tubeswell wrote:And won't the LTP sound better with a 12AT7 set up like in a blackface amp? (or at least a pair of 12AU7 cathode follower/buffers to drive the 6L6 grids? - since this is a bass amp)
I agree, 12at7 for the PI.
I think there is consensus on removing the 4th stage, too.
And rework your NFB loop.
After those are done I would revisit the Choke in place of the inductor/sag resistor.
Cool, thanks guys, I have a 12at7 lying around I will give it a try. NFB loop was making the distortion worse, when I removed it the signal cleaned up a ton. Maybe I had a wrong value on the resistor, will have to play with it.
I will see if the 4th stage is what is overdriving it by removing it.
tubeswell wrote:And won't the LTP sound better with a 12AT7 set up like in a blackface amp? (or at least a pair of 12AU7 cathode follower/buffers to drive the 6L6 grids? - since this is a bass amp)
I agree, 12at7 for the PI.
I think there is consensus on removing the 4th stage, too.
And rework your NFB loop.
After those are done I would revisit the Choke in place of the inductor/sag resistor.
Cool, thanks guys, I have a 12at7 lying around I will give it a try. NFB loop was making the distortion worse, when I removed it the signal cleaned up a ton. Maybe I had a wrong value on the resistor, will have to play with it.
I will see if the 4th stage is what is overdriving it by removing it.
Something is wrong if NFB made the distortion worse. look at a classic Fender ab763 circuit for proper implementation.
That what I was saying before. Any blackface LTP stage - (including the NFB) should work well. Pick the amp with the same number and type of output tubes
The tail resistor doesn't need to be 47k if you have unbalanced plate resistors. The blackface amps used 22k. You could go down to 10k. Usually with a 47k tail resistor you are getting into the balanced plate arena (i.e.: 2 x 100k). Also the coupling cap feeding the PI in those amps is around 1nF (since they are using a 12AT7)
billyz wrote:
I agree, 12at7 for the PI.
I think there is consensus on removing the 4th stage, too.
And rework your NFB loop.
After those are done I would revisit the Choke in place of the inductor/sag resistor.
Cool, thanks guys, I have a 12at7 lying around I will give it a try. NFB loop was making the distortion worse, when I removed it the signal cleaned up a ton. Maybe I had a wrong value on the resistor, will have to play with it.
I will see if the 4th stage is what is overdriving it by removing it.
Something is wrong if NFB made the distortion worse. look at a classic Fender ab763 circuit for proper implementation.
I hear ya Billy, that is what puzzled me the most. I kept reading about the feedback helping to keep the signal clean. I tried it off of the 8 and the 16 ohm secondary, and tied it in where it ties in on the Wrecks. I even fooled around with the resistor value to try and get it right. But it only seemed to add or subtract gain, not eliminate the distortion. I have a feeling I am doing something simple wrong, but I haven't figured out what yet. I know the lead dress is not good, I might have to start again from scratch. When I get some spare time I will bypass the last gain stage and fool with the NFB some more. I am going to take your suggestion on the first filter caps also, but I only have some old ones on hand that are iffy, so I am going to order new ones. Thanks so much for your help.
tubeswell wrote:That what I was saying before. Any blackface LTP stage - (including the NFB) should work well. Pick the amp with the same number and type of output tubes
The tail resistor doesn't need to be 47k if you have unbalanced plate resistors. The blackface amps used 22k. You could go down to 10k. Usually with a 47k tail resistor you are getting into the balanced plate arena (i.e.: 2 x 100k). Also the coupling cap feeding the PI in those amps is around 1nF (since they are using a 12AT7)
For better bass, use Big OT iron
The PI and LTP are still confusing the holy Sh*t out of me. I think that it is the final gain stage, lead dress or the PI that is causing what I am hearing. I am going to try the 12AT7 as you two have suggested and work forward from there. Thanks so much.
Last night before I went to bed I tried the old 12AT7 to replace the 12AX7 in the PI. The amp sounded great for a few minutes, then the gain slowly started to fade away and the distortion was all that was left. So I put the 12AX7 back in and then I still had no gain and fuzzbox distortion. WTF? Turned the amp off and went to bed.
This morning I fire it back up and it sounds good for a minute or two then again the gain fades and nothing but distortion. I figure it has to be something in the PI circuit so I decide to rewire the entire circuit and change the LTP values from 46K and 410 ohm to the more traditional 22K/470 ohms. I finish this step and go to fire the amp up and see a wisp of smoke near the power switch. Now I have no proper DC voltages. Break out the bulb limiter, won't even light the bulb.....?? Remove fried MOV, power back all looks good on the bulb limiter. Remove bulb limiter and hit power switch, only to see sparks fly from it. Should have went back to bed at this point.
Find loose wire on switch from PT primary, resolder. Bulb limiter, looks good. Remove limiter, fires up, voltages right, same distorted crappy sound with no gain. Get pissed, unplug tube amp. Plug 350 watt SS amp into Ampeg 4x10 cabinet. Same shitty sound, WTF? Plug SS amp into another cab, same sound. Switch chords, no difference. Replace 9v battery in bass, no change. try different bass without active Pup's, no change. Freak the fuck out and call my tech!!!!!!!!!! Load all of the equipment in truck and take it to his place. Cabinet works, SS head works, Tube amp works, and has nice clean signal on oscilloscope.
Must be something in the A/C line here. Plugged it in upstairs after I got back and it worked fine. Guess I have to haul it all back down the steps and try again.
Now I am afraid to put the 12AT7 back in......LOL, I'm going back to bed!!!
If I repeat someone elses info, I appologize, I scanned the responses.
I see few areas that I would look at.
1. 410ohm on the cathode of the PI is very small. I would start at 820 and go up from there.
2. Not related, but the 100K send pot is or will load the heck out of the tone stack. Especially as you have 1Meg pots for bass and treble. I would buffer it thru another triode stage,
3. the 1M over 100K tied to ground is a problem. Needs a cap and why throw away 21dB of gain after the tube?
Gary
Located in the St Croix River Valley- Afton, MN
About 5 miles south of I-94
aka K0GWA, K0 Glas Werks Amplification www.glaswerks.com
glasman wrote:If I repeat someone elses info, I appologize, I scanned the responses.
I see few areas that I would look at.
1. 410ohm on the cathode of the PI is very small. I would start at 820 and go up from there.
2. Not related, but the 100K send pot is or will load the heck out of the tone stack. Especially as you have 1Meg pots for bass and treble. I would buffer it thru another triode stage,
3. the 1M over 100K tied to ground is a problem. Needs a cap and why throw away 21dB of gain after the tube?
Gary
Thanks Gary, I changed the 46K in the LTP to a 20K and the 410 to 470R today and it seemed to help. Might go even higher than the 470 as you suggest.
I am going to add the buffer, Billy Z brought that to my attention, but I wanted to get the rest of the circuit stable first.
Have added a .1uF cap in front of the 1M/100K divider.
No good explanation for why I threw away 21dB. Cause I am still learning and I didn't totally understand all the small nuances of the design process, would be my best excuse. I was fighting distortion and thought I was overdriving the PI.
Here is a schematic of the present revision. After a hair raising day I can say I am happier than I have been so far with the sound of the amp.
The biggest change seemed to come when I rewired the PI circuit and changed the 46K to a 20K and the 410R to a 470R. That's when the nastiness went away.
Still no NFB, and I forgot to change the PI plate resistor back to a 82K when I lowered the 46K, so it is still two 100K plate r's, but I can finally turn it up and play it with some attack and still get a clean output. I will tackle the pre amp buffer and experiment with a choke and higher filter caps another day, as I have to play Friday and Saturday and don't want to bugger anything up before then.
Thanks a million for your input people, this place rocks.
You do not have the required permissions to view the files attached to this post.
Sorry, three more stupid questions. On the buffer, can I basically use the send side of an effects loop buffer? Would I break the signal from going to the PI and power tubes if the Pre is used? And if I do break the signal from feeding the PI and power tubes, do I still need a speaker load on the amp to protect the OT?
The buffer would be a cathode follower, similar to a send side of an effects loop. Basically, the output impedance of a cathode follower is under 1k depending on the tube type. If you drove the cathode follower from the output of the tone stack, you wouldn't load the tone stack down.
As it is you have a 100k pot loading down the output of the tone stack. You want to always drive a low impedance into a higher impedance to make sure you're not just losing signal due to loading. The tone stack output impedance is pretty high and you're driving it into a lower impedance 100k pot, which loads down and robs gain/frequency response.
Tap off the output of the treble control into a cathode follower, and that output of the cathode follower can be your ~600ohm direct out.
If you do it like this, you won't interrupt the signal from going all the way to the power amp. If you do want to wire it to interrupt the signal to the power amp you could, but you will always need to have a speaker load connected to the output of the power amp.
soma_hero wrote:The buffer would be a cathode follower, similar to a send side of an effects loop. Basically, the output impedance of a cathode follower is under 1k depending on the tube type. If you drove the cathode follower from the output of the tone stack, you wouldn't load the tone stack down.
As it is you have a 100k pot loading down the output of the tone stack. You want to always drive a low impedance into a higher impedance to make sure you're not just losing signal due to loading. The tone stack output impedance is pretty high and you're driving it into a lower impedance 100k pot, which loads down and robs gain/frequency response.
Tap off the output of the treble control into a cathode follower, and that output of the cathode follower can be your ~600ohm direct out.
If you do it like this, you won't interrupt the signal from going all the way to the power amp. If you do want to wire it to interrupt the signal to the power amp you could, but you will always need to have a speaker load connected to the output of the power amp.
Thank you for the excellent explanation. I have a couple of diagrams of buffers. One from Merlin and one from Dumble. They are pretty close for the most part. Do you think either would work for my needs, if so, which would you prefer?
You do not have the required permissions to view the files attached to this post.
They are both pretty similar. The 470k or 1M grid leak won't make much difference. They difference here is coupling cap size and the bias of the cathode follower. You'll need to play with these values to get what you want. I'd start with Merlin's and tweak from there.
a Blackface PI with a feedback loop and a 12AT7 would give you cleaner tone with tighter bass. You'll get no speaker damping with that open-loop power amp...
The Silverface PI with 330K grid resistors and 47K plateloads would also be a good choice; even cleaner and setup better for the 12AT7...
I'm sure your tired of making changes but that's what happens when you're building a prototype amp like you have here. Another reason lots of folks like making clones of amps... new designs can be a nightmare as you're finding out...
soma_hero wrote:that dumble schematic looks wrong. I would think a 15k bias resistor would not bias the stage right. Check out the dumbleator first stage:
They are both pretty similar. The 470k or 1M grid leak won't make much difference. They difference here is coupling cap size and the bias of the cathode follower. You'll need to play with these values to get what you want. I'd start with Merlin's and tweak from there.
Good catch, the Dumbleator uses a 1.8 and a 27K on the cathode. Don't know where I got that version, I thought it was from a thread I got here when searching for "buffers". Thanks again.
