm going to change the circuit up to cut frequencies off >15-20Khz at each stage, which should help. The original amp had a 1n cap across the load resistor of the fourth gain stage (C1 here http://www.thetubestore.com/lib/thetubestore/schematics/Marshall/Marshall-Class-5-Schematic.pdf) which rolled off a bunch of high frequencies (including a lot of audio band) - might have been why
     
  15. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Another quick update: I added a 470p cap from the grid to the cathode of the first gain stage to cut >20Khz frequencies out and swapped R12 (originally 470K, I changed to 10K) back for a 220K to cut HF at the 4th gain stage - still fails in the same way so didn't work
     
  16. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Tried the trick of touching lightly the signal side of R7 and that does consistently bring back signal. Not sure how there could be a fault in one of the caps but I'm going to replace C9 and C13 just to be sure. I don't have any 22ns on hand so this will take a few days
     
  17. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Ah wait, I think I solved it. The way I've wired the amp the volume pot goes to a cap, which goes to a grid stopper and then into the grid of the second gain stage. From a DC perspective it is floating, so when I turn up the gain, the positive waveform clips into the second stage, effectively making the DC voltage on the input to the second stage more and more negative. This gets so negative it is fully blocked, and since there is no path to raise the DC voltage back, it cannot unblock itself. If I'm right, if I remove C13 and jumper it that should fix it. That would mean there could be some slight DC across the pot, but hopefully not enough to be really scratchy. I will try this and report back
     
    DarrellV likes this.
  18. DarrellV

    DarrellV Murry Chrirstmers to earl! Silver Supporter Premium Member

    Messages:
    9,439
    Likes Received:
    9,946
    Joined:
    Nov 1, 2016
    funny, I was just thinking that.....

    FWIW, and I'm no expert it sounded in your clip like a cap was holding a charge more over time and not being able to drain it back fast enough... Add to that you said turning it off and draining everything fixed it...Now I see from your quote below that you have found what is prolly the answer!

     
  19. DarrellV

    DarrellV Murry Chrirstmers to earl! Silver Supporter Premium Member

    Messages:
    9,439
    Likes Received:
    9,946
    Joined:
    Nov 1, 2016
    Could you try a high value bleeder resistor there? Or even more crazy an inductor (choke)?
     
  20. bwest

    bwest Junior Member

    Messages:
    15
    Likes Received:
    7
    Joined:
    Oct 5, 2017
    Yep! Dropping C13 and jumpering it worked! Thank goodness - I've been trying to figure this out for weeks. Will do more tone tweaking later and finalize the circuit, for now I'm going to get a beer
     

Share This Page