Hi folks,
I have an odd one here for you - SRSE.
On first glance, it may look like a Morgan Jones beta follower, but it's a relative of the John Broskie SRSE.
The history on how this thing came about was to design something with the Octal PCB's with a pair of CCS pads on them (same configuration as the diyAudio CCS) that used 6SN7GT's. It was suggested I try this circuit from Patrick Turner. What I liked about this was the local NFB on the first stage to control the gain. So, in mad scientist tradition, I blended the two.
Here is the circuit and power supply:
http://bb.sicomm.us/diyAudio/6SN7_SRSE_linestage/6SN7GT_Linestage.gif
No tube rectifier or delay needed - a totem pole circuit won't "strip", since no path is made when cold to do such.
The SRSE has two impressive advantages over the SRPP in the same configuration - lower output impedance (less than half) and the output is isolated from shifting the gain of the bottom tube. The output impedance of this circuit is 680 ohms and F3 response is 2Hz to 130KHz. No, not a typo... 2Hz into the 22K static load resistors. I'm actually thingking of lowering the value of the coupling caps to the top tube to up that to 15Hz or so (those that were at the fest in Victoria will remember the interesting effect of too low an F3 when someone playing a LP walks across the deck
)
With a 2K load, this is 11Hz to 100KHz.
Voltage gain is 5 from input to output - about the same as a 12B4.
Sound is quite transparent. There is an increase in detail, especially on the top end, but this could be more due to the low output impedance driving my el-cheapo interconnects.
Pics of the completed unit can be seen here. I'm rather proud of how it turned out
Cheers!
I have an odd one here for you - SRSE.
On first glance, it may look like a Morgan Jones beta follower, but it's a relative of the John Broskie SRSE.
The history on how this thing came about was to design something with the Octal PCB's with a pair of CCS pads on them (same configuration as the diyAudio CCS) that used 6SN7GT's. It was suggested I try this circuit from Patrick Turner. What I liked about this was the local NFB on the first stage to control the gain. So, in mad scientist tradition, I blended the two.
Here is the circuit and power supply:
http://bb.sicomm.us/diyAudio/6SN7_SRSE_linestage/6SN7GT_Linestage.gif
No tube rectifier or delay needed - a totem pole circuit won't "strip", since no path is made when cold to do such.
The SRSE has two impressive advantages over the SRPP in the same configuration - lower output impedance (less than half) and the output is isolated from shifting the gain of the bottom tube. The output impedance of this circuit is 680 ohms and F3 response is 2Hz to 130KHz. No, not a typo... 2Hz into the 22K static load resistors. I'm actually thingking of lowering the value of the coupling caps to the top tube to up that to 15Hz or so (those that were at the fest in Victoria will remember the interesting effect of too low an F3 when someone playing a LP walks across the deck
)With a 2K load, this is 11Hz to 100KHz.
Voltage gain is 5 from input to output - about the same as a 12B4.
Sound is quite transparent. There is an increase in detail, especially on the top end, but this could be more due to the low output impedance driving my el-cheapo interconnects.
Pics of the completed unit can be seen here. I'm rather proud of how it turned out
Cheers!
Pretty cool Geek,
A friend and I messed around with this concept about 7 - 8 years ago and ended up with a topology that was eerily similar to yours, and per your observation it performed extremely well. My intent in experimenting with this topology was to evaluate the difference between a CCS in this location and a choke. We considered both approaches to be a variation or improvement on the standard mu follower as their operation is somewhat analogous.
I use a choke as the upper resistor in the classic srpp configuration which turns out to make a pretty darn good mu follower provided the H is high enough and the choke's r is appropriate for biasing or can be padded in series.
I use >100H with 6SN7 and get excellent results, eliminating two resistors and a cap in so doing. Larger chokes will extend the response downwards, IIRC with the choke the response is around -1dB at 25 - 30Hz or so. (A little HPF action going on there.)
I ended up with chokes because I preferred the simplicity of that approach, although the sensitivity to external magnetic fields and the need for substantial inductance makes this the less technically and economically attractive choice. (Maybe I have an irrational dislike of sand based devices.. LOL)
Oh, and I didn't think your homebrew chassis was all that ugly either.. (In reference to comments on that thread.) Mine doesn't look any better.
A friend and I messed around with this concept about 7 - 8 years ago and ended up with a topology that was eerily similar to yours, and per your observation it performed extremely well. My intent in experimenting with this topology was to evaluate the difference between a CCS in this location and a choke. We considered both approaches to be a variation or improvement on the standard mu follower as their operation is somewhat analogous.
I use a choke as the upper resistor in the classic srpp configuration which turns out to make a pretty darn good mu follower provided the H is high enough and the choke's r is appropriate for biasing or can be padded in series.
I use >100H with 6SN7 and get excellent results, eliminating two resistors and a cap in so doing. Larger chokes will extend the response downwards, IIRC with the choke the response is around -1dB at 25 - 30Hz or so. (A little HPF action going on there.)
I ended up with chokes because I preferred the simplicity of that approach, although the sensitivity to external magnetic fields and the need for substantial inductance makes this the less technically and economically attractive choice. (Maybe I have an irrational dislike of sand based devices.. LOL)
Oh, and I didn't think your homebrew chassis was all that ugly either.. (In reference to comments on that thread.) Mine doesn't look any better.
Hi Kevin,
Thanks!
LOL! Good ol' Choky (ZM, whomever now )
Hi SY,
I haven't had it chit-chat with the computer yet (I appear to have misplaced my box of "prized" NOS 6SN7's at the moment). On the 'scope, there was vaery little visibility as to the compression level (which began around 30V P-P out [SRSE does give less P-P]).
Cheers!
Thanks!
LOL! Good ol' Choky (ZM, whomever now
) Hi SY,
SY said:
I haven't had it chit-chat with the computer yet (I appear to have misplaced my box of "prized" NOS 6SN7's at the moment). On the 'scope, there was vaery little visibility as to the compression level (which began around 30V P-P out [SRSE does give less P-P]).
Cheers!
In the testing, I found an interesting bug. The lower (8.2K) input resistor causes this thing to be *extremely* susceptible to low impedance sources. I have revised this to fix that:
http://bb.sicomm.us/diyAudio/6SN7_SRSE_linestage/6SN7GT_Linestage_v1.gif
There's a cost though - reducing the NFB to give the same gain, the upper F3 is now only 33KHz. Square wave response is still similar, allowing for 5-7 harmonics.
If you drive the pre from Zout something ~> 5K, you'll not need change anything.
FFT measurements, new configurarion for THD (GE gray plate 6SN7GTA top tube, Silvertone 45 degree smooth black plate bottom) :
Av=1: 0.04%
Av=5: 0.07%
FFT measurements, new configurarion for THD (GE gray plate 6SN7GTA top tube, Westinghouse "tall bottle" black plate bottom) :
Av=1: 0.07%
Av=5: 0.16%
Not all 6SN7's are created equally
Yes. 1V (all my soundcard can take before THD goes through the roof)
That one is perfectly linear to level.
The new schema will give a -84dB 5th. (only visible harmonics are 2,3 and 5)
This forum still doesn't have a "flipping" smiley???
http://bb.sicomm.us/diyAudio/6SN7_SRSE_linestage/6SN7GT_Linestage_v1.gif
There's a cost though - reducing the NFB to give the same gain, the upper F3 is now only 33KHz. Square wave response is still similar, allowing for 5-7 harmonics.
If you drive the pre from Zout something ~> 5K, you'll not need change anything.
FFT measurements, new configurarion for THD (GE gray plate 6SN7GTA top tube, Silvertone 45 degree smooth black plate bottom) :
Av=1: 0.04%
Av=5: 0.07%
FFT measurements, new configurarion for THD (GE gray plate 6SN7GTA top tube, Westinghouse "tall bottle" black plate bottom) :
Av=1: 0.07%
Av=5: 0.16%
Not all 6SN7's are created equally
Yes. 1V (all my soundcard can take before THD goes through the roof)
That one is perfectly linear to level.
The new schema will give a -84dB 5th. (only visible harmonics are 2,3 and 5)
You do understand what motivated these questions, yes?
This forum still doesn't have a "flipping" smiley???
An externally hosted image should be here but it was not working when we last tested it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.