Designing / Building a custom amp, couple of questions...

[markbuckingham]

Hi-

I'm working on designing a new amp for myself...  This is either the 5th or 6th one I've built, but it's a bit more complicated the other ones.

Basically, it's two channels, channel 1 is an ef86 powered channel like a Rt 66.  Channel 2 is basically a marshall jcm 800.  Both channels will go through the same weber-style effect loop, and then spring reverb, with separate reverb level controls for each channel.  From there, it's going to look like the back end of a princeton reverb, including the phase inverter and the stage before it.

So, here are the questions:

1) I've read a few places now that you should use insulated jacks for better grounding, but I have a deluxe reverb chassis that has input holes that are too small for the insulated jacks.  I can ream them out if necessary, but I'm kind of wondering if it really *is* necessary.

2) Channel switching.  What I want to do is use relays to switch the channels, but I want to avoid popping.  I have to be able to switch preamps as well as which one of the two reverb level controls is active.  I was thinking I could do this one of two ways:

a) relay contacts right after the input jack, signal gets switched to the right preamp circuit.  The two channels will need a second set of contacts at the end of each channel before the effect loop.  Finally, I'll need a set of relay contacts to pick the right reverb control.

b) like 'a', but instead of the switch right after the input jack, I can always feed both channels with the guitar input, and maybe up the first stage gains on both channels to compensate.  (I'm thinking that this might save on popping because the only switching is later in the circuit.)

3) Would I be better off with a long tail inverter for any reason?  I like the sound of the concertina in my princeton, but if the long tail is better, I'm good with that too.

Any help is appreciated, thank you!

[tubenit]

Just some thoughts and ideas to consider.  PLEASE build the amp however you want, I have no agenda for you.

I would use a 5879 pentode (which has fewer microphonic issues) with another gain stage instead of the single EF86 into the FX & LTPI.
I think by doing this, there would be less of a volume drop from the JCM800 to the 5879 channel.  BTW, the 5879 & the 12AY7 is from Geezer's HoSo56 design which I consider a proven design.  (you could use a 5751, 12AT7 or 12AX7 for V4)

I personally would not isolate the input jacks.  I've never had a need for that.

I personally would use an LTPI.  I think it's a little easier to add reverb with an LTPI & have it sound like I want.  (I have done both LTPI and concertina)    And I think a concertina phase invertor is going to leave this amp with a lot of overdrive and grit.  I think the LTPI does a better job with a cleaner sounding overdrive.  Just an opinion.

Regarding the LNFB around V5-b.  I personally don't care for that and don't have that on my amps.

The one tube reverb is a good sounding reverb, IMO ................ IF you are not looking for surf reverb and prefer a subtle reverb flavoring.

The ExpressSCH schematics have a B+ rail on the second page to consider.

With respect, Tubenit

[markbuckingham]

Just some thoughts and ideas to consider.  PLEASE build the amp however you want, I have no agenda for you.

Wow- that was the most comprehensive answer to any question I've ever had  :-)  Thanks!

[tubenit]

Now if you did NOT mind using a mosfet CF (cathode follower), then you could reduce the preamp tube count.  I have found NO loss of tone using a mosfet CF vs. a triode CF ............. and I have done both.  My D'Mars ODS is currently using a mosfet CF.

This idea would be the JCM800 for crunchy sustained.   And the HoSo56 for a harmonically rich, chimey & touch sensitive approach which you could use for the "clean-ish".

It should be noted, that because of the CF,  you can simply use a triode recovery for the active FX.  Again, this is what I am doing with my D'Mars and I really like the tone I am getting.

Of the 3 ideas, I've posted here ............... this would be my personal 1st choice.   Just another option to consider.

With respect, Tubenit

[markbuckingham]

Looking at the valvewizard website, he mentions putting a triode in front of the pentode to help with microphonics. 

What if I used the the jcm800 side's first triode stage, after the .02uf plate cap, but before the gain control, fed that to a relay, then had the relay switch between the rest of the jcm800 circuit (i.e.: gain control then second triode) and the ef86?  I *think* I would just have a voltage divider before the pentode instead of a gain control, keeping the rt66 tone stack and volume control afterwards.

Does that sound workable?

[tubenit]

I think you should build this amp exactly how you want it.   Having said that, my experience with the 5879 pentode has been excellent whether it was in the first gain stage like in the HoSo56 ............ OR the 3rd gain stage like the TOS or the D'Mars. 

I do NOT have any experience with the EF86 pentode, but simply have read about lots of issues with microphonics so I have never tried one?   I've used the 5879, 6Au6 and 5654 all with success.

Tubenit

[markbuckingham]

I think you should build this amp exactly how you want it.   Having said that, my experience with the 5879 pentode has been excellent whether it was in the first gain stage like in the HoSo56 ............ OR the 3rd gain stage like the TOS or the D'Mars. 

I do NOT have any experience with the EF86 pentode, but simply have read about lots of issues with microphonics so I have never tried one?   I've used the 5879, 6Au6 and 5654 all with success.

Tubenit

Fair enough.  I have a new ef86 handy, so I'm going to go ahead and try it.  If I have too much trouble with it, I'll put a 5879 in there.

Anyway- I've redrawn the PI to be a long tailed pair, and made my reverb circuit look more like you had.  Thanks for your help!

[tubenit]

Do you have a schematic that you're willing to share/post?

With respect, Tubenit

[terminalgs]

1) I've read a few places now that you should use insulated jacks for better grounding, but I have a deluxe reverb chassis that has input holes that are too small for the insulated jacks.  I can ream them out if necessary, but I'm kind of wondering if it really *is* necessary.

Its all about eliminating a ground loop, or keeping noise that might be in that part of the chassis out of the pre-amp.  If you use Hoffman's board and follow his grounding scheme, insulated jacks seem to be unnecessary (I say that based on the years of reading this forum and lots of reported success).

If you do your own grounding scheme (like a star or something like that), I'd open the holes now for the possibility of switching to a isolated jack if needed.   Personally, on new builds, I use them, but I also lift the circuit ground off the chassis (ala any modern Vox).




regarding channel switching:  there is no need to switch anything before the first preamp stages.  let all channels triode/pentode grids get instrument signal all the time.  there is nothing wrong with a triode (or pentode) amplifying a signal that you later chose not to use.   putting switches, be them relays or toggles in front of the first gain stage is an open invitation to switch noise, popping, etc.  Your instrument signal is a sensitive/delicate 50mV to 500mV, pretend it is an orchid. 


you can do all the switching further down the line.   Don't set up any switches that leaves a cap (that sees instrument signal) open ended (i.e. without  a connection to anything on one end).  always give it a path to ground (5K-1M depending on the location and cap...),  this avoids  DC build up and reduces/eliminates pops on channel switching.

regarding an EF86:  I like the tube, and I've built a few amps with it.  I haven't had problems with microphonics.  I have had problems using commonly found EF86 circuits whereby the circuit was not well suited for guitar, produced too much gain, and was generally unstable (I suspect people mistakely blame the tube as being microphonic,  but its just that the circuit isn't well suited for guitar).    If you follow Merlin's EF86 advice for guitar Ef86 stages, you'll be okay.  If you follow the schematics of old Voxs you'll have problems. (old Voxs and the euro phonograph amps that used the same the EF86 circuits as the Voxs --- those designers perhaps used example design whitepapers released by Phillips/Mullard at the time?? ).


On a jcm800 (say a superlead 1959-std?),  use an EF86 (with Merlin's guidance) in place of V1a.  Instead of a 1M load (the vol pot), start with a ~2M load (1M resistor on top of a 1M pot,  or a 1.5M resistor on top of a 500K pot).  play with  the total load and the pot value to balance (1) stability and gain of the ef86 and (2) exit-gain to the 470K mix resistors.    in my experience, the EF86 likes to see a big impedance for the load, that's why I suggest 2M.   from there on, build the jcm800 pre as-is.  The tonestack follows V2, so no worries regarding the EF86.

3) Would I be better off with a long tail inverter for any reason?  I like the sound of the concertina in my princeton, but if the long tail is better, I'm good with that too.

dealers choice. depends on what you  want in terms of gain and/or clean headroom etc.

[markbuckingham]

Do you have a schematic that you're willing to share/post?

With respect, Tubenit

Sure- here you go...

[markbuckingham]

Its all about eliminating a ground loop, or keeping noise that might be in that part of the chassis out of the pre-amp.  If you use Hoffman's board and follow his grounding scheme, insulated jacks seem to be unnecessary (I say that based on the years of reading this forum and lots of reported success).

If you do your own grounding scheme (like a star or something like that), I'd open the holes now for the possibility of switching to a isolated jack if needed.   Personally, on new builds, I use them, but I also lift the circuit ground off the chassis (ala any modern Vox).

That sounds like a good idea.  I have both types of jacks either way, so I might just use the insulated ones and be done with it. 

I've thought about the floating ground thing too, but I was wondering if there were any safety issues involved with it?  (Obviously I'd still ground the chassis itself to the ground lug on the plug.)

regarding channel switching: 

(snip...)

That all sounds good, I'll adjust my design.

On a jcm800 (say a superlead 1959-std?),  use an EF86 (with Merlin's guidance) in place of V1a.  Instead of a 1M load (the vol pot), start with a ~2M load (1M resistor on top of a 1M pot,  or a 1.5M resistor on top of a 500K pot).  play with  the total load and the pot value to balance (1) stability and gain of the ef86 and (2) exit-gain to the 470K mix resistors.    in my experience, the EF86 likes to see a big impedance for the load, that's why I suggest 2M.   from there on, build the jcm800 pre as-is.  The tonestack follows V2, so no worries regarding the EF86.

I thought about doing that...  It would have the nice side effect of freeing up a triode to use later.  So- I could feed the ef86 from the front end, and put my channel switch after the plate cap and have that feed the channel switch?

Thanks!

Edit; Untangled quotes. Willabe

[tubenit]

I've tried opening the schematic on two different computers with NO success.

With respect, Tubenit

[Willabe]

I get them both to open but the preamp has no wires connecting the components.


                Brad   

[markbuckingham]

That was weird...

Anyway, here are the schematics as .png files.  I've updated them since the last time I was on anyway, to have the ef86 in front of both channels, like terminalgs suggested.

Thanks

[Willabe]

Um, I don't see them.


                 Brad   

[markbuckingham]

Um, I don't see them.


                 Brad   

I seem to be having technical difficulties.  Here they are.

[sluckey]

That's an odd looking (to me) tone stack on the top. Kinda reminds me of a James stack.

The reverb driver tube will require a grid leak resistor.

[markbuckingham]

That's an odd looking (to me) tone stack on the top. Kinda reminds me of a James stack.

The reverb driver tube will require a grid leak resistor.

Yeah- it's actually a rt66 tone stack.  I've used in an another amp I've built, and it works pretty well with the pentode.

I've added the grid leak resistor now on my working copy of the schematic.  (I figure there's no reason to repost that)  Thanks!

[tubenit]

Interesting amp & that could be a real winner.  Thanks for sharing the schematic.

If you are not happy with how the reverb turns out, I would move it closer to the entrance of the LTPI either before or after the FX loop.

I think you may be overdriving ALOT of reverb tone.  Instead of reverbing an overdrive tone.

with respect, Tubenit

[markbuckingham]

Interesting amp & that could be a real winner.  Thanks for sharing the schematic.

If you are not happy with how the reverb turns out, I would move it closer to the entrance of the LTPI either before or after the FX loop.

I think you may be overdriving ALOT of reverb tone.  Instead of reverbing an overdrive tone.

with respect, Tubenit

Good point, thanks.  I'll see if I can move things around a bit.