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.

Exploring Fender's Harmonic Tremolo

Discussion in 'Amp Tech Center' started by moosie, Dec 23, 2017.

  1. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    Yes. I did this in my Brownface-clone, worked like a charm, much more usable range.

    There were questions about how to calculate the frequency of the oscillator. You can't get there with simple RC time constant formulas, it's all about the phase shift which is easiest to figure out with two-dimensional [imaginary] numbers. I spent days working this all out in Excel and came up with a decent calculator.

    http://www.gofastforless.com/junk/PhaseShift.xls

    With the original Fender circuit values it estimated a lowest frequency of 3.1Hz, my actual circuit measured at 3.3Hz. When I switched to 0.02uF caps, the program estimated 2.0Hz, actual measurement was 2.1Hz. You can change any value and it'll tell you the frequency of the oscillator, how much gain is required to drive it, and also how much gain the tube will actually produce at that frequency. It'll light up red if the required gain exceeds the actual gain, which of course means that the oscillator will not be oscillating.

    Converting the Harmonic Tremolo circuit to a conventional bias-modulating tremolo can be done by simply adding a single-pole switch which disconnects the band-pass filters and runs the entire guitar signal though one side of the mixer (did this on mine). I also added a 1M resistor from ground to the leg than connects the pre-amp to the mixer. Without this resistor, that leg has no ground reference when in Harmonic Tremolo mode so it is likely floating around 280 Volts. I was getting a loud "POP" when switching to the bias-modulating tremolo mode since this voltage was being discharged through the grid of the mixer. The resistor pretty much eliminated the "POP" since that leg will be held at zero Volts DC in both modes now.

    The unit I built is just a Brownface pre-amp, tone-stack, and Harmonic Tremolo. I plug it into the signal return on my Randall head to utilize it's power amp and spring reverb.

    http://www.gofastforless.com/junk/brown_vibe_schem3.GIF
     
    Last edited: May 24, 2020
    Mexitele Blues and Bendyha like this.
  2. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    Heads-up to anyone who might have downloaded the PhaseShift calculator. I've made some improvements to it. You can use the same link above to download the lastest version.
     
  3. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    After building the Brownface Harmonic Tremolo I started researching actual "Vibrato" circuits to build. The concept is pretty simple, shift the phase back and forth. The various implementations are quite intriguing but not exactly simple. Way back in the beginning, Hammond used a big LC delay-line tank with a mechanical rotory switch. Magnatone's circuits were fairly simple but used unobtainium varistors. Somewhere in the middles is the Dorf phase-shift vibrato.

    The concept is quite ingenious. Split the signal and run it through two separate delay circuits. The delay of each circuit isn't a perfect phase-shift BUT the two imperfect phase-shifts are ~90 degrees from EACHOTHER. Fading back and forth between these two signals creates a smooth 90 degree phase-shifting signal output.

    I plotted this all out in Excel and realized something... Fender's Harmonic Tremolo DOES THE EXACT SAME THING! Well, okay, not exactly. The intricate RC divider network in the Dorf creates an almost flat frequency response and uniform 90 degree phase shift. The Fender has a very pronounced high-low-frequency seesaw tremolo effect but there is no denying the 90 degree phase shift modulation, even if it does progress in a wave rather than a uniform shift like the Dorf.

    [​IMG]

    Leo certainly gets a lot of flack for all the times he confused tremolo with vibrato. However, In the case of the Harmonic Tremolo circuit, I think it's fair to classify it as a "Vibrato", it clearly does phase-shift modulation. It's a "Tremolo" as well since there is also amplitude modulation. It's a magical blend. The alternate high-low-frequency amplitude swells perfectly compliment the phase-shift Doppler effect which can be rather bland on it's own. It really is as close to a Leslie speaker as you can get with just a handful of triodes.
     
  4. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    When I first put together my Brownface-clone, it had a POP-POP-POP drumbeat in time with the low frequency oscillator when the Intensity was turned up. I swapped out different tubes for the final mixer until I found one that didn't pop. That also seemed to take all the bottom end out of the guitar signal.

    I dove into it today with a scope to try and figure out why it was popping in the first place. My bias voltage was kind of high since I'm using a 6N2P in place of the 12AX7. Clipped a resistor substitution box to the cathode resistor and dialed down the bias. That only made it worse. Also tried clipping in a larger cathode bypass capacitor which made no difference at all.

    Finally, it occurred to me that the plate voltage is WAY over the datasheet maximum. That's how it is in the original schematic and my B+ was a bit higher still. The solution was to add another decoupling stage to feed the plates of the mixer. I used a 56k resistor and 22uF cap. That not only drops the voltage to a safe level but also has a better frequency cutoff to deal with the 2Hz signal from the LFO.

    No more popping, even with the "bad" tubes that didn't work before.
     
    moosie and tubegeek like this.
  5. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,879
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    My current project is a Hammond AO-15 conversion. All octal, 10 sockets. 2 1/2 of those sockets will be devoted to an octal implementation of the harmonic trem. You can bet I'm going to be begging you guys for help on this. I'm already smarter than I was 10 minutes ago, thank you so much.

    My first head start is, I have an Ampeg Rocket with a good 6SL7 oscillator, I'll be copying it (with one embiggened cap for slower trem) for this project.

    Has anyone here ditched the post-oscillator cathode follower yet? As mentioned much earlier in this thread (I think? There's also a related thread I was reading) that cathode follower and 10M pot make no sense so I don't know if I'll lose functionality if I skip it.
     
    Last edited: May 31, 2020
  6. tubegeek

    tubegeek Friend of Leo's

    Age:
    60
    Posts:
    3,879
    Joined:
    Jan 31, 2020
    Location:
    Brooklyn, NY
    Have you tried using the three-coil Hammond saturable reactor circuit? That's not the same one as the L/C you mentioned, right? It seems like it ought to be possible to reverse engineer that part, especially if you had an original to check into. There have been attempts to combine multiple modern varistors to get similar behavior to the Magnatone unobtaniums, not sure if anyone has claimed full success in that effort though.
     
  7. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    Next up was the Intensity control. With the reverse-audio pot it hit max intensity at about five or six on the dial. Also the intensity would drop drastically as the frequency was increased. I was getting 4.8Vp-p to the grid of the mixer at 2Hz and only 2.2Vp-p at 6Hz.

    The intensity circuit uses several capacitors to filter out harmonics. Plotting the response curve of the circuit you see that it peaks at 2Hz then drops off after that. I dug though the potentiometer drawer and found a 2M-linear pot, so I started with that then adjusted the other components in the circuit to better tailor the response curve to the 2-6Hz range of my LFO.

    [​IMG]

    Changes from the original circuit are; replace 4.7M resistor with 1M, replace 10M-RA pot with 2M-L, replace 0.1uF cap on pot wiper with 0.22uF.

    The results were just what I was hoping for. The Intensity control is useful throughout it's range and the intensity doesn't noticeably drop when the frequency is turned up. It's now getting 5.0Vp-p to the grid of the mixer at 2Hz and 4.5Vp-p at 6Hz.
     
    Bendyha, moosie and tubegeek like this.
  8. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    The 10M doesn't make sense with the cathode follower, I think it was just a carry over from the earlier model without a cathode follower. The oscillator is sensitive to load. I imagine Fender must have had issues otherwise they wouldn't have added the cathode follower. With the cathode follower there you have more options for intensity pots. It's easier to find a triode than it is to find a 10M-RA pot.
     
    tubegeek likes this.
  9. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    No I haven't played with that Hammond circuit, and no it's not the LC tank I was referring to. They had one with a long ladder of inductors and capacitor with multiple taps that were switched into the signal chain via a mechanical rotary switch driven by an electric motor.
     
    tubegeek and Digital Larry like this.
  10. mrriggs

    mrriggs TDPRI Member

    Age:
    41
    Posts:
    96
    Joined:
    May 24, 2020
    Location:
    Vancouver, WA
    I'm afraid that the lower plate voltage didn't cure the pitter-patter drum beat. It worked for a while but the noise eventually came back. Swapping tubes fixed it [for now] but I'd like to figure out what's causing it.

    To test the over-voltage plate theory, I put the pitter-patter tube in a curve tracer. Even at 450 Volts (as high as the curve tracer would go) there was no sign of shorting.

    Another thought is that the heater cathode voltage (~90 Volts) is close to the datasheet maximum of 100 Volts. I hooked the heater-cathode to the curve tracer and ran it up to 120 volts, even modulated the bias voltage. No signs of shorting.

    I'm kind of at a loss on what to check next.
     
  11. Bendyha

    Bendyha Friend of Leo's Silver Supporter

    Posts:
    2,524
    Joined:
    Mar 26, 2014
    Location:
    Northern Germany
    Back in post #19 I posted two PDF's by Richard Dorf about Vibrato circuits, and some more things in #80.
    I just came across another interesting article by him relating to the subject of vibrato, but approached in a slightly different manner, so I thought I would post it here for anyone else who may be interested.
    It's from the magazine Radio-Electronics from March 1957
     

    Attached Files:

    Last edited: Aug 24, 2020
  12. FenderLover

    FenderLover Poster Extraordinaire

    Posts:
    5,699
    Joined:
    Jun 11, 2009
    Location:
    Minnesota
    That sounds like something I said.

    See post #76. I've got a 6G2 chassis and want to replace the Trem with this modified Harmonic Vibe circuit for fun. Trying to fit a layout using a high voltage MOSFET for the LFO. I'll just have to add one tube. I figure that should give some stupid low LFO frequency, and I'm still convinced it should drive the high Zin of the split load circuit just fine. The blackface Princeton's LFO modulates the output grids just fine, I don't see much difference in loading.
     
    tubegeek likes this.
  13. trancedental

    trancedental Tele-Meister

    Posts:
    327
    Joined:
    Aug 15, 2010
    Location:
    London
    moosie likes this.
  14. trobbins

    trobbins TDPRI Member

    Posts:
    71
    Joined:
    Jan 21, 2011
    Location:
    Melbourne
    The linked doc looks at the Magnatone technique, and sets the historical scene a bit. The Maggie varistors have been cloned quite well, so there is a practical way to prepare those parts if you take some care. Of interest is that even the Maggie technique has now been cloned in 9V battery friendly solid-state format in a commercial pedal.

    I'm going to try and bench the LFO part and look at the harmonic levels for a range of circuit variations - that's pretty easy to do nowadays with a USB soundcard and REW software - so will add that to the article at some stage. The idea being that the harmonics can become annoying woofer excursions if their level gets too high - for a variety of trem and vib techniques.

    https://www.dalmura.com.au/static/Magnatone vibrato design.pdf

    Ciao, Tim
     
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.