Tube Emulation & EQ

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I still have the power amp section, can't remember what we used for a preamp.

I took this picture of what I had left before giving away or trashing about 2/3rds of my lifetime collection of stuff before moving 1200 miles. The little boards in various stages of completion are our modified version of the Plastic Tiger. More reliable, but still about 35 watts out. The larger boards are Universal Tigers, and the one oddball was a dual Plastic Tiger with some of the parts missing. We made those for stereo HiFi amps. Bridging them for higher power guitar amps led to smoking parts.

My guitar preamp was likely a clone of the Kustom 100 Watt head, or an Acoustic "120" or some combination of the two. I had traced both amps. I remember the numbers 2N3391 and 2N3393, but they were pretty common back then. The power amp section for both amps was taken directly from an RCA Transistor Manual, as was the Heathkit guitar amp. I got parts from the local Heathkit store.

I have another picture of a car audio power amp I made with the same boards.
 

Attachments

  • TigerAmps_2_A.jpg
    TigerAmps_2_A.jpg
    297.9 KB · Views: 309
  • MasterBlaster_A.jpg
    MasterBlaster_A.jpg
    250.8 KB · Views: 323
My gripe with it is from the usability end of things. Turning any one knob causes changes over the entire frequency spectrum. The response of the controls is also very nonlinear, a ten-degree rotation around the 7-o-clock position produces a very different change than the same ten-degree rotation around the 3-o-clock position.
Absolutely true. we learn to live with it, but yeah it's kind of messed up.

I've seen claims that the Hiwatt tone stack has smoother response as you turn the knobs, and can cop Fenderish and Marshallish vibes, as well as its own unique thing, but I have never tried it. It does have 4 extra components compared to Fender's stack. As usual, some say it's the cat's meow and others say don't bother.

I posted the LTSpice .asc file in post #426. Please feel free to download and tinker with it. Transistor swaps are just a few mouse clicks away!
I did just that, thanks. I compared BC550C/BC560C to 2N3904/2N3906. Nearly identical. You have to look really hard to see any difference at all.
 
I've seen claims that the Hiwatt tone stack has smoother response as you turn the knobs, and can cop Fenderish and Marshallish vibes, as well as its own unique thing, but I have never tried it. It does have 4 extra components compared to Fender's stack.
Merlin Blencowe's valve preamp book ( The Valve Wizard ) includes a short mention of the Voight tone control, originally designed before WWII to "lift bass" on early records, operating down below 50 Hz. It's a two-knob passive circuit that uses very few components, and is much better behaved than Leo's three-knob creation.

I took the schematic in Blencowe's book, and tweaked it extensively (some might say obsessively) in LTSpice, until I got a set of frequency responses that I thought would be well suited to guitar.

The two attached screen captures give you an idea; note that control response is fairly uniform over most of the range (though somewhat closer spaced near maximum cut settings), and there isn't much interaction between the two knobs.

The second image shows the result of stepping one knob, then stepping the second knob through its entire range, and repeating the process. The fact that the lines cluster into fairly distinct groups tells us that the controls are pretty independent - the treble control doesn't drag the bass response around with it, and vice versa.

I'm also attaching the .asc simulation files, if anyone wants to try running their own simulations with this circuit.

R6 (40k) is not part of the tone control, but represents the output resistance of the preceding 12AX7 gain stage - this is roughly what you get using the usual 100k anode resistor, combined with the internal anode resistance of the triode, which I took to be approximately 70k. So the frequency response curves you see are what you actually get in real life, with the tone control fed from half a 12AX7.

R2/R3 is a 500k (or 470k) logarithmic potentiometer, and is the bass control. R4/R5 is a 250k (or 220k) log pot, and controls treble. I derived the weird equations you see next to those four resistors myself - they describe the behaviour of the standard "10%" log audio pot. The variable "k" tells you how far the pot has been rotated; k=0 is full anti-clockwise, k=1 is full clockwise.

I derived those equations myself because I found at least one problematic LTSpice linear potentiometer model online, which produced the right voltage, but not the right source impedance; the wiper of a pot is not a pure voltage source, but has a Thevenin impedance that varies with the setting of the pot. This was missing in that model.

Having seen encouraging results from the simulations, I built this circuit a couple of years ago (take note, Benb!), and included it into a preamp I designed and built around a pair of 6JW8 triode-pentode valves.

I liked the result a lot - it doesn't produce Fendery sounds, since it doesn't have the deep notch in the mid frequencies, but I enjoyed its set-and-forget nature. In use, there is no perceptible interaction between the bass and treble controls, which is a Good Thing (TM).

It is, of course, quite possible to create a separate mid control elsewhere in the preamp. I didn't feel the need to do that myself; the small-signal "pentode" (really a beam tetrode) in the 6JW8 provided a bit of brightness and shimmer on its own, without needing a big mid scoop.

As usual, some say it's the cat's meow and others say don't bother.
And I'm sure the same would be true of this tweaked Voight circuit, if anybody else were using it. I think it meows just fine (and I have two constantly miaowing cats, so I know a miaow when I hear one!) But others may or may not share the sentiment.

I compared BC550C/BC560C to 2N3904/2N3906. Nearly identical.
Aha! I suspected as much.

And if you use four op-amps containing twenty five transistors apiece, you will probably produce almost exactly the same frequency response, but use a hundred transistors instead of four to do it!

Don't get me wrong, I love op-amps - they do so many things so very well. But I'm not convinced that they are the right solution in this specific case, where we might want a speaker emulation filter that overloads relatively gracefully, something op-amps are not known for doing.

-Gnobuddy
 

Attachments

  • Voight_Tone_Control_26_Jul_31_2016.png
    Voight_Tone_Control_26_Jul_31_2016.png
    63.9 KB · Views: 315
  • Voight_Tone_Control_30a_Jul_31_2016.png
    Voight_Tone_Control_30a_Jul_31_2016.png
    100.6 KB · Views: 313
  • Voight_Tone_Control_26_Jul_31_2016.asc
    2 KB · Views: 71
  • Voight_Tone_Control_30a_Jul_31_2016.asc
    2.2 KB · Views: 74
The Mosfets in the input, don't go to ground.
Makes sense. Since these appear to be enchancement MOSFETs, the source needs to be at lower voltage than the gate. Feeding all the sources from (-12V) allows the gates to be tied to ground, simplifying biasing.

There is another benefit, using a negative rail with a voltage much larger than |Vgs| stabilizes the operating point against MOSFET parameter spreads. Small changes in MOSFET Vgs are now riding on top of 12 V (rather than being applied directly across the source resistor), which reduces the percentage change in voltage across the source resistor.

I assume the (-12V) rail had to be there to power all the op-amps anyway.

The circuitry around IC3A, IC3B had me scratching my head for a while because they used the symbol for a standard op-amp. It looks like IC3A's output is a virtual ground, and that IC3B's output is therefore effectively shorte to ground. Then I found a datasheet for the M5201 online: M5201P datasheet(1/9 Pages) MITSUBISHI | GENERAL PURPOSE SWITCHING OPERATIONAL AMPLIFIER(DUAL INPUT, SINGLE OUTPUT TYPE)

From the datasheet, it looks as though either IC3A, or IC3B, can be enabled, while the other one stays off.

-Gnobuddy
 
A fancy switch, but why not? I wish the schematic were bigger for these eyes but better than nothing. I am back to carving wood, have to finish some obligations I got myself into. I wanted to stick this schematic with the other high voltage Mosfet circuit so I could delve into it at a later date. I am going back to work, :ill: but I can see putting my feet up for good in the not too distant future. :cloud9: Then I could play around.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.