Announcement

Collapse
No announcement yet.

Master volume and treble loss

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Helmholtz
    replied
    Originally posted by vintagekiki View Post
    only thing I don't understand is why Fender would make an guitar amps with the cutoff frequency < 4 kHz.
    That's why he used the bright cap at the vol. for compensation. But the frequency compensation is only perfect at mid setting of the pot.

    2nF capacitor between g1 and cathode
    In this example the source resistance is only around 10K (because of the low plate impedance of a 12AT7), so the cutoff frequency is around 8kHz. But as the LP filter is within the NFB loop, it's effect will be partly equalized.

    Connect a capacitor between the g1 (or anode) and cathode of the preamp tube and measure the cutoff frequency, and listen to guitar sound with and without 150pF.
    As said above I could verify the Miller theory by measuring input capacitances between 130pF and 150pF with my LCR meter.
    LP cutoff frequency varies with source resistance. You should be able to calculate it yourself for a given circuit.

    Effective source resistance between plate and cathode is typically only around 40k, so the cutoff frequency will be much higher than in my example with 300k.

    And yes, I know how a 4kHz LP filter changes the sound.
    Last edited by Helmholtz; 12-14-2020, 08:05 PM.

    Leave a comment:


  • vintagekiki
    replied
    Originally posted by Helmholtz View Post
    ... ...
    I'm free to notice that you didn't understand
    https://music-electronics-forum.com/forum/music-electronics/920070-master-volume-and-treble-loss?p=920246#post920246
    CLMV Question
    If the signal at point A as at the picture and level is U = 1V, what will be the signal at points ? (graphic form and how much is his the level).

    About instructive and practical
    Measured and practically checked everything I write.
    I do not base my comments on retelling someone else's experience.
    In order not to "heat hot water" (copy/ paste), I always list http where the original article can be downloaded.

    Leave a comment:


  • vintagekiki
    replied
    Originally posted by Helmholtz View Post
    ... ...
    I'm not really sure about the above math, only thing I don't understand is why Fender would make an guitar amps with the cutoff frequency < 4 kHz.

    (2nF capacitor between g1 and cathode OT)
    http://www.thetubestore.com/lib/thetubestore/schematics/Fender/Fender-Bandmaster-AD1269-Amp-Schematic.pdf
    http://www.thetubestore.com/lib/thetubestore/schematics/Fender/Fender-Bandmaster-Reverb-AA270-Schematic.pdf
    http://www.thetubestore.com/lib/thetubestore/schematics/Fender/Fender-30-Schematic.pdf
    http://www.thetubestore.com/lib/thetubestore/schematics/Fender/Fender-Dual-Showman-Reverb-AA769-Schematic.pdf

    (500pF capacitor between anode and cathode V1a / V1b / V2b / V4b)
    http://www.thetubestore.com/lib/thetubestore/schematics/Fender/Fender-140-Schematic.pdf
    http://www.thetubestore.com/lib/thetubestore/schematics/Fender/Fender-Twin-Reverb-II-Schematic.pdf

    I know that the road to hell has paved with good intentions, and so I refrain from comment about Miller and Fender, so that individuals are not found.
    Click image for larger version  Name:	idea.png Views:	0 Size:	134 Bytes ID:	920296
    It is not difficult. It only takes 2 wires and capacitor between 130pF and 150pF. Connect a capacitor between the g1 (or anode) and cathode of the preamp tube and measure the cutoff frequency, and listen to guitar sound with and without 150pF.

    Leave a comment:


  • Helmholtz
    replied
    Originally posted by vintagekiki View Post

    By reducing the MV pot (1M), the drive on g1 output tubes still exist only smaller.
    When the MV pot (1M) at minimum, the g1 output tubes are short-circuited, and since anodes PI are also short-circuited, there is no drive at the g1 output tubes, only idle (bias) current will flow through the OT.
    Even with zero signal BETWEEN the grids you can have a common signal on both grids wrt ground.
    Just apply the same signal to both grids and measure what happens.


    Just to notice
    The whole story about CLMV comes down to interpreting various simulations.
    It is not difficult. It only takes 2 wires and a quality 1M pot to move from theory to practice.

    Simulation is just a simulation, personal practical experience is something completely different, which is worth discussing.
    If you don't trust the simulations I recommend to get out your 2-channel scope and measure yourself.
    That should be most instructive and practical enough .
    Last edited by Helmholtz; 12-14-2020, 03:38 PM.

    Leave a comment:


  • Helmholtz
    replied
    Originally posted by vintagekiki View Post
    About Miller capacitance
    Miller capacitance is a story from the HF area (> 30MHz).
    Leo Fender, Jim Marshall and others probably haven't even heard of Miller capacitance, and they've made tons and tons of amps.
    Another fallacy!

    From the Miller theorem it follows that any capacitance between plate and grid appears multiplied by the gain of the stage between grid and ground.
    The internal capacitance of a 12AX7 is 1.7pF and the wiring typically adds another 0.5pF.
    So the input capacitance of the stage is around 2.2pF x 60 = 132pF for a typical gain of 60.

    I actually measured values between 130pF and 150pF.

    If this gain stage is driven from a source impedance of e.g. 300k (1M vol. pot at some middle setting), the cutoff frequency will be 4kHz ( not MHz ! ) or even lower.

    I am quite sure that Leo Fender was aware of this effect as he used bright caps of suitable values to compensate the Miller loss.

    https://en.wikipedia.org/wiki/Miller_effect

    Leave a comment:


  • vintagekiki
    replied
    Originally posted by pdf64 View Post
    Did you see the thread where Jon Snell raised the issue of a common mode positive bias caused by the type 3 / crossline master vol? https://music-electronics-forum.com/...olume-question
    I saw and read.
    #15
    I only know the theory as I have not recorded any results.
    Jon Snell fenced himself off correct.

    #13
    Consider drive of 10volts on each anode of the phase inverter, the drive is still there in this circuit wherever the control is set.
    By reducing the MV pot (1M), the drive on g1 output tubes still exist only smaller.
    When the MV pot (1M) at minimum, the g1 output tubes are short-circuited, and since anodes PI are also short-circuited, there is no drive at the g1 output tubes, only idle (bias) current will flow through the OT.

    Just to notice
    The whole story about CLMV comes down to interpreting various simulations.
    It is not difficult. It only takes 2 wires and a quality 1M pot to move from theory to practice.

    Simulation is just a simulation, personal practical experience is something completely different, which is worth discussing.


    Leave a comment:


  • Helmholtz
    replied
    Originally posted by vintagekiki View Post

    1) How high are the currents, if it is the bias voltage?
    2) If the signal voltage between PI outputs will be zero, how much is the signal voltage at g1 power tubes?
    3) In technique there is no "but you may still have", just have or not to have.
    4) I don't understand how there is plate dissipation when no output.

    Maybe I'll learn something I didn't know. I am especially interested:
    If the signal at point A as at the picture and level is U = 1V, what will graphic form of the signal at points ? and how much is his the level.
    1) Measure yourself. I would estimate up to 4mA peak depending on circuit values.
    2) Study this thread, especially the simulations by Chuck H: https://music-electronics-forum.com/...olume-question
    3) Refrain from indoctrinating phrases if you want answers from me! I meant what I said.
    4) Common mode grid signals produce in-phase plate currents. In-phase plate currents don't produce OT induction, only see low DCR and don't produce OT output.
    Last edited by Helmholtz; 12-13-2020, 11:17 PM.

    Leave a comment:


  • vintagekiki
    replied
    Originally posted by Helmholtz View Post
    A bright cap especially makes sense when a vol pot feeds a high gain triode stage, as there will be considerable Miller capacitance, typically >100pF.
    About Miller capacitance
    Miller capacitance is a story from the HF area (> 30MHz).
    Leo Fender, Jim Marshall and others probably haven't even heard of Miller capacitance, and they've made tons and tons of amps.

    Leave a comment:


  • vintagekiki
    replied
    Originally posted by Helmholtz View Post

    Well, you surely know that shorting kills voltage but may cause high currents. 1)
    So signal voltage between PI outputs will be zero2). But you may still have3) considerable common mode voltages wrt ground as Chuck H demonstrated in the other thread.
    Equal common mode grid signals cause plate dissipation but no output4).
    1) How high are the currents, if it is the bias voltage?
    2) If the signal voltage between PI outputs will be zero, how much is the signal voltage at g1 power tubes?
    3) In technique there is no "but you may still have", just have or not to have.
    4) I don't understand how there is plate dissipation when no output.

    Maybe I'll learn something I didn't know. I am especially interested:
    If the signal at point A as at the picture and level is U = 1V, what will graphic form of the signal at points ? and how much is his the level.

    Leave a comment:


  • Helmholtz
    replied
    Originally posted by vintagekiki View Post
    CLMV Question

    If signal at point A as at the picture, what will be the signal at points ?

    Click image for larger version  Name:	Question.gif Views:	0 Size:	53.8 KB ID:	920247
    Well, you surely know that shorting kills voltage but may cause high currents.
    So signal voltage between PI outputs will be zero. But you may still have considerable common mode (in-phase) voltages wrt ground as Chuck H demonstrated in the other thread.
    PI triodes will work at their limit to supply short circuit currents.

    Equal common mode grid signals cause plate dissipation but no output.
    Last edited by Helmholtz; 12-13-2020, 09:04 PM.

    Leave a comment:


  • vintagekiki
    replied
    CLMV Question

    If signal at point A as at the picture, what will be the signal at points ?

    Click image for larger version

Name:	Question.gif
Views:	656
Size:	53.8 KB
ID:	920247

    Leave a comment:


  • Helmholtz
    replied
    Originally posted by vintagekiki View Post
    Treble loss occur at resistive attenuators, so it is necessary to perform capacitive compensation (e.g. instrument probes).
    A resistor or resistive voltage divider cannot produce treble loss by itself. Rather treble loss is caused by a low pass filter which requires a series resistance and a shunt capacitance to ground.
    As the input capacitance of pentodes/tetrodes is very small, a dual pot MV itself will not produce any treble loss in the audio range.
    But there is some possibility, that long (shielded) wires to and from the MV have enough shunt capacitance to actually cause some audible treble loss.

    A bright cap especially makes sense when a vol pot feeds a high gain triode stage, as there will be considerable Miller capacitance, typically >100pF.

    For example, with 100k series resistance and 200pF shunt capacitance the -3dB frequency is 8kHz.
    The input capacitance of an EL34 is only 16pF.


    About CLMV
    MV 1MA pot simply mixes the two mirror image audio streams from the phase inverter together. The two streams will cancel each other out.
    If you mean signal currents by "audio streams", your interpretation is wrong. The 2 PI output currents do not cancel with a CLMV, rather they add/double and force excessive PI signal currents at low resistance/ CLMV settings. In fact a CLMV acts like loading/ shorting the plate outputs of a push-pull amp.
    Don't forget that PI plate voltages are opposite phase. So when one plate goes positive, the other goes negative. A load connected between the PI outputs sees doubled signal voltage. Consequently a CLMV causes increased PI currents. No cancellation effect at all.
    Last edited by Helmholtz; 12-12-2020, 11:22 PM.

    Leave a comment:


  • vintagekiki
    replied
    Originally posted by Helmholtz View Post
    The OT's MV doesn't cause treble loss either (except for some Fletcher- Munson effect at low volume). Don't forget pentode input capacitance is very small.

    And cross-phase MVs cause bass loss at low settings.
    Unfortunately, I do not share your opinion. Treble loss occur at resistive attenuators, so it is necessary to perform capacitive compensation (e.g. instrument probes).
    At amps the resistive attenuator is a volume control.
    To compensate treble loss in the process of volume control, in the assembly volume control, manufacturers install capacitors which raise treble (bright/ treble boost).
    At Fender old generation it's bright sw (120pF), Marshall (JMP, JCM) it's fixed treble boost (.005u), Vox AC30 it's fixed treble boost (100p), etc.

    About Fletcher-Munson curves
    The Fletcher-Munson curves (not effect) is some story from the Hi-Fi domain which indicate the ear’s sensitivity to different frequencies at various levels. That would not be copy / paste, whoever is interested can be informed here:

    https://en.wikipedia.org/wiki/Equal-loudness_contour

    https://www.teachmeaudio.com/recording/sound-reproduction/fletcher-munson-curves

    About CLMV
    MV 1MA pot simply mixes the two mirror image audio streams from the phase inverter together. The two streams will cancel each other out.
    Matchless and in some new Vox korgusa amp series are used CLMV

    Because of its simplicity, I use the CLMV schematics for years when the user requests to install an MV.
    It has performed well at old generation Cender, Marshall, Vox amp.
    Bass loss at low volume settings was not significantly noticed.

    https://music-electronics-forum.com/forum/amplification/guitar-amps/theory-design/12307-cross-line-master-volume

    https://schems.com/bmampscom/matchless/

    http://bee.mif.pg.gda.pl/ciasteczkowypotwor/%23SM_scena/VOX/Vox_AC30CC2_AC30CC2X%20(2005)%20SM.pdf

    http://dealers.korgusa.com/svcfiles/TB18C1.pdf

    Leave a comment:


  • J M Fahey
    replied
    Loved it.

    Leave a comment:


  • Boy Howdy
    replied
    Originally posted by pdf64 View Post

    Did you see the thread where Jon Snell raised the issue of a common mode positive bias caused by the type 3 / crossline master vol? https://music-electronics-forum.com/...olume-question
    I'm not terribly technical - I take it that this type of master is not a great idea?

    Here's a little spur of the moment thing I recorded of a Garnet PA I turned into Marshall that uses that Matchless type MV. The guitar is a Strat, the volume is at speech level. GARNET REBEL PA 90 BLUES by gene downs | Free Listening on SoundCloud

    Leave a comment:

Working...
X