My first build - AB763 Deluxe Reverb clone - opinions please

[tubeswell]

The more I read about "blocking distortion" the more I think that's what I have going on. What I've read describes symptoms similar to mine.

On the plus side it seems a simple and common fix is to just up the value of the grid stoppers, and I'm currently using damn near the minimum.

Yes you can easily go up to 10k -47k grid stoppers on an output tetrode or pentode without noticing any HF roll off, because of the low Miller capacitance enabled by the screen.

[Greg_L]

Yes you can easily go up to 10k -47k grid stoppers on an output tetrode or pentode without noticing any HF roll off, because of the low Miller capacitance enabled by the screen.

I think I'll try that next. Seems easy enough.

[Steve A.]

The '68 reissue has a 1.5K NFB resistor...

Slight derailing here... the "Custom 68" series are NOT reissues. Hence all the changes. These are not intended to be as spot- on as possible Like the 65 series are. Sure, they look the part, but they're not supposed to behave like a real 68 Fender. Otherwise, they'd have that "mixed bias" that supposedly everyone hates, etc.

I know it's a small point, but I feel it's worth making...

Carry on.

Justin

True enough, Justin... My own theory as to its name is that Fender decided to call it a "Custom '68" because as it was a silverface model it was not as valuable to collectors as the blackface models so it was often a candidate for mods. (As we all know for many years silverface amps were mistakenly considered to be total junk by the amp snobs... until we learned how easy it was to convert them to BF specs if so desired.)

So while not a reissue of something that Fender ever sold it might be a "reissue" (not really the best word) of an actual amp that had been customized... Probably not in the real world but it might have been, right?

Full-sized carry-on bags are not permitted on United Basic Economy flights!

Steve A.

[tubeswell]

With small signals in a Long Tail Pair inverter the signal is balanced, but as the signal gets bigger, the result of slight signal imbalance from the inverting triode driving the non-inverting triode gets more accentuated. If the LTP is biased on the warm side, the inverting (master) triode reaches grid-current-limiting sooner causing the input coupling cap to charge up, which in turn shifts the grid voltage more negative thus biasing the tube closer to cut-off. The signal on the master (input) triode plate gets clipped on the positive side, which shifts the whole output signal more negative to compensate. These negative signal peaks are passed into the slave (non-inverting) triode, which cause it to draw grid current, draining charge from its decoupling cap, which means the average bias on the slave triode remains more-or-less unchanged. But because the signal from the master (inverting) triode output is heavily clipped on the positive side, so it spends more time positive than negative, but the non-inverting output is doing the exact opposite. This means that one output tube gets driven harder than the other, which can also cause that side to red-plate under heavy signal conditions. Getting the LTP more centre-biased can help prevent this (along with using decent grid stoppers on the output tube grids).

[Greg_L]

With small signals in a Long Tail Pair inverter the signal is balanced, but as the signal gets bigger, the result of slight signal imbalance from the inverting triode driving the non-inverting triode gets more accentuated. If the LTP is biased on the warm side, the inverting (master) triode reaches grid-current-limiting sooner causing the input coupling cap to charge up, which in turn shifts the grid voltage more negative thus biasing the tube closer to cut-off. The signal on the master (input) triode plate gets clipped on the positive side, which shifts the whole output signal more negative to compensate. These negative signal peaks are passed into the slave (non-inverting) triode, which cause it to draw grid current, draining charge from its decoupling cap, which means the average bias on the slave triode remains more-or-less unchanged. But because the signal from the master (inverting) triode output is heavily clipped on the positive side, so it spends more time positive than negative, but the non-inverting output is doing the exact opposite. This means that one output tube gets driven harder than the other, which can also cause that side to red-plate under heavy signal conditions. Getting the LTP more centre-biased can help prevent this (along with using decent grid stoppers on the output tube grids).

Wow, thanks for that. Makes sense.

So how do I look for something wrong there, and what do I look for?

I have not had any red plating. This amp has been pushed hard in it's short life and nothing's gone bad yet.

[Steve A.]

With small signals in a Long Tail Pair inverter the signal is balanced, but as the signal gets bigger, the result of slight signal imbalance from the inverting triode driving the non-inverting triode gets more accentuated...

Thank you for making such an informative post! I just started a new folder for my Evernote cloud archive titled "___MEF.CLASSICS" right next to the existing "___MEF" folder... and your post was the initial entry (I saved the text of your post along with the URL. If you had uploaded any images I could have added them as attachments.)

One question: some guitarists like to replace the 12AT7 in the BF/SF phase inverter without changing the circuit. What effect would that have in regards to your explanation?

One comment/questions: the phase inverter in the Dumble ODS #124 works very well with its trim pot. Now that the circuitry of #124 has been verified and widely circulated has its PI been "officially" added to the list of cool mods for the SF/SF design? Also if connected to 4 or 16 ohm speaker tap what would you recommend to replace the 4K7/3W resistor? And finally what value fixed resistor would you suggest in lieu of the 2KB presence control?




Steve A.

P.S. Would anyone care to explain how to set the 5K bias trim pot?

[tubeswell]

Wow, thanks for that. Makes sense.

So how do I look for something wrong there, and what do I look for?

I have not had any red plating. This amp has been pushed hard in it's short life and nothing's gone bad yet.

Then that's probably not the cause of the problem.

How did the 220pF grid-to-cathode caps work out?

[tubeswell]

Merlin Blencowe covers this topic in the 2nd edition designing pre-amps book.

Its more of a problem in classic Marshall amps which run EL34s at high plate and screen (~500V) voltaqes.

If its a 12AT7, 470R-560R is a reasonable (shared) bias resistor value (for a cathode-biased stage). If its a 12AX7, 820R - 1k2 is more appropriate.

[Greg_L]

Then that's probably not the cause of the problem.

How did the 220pF grid-to-cathode caps work out?

Have not tried yet. Tomorrow is work-on-amp-day.

[Steve A.]

Have not tried yet. Tomorrow is work-on-amp-day.

I understand how you want to fix this one last issue (it is the last one, right?) and button up your amp but as a few people have mentioned here it might resolve itself on its own over time which can be more important than amp volume alone as various parts adapt to the circuitry as they sometimes do for mysterious reasons. I attribute it to a very wise and benevolent God but the amp guys here keep telling me that it has to do with science. Go figure!

I am a "fly by the seat of my pants" DIY hobbyist and if it were me I would add a AC30 style Cut control (which is great for matching your amp to whatever speaker you will be using) and if it solves the immediate problem be done with it for now. If you don't want to add or repurpose an existing hole for the control you can use use a trim pot for now or wire up a regular pot and wrap it with Scotch 33+ electrical tape. But I am assuming that as you use the amp you will want to make fine tuning tweaks, etc., and will be pulling the chassis out quite often as I would.

Good luck! When I saw these threads popping up about building amps with reverb and tremolo I did not realize that there were kits to build them priced very reasonably these days... half the retail price of a DRRI. God bless China!

Most of the old timers here would have never thought of trying to build a BF/SF amp with reverb and vibrato from scratch - especially as a first build!

As I think you might have suggested in one of your posts it would be extremely helpful for the vendor or affiliated user group to post detailed pictures showing the exact lead dress for successful builds including the wire under the boards as it can be extremely critical especially as we push these amps with sound and voltage levels never intended by Leo back in the 60's.

BTW one point that might not yet have come up in this discussion is that you sometimes need to shield the larger signal leads later in the circuit to keep them from broadcasting crap that is picked up by the more sensitive circuitry in prior stages. I like to stock a good amount of green waxed pushback wire for that purpose... to at least temporarily wrap around wires that I suspect might be causing interference in previous stages. If you ever do that you can try connecting it to different ground points to see if it makes a difference.

And don't forget about positive and negative feedback... each gain stage inverts the signal so if there is parasitic inference with the immediately previous stage you have negative feedback which will "calm" the signal to use the technical term. It is when you go back two stages (or any even number) that you can get positive feedback which can result in those nasty oscillations the amp guys keep talking about.

Steve A.

P.S. Are the supplied eyelet boards traditional fiberboard/fishpaper/whatever-it-is-is-called-these-days or something more modern? Just wondering...

[Leo_Gnardo]

Are the supplied eyelet boards traditional fiberboard/fishpaper/whatever-it-is-is-called-these-days or something more modern?

Depends on who is supplying the kit. Mojo AFAIK to this day still supplies Fender style kits with black cardboard - to be avoided at all costs! Ask Randall. Kit buyers should make sure they're getting fiberglass/epoxy or some other non conductive material.