Hi All,
I'm LTspicing my first tube amp dablings. 300B fantasies. I noticed the following:
When the input signal on the grid exceeds the cathode voltage, the 300b output tube clips.
On the 12ax7 input tube however, the grid signal does swing past the cathode voltage. Which is correct? Is this supposed to happen or is this a modeling discrepancy?
Probably common sense to you. Yet I haven't found the answer. Perhaps you would like to point me to the obvious truth. Thank You!
Kind regards,
Ruben
I'm LTspicing my first tube amp dablings. 300B fantasies. I noticed the following:
When the input signal on the grid exceeds the cathode voltage, the 300b output tube clips.
On the 12ax7 input tube however, the grid signal does swing past the cathode voltage. Which is correct? Is this supposed to happen or is this a modeling discrepancy?
Probably common sense to you. Yet I haven't found the answer. Perhaps you would like to point me to the obvious truth. Thank You!
Kind regards,
Ruben
Grid current when the grid is driven positive relative to the cathode resulting in turning on the virtual diode that exists between the grid and cathode.
Hint: Have a look at the cathode of U2B when you overdrive the 300B grid. The 300B is driven by the comparatively high source impedance , low current of the 12AX7A follower in your mu follower. (look at both voltage waveform and current waveform in your grid circuits.)
The grid of U2A is driven by a 100 ohm source. See what happens to the grid current as the grid goes positive relative to the cathode.
U2A cathode voltage should be at least 800mV for decent linearity. The 12AX7 is not particularly well suited to mu-follower topology, look for something with higher transconductance and a higher operating current.
Try commonly available 12AT7 in your mu-follower circuit running at perhaps 5mA.
I use 6E6P with cascode CCS plate load - not as a follower to drive my 300B. This stage runs at 18mA and ~200V on the plate. (CCS on 400V rail)
Re 300B - 330K grid resistance is probably asking for trouble. I use <= 100K.
Hint: Have a look at the cathode of U2B when you overdrive the 300B grid. The 300B is driven by the comparatively high source impedance , low current of the 12AX7A follower in your mu follower. (look at both voltage waveform and current waveform in your grid circuits.)
The grid of U2A is driven by a 100 ohm source. See what happens to the grid current as the grid goes positive relative to the cathode.
U2A cathode voltage should be at least 800mV for decent linearity. The 12AX7 is not particularly well suited to mu-follower topology, look for something with higher transconductance and a higher operating current.
Try commonly available 12AT7 in your mu-follower circuit running at perhaps 5mA.
I use 6E6P with cascode CCS plate load - not as a follower to drive my 300B. This stage runs at 18mA and ~200V on the plate. (CCS on 400V rail)
Re 300B - 330K grid resistance is probably asking for trouble. I use <= 100K.
Last edited:
Thanks!
I also used the 12at7 before but yielded a lower total voltage gain of only 19dB. The cathode voltage was about 1V2 though.
It seems these 12ax7 and 12at7 models does not have this virtual diode but the 300B does.
Im finding that the 300B is running out of grid voltage, running into the diode, before running out of plate current. This limits power to 5 Watts before clipping.
I suppose the only way to get more power is using an OPT with higher primary Z, id est higher turn ratio?
Also what trouble, if I may ask?
I also used the 12at7 before but yielded a lower total voltage gain of only 19dB. The cathode voltage was about 1V2 though.
It seems these 12ax7 and 12at7 models does not have this virtual diode but the 300B does.
Im finding that the 300B is running out of grid voltage, running into the diode, before running out of plate current. This limits power to 5 Watts before clipping.
I suppose the only way to get more power is using an OPT with higher primary Z, id est higher turn ratio?
Also what trouble, if I may ask?
Last edited:
The 12AX7A/ECC83 most assuredly has that pesky diode, although the model may not.
A 12AT7A/ECC81 mu follower properly executed should give you around 34 - 35dB of gain. You are missing a plate load resistor for the lower triode.
And you are not biasing your driver tubes correctly, please have a look at data sheets for the 12AX7A and 12ATT7A and chose a reasonable operating point.
If I understand your transformer model it is 3.2K primary to 8 ohm secondary, and this near ideal for a 300B, higher primary Z will result in lower power. 3.2K should be good for approximately 8W @ <5% thd. (not a typo) I use 5K: 8/16 in mine which are designed to drive horns. (A dedicated channel for each horn.) Thd is < 2% at 6.5W
A 12AT7A/ECC81 mu follower properly executed should give you around 34 - 35dB of gain. You are missing a plate load resistor for the lower triode.
And you are not biasing your driver tubes correctly, please have a look at data sheets for the 12AX7A and 12ATT7A and chose a reasonable operating point.
If I understand your transformer model it is 3.2K primary to 8 ohm secondary, and this near ideal for a 300B, higher primary Z will result in lower power. 3.2K should be good for approximately 8W @ <5% thd. (not a typo) I use 5K: 8/16 in mine which are designed to drive horns. (A dedicated channel for each horn.) Thd is < 2% at 6.5W
2 sections of a 6sn7 in +300V anode, cathode to 300B grid 100R, then 100K to -300V will drive into grid current linear.
Power can exceed 15 Watts, however you need to have an overcurrent detection which will under-bias the 300B with a regulator to prevent it from destruction.
You install a zener from 300B grid bypassing the coupling capacitor.
Power can exceed 15 Watts, however you need to have an overcurrent detection which will under-bias the 300B with a regulator to prevent it from destruction.
You install a zener from 300B grid bypassing the coupling capacitor.
R11 is too small, I think. 12AX7 are typically biased for 1 mA or so, that would give 82 mV at the cathode. Are you sure of that value?
As @jcalvarez wrote, R11 value is mistake.
Use -for example- 1k5.
Sample:
Even gain is satisfactory (even 210Vpp Ug swing with 1% THD), this 12AX7 stage has tremendous output impedance, which IMO makes it unsuitable to drive complex input impedance device.
If you use it as VAS stage and drive power tube via cathode/source follower, may be working.... but two stage (for example high gm trioded pentode driving 300B) amplifier is simplest.
Use -for example- 1k5.
Sample:
Even gain is satisfactory (even 210Vpp Ug swing with 1% THD), this 12AX7 stage has tremendous output impedance, which IMO makes it unsuitable to drive complex input impedance device.
If you use it as VAS stage and drive power tube via cathode/source follower, may be working.... but two stage (for example high gm trioded pentode driving 300B) amplifier is simplest.
The clipping you see at the grid of the output tube is clipping of the driver's output. That is where the clipping begins. No clipping on the input grid because it is the source signal.
BTW in this Amplimo toroid OPT datasheet there is a typo:
secondary DC-resistance 0.30Ω(4Ω), 0.48mΩ(8Ω)
The correct value is 0.48 Ohm.
secondary DC-resistance 0.30Ω(4Ω), 0.48mΩ(8Ω)
The correct value is 0.48 Ohm.
Thank you for your guidance. I like the idea of a two stage two tube per channel design.
These operating points with 12AT7 should be more reasonable. The output impedance of the 12AT7 stage is about 420 ohms now.
Still only 5Wrms of P_out before U1 grid voltage meets U1 cathode voltage.
Vo/Vi is 18.3 dB.
These operating points with 12AT7 should be more reasonable. The output impedance of the 12AT7 stage is about 420 ohms now.
Still only 5Wrms of P_out before U1 grid voltage meets U1 cathode voltage.
Vo/Vi is 18.3 dB.
Some comments.
1.) If you use cathode bias for power tube, the R8 grid stopper IMO is unnecessary.
2.) Why do you use R13? It's decreasing B+.
Cathode biased output stage requiring as big as possible B+ (up to allowed Uak, up to dissipation limit 80..85%).
3.) I'm not a fan of this simple voltage regulator for power stage (IMO unwanted). Good (calculated) CLC filtering could sound nicer.
4.) No separation between power stage and driver stage B+. If the power stage jerking the HV, it affects the driver stage HV too.
IMO if you use power stage from filtered HV and driver stage from regulator, the isolation will be realised.
5.) Also not a fan of Broskie's "hum injecting" technic (C11). Tried it a few times, but I always got it out.
6.) C6 value requires setup. If it too large, with Lp may causes bump at LF.
If you use power stage from higher B+ (for example 425V), and a little higher current (OPT limit is 85mA), the power at A1 limit is increasing.
With 425V B+, 800R cathode resistor the current would be about 82mA.
The estimated -simulated- parameters (with this Amplimo OPT):
Ugpp Power(W) Out THD%
134 V 6.36 3.92 A1
1.) If you use cathode bias for power tube, the R8 grid stopper IMO is unnecessary.
2.) Why do you use R13? It's decreasing B+.
Cathode biased output stage requiring as big as possible B+ (up to allowed Uak, up to dissipation limit 80..85%).
3.) I'm not a fan of this simple voltage regulator for power stage (IMO unwanted). Good (calculated) CLC filtering could sound nicer.
4.) No separation between power stage and driver stage B+. If the power stage jerking the HV, it affects the driver stage HV too.
IMO if you use power stage from filtered HV and driver stage from regulator, the isolation will be realised.
5.) Also not a fan of Broskie's "hum injecting" technic (C11). Tried it a few times, but I always got it out.
6.) C6 value requires setup. If it too large, with Lp may causes bump at LF.
If you use power stage from higher B+ (for example 425V), and a little higher current (OPT limit is 85mA), the power at A1 limit is increasing.
With 425V B+, 800R cathode resistor the current would be about 82mA.
The estimated -simulated- parameters (with this Amplimo OPT):
Ugpp Power(W) Out THD%
134 V 6.36 3.92 A1