Hi all:
Thought you guys might get a chuckle out of my latest adventure.
I just finished a remote control linestage using 6DJ8s. All of it in a really nice cabinet donated by a friend (Thanks Scott
) and made from some beautiful old-growth Yellow Cedar - what else for the BC Coast, eh!
I am listening to it as I write this and it sounds awsome - very open and detailed with excellet extension in both extremes.
About 9db gain. The 6DJ8s are running at a B+ of 160v with only 5V on the heaters. Input impedance is 250k and output impedance is 22K. I think they call the configuration 'cascode' but I am not sure excatly what that means. One tube per side, with only 2 Rs and 2Cs in each channel. Can't get very much simpler than that.
The schematic is
This is part of a full preamp I found on the web. I'm not sure where it comes from or who designed it - my thanks to whoever it is.
Here is a photo of it - just finished (sorry for the blurries).
A few more images to follow...
Jess
Thought you guys might get a chuckle out of my latest adventure.
I just finished a remote control linestage using 6DJ8s. All of it in a really nice cabinet donated by a friend (Thanks Scott
) and made from some beautiful old-growth Yellow Cedar - what else for the BC Coast, eh!I am listening to it as I write this and it sounds awsome - very open and detailed with excellet extension in both extremes.
About 9db gain. The 6DJ8s are running at a B+ of 160v with only 5V on the heaters. Input impedance is 250k and output impedance is 22K. I think they call the configuration 'cascode' but I am not sure excatly what that means. One tube per side, with only 2 Rs and 2Cs in each channel. Can't get very much simpler than that
.The schematic is
An externally hosted image should be here but it was not working when we last tested it.
This is part of a full preamp I found on the web. I'm not sure where it comes from or who designed it - my thanks to whoever it is.
Here is a photo of it - just finished (sorry for the blurries).
An externally hosted image should be here but it was not working when we last tested it.
A few more images to follow...
Jess
Next Photo:
This shows the layout of the inside.
Dont get alarmed by the bright RED - the top and bottom plated are made from a recycled road sign
An externally hosted image should be here but it was not working when we last tested it.
This shows the layout of the inside.
Dont get alarmed by the bright RED - the top and bottom plated are made from a recycled road sign
Now, the remote IO and Vol control board - this is a kit from DIY Gene.
It works a treat and cost only $45 for the 4-input version. It comes with a nifty little remote and gives you 4 inputs plus tape monitor, volume and mute. I call that neat!
An externally hosted image should be here but it was not working when we last tested it.
It works a treat and cost only $45 for the 4-input version. It comes with a nifty little remote and gives you 4 inputs plus tape monitor, volume and mute. I call that neat!
Last one:
Here it is in my system. Sounds very, very good and all up cost about $90 (thanks again, Scott).
Happy Listening to all
Jess
An externally hosted image should be here but it was not working when we last tested it.
Here it is in my system. Sounds very, very good and all up cost about $90 (thanks again, Scott).
Happy Listening to all
Jess
Cool jesse,,, it isn't a cascode... a cascode has the cathode of the top tube connected directly to the anode of the bottom tube, and the top tube's grid is grounded... what you have is an SRPP.
http://www.tubecad.com/may2000/
This circuit has a cascode in a long tail pair configuration
dave
http://www.tubecad.com/may2000/
This circuit has a cascode in a long tail pair configuration
An externally hosted image should be here but it was not working when we last tested it.
dave
Just wanted to add (been listening to this box for about 4 hours now) that the bass of this linestage is astonishing!
I have a Velodyne ULD12 sub-woofer aand I have never heard it working like this before. I put up the AT series 40 mike into the PC=based RTA to measure the bass lines from Nora Jones "Come Away With Me" and the new Allison Krauss & Robert Pland vinyl - "Rising Sand" - hot damn - I am getting solid response down to 25hz. And, I can hear the notes and texture of the electric bass.
Now this is a suprise - a very welcome one.
onward and upward!
I have a Velodyne ULD12 sub-woofer aand I have never heard it working like this before. I put up the AT series 40 mike into the PC=based RTA to measure the bass lines from Nora Jones "Come Away With Me" and the new Allison Krauss & Robert Pland vinyl - "Rising Sand" - hot damn - I am getting solid response down to 25hz. And, I can hear the notes and texture of the electric bass.
Now this is a suprise - a very welcome one.
onward and upward!
Hi JesseG. If you're ever in the mood to tweak, consider replacing the 47uF/150 ohm pair with an LED. Red will give you very close to the 1.6 volts required, easy to measure and match with a DMM. Personally I like to bypass the LED with 0.1uF or thereabouts to insure the cathode stays near ground to as high a frequency as possible but not everyone considers it required.
Edit: Also, that 22kohm output resistor is way too small. Into the nonimal 56kohm standard input impedance of most audio equipment the 0.47uF output coupling cap will be driving a ~15k ohm load, or -3dB at ~22 Hz. Interestingly that's also RC rollover frequency of the cathode load, not considering the tube. Unless that value was chosen to load the tube for a reason I don't see, or filter subsonic turn-on transients, it's just a dc bleed. It might be worth trying 200k ohm.
I would try the LED first, look for subsonic turn-on transients, then play with the output bleed resistor.
Edit: Also, that 22kohm output resistor is way too small. Into the nonimal 56kohm standard input impedance of most audio equipment the 0.47uF output coupling cap will be driving a ~15k ohm load, or -3dB at ~22 Hz. Interestingly that's also RC rollover frequency of the cathode load, not considering the tube. Unless that value was chosen to load the tube for a reason I don't see, or filter subsonic turn-on transients, it's just a dc bleed. It might be worth trying 200k ohm.
I would try the LED first, look for subsonic turn-on transients, then play with the output bleed resistor.
Thanks, rdf - that sounds like fun.
Right now I don't want to take the linestage out of my system 'cause I am having too mich fun listening to it.
However...I decided to use a little 5V/2A switch-mode PS from an old wall-wart fro the heater supply. heck, it was there and the power trafo didn't have a heater supply and... You can see it hanging off the angled part of the barrier plate, right beside the IEC box.
It works fine and the 6DJ8s seem to like it but I was perplexed by a small low level ticking buzz in the output, until I realized this is the sonic signature of a switch-mode PS.
Point is, I will have to make up a small metal cover for this so that it will keep its noise to itself.
you can count on it, Don
Hi soundbrigade: I really like the RC kit. It took me a little head-scratching to figure out the relay board and especially the indicator lights but turned out very straightforward in the end. I am really impressed with how quickly it responds and the degree of control it gives over the volume level. The pot supplied with mine is an Alps and doesn't sound too bad at all.
Well, as stated it was a convenience move - I can always replace it with a 6.3 trafo (here we go digging thru the wall-wart box again).
Huh?? - way byond my miniscule understanding of electronics here
- could you explain?
Nah, you just haven't seen it before. 'Junk' gets from the heater to cathode two ways. At audio frequencies in the vast majority of circuits it's by a mechanism which can be greatly reduced by increasing the DC voltage between the heater and cathode beyond a threshold level. At very high frequencies the small capacitance between the cathode and heater dominates and biasing becomes ineffective.
From your schematic the B+ voltage is 176 VDC, the lower tube's cathode sits at 1.6 VDC and the upper tube's at 96 VDC. Assume the negative end of your heater supply is tied to ground, placing the heaters at 5 VDC above ground. In that case the DC voltage between the heater and lower tube cathode is 5-1.6=3.4 VDC. This cathode is 'DC biased' 3.4 VDC below the heater. Similarly the upper tube's cathode is biased +91 VDC above the heater. The upper's tubes cathode is already heavily DC biased relative to the heaters and coupled noise is very, very unlikely to be a factor. The lower tube's cathode sits very close to the same DC voltage as the filaments, increasing the likelihood of coupled noise. Unfortunately, in your circuit it's also the lower tube which does all the amplification, noise as well.
Most tube data sheets specify limits for these values. The 6DJ8/ECC88 sheets are written in a manner I find a bit confusing but it appears the cathode of one triode (pins 6, 7 and 8) can be a maximum of 50 VDC either side of the heater and the second triode (pins 1, 2, 3) can be 150 VDC max positive (130 VDC max + 20 VAC peak from the footnotes.) First thing to confirm is the top tube is using the pins 1, 2, 3 side and the lower 6, 7 and 8 as the schematic shows!
Now consider replacing your 5 VDC heater supply with a 6.3 VAC centre-tapped transformer with the centre tap tied to ground and the other ends tied to the heater (lift all heater grounds.) Heater pins 4 and 5 swing an AC voltage either side of the centre tap which is tied to ground, but since there's no DC on the filament transformer secondary (it sits at ground = 0 VDC) the DC voltages between cathodes and heater are now 1.6 and 96 VDC. You may even find driving the filaments balanced to ground this way is quiet enough. What DC biasing does however is reference the filament transformer centre tap to a DC voltage between the cathode voltages instead of to ground.
Doing that is the easy part. Tie a 350Kohm and a 100Kohm resistor in series. Tie the 350K to B+ and the 100K to ground. The junction between them - assuming your supply is near 176 VDC - will rest just beneath 40 VDC. Now lift the above hypothetical AC filament transformer centre tap from ground and tie it to the resistor junction. The filament transformer secondary, and therefore the filaments, are now 40 VDC above ground. The upper tube's cathode is 96-40=56 VDC above the heaters, the lower tube's cathode 40-1.6=38.4 VDC below the heater. Both are DC biased far enough either side of the heaters that coupled noise should be inconsequential. If your B+ power supply voltage is significantly off the 176 VDC specified above tweak the top resistor to suit.
The last step is adding a capacitor from the resistor junction to ground. My preference is for a small high quality cap (~0.1uF or less) to assure a very good path to ground at very high frequencies. Others prefer 100uF+. The value isn't critical. A lot of text for two resistors and a cap!
From your schematic the B+ voltage is 176 VDC, the lower tube's cathode sits at 1.6 VDC and the upper tube's at 96 VDC. Assume the negative end of your heater supply is tied to ground, placing the heaters at 5 VDC above ground. In that case the DC voltage between the heater and lower tube cathode is 5-1.6=3.4 VDC. This cathode is 'DC biased' 3.4 VDC below the heater. Similarly the upper tube's cathode is biased +91 VDC above the heater. The upper's tubes cathode is already heavily DC biased relative to the heaters and coupled noise is very, very unlikely to be a factor. The lower tube's cathode sits very close to the same DC voltage as the filaments, increasing the likelihood of coupled noise. Unfortunately, in your circuit it's also the lower tube which does all the amplification, noise as well.
Most tube data sheets specify limits for these values. The 6DJ8/ECC88 sheets are written in a manner I find a bit confusing but it appears the cathode of one triode (pins 6, 7 and 8) can be a maximum of 50 VDC either side of the heater and the second triode (pins 1, 2, 3) can be 150 VDC max positive (130 VDC max + 20 VAC peak from the footnotes.) First thing to confirm is the top tube is using the pins 1, 2, 3 side and the lower 6, 7 and 8 as the schematic shows!
Now consider replacing your 5 VDC heater supply with a 6.3 VAC centre-tapped transformer with the centre tap tied to ground and the other ends tied to the heater (lift all heater grounds.) Heater pins 4 and 5 swing an AC voltage either side of the centre tap which is tied to ground, but since there's no DC on the filament transformer secondary (it sits at ground = 0 VDC) the DC voltages between cathodes and heater are now 1.6 and 96 VDC. You may even find driving the filaments balanced to ground this way is quiet enough. What DC biasing does however is reference the filament transformer centre tap to a DC voltage between the cathode voltages instead of to ground.
Doing that is the easy part. Tie a 350Kohm and a 100Kohm resistor in series. Tie the 350K to B+ and the 100K to ground. The junction between them - assuming your supply is near 176 VDC - will rest just beneath 40 VDC. Now lift the above hypothetical AC filament transformer centre tap from ground and tie it to the resistor junction. The filament transformer secondary, and therefore the filaments, are now 40 VDC above ground. The upper tube's cathode is 96-40=56 VDC above the heaters, the lower tube's cathode 40-1.6=38.4 VDC below the heater. Both are DC biased far enough either side of the heaters that coupled noise should be inconsequential. If your B+ power supply voltage is significantly off the 176 VDC specified above tweak the top resistor to suit.
The last step is adding a capacitor from the resistor junction to ground. My preference is for a small high quality cap (~0.1uF or less) to assure a very good path to ground at very high frequencies. Others prefer 100uF+. The value isn't critical. A lot of text for two resistors and a cap!
JesseG said:
rdf said:
Jesse,
Looks great! How much capacitance do you have in the power supply?
I've been thinking about something very similar. I played around with TubeCAD and came up with this 6DJ8 SRPP Tube Preamp Schematic. Like rdf mentions, 22k is way too small. I was toying around with the idea of 1M. At 1M, you can get away with a small capacitor as you can see 0,22uF is already overkill.
Cheers
Well - I asked for an explanatin and I got one! Thanks so much rdf. I get the voltage divider to the heater trafo CT layout and I understand that this sets the heaters relative to the B+. I think I got it, but I am going to study it till I fully grasp how this pushes noise down. I have a phono pre built from Thorsten's design in the 'dc phono' thread (1x12AX7 and 1x6DJ8). It uses AC heaters and is not very noisy at all, so I am not against AC heaters at all.
As this linestage is now, the heater supply is not grounded at all. The SM-PS gets its AC frm the same point as the other trafos, but after that the 5V DC is paralleled to the two heaters, pins 4 & 5. This cct is not grounded to the chassis at all.
Yes, my linestage is wired the same as the schematic I posted - pin numbers are the same. You make a good point tho' - the two triodes have somewhat different ratings. It did not occurr to me what the significance of that was.
As soon as I find a 6.3V CT trafo I will try exctly what you have described. BTW: the 6DJ8 heaters eat 375ma each. That adds up to 1.5A total drain. If I bias the CT at 40V B+ as you suggest, how much current will flow through the voltage divider? i.e. what wattage Rs will I need?
One more question. I notice that the seperator screen (pin 9) is shown in many schematics as grounded. Mine aren't. Will grounding them have any sonic impact?
Hi gmilitano
80uF - the B+ PS is a dead-simple: bridge rectifier to CRC with 40uF / 10K 5w/ 40uF.
Interesting schematic. I notice that is indicated 27db gain. This is waaaay to much for me. I find that the 8.5 or 89db of mine is very adequate for a linestage. I am using it to drive a 300B SET amp and even with this gain it will overdrive the 300B input is turned up too far.
Your schematic, if followed by a buffer stage, would almost be enough for a MM phono stage, no?
Jess
rdf said:
Hi rdf, can you describe your preference of LED biasing versus a resistor bypassed by a good cap in subjective sonic terms? I am rebuilding my amp with CCS LTP and such, and maybe my paralleled 12AY7 common cathode stage before the LTP, can benefit from LEDs and I have tens of reds, greens and amber ones to match. One further question: Will a string of 3 LEDs in series that I need sound worse than using just one?
P.S. I wanna apply NFB there too, will I still need a small resistor under the LED string to apply on that node?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.