My two 5 Watt builds (with trem and without trem)

[hesamadman]

Here is the start of two 5 watt amps that I have started. One with Trem and one without Trem. The one without Trem is complete all but installation in a cabinet. It will be installed using a Celestion Ten 30. Im incredibly happy with this amp!


This was based off of the Gretsch 6150 that a buddy of mine had me take a look at some months back. I made some minor changes to the circuit on my breadboard since I was trying to get more gain and low end out of the amp.


Changes made were:

• Series of two .005 coupling caps following first half of triode were replaced with a .01 as I was trying for a bit more low end
• The plate resistor on first half of the triode was originally a 270k. It was replaced with a 100k.
• Added a switch where the first half of triodes cathode was switching from a 45uf/1.5k and a 820 Ohm/no bypass cap

I ran the secondary of the OT outside on chassis as I wanted to keep these wires from being under my board since this is a small chassis.


The only thing I was hoping to get some input on is this. When fully cranked, the amp obviously sounds pretty gainy. However, I feel like the distortion is somewhat fuzzy. Kind of almost like a solid state practice amp. When I turn the volume knob down just a tad it takes away the full on gain tone and cleans up a lot. Just like a sweet tube distortion. Am I just maxing out the distortion of the tube where its squaring off the signal like in diode clipping? Im dont think this is a flaw by any means. I just dont necessarily prefer the tone but it can easily be dialed out.


Ill have more photos when progress is made as well as a video demo.


Thanks again.

[John]

The only thing I was hoping to get some input on is this. When fully cranked, the amp obviously sounds pretty gainy. However, I feel like the distortion is somewhat fuzzy. Kind of almost like a solid state practice amp. When I turn the volume knob down just a tad it takes away the full on gain tone and cleans up a lot. Just like a sweet tube distortion. Am I just maxing out the distortion of the tube where its squaring off the signal like in diode clipping? Im dont think this is a flaw by any means. I just dont necessarily prefer the tone but it can easily be dialed out.

I believe that yes, you're just going into the hard clipping there. I think Tubenit has said that in his experience, most times his best tone comes around 7'ish on the scale. FWIW, on my couple amps I've built I've found the same. Then again, I don't have them set up for maximum clean anyway, they start to break up earlier since I'm not playing clubs. :) Hope this helps, and hope I have not mis-quoted Tubenit.

[hesamadman]

Just curious. In comparison to another amp with a very similar preamp circuit such as the tiny terror, how is it that the TT has a very heavy gain.  The TT is driven by two EL84 and obviously a phase inverter. Just like an AC15. Does this power amp circuitry add to the amount of gain in that amp? This amp also has a dual pot that controls the signal before it enters the 2nd triode as well as before the PI. Curious what that really does.


(Tiny Terror http://www.diyitalia.eu/forum/download/file.php?id=6441&mode=view)

[AZJimC]

The tiny terror has the first two triodes as full gain triodes, rather than two heavily loaded stages, and the dual gain pot actually changes the effective output of the previous triode. This keeps the back to back gain triodes from overdriving all the time. There is no preamp tone loading, and that lends to more total gain. The long tailed pair phase inverter has a substantial gain by it's self, and a strong output, and that allows easily overdriving the relatively easily pushed EL84's. The volume on the amp would be called a Post Phase Inverter Master Volume if installed on another amp.

With this circuit, your first gain control is the guitar volume knob, then the gain control. With this, you can overdrive the preamp tubes or not. Leaving the volume down would mean the preamp tubes would distort, but not the output tubes. The other choice could be to adjust preamp gain to a good strong clean tone, and then take the volume control up until output tubes are overdriven, and, as you stated, the loudness would then be similar to an 18w cranked. You find these are two quite different overdrive sounds. Of course a third option would be to overdrive the input, and crank the volume for both preamp and output distortion.

Nice amps you've got going there,
Jim

[hesamadman]

Hi Jim. Thanks For the reply. I have a couple quick questions for you.

The tiny terror has the first two triodes as full gain triodes, rather than two heavily loaded stages

Whats the difference in the two?


I dont know what the B+ is like in the tiny terror, but the PT I use in my SE circuit is a 500v. I have another transformer that is 650v. But wouldnt a higher B+ yield more gain as well? Im also wondering in taking my SE circuit and driving it with a pair of EL84's on my breadboard. In which case would require an additional triode for a phase inverter as well as the two EL84's. Im wondering if my Hammond 290AX (650vCT 81MA) would support all of this. IT has a 5V tap too so maybe ill try a rectifier tube as well. Ive only done SS rectifiers to cut cost in my builds.

[AZJimC]

Hi Jim. Thanks For the reply. I have a couple quick questions for you.

The tiny terror has the first two triodes as full gain triodes, rather than two heavily loaded stages

I dont know what the B+ is like in the tiny terror, but the PT I use in my SE circuit is a 500v. I have another transformer that is 650v. But wouldnt a higher B+ yield more gain as well? Im also wondering in taking my SE circuit and driving it with a pair of EL84's on my breadboard. In which case would require an additional triode for a phase inverter as well as the two EL84's. Im wondering if my Hammond 290AX (650vCT 81MA) would support all of this. IT has a 5V tap too so maybe ill try a rectifier tube as well. Ive only done SS rectifiers to cut cost in my builds.

Whats the difference in the two?

The gain of a stage is set (within the capabilities of the tube) by the plate voltage, and the plate resistor, which I like to think of as allowing the plate voltage to swing up and down, and, the resistance (lets ignore the capacitor for now) on cathode that allows it's voltage to be pulled upward in relation to the grid, and the loading that follows that stage, such as a 220K, 100K, 470K, or whatever to ground.

Other schematics that have sequential may have a 1.5K cathode resistor on first triode, and a 5-10K on second triode. This would make the second triode much lower gain.

If I have a triode that has a 3K resistor on cathode, and another that has 1.5K on it, the later would have more gain. If one had a 100K resistor to ground, from plate, (after the coupling cap normally) the other had a 470K, the later stage would have a greater voltage gain capability. In the case of the TT, it's gain stages are not heavily loaded (again after the coupling cap), and each have 1.5K cathode resistors, so they are back to back full throttle. The dual gain pot allows adjusting of the effective loading before signal is sent to the next stage.


About the voltage, if you're meaning 600+VDC, yes, you'd get more gain, until the tube goes ablaze with sparks. See your tube data charts, and don't exceed their max conditions by much if any. (Voice of experience) By all means, play with it, but be careful, and don't use your best tubes to do it.

I doubt the 81ma transformer would carry a pair of '84's in push-pull, and certainly not at 18w. Also be aware that adding a load to the 5v circuit, even within spec, adds heat to the transformer total as well.

[HotBluePlates]

... wouldnt a higher B+ yield more gain as well? ...

Yes, to some degree. But that's not the first place to go messing. Transformers are expensive, and the first thing to do is look at how the circuit throws away signal strength or has other tradeoffs.

Let's go through your circuit & the Tiny Terror in a back-of-envelope way (i.e., no loadlines which might give a little more accuracy but take a lot more time).

Your Amp:
1st Gain Stage:
12AX7 triode with 270kΩ plate load and bypassed cathode resistor. After this stage is a 470kΩ to ground then a 500kΩ Volume pot; above the frequency where the coupling caps roll-off, these are in parallel with each other and with the 1st stage's plate load resistor (and load the 1st stage, reduce gain). The resulting effective load resistance to a signal is 270kΩ in parallel with 470kΩ and 500kΩ, or ~128kΩ. Assuming internal plate resistance of the 12AX7 is 60kΩ as-operated, gain to the plate output is ~68 times.

Volume control: reduces signal strength for any setting below full.

2nd Gain Stage:
Sees at least a 470kΩ load from the grid reference of the 6V6. It also sees at least an additional parallel 500kΩ due to the tone control (more loading when the tone control is set for a bass-ier sound). These are in parallel with the 270kΩ plate load, which would give gain of 68 as in the first gain stage. However, the cathode resistor in not bypassed, and gain is cut by at least half due to local negative feedback to 34. But due to the relatively large 3.9kΩ cathode resistor, gain is reduced to ~22 (bigger unbypassed cathode resistors equal more local negative feedback equal less gain).

Total gain of preamp with Volume control at Max = 68 * 22 = 1496 (this assumes no loss due to tone control setting, and does not consider the different frequency roll-offs caused by the caps).

Your amp also has a 6V6, which in the same circuit has a larger bias voltage than an EL84. This means while you can get the same power output with each tube type, it takes more drive signal to reach that power output with the 6V6, making the circuit seem less sensitive or "gainy".

Tiny Terror
1st Gain Stage:
100kΩ plate load and bypassed cathode resistor. We don't know the resistance of the Gain pot, but I'll assume 1MΩ. R5 & R6 form a voltage divider, but the previous tube sees the 68kΩ in series with the parallel combination of the 470kΩ and the (assumed) 1MΩ resistance of the pot, or ~320kΩ. The effective load of the plate load resistor in parallel with this is 76.2kΩ. Assuming internal resistance of 60kΩ as before, gain is 56.

Voltage divider after 1st gain stage reduces signal, giving an effective gain of 0.87. Total gain to the top of the Gain pot is 56 * 0.87 = ~48.8.

2nd Gain Stage:
100kΩ plate load and bypassed cathode resistor. R9 & R10 form a voltage divider, with R10 (220kΩ) in parallel with the 2nd half of the Gain pot. Again assuming the pot is 1MΩ, the pot & R10 have a total resistance of 180kΩ. The previous triode sees the 68kΩ in series with this (248kΩ) and this total resistance in parallel with the 100kΩ plate load. Effective load is 71.3kΩ, for a gain of 54.

The 2nd voltage divider has an effective gain of 0.73, giving a total gain for the 2nd stage of 54 * 0.73 = 39.4.

Assuming Gain control is full-up, total gain to this point is 48.8 * 39.4 = ~1923. That's almost 30% higher than the preamp of your amp.

The Long-Tail Inverter gives a gain to one output of about half that of the same triode in normal preamp use. For an a.c. signal, the 100kΩ plate loads are in parallel with the 500kΩ master volume pot resistance and the 220kΩ EL84 grid reference resistor, for a total load of 58.8kΩ. The gain in a normal triode stage would be ~49-50, so the gain to one output is ~25.

So gain from input jack to the EL84 grid is 1923 * 25 = 48,075. That's 32 times as much gain as your amp to the same point in the circuit. And the push-pull EL84's probably have lower bias voltage than your SE 6V6 (though possibly more than a SE EL84), so the effect is compounded a bit more.

So despite seeing the Tiny Terror as having a preamp made of 2 12AX7 stages, it really has "2 and a half" due to the amplification of the long-tail inverter, plus extra gain due to bypassing the 2nd gain stage, plus an easier-to-drive output tube. Which is why it has more distortion than your amp.

[HotBluePlates]

The gain of a stage is set (within the capabilities of the tube) by the plate voltage, and ...

Maybe I should have written this in my previous post...

The formula used to determine gain of a triode (when not using the graphical procedure of loadlines) is:
Gain = Mu * [Load Resistance / (Load Resistance + Internal Plate Resistance)]

In a simple "1st Approximation" the "Load Resistance" is the value of the plate load resistor. But if there is any resistance to ground or B+ after the plate load resistor, then that resistance is effectively in parallel with the plate load resistor. So all that parallel resistance increases loading and reduces gain.

The other major variable component of that equation is internal plate resistance. I assumed a fixed value in all my calculations above. If you increased supply voltage such that plate current increased, the internal plate resistance would go down somewhat (you might even rationalize that as "internal resistance went down, so current went up" though that's not the whole story).

But there's a catch: Tube Mu or Amplification Factor. According to the definition of Mu, a 12AX7's mu of 100 really means "an increase in the grid-to-cathode voltage is 100 times more effective at increasing plate current than an increase in plate voltage." So jiggering the cathode resistor up/down (and the resulting change in grid-to-cathode bias voltage) will be much more effective in changing tube current (and lowering plate resistance) than altering the supply voltage.

Which is why I said, "... that's not the first place to go messing."

[AZJimC]

... wouldnt a higher B+ yield more gain as well? ...

Yes, to some degree. But that's not the first place to go messing. Transformers are expensive, and the first thing to do is look at how the circuit throws away signal strength or has other tradeoffs.

Let's go through your circuit & the Tiny Terror in a back-of-envelope way (i.e., no loadlines which might give a little more accuracy but take a lot more time).

Your Amp:
1st Gain Stage:
12AX7 triode with 270kΩ plate load and bypassed cathode resistor. After this stage is a 470kΩ to ground then a 500kΩ Volume pot; above the frequency where the coupling caps roll-off, these are in parallel with each other and with the 1st stage's plate load resistor (and load the 1st stage, reduce gain). The resulting effective load resistance to a signal is 270kΩ in parallel with 470kΩ and 500kΩ, or ~128kΩ. Assuming internal plate resistance of the 12AX7 is 60kΩ as-operated, gain to the plate output is ~68 times.

Volume control: reduces signal strength for any setting below full.

2nd Gain Stage:
Sees at least a 470kΩ load from the grid reference of the 6V6. It also sees at least an additional parallel 500kΩ due to the tone control (more loading when the tone control is set for a bass-ier sound). These are in parallel with the 270kΩ plate load, which would give gain of 68 as in the first gain stage. However, the cathode resistor in not bypassed, and gain is cut by at least half due to local negative feedback to 34. But due to the relatively large 3.9kΩ cathode resistor, gain is reduced to ~22 (bigger unbypassed cathode resistors equal more local negative feedback equal less gain).

Total gain of preamp with Volume control at Max = 68 * 22 = 1496 (this assumes no loss due to tone control setting, and does not consider the different frequency roll-offs caused by the caps).

Your amp also has a 6V6, which in the same circuit has a larger bias voltage than an EL84. This means while you can get the same power output with each tube type, it takes more drive signal to reach that power output with the 6V6, making the circuit seem less sensitive or "gainy".

Tiny Terror
1st Gain Stage:
100kΩ plate load and bypassed cathode resistor. We don't know the resistance of the Gain pot, but I'll assume 1MΩ. R5 & R6 form a voltage divider, but the previous tube sees the 68kΩ in series with the parallel combination of the 470kΩ and the (assumed) 1MΩ resistance of the pot, or ~320kΩ. The effective load of the plate load resistor in parallel with this is 76.2kΩ. Assuming internal resistance of 60kΩ as before, gain is 56.

Voltage divider after 1st gain stage reduces signal, giving an effective gain of 0.87. Total gain to the top of the Gain pot is 56 * 0.87 = ~48.8.

2nd Gain Stage:
100kΩ plate load and bypassed cathode resistor. R9 & R10 form a voltage divider, with R10 (220kΩ) in parallel with the 2nd half of the Gain pot. Again assuming the pot is 1MΩ, the pot & R10 have a total resistance of 180kΩ. The previous triode sees the 68kΩ in series with this (248kΩ) and this total resistance in parallel with the 100kΩ plate load. Effective load is 71.3kΩ, for a gain of 54.

The 2nd voltage divider has an effective gain of 0.73, giving a total gain for the 2nd stage of 54 * 0.73 = 39.4.

Assuming Gain control is full-up, total gain to this point is 48.8 * 39.4 = ~1923. That's almost 30% higher than the preamp of your amp.

The Long-Tail Inverter gives a gain to one output of about half that of the same triode in normal preamp use. For an a.c. signal, the 100kΩ plate loads are in parallel with the 500kΩ master volume pot resistance and the 220kΩ EL84 grid reference resistor, for a total load of 58.8kΩ. The gain in a normal triode stage would be ~49-50, so the gain to one output is ~25.

So gain from input jack to the EL84 grid is 1923 * 25 = 48,075. That's 32 times as much gain as your amp to the same point in the circuit. And the push-pull EL84's probably have lower bias voltage than your SE 6V6 (though possibly more than a SE EL84), so the effect is compounded a bit more.

So despite seeing the Tiny Terror as having a preamp made of 2 12AX7 stages, it really has "2 and a half" due to the amplification of the long-tail inverter, plus extra gain due to bypassing the 2nd gain stage, plus an easier-to-drive output tube. Which is why it has more distortion than your amp.

^^^^^^^^^^ What he said ^^^^^^^^ 
 I love it when he talks like that.

[PRR]

? ... wouldnt a higher B+ yield more gain as well? ...

Hardly any. And it may take *more* input to get into clipping.

See chart below for 12AU7 and 12AX7 amplifiers.

See 12AX7, use Rp=0.1Meg (100K) Rs=0.24Meg (240K), blue underline, because these are typical g-amp values.

Note they give three values for B+ ("Ebb"): 90V, 180V, 300V.

Compare the gains. 38, 46, 50.

You won't hardly hear a difference 46 or 50. So B+ can be 180V or 300V with nearly-no change of gain. (And going to 450-500V supply still doesn't buy enuff gain to notice.)

OTOH.... what input does it take to make a stage distort? This is roughly "Eo" (maximum clean output) divided-by Gain.

6.9V/38= 0.18V
20V/46= 0.43V
40V/50= 0.8V

The low-low voltage stage will overload *four* times easier than the high B+ stage.

[hesamadman]

The gain of a stage is set (within the capabilities of the tube) by the plate voltage, and ...

Maybe I should have written this in my previous post...

The formula used to determine gain of a triode (when not using the graphical procedure of loadlines) is:
Gain = Mu * [Load Resistance / (Load Resistance + Internal Plate Resistance)]

In a simple "1st Approximation" the "Load Resistance" is the value of the plate load resistor. But if there is any resistance to ground or B+ after the plate load resistor, then that resistance is effectively in parallel with the plate load resistor. So all that parallel resistance increases loading and reduces gain.

The other major variable component of that equation is internal plate resistance. I assumed a fixed value in all my calculations above. If you increased supply voltage such that plate current increased, the internal plate resistance would go down somewhat (you might even rationalize that as "internal resistance went down, so current went up" though that's not the whole story).

But there's a catch: Tube Mu or Amplification Factor. According to the definition of Mu, a 12AX7's mu of 100 really means "an increase in the grid-to-cathode voltage is 100 times more effective at increasing plate current than an increase in plate voltage." So jiggering the cathode resistor up/down (and the resulting change in grid-to-cathode bias voltage) will be much more effective in changing tube current (and lowering plate resistance) than altering the supply voltage.

Which is why I said, "... that's not the first place to go messing."

Thats awesome. Thank you for all that information. I spent a lot of time reading over it all and making sure I understood the math. I was reading your first post and figuring the math. I figured out the calculation of all the resistance to where you got 128k but I didnt know where to go from there......but then I read your second post and saw the formula. Thanks again. Im going to grab a copy of "Designing Preamp Circuitry" and study up on all that. I browsed through the chapter thats available on Valve Wizard website and thats what your post reminded me of.

[hesamadman]

Thanks everyone for your help. These builds turned out great. Built them for two very good friends. Amps sound great!

[hesamadman]

Blue amp no tremolo. White amp has tremolo. Both loaded with a Celestion ten30

[HotBluePlates]

Fine-looking amps!!