The 12AT7 has improved the THD at 1 Wrms (-50% for the EL84 version and -25% for the 6550) and the maximum power at 5% THD, plus has eliminated the need of a negative voltage for the resistors of the shunt cascode. Being also the bias positive, the only negative voltage is the -12V for the CCS of the phase splitter.
@Tubelab_com and @Rod Coleman ,
I would like to have your comments on the evolution below, its stability and any drawback that is clear to your eyes and not to mine.
Thinking on how to simplify this circuit further more I came with this solution and I would like to ask your opinion on it.
Preamp and poweramp are now DC coupled, with a dedicated stabilized power supply for the preamp using buffered LR8.
The bias is set by chosing Vpi through the pot U5, so the current flowing through R12 and R20, then balanced through the pot U2 (that I need to bypass with caps to permit changing the bias of the two tubes while avoiding AC unbalance).
The pot U7 sets the voltage at which the plates of the 12AT7 will work at.
Thanks in advance,
Roberto
I would like to have your comments on the evolution below, its stability and any drawback that is clear to your eyes and not to mine.
Thinking on how to simplify this circuit further more I came with this solution and I would like to ask your opinion on it.
Preamp and poweramp are now DC coupled, with a dedicated stabilized power supply for the preamp using buffered LR8.
The bias is set by chosing Vpi through the pot U5, so the current flowing through R12 and R20, then balanced through the pot U2 (that I need to bypass with caps to permit changing the bias of the two tubes while avoiding AC unbalance).
The pot U7 sets the voltage at which the plates of the 12AT7 will work at.
Thanks in advance,
Roberto
Yesterday I thought a bit more on how to make it simple to be set, and I came out with the following.
U1, the trimmer on the cathodes of the phase splitter, will balance the bias of the two output tubes (gain seems not to be affected on simulation) while U6 will set the output tubes' bias once balanced. 12AT7 voltages are fixed at 220V by U2, safe point for the FQPF9P25 maximum allowed voltage swing of 250V.
I will start building the preamp section and will connect an Arduino connected to a PC to store the data of the output DC voltage with input shorted.
This will help me understand what is the stability of the biasing voltage for the output tubes, so the possibility to have a DC coupled preamp and power amp.
PS
P-Mosfets here are IRFP9240 because on this computer I haven't the FQPF9P25 model.
PPS
Reading more information on the LR8, I will need a zener to lower the input voltage of the first one below 450V, and some capacitors to lower its noise.
U1, the trimmer on the cathodes of the phase splitter, will balance the bias of the two output tubes (gain seems not to be affected on simulation) while U6 will set the output tubes' bias once balanced. 12AT7 voltages are fixed at 220V by U2, safe point for the FQPF9P25 maximum allowed voltage swing of 250V.
I will start building the preamp section and will connect an Arduino connected to a PC to store the data of the output DC voltage with input shorted.
This will help me understand what is the stability of the biasing voltage for the output tubes, so the possibility to have a DC coupled preamp and power amp.
PS
P-Mosfets here are IRFP9240 because on this computer I haven't the FQPF9P25 model.
PPS
Reading more information on the LR8, I will need a zener to lower the input voltage of the first one below 450V, and some capacitors to lower its noise.
The driver looks simple enough, and you could simply build it and measure; the required output voltage swing is not so demanding as the DHTs I usually drive, so good results should be possible.
Noise from 3-terminal regulators like the LR8 is usually scaled up with the output voltage. It may need more than a capacitor to control it!
Noise from 3-terminal regulators like the LR8 is usually scaled up with the output voltage. It may need more than a capacitor to control it!
First of all I will check the drift as I wrote here:
Then, being the amp quite simple, I will check if an autobiasing system is required to be implemented.
Off-Topic with this thread, but In-Topic with your brand new GU72:
if it is of your interest, I've opened a thread with the same concept applied to Single Ended:
https://www.diyaudio.com/community/...t-cascoded-and-kt88-in-ul-unset.389721/latest
Last edited:
Just remembered, a few decades back i used some serial connected 2 legged TAA550 ICs instead of zeners, A/B/C types fŕom 30V-36V , dynamic resistance below 25 ohm, tempco -1,6mV +3,2mV/°C
You can still find em if you search the internet
Last edited: