6L6GC EL34 Single Ended Tube Amplifier

[Mars]

The amp is playing in the study and sounding very good. It has a very full sound with a lot of bass, especially with the switches on the "pentode" setting. The screen grids are tapped at 25%, so the sound is 75% pentode. I am also using the OPT secondary winding for cathode feedback.

The power supply settles at 400V with about 368VDC on the 6L6GC between anode and cathode. The 6L6's are running at 25watts dissipation (32V over the 470 ohm cathode resistor). The 6DJ8's have about 220V on top of the totem pole.

I hear no hum. The sound is crystal clear, but the amp needs a few hours of good music to get to that smooth sound I like so much. The cathode caps have not been by-passed (every other electrolytic has been bypassed with polyprop caps). I am waiting to do this later, to hear the difference.

[Mars]

After a few hours it is becoming apparent that the 6L6GC's bass is a bit flabby. The EL34's sound better, and the KT100's best. The KT100 make the PT run too hot. I think the voltage is too high for the 6L6GC's. I am going to reduce the first cap in the power supply and see if I can drop the voltage to about 360V.

[Mars]

Some photos from yesterdays build progress:

[Mars]

The amp is now with her new owner. The demo went well, especially the thunderous loudness that a single ended amp can deliver through high sensitivity speakers. I do not know how much dB were emitted, but it felt like being close to the stage at a live electric music show.

To recap: I did some simulations on SEcad of Glassware, and realized that the flabby bass I heard was probably 2nd order distortion. Running a tube at 100% is not a good idea anyway. I have to still hear the tube that sounds best at 100%. The simulations indicated that much lower 2nd order distortion can be had at lower voltages. At 250V the distortion drops to about 1%, but you can only get mWatts into the speakers. Most people seem to opt for just over 300V on the anodes. Keep in mind that I am using 5K:8ohm OPT.
.
The first tweak was to change the power supply to choke input filter. Like magic the voltage dropped to 280V.The choke started to sound like a diesel generator in the distance. Choke input filters probably like big cores to do the job.

At this setting the sound improved dramatically, but the bass was a bit thin. Cathode resistors were changed to 220ohm and the sound was starting to get close to audio nirvana. Dissipation were now at 69%. 

Next I added a 1uF 630V polyprop cap as first cap in the pi filter (power supply). Voltage went up to 290V. The sound was even better. Eventually I settled on 2uF for the first cap value. Voltage is now 310V and the choke is silent. This was the configuration with which the amp shipped. (BTW the 6DJ8 were at 230V on the totem pole with the original 330R values on cathodes)

[Prince]

Wow - I wish I could wire so neatly one day!..

-Brilliant job Mr Mars!!

[Mars]

Thanks Prince  Some of the last pics I took, before closing up. At the last minute I removed one of the yellow 1uF 630V caps to drop the dissipation from 86% to below 80%.



This tube kicks! (6DJ8 Sylvania)

[Rudi]

are you using the Jimson as coupling caps?

[Mars]

Jip... do you like them too? 

[Mars]

I did some research on the latest news on coupling caps, mostly dwelling here http://www.laventure.net/tourist/caps.htm

From my experience the Jensen Copper Foil in oil is the Holy Grail, but it seems there are now quite a few holier grails to be found.

The teflon caps are probably the best quality and the Solen/Auricap the best value for money. The polypropylene in oil caps from Angela et al seems to be best for power supply.

I think it is time to build a SE with more audiophile grade parts. I suspect the OPT's I rolled are pretty good, and it's time to refine the rest.

BTW I think that a bulkier choke will work better when a tube rectifier is used especially in conjunction with polyprop filter caps.

[Mars]

I had some feedback from the customer.

He started the amp in triode mode, and she sang! Bass was full through his Full range drivers. The sound seems to be up to par. He did mention though that there is a perceivable hum coming from the amp and the loudspeakers. I am quite confident that nothing is busy self destructing, but that larger iron in the power supply is probably called for. Or maybe I should get the power transformer glued together more solidly that just waxing it.

A temporary remedy would be to explore the 6V6GT as a possible output tube that might harmonize better with the current power supply design. An other option would be to see how the EL34 fare if the higher voltages of 380V are used.

More to follow.

[Rudi]

are you using the Jimson as coupling caps?

only used them in power supplies. will test one on the preamp.

[handsome]

For choke input supplies you have to consider the sum of the DC current and the AC current. Because the choke is first in line it sees the full ac voltage across itself from the power transformer. The formula for DC current in a choke loaded PSU is Vin/(1155*L) Vin(rms) is your transformer secondary and L is the choke value in henries. This figure must be added to your DC current draw to see the total current that the choke sees. If your choke's current rating is too low it will rattle and hum.

Regarding distortion diddling with SECad shows that distortion drops as bias current rises and output power rises as voltage rises. Personally, if you're not operating hideously expensive DHTs i'd whack up the current (and dissipation) for the sound. 6L6s aren't hugely expensive or rare......of course those sylvania 6dj8s are another story

Regarding hummm well that's the problem with SE amps; they have virtually no PSRR compared to PPs. Proximity of OPTs to choke (if it is choke loaded) and power transformers are the usual culprits here. If you suspect AC standing the heaters on 30-70V via a pot across the heater lines will help.

[Mars]

Thank for the input handsome. My choke is a 4H 200mA choke (2500 0.3mm Cu turns on a 1"x1" bobbin with a 1mm airgap). I could go for thicker copper (0.4mmCu), and drop the H to 1, but that would leave very little regulation. I am limited to the amount of capacitance I can add due to peak max current of the tube rectifier.

According to the formula you mention, I should aim for at least 260mA capacity in the choke. I'll probably be able to get about 8 H from a 32mmx32mm bobbin wound with 0.4mmCu.

I noticed the see-saw around distortion/current vs. voltage/output power

The heaters are floating at 20V at the moment (tied to the cathode resistor of one 6L6GC)

[handsome]

as long as the capacitance is after the choke you are not limited at all really. The output of the rectifier is AC to which the choke displays a very high impedance. Capacitance directly after a rectifier has to be limited. Ideally you would want to go a bit higher in henries for an input choke, usually 10H - 20H. In the (good?) old days input chokes were 'swinging' chokes presenting high inductance at low currents and lower inductance at high currents.

Using a lot of capacitance after the choke lowers the resonant frequency of the filter which is: 1/(2*pi)*square root of (L*C).
The Q of the filter is: 1/R* square root of (L/C). You can add series resistance to damp the filter.

Bass resonance and high Q may be responsible for the bass you are hearing.

[Mars]

I got a photo from the owner: