5F1 Champ w/ Capacitor Coupled One-Tube 6BM8 Reverb (Eric Barbour/Ampeg-style)

[dchang0]

Thanks, guys, for all the help imagining the Eric Barbour one-tube 6BM8 reverb.

Here's the reverb circuit inserted into a Champ 5F1 circuit.

I do have a few questions:

1) Please let me know if you see any errors in the schem, especially anything that involves unsafe miswiring.

2) Should I insert the reverb before or after the Volume pot? I'm thinking before, which is similar to Bill M.'s mods on the Blues Jr. TheTinMan posted a link to some of tubenit's schematics showing Psychonoodler's Champ with reverb, and Pscyhonoodler put the reverb AFTER the Volume pot.

Technically, the outcome should be the same, but one way might result in more interference between the reverb circuit and the 5F1 circuit than the other.

3) Any recommendations on the values of the insert/bridge RCs? Bob-I over at ampgarage recommended a 470K balance resistor to start with the 500pF bright pass cap. He said that I could raise the 470K to 1M or as high as Leo Fender's 3.3M to get more Fender surf (killing the dry tone) and put a 10pF cap in parallel to the resistor to pass the highs.

Psychonoodle uses a lower 100K balance resistor with a 1nF bright pass cap.

Tubenit has often used a 220K balance resistor with 1nF bright pass.

I figured I'd start with 470K and go up from there towards Leo's insane 3.3M with 10pF in parallel. 470K seems kinda "safe" as a starting point.

Here's the schem: Rev. 1.0e.

[TheTinMan]

I think you need to insert the reverb between the volume pot wiper and the grid of V1b just like phsyconoodler did.

Wipe out the reverb circuit but keep in your mixing resistor. What do you have? A 470K resistor above your 1 meg volume pot. You could never go to full volume, let alone "11"! (see note at bottom though... )

If you put the reverb after the volume pot, the reverb pot is in parallel with the volume pot. That reduces the effective grid return (or "grid leak") resistor for V1b to the parallel combination of the 1 meg reverb pot and the ground leg of the volume pot (i.e. whatever is below the wiper), which I think would reduce the gain of V1b a little bit. Just turn up the volume a little bit. You've also added a 470K grid stopper on V1b where there wasn't one before. Slight low pass filter, cutting high end. How much? Don't know exactly.

I do know that Phsyconoodler reported great success with a 100K mixing resistor and the reverb after the volume pot. Here's the development thread: last little bit in 5F2A princeton with reverb and VVR

There is a difference in that I don't know if there is some minimum signal level needed to drive the pentode side of the 6BM8 which is higher than what a 12AX7 needs. If you use a 100K mixing resistor, the effect on the total volume may not be too much. Also, you could bump the gain of the second stage in a variety of ways if you needed to (bigger plate resistor, cathode bypass cap, smaller NFB resistor, etc.)

Hope that helps,

Chip

[dchang0]

Keen eye, Chip! And very well explained. I'll go back in and work it out! Thanks very much for the tips and the link--I've lowered the mix resistor to 100K based on the strength of your argument.

One thing that psychonoodler said concerns me: the reverb drops out at low volumes. Considering that I often play at very low volumes (middle of the night--can't sleep), having the reverb there would be nice. Perhaps I need a more sophisticated way of mixing the two signals together.

Here's Rev 1.0f, which is based on your suggestions. I added a cathode bypass cap to boost the gain a bit. Just seems right since the other two stages have one. It might allow me to bump the 100K mixing resistor up to 220K to try to get more reverb.

[txstrat]

I've used "Tubenits reverb" with one 12AX7 in several amps now and I'm very satisfied with the result.
I usually insert it the way you have in Rev. 1.0f of your schem. (right before the PI since my amps are PP). Even at veeery low volume settings I can dial in the amount of reverb I like (have to turn it to ten sometimes for those midnight alone in the basement rehearsals). If I needed even more reverb I would consider raising the balance resistor to 220k as well.
Just my 2c

[dchang0]

Sweet! Thanks for letting me know and confirming that this is the right location for the insert--100K it is!

Finally started stuffing the 5F1 board. I'll post up photos of the double-decker eyelet board when I get it going. It's gonna be challenging--the board will have to be cut smaller than the main board because of the tube sockets, pots, and Cliff jacks.

[dchang0]

Update:

With the help of several forum veterans, I built this amp and tweaked several values to make the amp actually work and sound good. Here's the schematic from all that experimenting:



So yes, this is a tested schematic. It works.

[TheTinMan]

dchang0 - First off, thanks a bunch for sharing the results of your experiments. Glad you got it working well!

I'm also curious about two things in that circuit design relating to the cathode bypass cap on the second preamp triode. Doesn't the cathode bypass cap result in a really strong signal hitting the 6V6? Not saying Fender was always right, but he didn't put the bypass cap on that gain stage until the AA764 circuit with a relatively high-loss TMB tone stack. I have seen other designs with that bypass cap switchable as a gain boost though.

My second question is sort of theoretical. If you insert NFB directly at the cathode and there's a cathode bypass cap, isn't that the same as no NFB at all? IOW it looks to me as if all of the NFB in the audible range would go to ground. Again, the AA764 has a resistor at the "bottom" of the normal bypass cap/cathode resistor network to create a voltage divider. So does the Princeton Reverb, etc.

Please don't take this as criticism of your design - I just want to understand how/if NFB works in this configuration.

Cheers,

Chip

[dchang0]

Hey, no worries, man. Constructive criticism is always welcome.

It's funny that you mention that particular cap. My plan has been to remove it when I can get to it (it is under the reverb daughterboard). I originally added it for the purpose of increasing gain to recover from the high-value mixing resistor. It seems to work well enough--the extra bite it adds is welcome. But in the interest of taming the noise and microphonics, I'm "softening" the stages after the reverb driver. It is the only change I plan to make to the motherboard.

As for NFB, I hadn't taken that into account here, and you are probably right in that I've wiped out NFB inadvertently. But I don't know much about NFB theory, so I really couldn't say. I'll have to read up on it online.

There will likely be two more schematic updates anyway, and the removal of that particular cathode bypass cap will be in one of them, after the actual removal from the amp.

[dchang0]

ALL DONE! This is confirmed to work, sounds good (albeit a whole lot cleaner than Rev. 1.0h), and is very low noise.

Removing the 25uF cathode bypass cap on the 6V6 reduced the gain dramatically AND reintroduced NFB, cleaning up the tone all the way up to 10/12 on the volume dial.

Rev 1.0i:



A brand new super-compact layout on a 2" x 5" eyelet board and improved buss grounding removed 99.9% of the noise and hum AND also any trace of low-frequency oscillations. It probably helps that the recovery stage is lower gain AND lower voltage than before too.

[txstrat]

Would you mind sharing the layout?

[dchang0]

I never drew up a layout, so it's an exercise best left for the reader. It was easier to just tape the loose components to the board, then drill holes where the leads crossed.

But here, I'll describe it for you:

5"W x 2"H eyelet board. All components vertically arranged in columns, with either two eyelets or a vertical column of three eyelets. Top component (on the control side of the chassis) and bottom component (tube side) as indicated. Each number below is one column. Starting left (near pilot light) to right (near input jack):

1) 50uF+560 in parallel
2) 50uF+1.5K in parallel

3) TOP: 100K (mix resistor) BOTTOM: 500pF
4) TOP: empty BOTTOM: 220K (between volume pot and pin 3)

5) TOP: 0.022uF coupling cap BOTTOM: 220K plate resistor, recovery side
6) TOP: room for a spacer screw BOTTOM: 22K grid leak, recovery side

7) 5K plate resistor, driver side
8) 0.470uF coupling cap, driver side

Fits pretty tight in there. Leave room on the right side for your wrench to fit for the chassis screw, and leave as much room on the left as you can to put space between the reverb components and the filter caps on the motherboard. I left 0.5" on both sides clear, so I really only needed a 2"H x 4"W daughterboard.

Make sure your spacer screw holes are drilled when BOTH boards are empty so that nothing will get shorted out on either the motherboard or the daughterboard. Draw circles around the screw holes to make sure no components get too close to the screws.

[TheTinMan]

DC - glad you got it going the way you like it!

From the your thread focused on the reverb:

have you tried higher value pots for the verb level pot? I usually would use a 1 meg in that location.

jamie

Right now, there's a 1:1 voltage divider between the 100K mixing resistor and the 100K reverb pot. Also, the 100K reverb pot is the grid return (or grid leak) for the second/last gain stage. IIRC raising the value of the grid return resistor increases the gain of a common cathode gain stage. If you want to get more gain out of the second/last gain stage, raising both the mixing resistor and the reverb pot to 500K. Also, the 22k/1.5K NFB ratio may not be ideal for this lower gain stage (relative to a 5F1 turned up a bit or more). You might try raising the 22K to 27K or more. OTOH you may like it clean all the way up to "11".

Cheers,

Chip

[dchang0]

LOL-- man, Chip, you totally nailed it. This morning, at breakfast, I was mulling over moving back towards the 500KA pot and 1M mixing resistor I was using before to get back that top-end bite and overall higher volume that's now missing.

But the super-low noise floor that I've got now sure is nice, especially since this is a home recording amp, so I can't quite make up my mind. I think I'll jam on the amp for a while and see what comes of it. It's not like this sucker can't scream (past 11 out of 12); it's just that I've heard it scream better.

You definitely helped though, by pointing out the 1:1 voltage divider. It is obvious now that the 1:2 setup I had before was problematic. I was basically shoving signal through the noisy reverb section and then having to tame it with a 470K "interstage attenuator" resistor running off of the top leg of the 500KA reverb pot into the grid of the 6V6. 20-20 hindsight: the 470K + 500KA pot = 1M mixing resistor =~ 1:1 ratio.

It might be worth trying a 250KA pot, 250K mixing resistor, and a 100K "return leg" resistor to form a voltage divider that slightly favors the dry signal and brings overall gain up a bit.

[dchang0]

After six months of tweaking using 500K trimpots, etc., I came up with this new final-final version of this schematic, 1.0j.

Actually, this exact schematic is not what I'm running right this instant; I am using a 470K reverb return grid leak resistor instead of the 270K as shown, but that produces an unreasonable amount of reverb for most people, and I found the 270K to produce the widest useful range on the reverb level (mix) knob.

I did find the 33K on the return leg of the pot to be necessary to cut off the dip in the dry signal between 8/10 and 10/10 on the reverb knob.

This is 1.0j. It has been built, tested, and sounds phenomenal. At some point, I'll post sound clips of it as compared to a Pro Jr. a buddy and I added the Tubenit single-12AX7 reverb to, which also sounds amazing. My 6BM8 design is much stronger in the drive side--past Fender surf levels. Tubenit's design is much more "reasonable."

[pietro_moog]

very nice. i have almost all the parts.
what about the tank? did you use the 800Ω impedance intentionally or it's spare part?
do you think i could use the 2500Ω one? i can find it at a good price, while the 800Ω one must come from oversea..