Reverse Engineering... questions about amp design

Discussion in 'Amps/Cabs Tech Corner: Amplifier, Cab & Speakers' started by Beam Tetrode, Jun 27, 2006.


  1. Beam Tetrode

    Beam Tetrode Member

    Messages:
    1,425
    Joined:
    Nov 16, 2005
    Location:
    The Moon
    Hi,

    I'm reverse engineering an old Fender 5D8 Twin. I'm hoping to build a similar amp, with some mods to enhance efficiency. Here's the stock circuit:

    http://www.ampwares.com/ffg/schem/twin_5d8_schem.gif
    http://www.ampwares.com/ffg/schem/twin_5d8_layout.gif

    Since both low gain inputs are pretty much useless, I would love to omit them. In doing so, I can eliminate one whole 12AY7 and its corresponding circuitry (plate resistors, etc).

    With such a modification, would I need to change any component values to maintain spec voltages and biases throughout? If so, please explain why.
    How are the proper values calculated?

    Any tips, info, pointers, and/or educational references would be greatly appreciated. Thank You!
     
  2. donnyjaguar

    donnyjaguar Member

    Messages:
    4,234
    Joined:
    Nov 4, 2005
    Location:
    Canada
    I suspect you'd want to change the 10k resistor in the power supply chain to something a little higher in value. Guessing 15k here.

    The value of this resistor is determined by how you've engineered the various sections that are being powered. You add up the currents, subtract the voltage from your feed voltage and then divide this by the current. This will give you the right resistor value. If there are other sections up-stream, you'll have to factor this in as well.

    This is a simple answer, but I'll leave it at that for the time being.

    DJ
     
  3. John Phillips

    John Phillips Member

    Messages:
    13,080
    Joined:
    Mar 17, 2002
    Location:
    Scotland
    You also must change the cathode resistor on the first tube from 470 ohms to 820 or 1K - it's shared by both tubes in the original circuit, and if you halve the current going through it you will also halve the bias voltage developed, which will then run the tubes at the wrong operating point.

    The original circuit is very wasteful in a couple of ways - not only having a 12AY7 which is not really necessary (although in fact, there are no low-gain inputs - they're all high-gain, which is why it's done with a separate tube stage for each jack), but the way the two tube halves are mixed using four 100K resistors... you could easily use one 50K and parallel the plates.

    If you're using a single 12AY7 and still retaining both channels, all you need is one 100K resistor per plate of the 12AY7. Copy the input section of a 5F6 Bassman if you can't picture it :).
     
  4. donnyjaguar

    donnyjaguar Member

    Messages:
    4,234
    Joined:
    Nov 4, 2005
    Location:
    Canada
    At the risk of sounding like a pedant, this may not be a good idea. It would certainly be safe to do it this way but you'd lose gain.

    The more I look at the schematic the more I see much room for improvement. But of course one has to decide between optimisation and "killer tone". :)

    DJ
     
  5. John Phillips

    John Phillips Member

    Messages:
    13,080
    Joined:
    Mar 17, 2002
    Location:
    Scotland
    Ah, but although you lose gain by halving the plate resistors, you then don't by avoiding mixing the two signal via the other two 100K resistors :).

    I'm pretty sure that you actually come out exactly equal with either scheme (6dB loss compared to a single 100K stage either way, I think).
     
  6. Beam Tetrode

    Beam Tetrode Member

    Messages:
    1,425
    Joined:
    Nov 16, 2005
    Location:
    The Moon
    I really don't want to alter the gain/tone in any way. I'd like to follow the original circuit as close as possible. The two channels will remain separate.

    Here's a link to an early Bassman circuit. Why don’t I follow this, and add 100K resistors between the plates and .02 caps?


    http://www.ampwares.com/ffg/schem/ba...e6-a_schem.gif

    I knew that I'd have to change the cathode resistor. However, I don't know how to calculate the proper value.
    Is the cathode resistor value always directly proportonal to the voltage on the cathode?


    Using this Bassman circuit, would I still have to tweak power suppy values? I'd like to maintain the 5D8's spec plate voltages. I am not too familiar with the math side of amp design...
     
  7. donnyjaguar

    donnyjaguar Member

    Messages:
    4,234
    Joined:
    Nov 4, 2005
    Location:
    Canada
    JP - Good point. What was Leo thinking? :)
    IT - If you consider this schematic you'll see Fender didn't bother to change the 10k resistor to something lower when there was only one tube. There seems to be some folk' who believe that lower voltage on the first section sounds better, and I won't disagree with that. If anything, I'd change it to 12k or 15k.

    The choice of plate and cathode resistor are entirely up to the designer. Not sure how to explain this in ten words or less. In analogue electronics tradeoffs abound. You have to consider how much gain you need vs how much distortion you can handle while factoring in current limitations of your power supply, and add to that how much life you want to get out of the tube itself. This is where those conduction charts in the tube manual come in handy. There's no right or wrong value for the resistors. 100k is a popular plate resistor value for a single 12AX7, but I have a stereo preamp that uses 301k plate resistors and it works fine too.

    Hope this helps. If you want me to expand on how to design a simple triode voltage amplifier section let me know.

    DJ
     

Share This Page