Greetings, friends. I've been quite happy with my Lacewood 2.0 build for some time. I built it just as designed over at the Cascadetubes.com page and put it in a nice Hammond chassis. Pics and build log here:
Lately, I've gotten curious about that SET sound, and since I have most of the parts on hand I want to look into re-building this into a 6B4G amp driven by a 6SL7. This would be my first attempt at a DHT build. While I do realize most 2A3 output tubes are driven by more than a single gain stage, I'm reluctant to modify this chassis, so I'm stuck with 3 octal sockets, and 6SL7 is the most appropriate octal-base tube I know of. I guess I'd need a preamplifier to get full power from this amp. I have a pair of Edcor 3.5K 10w OPTs I'd like to use - not the 25W that I should be using, but bought and paid for and in need of a good home.
I'd be using the PSU that Matt designed for the LW2.0:
Removing the 350R resistor and switching to a 5AR4 may add as much as 90v to b+ but I still won't have enough volts to do a directly-coupled driver, so I'll need use coupling caps?
There's an old kit amp called Get*Set*Go that seems to have everything I need, even rectifiers for the heaters:
I believe the PT's 6.3v@5A can heat everything in DC with power to spare, and at 60mA per channel I should have enough b+ power as well.
But if I'm reading that right they are biasing the cathodes at 2.7k and that seems high, I think I want to use a 750R cathode resistor on the 2A3s, assuming I can get 300v to the plates. Hum Pots? I'm not committed to this design, it just has all the parts I need, I'd appreciate any advice.
thanks!
Lately, I've gotten curious about that SET sound, and since I have most of the parts on hand I want to look into re-building this into a 6B4G amp driven by a 6SL7. This would be my first attempt at a DHT build. While I do realize most 2A3 output tubes are driven by more than a single gain stage, I'm reluctant to modify this chassis, so I'm stuck with 3 octal sockets, and 6SL7 is the most appropriate octal-base tube I know of. I guess I'd need a preamplifier to get full power from this amp. I have a pair of Edcor 3.5K 10w OPTs I'd like to use - not the 25W that I should be using, but bought and paid for and in need of a good home.
I'd be using the PSU that Matt designed for the LW2.0:
Removing the 350R resistor and switching to a 5AR4 may add as much as 90v to b+ but I still won't have enough volts to do a directly-coupled driver, so I'll need use coupling caps?
There's an old kit amp called Get*Set*Go that seems to have everything I need, even rectifiers for the heaters:
I believe the PT's 6.3v@5A can heat everything in DC with power to spare, and at 60mA per channel I should have enough b+ power as well.
But if I'm reading that right they are biasing the cathodes at 2.7k and that seems high, I think I want to use a 750R cathode resistor on the 2A3s, assuming I can get 300v to the plates. Hum Pots? I'm not committed to this design, it just has all the parts I need, I'd appreciate any advice.
thanks!
I have a pcb but have not built it (yet). I believe GSG has a solid reputation. If you have the bits why not try it.
R15/16 a, b & c are three 2.7k in parallel, i.e. 900 Ohms. Hum pots - this is what R11, 13 and R12, 14 try to do. I just noticed R11, 13 are different values, not sure why; probably empirically determined.
R15/16 a, b & c are three 2.7k in parallel, i.e. 900 Ohms. Hum pots - this is what R11, 13 and R12, 14 try to do. I just noticed R11, 13 are different values, not sure why; probably empirically determined.
Last edited:
heh oh yea 900R. I guess R11 & R13 mimic a 100R pot at center and with DC heating we don't need a pot?
When dealing with DHT, ex 6B4G or 6C4C don’t forget that you need one 6.3V winding for each tube, they must be independent.
Eric
Eric
I believe so. The 6.3Vdc is simply to keep the amp quiet. If using a full bridge rectifier to get the DC remember that the DC current will only be 0.62 x the AC. Ex a 6.3Vac, 5A rectified with provide a max of 5A x 0.62 therefore about 3A.
Last edited:
More an independent bias issue of the output tubes. Tubelab uses a single regulated heater supply for both channel’s output tubes on the TSE design, but that is control grid biased. The RSG originally had the Dynaco PA774 transformer specified, which has two independent 6.3 windings.
What you need for the 6B4G or 6C4C is a pair of Rod Coleman V9 regs, which you can get from Lyrima in the UK. The best filament regs you can find. Very popular with the DHT crowd. I've been using them for years. Very good value and all information comes with the kits. The regs need 3.5V headroom so you need to feed them with 10V. That looks like a 9+9V transformer, at least 30VA or 50VA better. Schottky diodes and step-down resistor, good size wire-wound. If there's no room in the chassis you can always take it outboard. The regs need to be mounted on a decent size heatsink as well which would go on the back or on the top. That's exactly my own setup.
You should have a very nice amp! If you don't need too much gain you could use a 6SN7 input, or a 6N1P. Both good.
You should have a very nice amp! If you don't need too much gain you could use a 6SN7 input, or a 6N1P. Both good.
Last edited:
Dubadub,
Nice idea! 6B4G
1. The Output tubes have separate self bias networks.
Unless you use Super Well Matched output tubes, you need two (2) separate DC filament supplies, one for each output tube.
Not much chance to find such well matched NOS or used output tubes of that vintage.
Do use separate DC filament supplies. What if one tube dies, now find an exact match.
Separate self bias circuit for each tube works.
You only have to find a second filament transformer, if you power transformer does not have a second one.
And one can power both one output tube, and one driver tube (the driver tube will not mind if its filament supply is elevated to the self bias voltage (+45 to +60V).
2. The 47 Ohm and 51 Ohm filament resistors are there to make the voltage at the ends about 6.3V/2 = 3.15V.
That makes one end of the filament 3.15v higher than the self bias voltage; and the other end of the filament be 3.15V lower than the self bias voltage.
The reason is the negative end of the DC powered filament draws More plate current than the positive end of the DC powered filament.
There is more total current in the 47 Ohm resistor; versus the total current in the 51 Ohm resistor.
Empirical? . . . Not really.
All of that is to place the middle length of the filament (only accessible inside the tube) at exactly the self bias voltage.
If you swap the 47 and 51 Ohm resistors, versus the + and - 6.3VDC supply, you will get unbalanced voltages versus the self bias voltage.
3. Be sure to use Schottky diodes in the DC filament supplies.
The more voltage from the diodes, the larger the resistor resistance can be between the two (2) 4,700uF caps can be. That makes for lower ripple on the filament DC supplies.
When determining the final resistance there, be sure to wait for at least a minute when measuring the 6.3V across the filaments.
That is because the DC filament supply circuit has Intrinsic slow warm up time.
When I made my 6A3 (1A DC) and 300B (1.25A DC) supplies, I used Schottky diodes, and the two caps were 22,000uF each . . . I got 1mV ripple.
The only "heat sink" was the Schottky bridge was mounted directly on the chassis.
4. 3.5k output transformers, slightly lower than the textbook 2.5k, but . . .
Event thought 3.5k has slightly less power output, The advantage is: Lower distortion and Higher damping factor.
Have fun building and listening!
P.S. I have a question, was the original amplifier using 6V6, 7591, or other output tubes?
Thanks!
Nice idea! 6B4G
1. The Output tubes have separate self bias networks.
Unless you use Super Well Matched output tubes, you need two (2) separate DC filament supplies, one for each output tube.
Not much chance to find such well matched NOS or used output tubes of that vintage.
Do use separate DC filament supplies. What if one tube dies, now find an exact match.
Separate self bias circuit for each tube works.
You only have to find a second filament transformer, if you power transformer does not have a second one.
And one can power both one output tube, and one driver tube (the driver tube will not mind if its filament supply is elevated to the self bias voltage (+45 to +60V).
2. The 47 Ohm and 51 Ohm filament resistors are there to make the voltage at the ends about 6.3V/2 = 3.15V.
That makes one end of the filament 3.15v higher than the self bias voltage; and the other end of the filament be 3.15V lower than the self bias voltage.
The reason is the negative end of the DC powered filament draws More plate current than the positive end of the DC powered filament.
There is more total current in the 47 Ohm resistor; versus the total current in the 51 Ohm resistor.
Empirical? . . . Not really.
All of that is to place the middle length of the filament (only accessible inside the tube) at exactly the self bias voltage.
If you swap the 47 and 51 Ohm resistors, versus the + and - 6.3VDC supply, you will get unbalanced voltages versus the self bias voltage.
3. Be sure to use Schottky diodes in the DC filament supplies.
The more voltage from the diodes, the larger the resistor resistance can be between the two (2) 4,700uF caps can be. That makes for lower ripple on the filament DC supplies.
When determining the final resistance there, be sure to wait for at least a minute when measuring the 6.3V across the filaments.
That is because the DC filament supply circuit has Intrinsic slow warm up time.
When I made my 6A3 (1A DC) and 300B (1.25A DC) supplies, I used Schottky diodes, and the two caps were 22,000uF each . . . I got 1mV ripple.
The only "heat sink" was the Schottky bridge was mounted directly on the chassis.
4. 3.5k output transformers, slightly lower than the textbook 2.5k, but . . .
Event thought 3.5k has slightly less power output, The advantage is: Lower distortion and Higher damping factor.
Have fun building and listening!
P.S. I have a question, was the original amplifier using 6V6, 7591, or other output tubes?
Thanks!
Last edited:
Thank you for all that great info!
right now the amp uses all RCA tubes: a pair of 6V6GT, a 6SN7GT driver and a JAN 5U4GB.
planning to use 1N5821 diodes to rectify heater supplies
w
E: Can I run both heater supply networks in parallel from the single 6.3v supply of the main PT or do I need a separate transformer AND rectifier for each power tube?
right now the amp uses all RCA tubes: a pair of 6V6GT, a 6SN7GT driver and a JAN 5U4GB.
planning to use 1N5821 diodes to rectify heater supplies
w
E: Can I run both heater supply networks in parallel from the single 6.3v supply of the main PT or do I need a separate transformer AND rectifier for each power tube?
Last edited:
If not empirically determined how do you calculate the value of these resistors?
The latter statement is correct, as 6A3sUMMER explained in point 1. in post #9. Since you have one 6.3V supply coil already, you need to acquire one more 6.3 Vac transformer.
Last edited:
Not sure what you mean. On the secondary side, in addition to the HV coil winding with center tap, I see one center-tapped 6.3 V and one 5 V in your power supply sketch.
Be aware that if you use the 6.3 V winding for an output tube filament heater, you can not ground the center tap as shown. The 900 ohm bias resistor forms the ground connection.
Be aware that if you use the 6.3 V winding for an output tube filament heater, you can not ground the center tap as shown. The 900 ohm bias resistor forms the ground connection.
Last edited:
Believe me - I've build quite a few SE 6B4G amps and use one myself. There's a BIG difference in sound quality between Rod Coleman V9 filament regs from Lyrima and basic voltage regulators. Rod has been perfecting his regs for 12 years and this is the 9th upgraded version. As stated you will need 10V for each reg, which you can get from a 30VA or 50VA 9+9V transformer.
You also have the AC option being discussed
I need 2 completely separate heater supplies for this amp, and that's what's drawn on many DHT schematics, 2 transformer coils.
I'll remove that center tap from the 6.3v windings on the main PT, and use that to power one 6B4G and the 6SL7 heaters. Then I'll get a Hammond 166L6 to provide another 6.3v supply for the other tube 👍
And I will certainly look into Mr Coleman's regulators, but as that will require 2 international shipments, I'll look to a domestic solution that may have me done by Friday.
thanks!
I'll remove that center tap from the 6.3v windings on the main PT, and use that to power one 6B4G and the 6SL7 heaters. Then I'll get a Hammond 166L6 to provide another 6.3v supply for the other tube 👍
And I will certainly look into Mr Coleman's regulators, but as that will require 2 international shipments, I'll look to a domestic solution that may have me done by Friday.
thanks!
dub - this post by Ed Brown (RIP) is a great read - see the tubelab link too
http://www.diytube.com/phpBB2/viewtopic.php?t=4551
http://www.diytube.com/phpBB2/viewtopic.php?t=4551
Francois G,
I can not tell you if the 47 and 51 Ohm resistors are correct or not.
With 6.3VDC, one 6A3 might have 0.95A of filament current, another 6A3 might have 1.05A filament current (in that example, 0.1A, or 100mA difference in filament currents of two tubes).
Start with both filament pseudo-center-tap resistors = 50 Ohms.
Measure and compare:
One end of the filament to the self bias voltage.
The other end of the filament to the same self bias voltage.
The filament is 1A (approximately).
Suppose the current through the self bias resistor is 60mA (the same as the plate current).
Approximately 2/3 of the 60 mA plate current comes through the more negative end of the filament.
Calculate the resistors to make up for the measured difference in voltage across the two 50 Ohm resistors (in this example due to 20 mA and 40 mA through the two 50 Ohm resistors).
As stated above, you have to build first, measure, then calculate to find the resistors that make up for the initial filament end voltage difference (due to the plate current, and the difference of the un-balanced plate current through the filament).
The 6.3V filament relative voltage along the length of the filament wire, is 0V to 6.3V DC from one end to the other.
The fractional difference of the 5500 uMhos transconductance along the filament wire is:
More at the negative end of the filament, and less at the positive end of the filament.
I hope that explains it.
I can not tell you if the 47 and 51 Ohm resistors are correct or not.
With 6.3VDC, one 6A3 might have 0.95A of filament current, another 6A3 might have 1.05A filament current (in that example, 0.1A, or 100mA difference in filament currents of two tubes).
Start with both filament pseudo-center-tap resistors = 50 Ohms.
Measure and compare:
One end of the filament to the self bias voltage.
The other end of the filament to the same self bias voltage.
The filament is 1A (approximately).
Suppose the current through the self bias resistor is 60mA (the same as the plate current).
Approximately 2/3 of the 60 mA plate current comes through the more negative end of the filament.
Calculate the resistors to make up for the measured difference in voltage across the two 50 Ohm resistors (in this example due to 20 mA and 40 mA through the two 50 Ohm resistors).
As stated above, you have to build first, measure, then calculate to find the resistors that make up for the initial filament end voltage difference (due to the plate current, and the difference of the un-balanced plate current through the filament).
The 6.3V filament relative voltage along the length of the filament wire, is 0V to 6.3V DC from one end to the other.
The fractional difference of the 5500 uMhos transconductance along the filament wire is:
More at the negative end of the filament, and less at the positive end of the filament.
I hope that explains it.
Francois G,
You are correct, the solution is empirical.
I was wrong.
Just remember, anything that has a variable in it will require some empirical solution.
Such as . . .
Very few 6B4G tubes have exactly 1 Amp filament current.
Very few 6B4G tubes have exactly 5,250 uMhos at 60mA plate current.
And then there is the exact nature of the variable point to point transconductance from one end of the filament to the other (it is not exactly linear; many do not know that).
The approximate 5,250 uMhos, is divided up, point to point along the filament (5,250 is the integral of the whole filament).
An exact science with in-exact parts?
'Nuf said?
You are correct, the solution is empirical.
I was wrong.
Just remember, anything that has a variable in it will require some empirical solution.
Such as . . .
Very few 6B4G tubes have exactly 1 Amp filament current.
Very few 6B4G tubes have exactly 5,250 uMhos at 60mA plate current.
And then there is the exact nature of the variable point to point transconductance from one end of the filament to the other (it is not exactly linear; many do not know that).
The approximate 5,250 uMhos, is divided up, point to point along the filament (5,250 is the integral of the whole filament).
An exact science with in-exact parts?
'Nuf said?