I’m building a small single ended zero feedback amp based on 6L6GC tubes.
The goal is to achieve as low noise and hum as possible - after experience with commerical kit using 6L6GC where some noise is still audible (with my 98 dB horns) - 6L6 SE Tube Amplifier Kit 10+10W (Stereo)_Power Amplifier Kit_Tube Amplifier Kit_Analog Metric - DIY Audio Kit Developer
Obviousky I will pay a lot of attention to ground loops, placament of transformers etc.
However one key decision is which circuit to build to achieve minimum noise too.
In general I found two different concepts for the driver stage:
1. simple with a single triode - eg:The 6V6 Marblewood Amp | Cascade Tubes or DIY Single-Ended (SE) KT88 / 6L6 / EL34 / 6CA7 Tube Amplifier
2. more comples with “SPRR” and two input tubes -eg: DIY Tube Amplifier Kit 6L6 + 6N8P Single Ended Tube Power Amplifier Kit | eBay or 6L6 SE Tube Amplifier Kit 10+10W (Stereo)_Power Amplifier Kit_Tube Amplifier Kit_Analog Metric - DIY Audio Kit Developer
What is the benefit of going with more complex driver stage (2)? Linearity, frequency range? Does it also help with noise?
If I go with the simple driver stage using a single triode (1) is there an advantage if using a selarate tube for each channnel? 2x 12AX7 while only 1/2 will be used for each of R and L channels? Will it help with crosstalks?
Thanks
The goal is to achieve as low noise and hum as possible - after experience with commerical kit using 6L6GC where some noise is still audible (with my 98 dB horns) - 6L6 SE Tube Amplifier Kit 10+10W (Stereo)_Power Amplifier Kit_Tube Amplifier Kit_Analog Metric - DIY Audio Kit Developer
Obviousky I will pay a lot of attention to ground loops, placament of transformers etc.
However one key decision is which circuit to build to achieve minimum noise too.
In general I found two different concepts for the driver stage:
1. simple with a single triode - eg:The 6V6 Marblewood Amp | Cascade Tubes or DIY Single-Ended (SE) KT88 / 6L6 / EL34 / 6CA7 Tube Amplifier
2. more comples with “SPRR” and two input tubes -eg: DIY Tube Amplifier Kit 6L6 + 6N8P Single Ended Tube Power Amplifier Kit | eBay or 6L6 SE Tube Amplifier Kit 10+10W (Stereo)_Power Amplifier Kit_Tube Amplifier Kit_Analog Metric - DIY Audio Kit Developer
What is the benefit of going with more complex driver stage (2)? Linearity, frequency range? Does it also help with noise?
If I go with the simple driver stage using a single triode (1) is there an advantage if using a selarate tube for each channnel? 2x 12AX7 while only 1/2 will be used for each of R and L channels? Will it help with crosstalks?
Thanks
If cross talk is a concern, a 6CG7 or 6DJ8 has a shield between sections.
Noise for what you're planning to do will come to down good layout practices and adequate power supply filtering.
The 6N8P/6L6 amp you linked to has a cascode driver with both a local and global feedback loop. This would be a choice that you'd make if you wanted to maximize power out of the amp. The feedback loops should also make the amp quieter. My suspicion is that you don't care about maximizing power, so you're more likely to go with triode strapped 6L6s instead (consider the 6AV5 while you're at it).
Noise for what you're planning to do will come to down good layout practices and adequate power supply filtering.
The 6N8P/6L6 amp you linked to has a cascode driver with both a local and global feedback loop. This would be a choice that you'd make if you wanted to maximize power out of the amp. The feedback loops should also make the amp quieter. My suspicion is that you don't care about maximizing power, so you're more likely to go with triode strapped 6L6s instead (consider the 6AV5 while you're at it).
The amp you built has a high gain 12AX7 in SRPP, and a 6L6GC in either Beam Power mode, or in Ultra Linear mode.
There is lots of open loop gain there, which will amplify any noise of the input tube, and any ground loop noise at the input RCA jacks.
With closed loop gain, some of those problems still exist.
Open loop gain also amplifies any noise from ground loops related to: the power mains of the signal sources and the amp; the power supplies of the signal sources, cabling from the signal sources to the amp, and the input circuits of the amp.
The input circuit also amplifies any ground loops of the amp B+ power supply as they interact with the input circuit ground loops.
There may be magnetic coupling from power transformer and choke to the output transformers.
Steel chassis, not enough spacing of the output transformers to the power transformer and choke can cause hum. Bad rotational orientation of the power transformers to the power transformer and choke can also cause hum.
Negative feedback sometimes reduces noise and hum, but mostly it reduces gain, so the volume control has to be turned up (which in some cases brings noise and hum up again).
How close are you to your 98dB sensitive speakers?
Do you need 2 Watts, or 10 Watts?
What kind of hum and noise does your amplifier have . . . power mains frequency, 2X power mains frequency, mid frequency noise and grunge, high frequency noise and grunge?
Unfortunately, I do not have a software plug-in to allow me to see the schematic of your amplifier on the link you put up.
I have never had left and right channel separation problems with 9 pin dual triode tubes.
The one time I had a separation problem, it was -40dB from channel to channel (and it was across the complete audio frequency range).
At -40 dB, it sounded real good (an excellent phono cartridge has -30dB at mid frequencies, and only if installed extremely carefully).
The amp chassis was small, but had a power transformer, choke, and 2 output transformers. The cause was only having one filament winding and two type DHT 45 tubes on that single winding. The signal currents of both tubes went through 25 Ohm hum balance resistors at the ends of the filaments, and the center of those resistors went to the self bias resistor and its bypass cap. Basically, the signal current was effectively driving 12.5 Ohms, the resultant voltage became signals from/to both channels.
There is lots of open loop gain there, which will amplify any noise of the input tube, and any ground loop noise at the input RCA jacks.
With closed loop gain, some of those problems still exist.
Open loop gain also amplifies any noise from ground loops related to: the power mains of the signal sources and the amp; the power supplies of the signal sources, cabling from the signal sources to the amp, and the input circuits of the amp.
The input circuit also amplifies any ground loops of the amp B+ power supply as they interact with the input circuit ground loops.
There may be magnetic coupling from power transformer and choke to the output transformers.
Steel chassis, not enough spacing of the output transformers to the power transformer and choke can cause hum. Bad rotational orientation of the power transformers to the power transformer and choke can also cause hum.
Negative feedback sometimes reduces noise and hum, but mostly it reduces gain, so the volume control has to be turned up (which in some cases brings noise and hum up again).
How close are you to your 98dB sensitive speakers?
Do you need 2 Watts, or 10 Watts?
What kind of hum and noise does your amplifier have . . . power mains frequency, 2X power mains frequency, mid frequency noise and grunge, high frequency noise and grunge?
Unfortunately, I do not have a software plug-in to allow me to see the schematic of your amplifier on the link you put up.
I have never had left and right channel separation problems with 9 pin dual triode tubes.
The one time I had a separation problem, it was -40dB from channel to channel (and it was across the complete audio frequency range).
At -40 dB, it sounded real good (an excellent phono cartridge has -30dB at mid frequencies, and only if installed extremely carefully).
The amp chassis was small, but had a power transformer, choke, and 2 output transformers. The cause was only having one filament winding and two type DHT 45 tubes on that single winding. The signal currents of both tubes went through 25 Ohm hum balance resistors at the ends of the filaments, and the center of those resistors went to the self bias resistor and its bypass cap. Basically, the signal current was effectively driving 12.5 Ohms, the resultant voltage became signals from/to both channels.
Last edited:
I’m close to my speaker and so no need for a lot of power. My original issue was a hum (150 Hz) which I fixed by adjusting the position of transformers. What I can hear now is a high frequency noise (like a white noise) and also some buzz (300 Hz) from one speaker only - my SET amp is without a global feedback now.
Reading your text it seems that I don’t need that high gain using SRPP (12AX7). Is there any other benefit other than higher gain while using SRPP? Otherwise would it be better to use simpler single triode for input stage only? (No SRPP)
Reading your text it seems that I don’t need that high gain using SRPP (12AX7). Is there any other benefit other than higher gain while using SRPP? Otherwise would it be better to use simpler single triode for input stage only? (No SRPP)
A while back I was playing around with a triode wired 7C5 (loctal 6V6GT) SE amp using 6SJ7 as the input stage. It had plenty of drive and loudness with my 89dB speakers and was one of the best amps I’ve heard. I imagine it would work nicely with 6L6 too, which I’m about to try and will post back.
I once made a SE 6L6GC amplifier with James OPT and a single 12AX7 driver. Schematic here: Attempt at giving back: 6L6GC SE amplifier . However I've have started to like the 5670 / 2C51 / 6N3P-E / 6N3P-EV (last two of only Russian heritage) instead. More linear, robust, has a shield between both sections and both sections are supposed to be identical. Also can be biased with more current. A lesser known fact is that one section of the 12AX7 has a higher noise floor than the other.
Last edited:
So if I understand it right the main benefit of SRPP is higher gain (to be used with negative feedback) and maybe linearity?
I assume using dual triodes in the driver stage (12AX7/12AU7/6SN7) what is the best approach:
- 1 tube with 1/2 of it for each L / R channel
- 2 tubes with 1/2 of each for each L / R channel (2 triodes unused)
- 2 tubes while both triodes will be used for each channel (eg is SRPP, dual stage, parallel, … or any other?) comfiguarion?
I don't need a lot of power (98dB horns with near field listening position) but I want a "dead quiet" amp (minimum noise) and I possible an option to with / without global negative feedback. Output tubes are RCA 6L6 GC in triode mode.
I assume using dual triodes in the driver stage (12AX7/12AU7/6SN7) what is the best approach:
- 1 tube with 1/2 of it for each L / R channel
- 2 tubes with 1/2 of each for each L / R channel (2 triodes unused)
- 2 tubes while both triodes will be used for each channel (eg is SRPP, dual stage, parallel, … or any other?) comfiguarion?
I don't need a lot of power (98dB horns with near field listening position) but I want a "dead quiet" amp (minimum noise) and I possible an option to with / without global negative feedback. Output tubes are RCA 6L6 GC in triode mode.
Input stage:
Take a common cathode stage (signal input between grid and ground) with a self bias resistor and a bypass cap. The cathode is held to AC ground.
Any filament to cathode leakage (like from the AC filament power), is shunted to ground by the bypass cap.
Take an SRPP stage. The top cathode is not bypassed to ground. Any filament to cathode leakage will appear as a signal. Hmmm, . . . possibly.
If you want similar gain and linearity of an SRPP, then an alternative is to use a triode and an IXYS current source as the plate load (I prefer the 900V rated part instead of the 450V rated part).
Any hum and noise in the input stage will appear all the way to the output.
Output Stage:
Single ended?
6L6 triode wired mode. Yes, seems like a good idea because you are not using negative feedback.
Using a larger plate primary impedance will lower the gain (any residual hum in the driver will be amplified less too).
If you already have the output transformers, then just connect the loudspeaker to the next lower tap (16 Ohm speaker on 8 Ohm tap, 8 Ohm speaker on 4 Ohm tap). That will lower any hum, but will lower the output power too.
If there is not enough inductance in the output transformer, it will also reduce the amplitude of the extreme low frequency bass.
You will only get about 2 to 3 watts with the 6L6GC wired in triode mode.
It probably will be enough for your sensitive speakers, especially since you are close to them.
You need to decide if you want a single ended amp, or a push pull amp.
I will describe a push pull amp that I most recently designed and built in my next post (if I can remember to do that).
Take a common cathode stage (signal input between grid and ground) with a self bias resistor and a bypass cap. The cathode is held to AC ground.
Any filament to cathode leakage (like from the AC filament power), is shunted to ground by the bypass cap.
Take an SRPP stage. The top cathode is not bypassed to ground. Any filament to cathode leakage will appear as a signal. Hmmm, . . . possibly.
If you want similar gain and linearity of an SRPP, then an alternative is to use a triode and an IXYS current source as the plate load (I prefer the 900V rated part instead of the 450V rated part).
Any hum and noise in the input stage will appear all the way to the output.
Output Stage:
Single ended?
6L6 triode wired mode. Yes, seems like a good idea because you are not using negative feedback.
Using a larger plate primary impedance will lower the gain (any residual hum in the driver will be amplified less too).
If you already have the output transformers, then just connect the loudspeaker to the next lower tap (16 Ohm speaker on 8 Ohm tap, 8 Ohm speaker on 4 Ohm tap). That will lower any hum, but will lower the output power too.
If there is not enough inductance in the output transformer, it will also reduce the amplitude of the extreme low frequency bass.
You will only get about 2 to 3 watts with the 6L6GC wired in triode mode.
It probably will be enough for your sensitive speakers, especially since you are close to them.
You need to decide if you want a single ended amp, or a push pull amp.
I will describe a push pull amp that I most recently designed and built in my next post (if I can remember to do that).
Last edited:
Information from the data sheets is that a 6L6 in triode mode with 250 volt plate supply biased at -20 volts produces a 1.4 watt output. A 6V6 in the same circuit biased at -15 volts produces 1.0 watt output. Higher output with higher B+ can be had from both tubes.
Tubelab's comments about the 6V6 in triode compare it favorably to a type 45 tube.
I have a bi-amp system with an ESS AMT in my workshop that I listen to from less than 10 feet away. I have measured the response of my AMT and they produce 98db with one watt input.
I used a 6SN7 to triode wired 6V6 and they sounded excellent and would drive the AMT louder than I would listen to. I got a too good to pass up deal on a 300B amp and now that drives the AMTs.
You might consider trying the 6V6. Easier drive requirement and potentially better sound quality.
Steve
Tubelab's comments about the 6V6 in triode compare it favorably to a type 45 tube.
I have a bi-amp system with an ESS AMT in my workshop that I listen to from less than 10 feet away. I have measured the response of my AMT and they produce 98db with one watt input.
I used a 6SN7 to triode wired 6V6 and they sounded excellent and would drive the AMT louder than I would listen to. I got a too good to pass up deal on a 300B amp and now that drives the AMTs.
You might consider trying the 6V6. Easier drive requirement and potentially better sound quality.
Steve
+1
My first impression of the 6SJ7 into triode-wired 6V6 was that it sounded better than my 300B SETs.
I want to see if 6V6 can sound good as a single-ended pentode. Would certainly require feedback.
My first impression of the 6SJ7 into triode-wired 6V6 was that it sounded better than my 300B SETs.
I want to see if 6V6 can sound good as a single-ended pentode. Would certainly require feedback.
Before I switched to triode strapped I used a 6V6 with local (schade style) feedback.
I got 4 watts output and flat response to 45khz. But, I never listened to them at full power because that was too loud for me. After reading Tubelab's article about them I switched to triode strapped.
I really couldn't tell the difference between the triode version and the local feedback version. They both tested very well on the bench.
YMMV. I am 67. I had my hearing tested last year and I don't hear beyond 14khz unless it is at a really high SPL.
I got 4 watts output and flat response to 45khz. But, I never listened to them at full power because that was too loud for me. After reading Tubelab's article about them I switched to triode strapped.
I really couldn't tell the difference between the triode version and the local feedback version. They both tested very well on the bench.
YMMV. I am 67. I had my hearing tested last year and I don't hear beyond 14khz unless it is at a really high SPL.