New amp build.....

[Jack Hester]

I'm making a return to tubes, amplifiers in particular, after a few decades away from them. It's been a slow re-learning process for me, but most enjoyable. I've built a few small SE amps, chassis only, to get back into the swing of just building something electronic. I have an old PA amp, that I purchased to restore. Just for the fun of doing so. It was not packaged very well, so the power transformer ripped itself loose and destroyed tubes and pulled out components. As the amp had no identification as to make/model, and no schematic, I chose to strip it and build from scratch. I have included a drawing of the as-built amp. It is in PDF format. The amp works fine, and am pleased that it does so, as it was a rather difficult build for me (point-to-point). Mainly because I put a lot more into it than it had before (components).

The reason I'm posting is that it is very weak for what I think 6L6's should produce. My friend from Church, Don, says that it is 'timid', and at full volume, no one has to leave the room. I have a '54 Gibson GA-40 with 6V6's that will blow it away. Please look over my drawing and critique as you see fit. Please make recommendations and corrections. Thanks for all help.

Jack

[MWJB]

The first things that stand out to me are the 6SN7 at the cathode follower & PI. Try 6SL7.

R30 & R31 should be matched to balance the signals sent to each power tube.

[MWJB]

Add a 1.5k between the PI tube's cathode & R30, ground R29 to the junction of the 1.5K & R30.

R38 at 820ohms is far too small. Try 33K.

Difficult to follow due to part numbers on the schem & having to refer to chart, can you paste values onto the schem?

[Jack Hester]

The first things that stand out to me are the 6SN7 at the cathode follower & PI. Try 6SL7.

R30 & R31 should be matched to balance the signals sent to each power tube.

Thanks for the really fast response. Building from scratch is a challenge for me. I used the 6SN7 at the CF, because I didn't know what kind of current draw would take place. Figured this tube could handle all, until I learned better. I used the same at the PI, as I have several on hand, and fewer of the 6SL7's. I will put a 100K on R30/R31, as a starter, and go from there. Will put one there when the changes are made.

Add a 1.5k between the PI tube's cathode & R30, ground R29 to the junction of the 1.5K & R30.

R38 at 820ohms is far too small. Try 33K.

Difficult to follow due to part numbers on the schem & having to refer to chart, can you paste values onto the schem?

Thanks ever so much! Sorry for the separate sheets. I'm beginning to add values to my future builds. Will do so on the next revision of this drawing. I'm at work and can hardly wait to get home to try your fixes. And, today is going to be a long day at work, but I will make the changes before going to bed. It may be tomorrow morning before I report back. Have a good one.

Jack

[MWJB]

Hi Jack,

The 2 biggest issues are probably the unbalanced PI and R38 - to flesh things out a bit, NFB loops usually have ratios from around 10:1 up to 70:1, meaning for every volt that develops at the speaker/OT secondary 1/10th (0.1v) to 1/70th (.015v) is fed back to an out of phase point in the preamp/PI. As it stands at the moment, your design feeds back .65 volts for every volt developed at the speaker/OT, or a ratio of 1.5:1, so it's having a significant effect on your output.

[Jack Hester]

Hi Jack,

The 2 biggest issues are probably the unbalanced PI and R38 - to flesh things out a bit, NFB loops usually have ratios from around 10:1 up to 70:1, meaning for every volt that develops at the speaker/OT secondary 1/10th (0.1v) to 1/70th (.015v) is fed back to an out of phase point in the preamp/PI. As it stands at the moment, your design feeds back .65 volts for every volt developed at the speaker/OT, or a ratio of 1.5:1, so it's having a significant effect on your output.

You're the best! Can't wait to make the changes and give it a try. I'll post the results, soon after this evening, I hope. Have a good one.

Jack

[Jack Hester]

Didn't get out to my shop until around 06:30pm. Changed out R38 (NFB) to a 33K. I changed R31 (PI plate) to a 100K. Made R30 a 100K. Tied R30 to signal ground and put a 1.5K (no tagname assigned, yet) between R30 and the PI cathode. I connected R29 (PI grid) to the junction of the 1.5K and R30.

Warmed up the amp, using the existing tube lineup, and noticed an increase in volume, from before I made the changes. Removed the 6SN7 tubes from V3 and V4, and swapped them with 6SL7 tubes. Warmed it up, again, and this time it was very loud as compared to using the 6SN7's. Channels 2 & 3 are not uncomfortably loud at full volume. But, I think that in my living room, at the house, it would be plenty loud. Channel 1 is very loud, compared to the other two channels. And, above 6 on the dial, it distorts quite a bit.

So, your suggestions made quite a change. Thank you very much for this big help.

Now, I have a question about biasing. I have an adapter that fits in between the 6L6 tubes and the socket, with a selector switch and leads to connect to a millivolt meter. The current flow does not match for each tube. When I set the current for 37mA on the highest tube, the other is only about 10mA. I neglected to pay attention to which one was the highest, but I believe that it was V5, connected to the cap coming from the PI plate. How can I balance these current flows to be closer? I thought that they would when I made the PI plate and cathode resistors the same. But then, the 1.5K has me puzzled. If you would, please explain the purpose of the 1.5K, and why I wouldn't just have R30 go straight to signal ground, as well as R29.

Thanks ever so much for your help with my learning process and troubleshooting. Have a good one.

Jack

[MWJB]

The PI plate & cathode resistors have no influence on dc idle current of the power tubes, they balance the AC sent to the power tubes. Your power tubes (assuming plate, screen & negative grid voltages match accross both power tube sockets, swap tubes in their sockets to ascertain this) are simply unmatched. Buy a pair that are matched to +/-5mA @ 35mA.

Some designs do simply have the plate & cathode resistors. Adding the 1.5K biases/voices the tube more similarly to your other preamp stages.

[Jack Hester]

The PI plate & cathode resistors have no influence on dc idle current of the power tubes, they balance the AC sent to the power tubes. Your power tubes (assuming plate, screen & negative grid voltages match accross both power tube sockets, swap tubes in their sockets to ascertain this) are simply unmatched. Buy a pair that are matched to +/-5mA @ 35mA.

Some designs do simply have the plate & cathode resistors. Adding the 1.5K biases/voices the tube more similarly to your other preamp stages.

This is really good information for me to file away. My 6L6's are mostly new old stock metal case tubes, pretty much guaranteed to not be matched. As time permits, I'll do some tube swapping and monitor the current on each. Looks like the amp may be at a good stopping point, and give it a good workout. It means a great deal to me, to have your comments to guide my build. I will post my changes and their results. More to come. Thanks a bunch.

Jack

[Jack Hester]

Well, this build turned out really well, because of your help. Being as I built it in an old PA amp chassis, I had to be very creative with the layout. As of now, it is loud enough that I believe that it could really benefit from a master volume, downstream before the power section. But, there is absolutely no real estate left for one. My Channel 1 has way lots more gain than Channels 2 & 3. And, gets a bit dirty above 3 or so on the dial. Channels 2 & 3 are very clean, all the way up to around 8, then break ever so slightly. I can now hear the brighter sound to Channel 3, like I had hoped.

Thanks, again, for the quick response and the great advice.

Jack

[MWJB]

You're welcome Jack.

Mark.

[Jack Hester]

The finished amp...........

I made a modification to Channel 1. I put a 200K resistor is series with the 270K that tied Channel 1 volume pot wiper to pin 1 of the 2nd gain stage (V3a). This attenuated the signal just the right amount to clean up that channel up to about 6 or 7 on the volume. Then it begins to break up some, but is way to loud for me to stand in front of the speaker, like I've been doing as I work. I also put a 120pf cap across the signal input to the pot and the wiper. Brightened it up a bit, though not as much as it did on Channel 3. I'm satisfied that lead dress and the PT are the main sources of my background hum, though satisfied is the key word for the whole amp, now.

Here are some pictures of the old amp:







Thanks, again Mark, for all the help.

Jack

[mooreamps]

I gave the drawing a once over... Looks OK to me... I do like the way it's drawn.. Very good job on that....

-g

[Jack Hester]

Here is the actual 'as-built' drawing................

I gave the drawing a once over... Looks OK to me... I do like the way it's drawn.. Very good job on that....

-g

Thanks for taking a look at it. The drawing is my struggling effort to make a good schematic, that others can read. I tend to draw literally from the actual tube layout, and the drawing wouldn't flow as easily as a typcal schematic. So, I'm trying one this way. Having to build a different symbols library to do so. My other tubes are drawn straight from the data sheets.

This one incorporates some mods to the PI, and a change in NFB resistor value, that Mark (MWJB) suggested. I decreased the plate and cathode resistor values to change the voltages a bit. Here's the finished product, and it sounds good! And, the amp is really loud. Not that I need it that way. But, I feel that I'm getting the kind of output that the 6L6's should be producing.

Jack

[Jack Hester]

Can't seem to keep my hands off of it............

Cleaned up my voltage chart and added some fresh readings. Also, remembered to record my B+ voltages, along with the transformer secondary AC voltages. Recorded the Bias AC voltage, and the negative DC out from the Bias Pot with 37mA current at the cathodes of the 6L6 tubes. That being said, I don't have a matched pair, so I set the highest tube for 37mA. The other is quite a bit lower, but doesn't matter at this point. I haven't been through any of my collection of these old metal cased tubes, to see if I match up a set.

Replaced one twisted pair filament feed with a shielded twisted pair, on V4. Ever so slight reduction in the background hum. Hum is not really noticeable, unless you are in an absolutely quiet room, or have your ear close to a speaker. This amp sounds really good, with three very different channel voicings. I put the PI resistor values back to where they were, from Mark's original suggestion. I liked the voltages better.

So, I've attached the what I plan to be the last update for this amp. It's just too good to mess with anymore.

Jack