Very beautiful construction!
I like the use of Microcontroller or Arduino for control. Does the software use low-power modes (SLEEP etc) to reduce electromagnetic emissions?
I like the use of Microcontroller or Arduino for control. Does the software use low-power modes (SLEEP etc) to reduce electromagnetic emissions?
Thanks Rod!
The software permanently runs, to keep the current constant through the LDRs, but the digital and analog power supplies and circuits are completely separated, on different power transformer windings... And no relay coil is powered during normal use, only during calibration.
The software permanently runs, to keep the current constant through the LDRs, but the digital and analog power supplies and circuits are completely separated, on different power transformer windings... And no relay coil is powered during normal use, only during calibration.
Wow...lovely build! I see a GM70 in your profile pic. Do you have some pictures and or schematic for your gm70 amps?
Thanks Bas!
The pics of The GM-70 are here: http://www.diyaudio.com/forums/tubes-valves/71300-photo-gallery-642.html#post4117487 and the schematic here: http://www.diyaudio.com/forums/tubes-valves/71300-photo-gallery-647.html#post4167763
Very nice build Vincent !
I really like the idea of not using filament bias.
I'll also use a SSHV2 for my 4P1L preamp, so I can try it too, no need for hot smoking cathode resistors 🙂
Cool, Vincent.
I wonder what are those F1, F2 and FE1, FE2, FE3 elements in the schematic?
Br
Evgeny
Are those that Rod recommends?
28C0236-0BW-10 - LAIRD TECHNOLOGIES - FILTER, 6X10-15MM, 100MHZ, 835R | Farnell element14 Россия
28C0236-0BW-10 - LAIRD TECHNOLOGIES - FILTER, 6X10-15MM, 100MHZ, 835R | Farnell element14 Россия
The Laird beads are for high frequency filtering of power supplies.
For use in gates of FETs and triode grids (to prevent oscillation), you can use smaller beads - multi-turn beads are not needed here. Small size is important, instead.
For use in gates of FETs and triode grids (to prevent oscillation), you can use smaller beads - multi-turn beads are not needed here. Small size is important, instead.
Maybe something like this:
28L0138-10R-10 - LAIRD TECHNOLOGIES - FILTER, 3.5X4.5-59MM, 100MHZ, 75R | Farnell element14
It mostly needs to have 20+ ohms of impedance at 10MHz (Real part of complex number).
28L0138-10R-10 - LAIRD TECHNOLOGIES - FILTER, 3.5X4.5-59MM, 100MHZ, 75R | Farnell element14
It mostly needs to have 20+ ohms of impedance at 10MHz (Real part of complex number).
I agree with Bas, WOW. In fact, WOW, cubed.
Magnificent Vincent! Just magnificent.
You have created the ultimate expression of the 4P1L art.
I need to get back to work on mine with a steady gaze over your shoulder, not that I am capable of coming close to the beauty of your build but I have been inspired.
Magnificent Vincent! Just magnificent.
You have created the ultimate expression of the 4P1L art.
I need to get back to work on mine with a steady gaze over your shoulder, not that I am capable of coming close to the beauty of your build but I have been inspired.
I recently changed my 4P1L preamp output transformers to the LL1671/30mA (I was using the LL1692A/18mA) with the tubes running at 25mA (no change in bias current) and the improvement in sound quality was really many times more than I expected. I reasoned that a low plate impedance tube like the 4P1L could be loaded with a lower impedance transformer without loosing any bass response and the sound might benefit from the lower DC resistance and reduced amplitude of resonances over 25KHz. Boy, was I surprised with how much improvement!! I am not saying that the DC resistance drop and lower secondary capacitance was responsible for the sonic changes, just that they were the reason to try the LL1671/30mA. Having heard the results, any excuse would have justified the effort!
The rest of my line stage is choke loaded B+ with a common ss shunt regulator feeding both channels 155VDC and a choke loaded filament supply with 2 of Rod's V5 regulators. The bias is provided by a cathode resistor with filament current coursing through it.
The rest of my line stage is choke loaded B+ with a common ss shunt regulator feeding both channels 155VDC and a choke loaded filament supply with 2 of Rod's V5 regulators. The bias is provided by a cathode resistor with filament current coursing through it.
I'm curious, has anybody run any frequency response plots with RMAA? I built mine capacitor coupled, just wondering what i'm missing out on 🙂
Strogly agree with Kevin. I like to run my 4P1Ls very hot and LL1671/30mA is a great fit and sounds extremely good.
Hi Kevin
I know you are one of the most experienced with Lundahl iron, still, wouldn't the previous 18mA transformer saturate with the 25mA standing current?
I know you are one of the most experienced with Lundahl iron, still, wouldn't the previous 18mA transformer saturate with the 25mA standing current?
That's a good question that is asked from time-to-time about gapped transformers used in SE circuits. In short, it's a matter of how much deviation from the design bias current you are operating and how much signal level you can get before waveform clipping (due to saturation). Some time ago I asked Per Lundahl to succinctly describe the trade-off between bias current through the transformer primary winding and the maximum output level achievable based on the way Lundahl Transformer specifies their transformers. Here's the answer:
"In a transformer with a core air-gap, the magnetization curve is dominated by the airgap. The magnetization properties of the core itself can be ignored except at around 0 Tesla and at levels close to core saturation. In a typical iron-based C-core with a max magnetization density of around 2 T, the curve is basically a straight line at least between 0.2T and 1.6T. For this reason, we set the operating point for most SE transformers to 0.9T, half way between 0.2T and 1.6T
How important is it to run the transformer at precisely at the operating point, and what is the drawback if I don’t?
In most of our data sheets, in addition to operating point (mA) we present recommended maximum signal swing. If we have done our calculus right, the maximum swing (at 30Hz) should match the 1.6T point.
Example: You have an LL1689/18mA and you wish to run it at 25mA in configuration “Alt Q” (18:2). Data sheet indicates max output voltage at 30Hz is 28V rms.
At 25mA your operating point will be 0.9T * (25/18) = 1.25T
Your available signal swing at 30 Hz will be 28V * (1.6T – 1.25T) / (1.6T – 0.9T) = 14V
It is up to you to decide if this is enough signal swing."
I hope this provides a simple quantitative way to decide how far you can stray from the design current in your own design.
"In a transformer with a core air-gap, the magnetization curve is dominated by the airgap. The magnetization properties of the core itself can be ignored except at around 0 Tesla and at levels close to core saturation. In a typical iron-based C-core with a max magnetization density of around 2 T, the curve is basically a straight line at least between 0.2T and 1.6T. For this reason, we set the operating point for most SE transformers to 0.9T, half way between 0.2T and 1.6T
How important is it to run the transformer at precisely at the operating point, and what is the drawback if I don’t?
In most of our data sheets, in addition to operating point (mA) we present recommended maximum signal swing. If we have done our calculus right, the maximum swing (at 30Hz) should match the 1.6T point.
Example: You have an LL1689/18mA and you wish to run it at 25mA in configuration “Alt Q” (18:2). Data sheet indicates max output voltage at 30Hz is 28V rms.
At 25mA your operating point will be 0.9T * (25/18) = 1.25T
Your available signal swing at 30 Hz will be 28V * (1.6T – 1.25T) / (1.6T – 0.9T) = 14V
It is up to you to decide if this is enough signal swing."
I hope this provides a simple quantitative way to decide how far you can stray from the design current in your own design.
Hi Kevin,
many thanks for your reply and the explanation. On the RDC: I would also think that lower RDC would improve the performance. At the moment I am contacting some winders for "limited range" output transformers, to be used in my tri-amplified system: the midrange will play from 400Hz and up, so my reasoning is that I don't need all the inductance of fullrange transformers, and could go with an OPT with less inductance and consequently less RDC.
But this is going very slowly 🙂
Best regards, Erik
many thanks for your reply and the explanation. On the RDC: I would also think that lower RDC would improve the performance. At the moment I am contacting some winders for "limited range" output transformers, to be used in my tri-amplified system: the midrange will play from 400Hz and up, so my reasoning is that I don't need all the inductance of fullrange transformers, and could go with an OPT with less inductance and consequently less RDC.
But this is going very slowly 🙂
Best regards, Erik
Hello,
I took some more measurements of my 4P1L preamp and I have to report that the way I used the SSHV2 to bias the tubes is not optimal.
Indeed, the impedance of the shunt regulator becomes insufficient at high frequencies and it starts injecting signal into the cathode.
So, filament bias is still the best way to do it.
It sounded very good like I did it, though...
I took the schematic offline, it contained a very embarassing error. 🙁
Also, the LL1689 is not recommended: it rings a lot when only loaded with the high input impedance of the following amplifier. When I load it down with a 330R resistor, it behaves well, but I lose the gain I need because of the highish output impedance.
The LL1689 is a medium/high impedance transformer designed more for tubes like the 6SN7GT, rather than low impedance tubes like the triode-connected 4P1L. One could mostly damp the resonances with a series RC load across the secondary without losing the gain that results from a low resistance load, but a better solution would be a lower impedance transformer, like the LL1671 that I discussed in this thread several days ago.