Candidate for jean-paul's demand of 1 volt rms for full output, here is wonderful EC86
http://www.r-type.org/pdfs/ec86.pdf
Regards
http://www.r-type.org/pdfs/ec86.pdf
Regards
Not my demand, I plead for standardization as it is severely lacking in this field. The whole issue is of course because of the reverse way of engineering. If the design parameters are first defined and then the tube would be chosen things would be way less complicated.
Grabbing a random tube and then have and accept (!) non standard parameters is the normal way of working however in the tube DIY field. This creates more issues than it solves as it is a choice driven by convenience. It is very annoying when device A only can work with device B and not with device C. Unnecessary, stupid and laughable because it is almost exclusively in this field. If semi guys would design a DAC with 8V rms output voltage then it will not be accepted by the same people that defend their own non standard stuff.... Either very high input sensitivity of 100 mV or the other extreme of several Volts rms.
It is easily solved by assuming the input/volume section to be a passive black box with let's say 25 kOhm input impedance (10 kOhm might be a bridge too far) and 1V rms for full output as a standard. A bit like the rest of the audio world does for decades. With 2V rms of the sources as a standard things will then always work out OK. Then the necessary gain and calculations are up to the amplifier designer.
Beware, it would make the often already unnecessary high gain preamps definitely obsolete at once. Second risk is simpler audio chains with less distortion, less noise etc. Third risk is that everything would work together perfectly. These sacrifices are a high price to pay 😀 Seriously, if someone can point me to negative sides of standardization I am the first that wants to know.
Grabbing a random tube and then have and accept (!) non standard parameters is the normal way of working however in the tube DIY field. This creates more issues than it solves as it is a choice driven by convenience. It is very annoying when device A only can work with device B and not with device C. Unnecessary, stupid and laughable because it is almost exclusively in this field. If semi guys would design a DAC with 8V rms output voltage then it will not be accepted by the same people that defend their own non standard stuff.... Either very high input sensitivity of 100 mV or the other extreme of several Volts rms.
It is easily solved by assuming the input/volume section to be a passive black box with let's say 25 kOhm input impedance (10 kOhm might be a bridge too far) and 1V rms for full output as a standard. A bit like the rest of the audio world does for decades. With 2V rms of the sources as a standard things will then always work out OK. Then the necessary gain and calculations are up to the amplifier designer.
Beware, it would make the often already unnecessary high gain preamps definitely obsolete at once. Second risk is simpler audio chains with less distortion, less noise etc. Third risk is that everything would work together perfectly. These sacrifices are a high price to pay 😀 Seriously, if someone can point me to negative sides of standardization I am the first that wants to know.
Last edited:
jean-paul......I am grateful to you for your valueable input. I 100% agree with your proposition. Wonder why many of us (Diyers) don't act accordingly. A good read about gain structure What is Gain Structure?.
Regards
Regards
Your attachment looks almost identical to the circuit I and others came up with here: 45 amp build direct coupled
Can you explain what's wrong with this simple math (i.e., why gain=15 is not enough to drive a 45 tube biased between -40 to -50 V):
2.82 Vpk x 15 = 42.3 Vpk
A very good read indeed! The conclusion of this write up is this:
It is probably not surprising that I am in the camp for "low-gain amplifiers".
Last edited:
Hello,
I am not anderstand all but : If you need 100 Vpp for the swing of the final.
100 / 2,8 = 35 Vrms / 1 Vrms = 35 of AF
C3g in triode no microphonic low hum, AF 40 and output impedance around 2 k is depend of the load.
I'm not if that is the topic.
I am not anderstand all but : If you need 100 Vpp for the swing of the final.
100 / 2,8 = 35 Vrms / 1 Vrms = 35 of AF
C3g in triode no microphonic low hum, AF 40 and output impedance around 2 k is depend of the load.
I'm not if that is the topic.
Low gain and improper drive are two different things. Sorry if you don't want to understand I can't make you understand. I give up!
Regards
Regards
EBA....earlier I was taking about C3m. My bad, you are right, it should be c3g instead of C3m 'cause mu of c3m is about 20 and mu of c3g is about 40.
Regards
Regards
Yes, I was one of "others", and then I tested for personal use this sketch, and the own verdict was "unsatisfactory gain" 😛
As you can see I wrote in #76 post:
"At 2V RMS (2.82 V peek) no enough swing even to A1 border (at -47V bias)"
47/2.82= 16.66
If you choose lower bias (thus lower output power), the gain 15 would be enough .... for reaching A1 ... without any headroom.
My simple method of reverse researching of amplifier:
- choosing for requiring SPL a loudspeaker;
- choosing for this loudspeaker the adequate power;
- choosing the type of amplifier for this power and loudspeaker parameters;
- choosing matching OPT (depends of power and loudspeaker parameters);
- choosing tube for OPT (and power, and loudspeaker parameters, and ....);
- choosing B+ and operating point depending of tube, power, reflected load ...;
- choosing grid swing for requiring power (depending of load);
- choosing VAS stage swinging capacity at least +3dB larger, than calculated, requiring for proper headroom !!!!
16.6 is not far off from 15.
Why is it useful to design the amp such that it will go into clipping?
@ euro21,
2 VRMS is 5.86 Vpk-pk.
The 45 tube working in class A1 would need 94 Vpk-pk in your sim.
A gain of 15 in the driver tube will give some 88 Vpk-pk.
To drive the 45 in A1, the input sensitivity would be 1.07 VRMS with a 6H30 tube and gain of 15.
Very close to the "1 VRMS" target...
2 VRMS is 5.86 Vpk-pk.
The 45 tube working in class A1 would need 94 Vpk-pk in your sim.
A gain of 15 in the driver tube will give some 88 Vpk-pk.
To drive the 45 in A1, the input sensitivity would be 1.07 VRMS with a 6H30 tube and gain of 15.
Very close to the "1 VRMS" target...
Thank you, it is indeed meant constructively and of course also to make coupling of tube devices with semiconductor devices that have standardized line levels/impedances (and vice versa) also less troublesome. The link you gave can not be repeated enough, I share it myself regularly.
Last edited:
I believe this should read "To drive the 45 in A1, the input sensitivity would be 2.07 VRMS with a 6H30 tube and gain of 15. Very close to the "2 VRMS" target."
Correct?
It is not about overdriving, but about minimizing driver distortion at maximum swing.
If you design the driver section exactly to required capacity, the driver would be clipping earlier than power section at random overdriving.
Commonly the driver section clipping is "harder" (mostly more odd components) than power section clipping, so clipping power tube/s/ earlier than drivers results more pleasant sound.
Agree....for some reason I was still in 1:2 input transformer mood (1:2 stepup input transformer, tube amp.).
1 VRMS is the correct target for input sensitivity.
1 VRMS is the correct target for input sensitivity.
Last edited:
I am happy that we have reached the correct conclusion. Being equipped with this we may select the right driver, not a big deal. Now ccs load resistor load choke load fixed bias auto bias filament bias battery bias all are personal calls.
I like filament bias for signal dht and choke load for idht driver for output dht. By the way we can filament bias idht driver also, anybody tried this yet?
Regards
I like filament bias for signal dht and choke load for idht driver for output dht. By the way we can filament bias idht driver also, anybody tried this yet?
Regards