Adding EF86 Channel to Hoffman AC30

[fiftynine]

Gents,


I'm building a Hoffman AC30 and want to add the EF86 channel from the AC30/4. So that's Normal, Top Boost and EF86 all in one AC30.


But I read you can't just keep adding channels via 220k mixing resistors as it screws things up http://www.valvewizard.co.uk/resistive.html


If the EF86 channel had a switch just prior to the 220k mixing resistor (to take it out of the circuit while not in use), I'd have the stock Hoffman AC30. But I'd need another switch to take the Top Boost circuit out while using the EF86 to simulate the AC30/4 with Vib-Trem switched off. Also, I'd have to be continuously aware of only having 2 channels (2 x 220k's in circuit) at once.


Any ideas? Or am I worrying about nothing? i.e. just stick all 3 channels through 3 x 220k mixing resistors.


Cheers. I'm keen builder but a novice.

[sluckey]

I would just connect the EF86 channel to the unused input of the phase inverter just like the AC-15. No need for another mixing resistor. Hoffman's board layout makes this easy. See pic...

[kagliostro]

This may be one other way


K

[terminalgs]

I would just connect the EF86 channel to the unused input of the phase inverter just like the AC-15.

+1.

[terminalgs]

Here's an idea for the EF86 that came up in another recent AC30 thread:





The top channel has two inputs: the ef86 by itself, and then another input to feed a triode to drive the ef86 for more total gain. The bottom channel is the regular top boost channel.  None of it is really drawn in any full detail, but thats the idea.


See valve wizard's pentode preamp page for details about the ef86.  I've used his suggested ef86 design with great results (much better results than using the design found on original vox ef86 designs, like an AC4 or AC10).  I found that a load of 1.5M to 2M worked well with a Merlin's ef86 (although I used a 68K plate  resistor instead a 47K, based solely on part of hand availability).

[MakerDP]

See valve wizard's pentode preamp page for details about the ef86.  I've used his suggested ef86 design with great results (much better results than using the design found on original vox ef86 designs, like an AC4 or AC10).  I found that a load of 1.5M to 2M worked well with a Merlin's ef86 (although I used a 68K plate  resistor instead a 47K, based solely on part of hand availability).

What do you like about Merlin's design over the Vox for the EF86 channel?

[terminalgs]

 

What do you like about Merlin's design over the Vox for the EF86 channel?

I built an amp using Vox's EF86 design.  The amp squealed at high volume settings and it would randomly misbehave at other settings.  At first, I did what a lot of people do: blame the tube for being microphonic.

The amp I built, by the way, was an AC4. The first build was 100% stock, component by component, matching the original.  Besides the squeal, the amp was too quiet, it was 3W --maybe 4W-- but no more than that.  The next version of the amp  went:  input-> 12AX7 with a gain stage and a DC-coupled CF ALA a top boost -> vol pot -> ef86 -> el84.  the stabliity issues continued.  I tweaked grid-stoppers, I tweaked the EF86's screen-grid supply and the screen-bypass cap (orig used a 1M resistor, I played with various R's, including higher ones like the 4.7M that the recent AC30HW used, also tweaked .1uf to smaller and larger values...)

Anyhow, I never got the EF86 to behave until I read Merlin's article.  The key (as he points out) is lowering the gain the EF86 produces. I speculate that perhaps Denney originally used some Philips or Mullard's design sheet components AS-IS with a 220K plate-R and a 2.2K cathode R.  These might have worked well with low dBV inputs like a phonograph or a microphone.   I also played with different AC signal loads (post coupling cap) and found the best performance was with a total of 2M (1.5M + a 500K level pot).   Of course, the EF86 by itself (even in the higher-gain trim of the original Vox design) doesn't produce enough umph to make it a fun guitar amp, so I still employed a 1/2 12AX7 for added gain.


In all of the previous failed attempts, I tried 3-4 different EF86's, include new ones and an NOS Valvo.  they all acted the same...


The next amp I'm planning is a Marshall 18W using that top channel in the above posted design.  I'll see how it goes, and post results here (of course!)

[MakerDP]

Thanks for the explanation.

I was wondering because I built a 15 watt clone of a Matchless DC30 and thought maybe I could tweak it a bit. But it sounds like maybe that would be a bad idea. The EF86 channel has none of the issues you described in your build.

[fiftynine]

Thanks very much for the interesting suggestions. I will try them out and post up how I got on. It was bnwitt's comments 5 posts down here that also got me thinking.

To add to terminalgs's post, I've never played through a Vox EF86 channel that didn't squeal at high volumes and always blamed the tube. The DC30's have been fine.

[MakerDP]

To add to terminalgs's post, I've never played through a Vox EF86 channel that didn't squeal at high volumes and always blamed the tube. The DC30's have been fine.

The difference between the DC30 and the AC30 is the plate resistor (330k Matchless vs 220k Vox) and the pin 9 grid resistor (2.2M vs 1M). Funny how such seemingly tiny changes make that much of a difference.

[fiftynine]

Sluckey, I've been working to the schematic attached with the AC30 power stage. It seems this particular AC15 doesn't have any mixing resistors. Normal/Vib-Trem and EF86 just straight in via C27 and C28 like you suggest with my EF86. I'm confused as Hoffman adds the mixing resistors. If they're just to minimise channel interference but are inefficient, I'm thinking a rotary switch just before the PI for all 3 channels to meet.


So what, you can't run two guitars but who does that nowadays anyway?


Or, do the inefficient mixing resistors give less pre-amp gain and more power amp clipping? You know, what everyone raves about.

[sluckey]

Leo used two 220K mixing resistors in most of his two channel Fender amps. That's probably why Hoffman chose to use two resistors. There's nothing inefficient about using them and they do provide adequate (not perfect) isolation between channels.

That AC-15 schematic is the one I used to build my AC-15. It has two preamps and the PI has two inputs. Just seems logical to mix the channels using the two PI inputs. It's very efficient and the channel isolation is nearly perfect (not just adequate).

But you want THREE preamps. You can certainly use a 3 pole rotary switch to select only one preamp at a time without using any mixing resistors. But doing so you will give up the ability to mix channels together or use an AB/Y footswitch to select one, or the other, or both preamps.

Or, you could use three mixing resistors. Lot's of easy ways to do it. Personally, I'd do it exactly as I showed above.

[kagliostro]

Merlin dedicated some time to talk about mix resistors


may be worthwhile to give it a look



http://www.valvewizard.co.uk/resistive.html


K

[sluckey]

Merlin dedicated some time to talk about mix resistors


may be worthwhile to give it a look



http://www.valvewizard.co.uk/resistive.html


K

That article was the reason for this discussion.

[kagliostro]

My memory has always more laks 




Franco

[jojokeo]

Interesting comments regarding ef86 experiences. For me it's been the opposite when it comes to load resistors, that is I lower my values to lessen gain which addresses microphonic issues. I've combined them w/ 12ax7s & another ef86, & so on but my favorite is into a cf. That will get you going to match your other channel idea (in signal level) & I've used voltage dividers on the outputs to keep things under better control. No matter how much you plan this out you'll have to tweak it once it's built so don't spend too much time worrying about it.

[roseblood11]

Here's an idea for the EF86 that came up in another recent AC30 thread:





The top channel has two inputs: the ef86 by itself, and then another input to feed a triode to drive the ef86 for more total gain. The bottom channel is the regular top boost channel.  None of it is really drawn in any full detail, but thats the idea.


See valve wizard's pentode preamp page for details about the ef86.  I've used his suggested ef86 design with great results (much better results than using the design found on original vox ef86 designs, like an AC4 or AC10).  I found that a load of 1.5M to 2M worked well with a Merlin's ef86 (although I used a 68K plate  resistor instead a 47K, based solely on part of hand availability).

Very interesting! Do you remember all the parts values that you used?

I'm planning to add a similar EF-86 channel to my AC-30 TB built. The jacks will be wired like it's done in the JCM800: https://elektrotanya.com/PREVIEWS/63463243/23432455/marshall/marshall_jcm800_2204-1-50w.pdf_1.png

regards, Immo