I am now presenting my KT150 SE design for your comment. I have turned away from the 300B all dc coupled to kt150 se as the 300b I design would be quite massive and I want my first build to be manageable. I still keep the dc couple design by using 2 power rail for output and preamp circuit. Nevertheless I added the 2nd stage preamp which is ac coupled from the 1st stage for user to meet situation where extra gain is required by toggle a 2p2t switch (for each channel).The ECC88 along can provide 60V swing required by the kt150 from 5v p-p input which is around 1.77Vrms input signal. If 80V swing is required for driving 300B than the 2nd stage is certainly required unless your line level has at least 2.4Vrms output. Now just let me provide some key figures of the amp:
1st stage : 1/2 ECC88
Rk=370ohm (This value has been compared to , and should be close enough, to the value chosen from referencing the loadline (Ra+Rk=10k), finding the intercept of Vg and calculate the ratio of dVg/dIa in order to determine the value which hope the Rk can fully compensate the Vk to maintain the Va due to tube ageing or minor dc offset from input signal. I think a stable Va is of much importance to the next dc coupled stages as that affect the Vg indeed. Please share your view on this.). A lower resistance volume pot (20~25k) is chosen to meet -0.1db at 20khz for this stage after consider the miller capacitance, and still high enough for low impedance line level output (600ohm) from modern source.
2nd stage : 1/2 ECC88
This stage has unbypassed cathode resistor to lower the gain and increase headroom for the input from 1st stage. The output impedance will be higher compared to the one with bypassed cathode resistor but still the value is <6.3k, which is enough to meet the -0.1db at 20khz for driving the kt150 (assume the total miller capacitance is 170pf).
Output stage: KT150
No grid leak resistor is required I think as the dc coupled previous stage has DC resistance <25k which is much less than the min. requirement of 51k for fixed bias. The OPT is 4.2k ohm which is suitable for KT150 and also 300B, allow subsequent modification (320v tap is therefore prepared for 300b or other lower power tube). Ultralinear tap will be there on the OPT and wired as option for user to choose triode or ultralinear mode by a 2p2t switch (both channel). The KT150 will be at Va~540v Ia=120mA, Vg=-60. The "fully compensated" Rk is around 650ohm but it is way too hot and also not necessary to maintain Va as it is the last stage so the 200ohm is chosen will still provide some compensation and with acceptable power consumption.
The operating point is initially set for triode strapped but I look for the ultralinear curve and note that the operating point should also be ok under ultralinear mode.
Heater:
The preamp stages heater is now wired as AC powered but it is likely to switch to DC powered to minimize heater hum for these sensitivity stages. The kt150 heater will be AC powered with artificial ground and elevated but not a hum pot, which I guess is already enough for low amplification IDHT.
Tube rectifier:
6CJ3 is selected given its current output capacity for meeting the KT150 as will as its better look (no connection tap at the top of tube like 6d22s). The input capacitors are limited to 47uF which is lower than the limit of 6CJ3 and 6CA4. Capacitor value after that has no special limit as I can see from various materials and so they are chosen to achieve the desired ripple voltage target using PSUD2. A 47uF capacitor is connected from the filtered output tube power rail to the output tube cathode resistor as I learned this tech from an article from TubeCAD which can offset the B+ ripple on the output tube. Nevertheless I think a better filtered power supply is always superior to this tech as this induced B+ ripple to the cathode may mix with the signal and create intermodulation product.
NFB:
I think it worth the effort to provide option of adjustable GNFB from the speaker terminal to the 2nd stage preamp, as there is only DC coupled in between which should allow min phase shift problem when dealing with NFB.
Remarks:
I guess a potential divider may be required before c7 to lower the signal so that the volume will not jack up to much when toggle on the 2nd stage and also allow better volume adjustment.
Please share your view on the circuit, many thanks.
1st stage : 1/2 ECC88
Rk=370ohm (This value has been compared to , and should be close enough, to the value chosen from referencing the loadline (Ra+Rk=10k), finding the intercept of Vg and calculate the ratio of dVg/dIa in order to determine the value which hope the Rk can fully compensate the Vk to maintain the Va due to tube ageing or minor dc offset from input signal. I think a stable Va is of much importance to the next dc coupled stages as that affect the Vg indeed. Please share your view on this.). A lower resistance volume pot (20~25k) is chosen to meet -0.1db at 20khz for this stage after consider the miller capacitance, and still high enough for low impedance line level output (600ohm) from modern source.
2nd stage : 1/2 ECC88
This stage has unbypassed cathode resistor to lower the gain and increase headroom for the input from 1st stage. The output impedance will be higher compared to the one with bypassed cathode resistor but still the value is <6.3k, which is enough to meet the -0.1db at 20khz for driving the kt150 (assume the total miller capacitance is 170pf).
Output stage: KT150
No grid leak resistor is required I think as the dc coupled previous stage has DC resistance <25k which is much less than the min. requirement of 51k for fixed bias. The OPT is 4.2k ohm which is suitable for KT150 and also 300B, allow subsequent modification (320v tap is therefore prepared for 300b or other lower power tube). Ultralinear tap will be there on the OPT and wired as option for user to choose triode or ultralinear mode by a 2p2t switch (both channel). The KT150 will be at Va~540v Ia=120mA, Vg=-60. The "fully compensated" Rk is around 650ohm but it is way too hot and also not necessary to maintain Va as it is the last stage so the 200ohm is chosen will still provide some compensation and with acceptable power consumption.
The operating point is initially set for triode strapped but I look for the ultralinear curve and note that the operating point should also be ok under ultralinear mode.
Heater:
The preamp stages heater is now wired as AC powered but it is likely to switch to DC powered to minimize heater hum for these sensitivity stages. The kt150 heater will be AC powered with artificial ground and elevated but not a hum pot, which I guess is already enough for low amplification IDHT.
Tube rectifier:
6CJ3 is selected given its current output capacity for meeting the KT150 as will as its better look (no connection tap at the top of tube like 6d22s). The input capacitors are limited to 47uF which is lower than the limit of 6CJ3 and 6CA4. Capacitor value after that has no special limit as I can see from various materials and so they are chosen to achieve the desired ripple voltage target using PSUD2. A 47uF capacitor is connected from the filtered output tube power rail to the output tube cathode resistor as I learned this tech from an article from TubeCAD which can offset the B+ ripple on the output tube. Nevertheless I think a better filtered power supply is always superior to this tech as this induced B+ ripple to the cathode may mix with the signal and create intermodulation product.
NFB:
I think it worth the effort to provide option of adjustable GNFB from the speaker terminal to the 2nd stage preamp, as there is only DC coupled in between which should allow min phase shift problem when dealing with NFB.
Remarks:
I guess a potential divider may be required before c7 to lower the signal so that the volume will not jack up to much when toggle on the 2nd stage and also allow better volume adjustment.
Please share your view on the circuit, many thanks.
Attachments
Last edited:
I would give some consideration to whether you plan to wire the KT-150 as a triode, pentode, or in UL. You'll need a different amplifier design (and possibly a different OT) for each configuration.
I know the OPT should got the ultralinear tap and the optimum operating point for the 2 modes might be different. It is initially designed for triode mode but I just look at the ultralinear mode curve again and think the operating point should be ok. What kind of different design consideration is required if it is wired as UL mode besides the different Rk and Va for setting a different the operating point ?
If you wire for UL or pentode operation, feedback becomes a necessity and some extra gain can be helpful. Triode operation is by far the easiest on the output transformer as well. A 10W OT would be fine for that, but if you go for more power you'll want bigger iron.