Thanks Michael. I gather you got as far as putting together an initail build for the 2SK60 that you were tweaking - do you still have the schematic or any other material - perhaps it would be worth resurrecting it? I'm no EE but there are some very helpful and knowledgable chaps here abouts.
Thanks for coming back Michael.
I'm assembling a MoFo and Ben Mah's 2SK180 choke-loaded follower;
https://www.diyaudio.com/forums/pas...-choke-loaded-2sk180-lamp-11.html#post6601437
(I plan to use them for driving the 'full-range' speakers in a Pass SLOB project, rolling them off at the bottom around 180Hz via tweaks to the input/output caps and with a smaller choke.) Ben Mah uses a Luminaria to drive his followers.
My sources are digital only, DSD decoders (ValveDAC and DSC2) which are a little lower than normal with their outputs at around 1V RMS - I've assumed I need something like 20dB of gain.
I also have some LU1014Ds coming via the GB facilitated by Papa's gift to the community so may have a look at X's LuFo down the line.
I am also looking for a LuminAria with 2SK60 or may be 2SJ18, especially with view on the significant lower input capacity. Choke loaded is wellcome. I guess, half the current will work with still low output impedance. Maybe no input buffer is neccessary. I am only a little hesitant because I am not realy familiar with seeking for the sweet spot.
Michael, just wondering if this post implied you might be resurrecting your 2SK60 Luminaria variant?
Thank you
Ray
Another LuminAria Comes to Life!
Hi Everyone,
Just finished building the LuminAria. I made a few changes to Mike's original:
1. Used a separate neg bias circuit. Similar to L'amp Part II.
2. Source connected to ground.
3. Used a CCS as the load instead of a resistor.
4. Using Salas SSHV for regulation.
Bias is about -18V and CCS set to 50 mA.
It sounds just a little edgy. Underneath the edginess, the sound is quite good. Especially the mids and bass, I'm hearing things I hadn't heard before. Soundstage is some of the best.
I'm driving it directly, i.e., without buffer, from an RME ADI-2 Pro, but wondering if a buffer is needed as Mike has suggested. I guess I could give it a shot and see if it makes a difference.
What have others found when driving the 2sk82? MOS follower buffer made a difference, or not?
Hi Everyone,
Just finished building the LuminAria. I made a few changes to Mike's original:
1. Used a separate neg bias circuit. Similar to L'amp Part II.
2. Source connected to ground.
3. Used a CCS as the load instead of a resistor.
4. Using Salas SSHV for regulation.
Bias is about -18V and CCS set to 50 mA.
It sounds just a little edgy. Underneath the edginess, the sound is quite good. Especially the mids and bass, I'm hearing things I hadn't heard before. Soundstage is some of the best.
I'm driving it directly, i.e., without buffer, from an RME ADI-2 Pro, but wondering if a buffer is needed as Mike has suggested. I guess I could give it a shot and see if it makes a difference.
What have others found when driving the 2sk82? MOS follower buffer made a difference, or not?
No edgy sound from my system with the Luminaria in it.
My Luminaria is also fixed bias with no source resistor, 660 Ohm resistor loaded, Iq=50mA, Vds=88V. I dropped the voltage to the 2SK82 to account for the removal of the source resistor. Power supply is MR Luminaria design. Input buffer is MR Luminaria design. The input buffer is hard wired in place with not option to bypass it. So I have no experience with no input buffer.
What is your Luminaria Vds? Perhaps you need to do some tweaking of the operating point, and do some harmonic distortion measurements. My Luminaria is second harmonic dominant, although it does not measure as low as MR's numbers.
RME ADI-2 Pro specifications show that its RCA output impedance is 100 Ohm so if you are using the RCA outputs, I think that you can get away with not having an input buffer. High source impedance generally reduces low frequency response.
The edginess may be due to something other than your Luminaria, and your Luminaria is exposing it?
My Luminaria is also fixed bias with no source resistor, 660 Ohm resistor loaded, Iq=50mA, Vds=88V. I dropped the voltage to the 2SK82 to account for the removal of the source resistor. Power supply is MR Luminaria design. Input buffer is MR Luminaria design. The input buffer is hard wired in place with not option to bypass it. So I have no experience with no input buffer.
What is your Luminaria Vds? Perhaps you need to do some tweaking of the operating point, and do some harmonic distortion measurements. My Luminaria is second harmonic dominant, although it does not measure as low as MR's numbers.
RME ADI-2 Pro specifications show that its RCA output impedance is 100 Ohm so if you are using the RCA outputs, I think that you can get away with not having an input buffer. High source impedance generally reduces low frequency response.
The edginess may be due to something other than your Luminaria, and your Luminaria is exposing it?
Thanks Ben! That gives some perspective. But I am thinking the RME does not have enough current to charge/discharge the gate to source capacitance of the VFET, which is pretty high I believe. This is why you need a buffer.
I did run a few measurements. Getting flat out to 20 kHz. I undersized the input cap because of what I had on hand, so it is about a db down at 20 Hz. I too am getting second dominant and a good bit of it, about -60 db, though I failed to check the Vout at that level. Will do so later.
No, it is not the rest of the system, it is the preamp. I have a good check in a 300B preamp. Well, it is the same as the 01/26/10Y pre, but it has 300B tubes in it right now.
One thing to check is the distortion against frequency. If there is problem driving the GS capacitance, I guess it would show up there as higher distortion in the HF. I am pretty sure it will because that's what I'm hearing.
I did run a few measurements. Getting flat out to 20 kHz. I undersized the input cap because of what I had on hand, so it is about a db down at 20 Hz. I too am getting second dominant and a good bit of it, about -60 db, though I failed to check the Vout at that level. Will do so later.
No, it is not the rest of the system, it is the preamp. I have a good check in a 300B preamp. Well, it is the same as the 01/26/10Y pre, but it has 300B tubes in it right now.
One thing to check is the distortion against frequency. If there is problem driving the GS capacitance, I guess it would show up there as higher distortion in the HF. I am pretty sure it will because that's what I'm hearing.
Yes, your low frequency response does need a boost.
Too much distortion is audible and can be grating to the ears. But I think it needs to be much higher than -60dB to be bad sounding.
Here is the my Luminaria's distortion 1.0Vp output (approximately -60dB):
Too much distortion is audible and can be grating to the ears. But I think it needs to be much higher than -60dB to be bad sounding.
Here is the my Luminaria's distortion 1.0Vp output (approximately -60dB):
Attachments
There is one big change I made. I swapped out the Salas shunt regs for a 15H choke. So, now the supply looks like a CRCLC and into the CCS load. My mind may be playing tricks, but this sounds a lot better.
Okay, on to the measurements. It is not really that far off on the low end. One db down at 20 Hz is fine with me. I knew I was going to get that with a small cap.
THD at 1 kHz looks similar to yours. Not a problem there, except the PS harmonics.
THD versus frequency at 1Vrms (higher output than 1Vpk) looks pretty decent also. Now, the problem might be above 20 kHz.
I'll do some more listening with the choke supply. And maybe try out the buffer too.
Okay, on to the measurements. It is not really that far off on the low end. One db down at 20 Hz is fine with me. I knew I was going to get that with a small cap.
THD at 1 kHz looks similar to yours. Not a problem there, except the PS harmonics.
THD versus frequency at 1Vrms (higher output than 1Vpk) looks pretty decent also. Now, the problem might be above 20 kHz.
I'll do some more listening with the choke supply. And maybe try out the buffer too.
Attachments
Last edited:
Yeah, the L is better but the edginess remains. I am not seeing anything in the measurements right now. I gotta try the buffer.
With a choke supply, the first MOSFET in the CCS takes a lot of heat. Earlier, the shunt reg dropped a good 50V. It needs that much anyway for good operation. But now, all of that is across the first FET in the CCS. It has a little clip-on heatsink, which is not going to do the job.
The operating point is pretty close to Mike's, with a bias of about -17V and 50mA through the CCS, resulting in about 75-80 volts on the drain.
With a choke supply, the first MOSFET in the CCS takes a lot of heat. Earlier, the shunt reg dropped a good 50V. It needs that much anyway for good operation. But now, all of that is across the first FET in the CCS. It has a little clip-on heatsink, which is not going to do the job.
The operating point is pretty close to Mike's, with a bias of about -17V and 50mA through the CCS, resulting in about 75-80 volts on the drain.