5879 SE redo

[firemedic]

Okay, I'm going back in.
My little SE 6v6 amp with 5879 input and paralleled 12AX7 triodes for 2nd preamp stage, well it has a little unwanted fuzz riding over the notes. Otherwise sounds great. I built it with screw-down terminals, so I'm gonna start there and use solder terminals.
I'm also going to change the paralleled 12AX7 to the input, into a FMV stack, to the 5879 (low gain setup), to the 6V6 output.

Any special considerations concerning the load impedance (or slope resistor, or what have you) of 1) the paralleled triode to the stack and 2) the 5879 to 6V6 output? I see that some like to insert a CF after pentodes, is there that great of a Z mismatch?

[jojokeo]

It's all up to you and your individuality & creativity. I'd suggest using the same values as a starting point and get a nice direct comparrison from what you had to judge by. You can always change values around later if tweaking needs to be done.

[firemedic]

Here is the 1st version....

[firemedic]

Here is the new version.
It still may need a little tweaking. I moved the parallel triode stage to the input, & switched the grid of one triode. In the old version, the amp sounded sort of lame with just one triode, so I had to keep the "fat" switch on all the time. Now, both positions provide a usable tone and are nice & chimey.

I reduced the gain of both PA stages and increased the output load on the 5879. I wasn't sure what level the 6V6 likes so I guessed on the 150k grid leak there. It seems to have helped, along with a 4.7k grid stopper. I must have had some blocking distortion before.

Also beefed up the power supply.

On this project I tore almost everything out and started over with new terminal strips, it took me all day and when I finally fired it up-
Nothing was wrong. It worked the first time. Minimal hum, good tone. Color me surprised.
So, I'll post voltages as soon as I'm satisfied this is the end version. They do seem quite a bit higher now, is it the higher value main cap perhaps?

[tubenit]

it has a little unwanted fuzz riding over the notes

Here's what I would personally try to resolve that issue.  Just tossing out another idea to consider?

With respect, Tubenit

[jojokeo]

Hi tubenit,

I just was curious to say that with the two gain stages going into the CF/V1b w/out any kind of signal attenuation (especially from a pentode too), it seems that there could be some possible blocking distortion when the dials are getting maxed? I'm sure you've used something like this on your TOS or COS but I wonder if there's something about a CF that prevents this by it's nature? (Merlin???)

[tubenit]

some possible blocking distortion when the dials are getting maxed

My guess is that there could or even would be?  If I were going to build this, I'd start with the schematic I drew up and then begin experimenting with it to see what was needed? 

A 500k volume might be a better choice or a larger grid resistor like 220k bypassed with a 500p going into the pentode?

The fact that it has a volume going into the pentode/CF and a tone stack w/master volume coming out should help control some concerns.

I am personally convinced that the CF following the 5879 helps with smoothness in the tone.

With respect, Tubenit

[jojokeo]

I am personally convinced that the CF following the 5879 helps with smoothness in the tone.

Do to your help and conviction w/ this, I have a similar design ready to build w/ my first opportunity. I is paralleling the first two stages w/ a 12ax7 & an EF86 and also cascading the triode into the pentode but was to have a tone stack then going into the remaining triode. But, now all I have to do to switch the triode & tone stack and whallah, I've got the pentode-CF going.  :smile:

[ncusack]

If anyone gets this little amp up and running Id love to hear some sound clips. I have an old champ build that could very easily be switched over to something like this.

[firemedic]

I'll try it- the new design still has that blocking distortion- I think. Does it kind of sound like paper being balled up when notes are played? I think the 150k 6V6 grid leak may be sucking a little too much tone out.
Anyhow, the tone is pretty decent with a single triode input stage. So that frees up the other triode for a CF & I'm sure it will help with my issues. I had thought of it before but I REALLY wanted that paralleled input. Time to let it go.

[firemedic]

Whatever the problem was, the CF fixed it. Grossly mismatched 5879->6V6 impedance, probably. Thanks tubenit! I'll work the CF into my TOS at some point.
I didn't follow your schem exactly, a couple of uF/R values were different. The 1M volume pot is a good value. If I want to increase headroom I'll pop in a 5751 or 12AY7, or even 12AU7.

With a 12AX7 this is a real growler when turned up, with very smooth overdrive. Real nice clean tone at middle volume. And every tone setting sounds good- without the CF it was kind of dull with treble down and harsh with treble dimed. Now it sustains forever, even clean, and the attack is hard to describe, chimey I guess is the word. And I get some nice feedback out of my little 8" speaker! Imagine that.

[tubenit]

Whatever the problem was, the CF fixed it. Grossly mismatched 5879->6V6 impedance, probably. Thanks tubenit! I'll work the CF into my TOS at some point.
I didn't follow your schem exactly, a couple of uF/R values were different. The 1M volume pot is a good value. If I want to increase headroom I'll pop in a 5751 or 12AY7, or even 12AU7.

With a 12AX7 this is a real growler when turned up, with very smooth overdrive. Real nice clean tone at middle volume. And every tone setting sounds good- without the CF it was kind of dull with treble down and harsh with treble dimed. Now it sustains forever, even clean, and the attack is hard to describe, chimey I guess is the word. And I get some nice feedback out of my little 8" speaker! Imagine that.

Wow!    Very cool.  Thanks for reporting back on that.  I'm a handful of hrs away from my mini-TOS being completed and I'm excited about it. Hope it works like I'm thinking it will.

With respect, Tubenit

[firemedic]

Yeah, I think so. 

[ncusack]

Any chance of getting an updated schematic with the new working values and voltages?

[firemedic]

Here is the new schem. I'm bad about voltages, I was aiming for @ 250v plate & @80v screen on the 5879, that's all I really paid attention to since the other stuff already worked. Note the changed power rail as well.
I think I have a little parasitic oscillation in there somewhere, probably poor lead dress. It's very slight though.

[ncusack]

I'm Not sure you posted the updated schematic. This "new one" doesn't appear to have the CF like Tubenits does.

[firemedic]

Ooops. I'm on the wrong computer to send a schem. Look for it tomorrow, sorry!

[firemedic]

Let's try that again. This is the latest version, I plan on going in today & placing a bright cap across the volume pot, and a PAB to lift the tonestack 680k(just guessing) off the ground. Also, I think tubenit is correct that a lower value volume pot will work better; at high volume it still has a little ugly distortion riding on top, but only when I hit the strings hard.
In fact, I will also do a large bypassed gridstopper on the 5879 for good measure.
I have not included a master volume so hopefully this will control any blocking distortion generated.
As usual, I'm leaning heavily on tubenit in the design of this thing.

[ncusack]

Right on thats looking more like it. After you make your adjustments let us know how she sounds. Im going to be doing mine as a scratch build using parts from an earlier 5F2A head build so having you work out some of the early tweaks will really help me get a good staring point.

[tubenit]

Since your 5879 has a 56k on the plate, try a 47k or 39k on the cathode of the CF & see if that smooths the tone even more?

I have about 2 hrs left on my mini-TOS to fire it up.

With respect, Tubenit

[ncusack]

Ok so using a previous schematic idea here is what my setup will theoretically provide. (see attatched schematic)

The voltages are what Duncan PSU spit out given 2.7mA to the CF and 5879 and 1.4ma to the first triode. The power tube current was stepped from 35ma-1.5ma on the plates and 2.9ma-35ma on the screens. Im not sure exatly sure what kind of voltage the 5879 will be seeing because im not sure of the current through each branch feeding it(68k and 330k). I know an ef86 typically draws like 1.5ma in examples like these so I used that. Let me know if the suggested voltages look like a decent starting point here.

[firemedic]

If you have the theory correct you'll prob. save yerself all the hassle of trial & error, which is my default method. I'll be making the changes tonite & I'll try to jot down the voltages too. Which reminds me, my voltages will probably still be higher than yours so your results will vary.
Tubenit, I will try a little lower cathode value on my CF in addition to the other stuff.

[ncusack]

Yeah my "theory" work is still only speculation thats why I was hoping someone else who knows the voltages and such could comment. I know the 12A_7  tube in V1 would be ok with the stated voltages its just what the 5879 should be seeing im curious about.

[firemedic]

As far as the 5879 voltages go, I really tried to keep them within max specs per the datasheet. Thus the somewhat big dropping resistor to D on the B+ rail.

I made the changes, had it all tied up then realized, looking at the 6V6, that the side of the plate had this red GLOW. And it must have been overbiased since day 1. So after some T&E I came up with a 220R before the main cap to drop the overall voltage @ 10VDC, and a 740R cathode resistor (270/470 in series). So it's now at max dissipation but not redplating.
I know I sould prob. use a 6L6 with these voltages but I like the 6V6 for this amp.

And I don't know if that is correct power supply design, anyone who believes it may not be, please chime in....

I used a 56k grid stopper to the 5879, bypassed with a 390p cap. Tried a 220k at first but it really sounded thin.

Increased the CF cathode to 47k.
I placed a 1M resistor across the outside terminals of the volume pot, I believe that makes it a 500k for all intents & purposes. And a 68p bright cap.

The "PAB" switch really fattens up the sound, but both settings have a pretty nice tone. I found the Mid resistor (6.8k to ground on the bass pot) used in standard Fender stacks drained the tone right out. Since I like a little mid scoop anyway, I settled on 22k for that value but could easily go higher if necessary.

Finally, I realized that I had no 6V6 screen resistor all this time! So note the added 1.5k screen resistor.
Aren't you guys sposed to call newbies like me out on this stuff?

[tubenit]

You may have it exactly where you want it? 

Having said that, when I experimented tone wise, I liked the 5879 to have 140-155v range on it's plates. And closer to the 140-145v range. That sounded warmer to me.

With respect, Tubenit

[ncusack]

I guess in my hap hazard design attemp here im looking to confirm the current draw values for a setup like this. I know a decent rule of thumb value for 12ax7 tube is like 1mA and for the 6V6 the plates should be around 35mA and screens like 3mA. The question im looking to answer is the 5879 CF section. Judging from Firemedics voltages that section is pulling around 6mA if I have my theory correct 2.6mA of that being the 5879 itself. Im just wondering if the numbers im using with the Duncan PSU are anywhere near what they should be.

[firemedic]

You may have it exactly where you want it? 


Well, maybe. It was a nightmare last night dealing with the redplating, it just went on & on. I'm tired of messing with it, too many times I've kept tweaking an amp for no good reason when it sounded great to begin with.
Now as I understand it, a little pentode's characteristics are controlled to a large degree by the screen voltage, which is a nice low 74v in this amp. I'm going to leave it alone since the blocking distortion seems to be taken care of, which was the issue I had in the first place.
Inevitably I will go in again & then I'll drop the 5879 voltage, since you are almost always correct in your long-distance assessments of my little projects.

Any thoughts on the 220R B+ dropping resistor, right after the recto? Think it will be stable? It gets warm but not too hot.

[jojokeo]

Im just wondering if the numbers im using with the Duncan PSU are anywhere near what they should be.

Any thoughts on the 220R B+ dropping resistor, right after the recto? Think it will be stable? It gets warm but not too hot.

Yes for both. If my math & thinking are correct 10v dropped from the 220r shows 0.045A going through it.
For minimal power rating of resistor: 401-391=10v, 10²/220=0.45W x 2 (for safety) =  0.91W rating to use. You "should" be okay using a 1 watter but if your 10 watter is warm stay w/ the less heated part. It should barely feel warm being this large.

[moonbird]

Firemedic --

What are the specs for the PT you are using in this amp? thx.

[firemedic]

It's the 125P1B from Doug's store. 

I turned it up last nite.....
You know how some amps make you really sound like you know how to play guitar?
The 5879 chime is unmistakable, even with treble turned way down. What a beautiful tone. It makes a big difference, having something driving that pentode vs dry guitar signal. 
It still needs minor, minor tweaks, I'm hearing just a trace of blocking distortion when it's turned up & the tonestack is lifted from ground (PAB). It's barely there though. Prob. just a 330k or so instead of 680k lift, will do it. As it is there is a HUGE gain increase with the lift. If that fails I'll try a 250k volume pot.
I'll also try a 33k mid resistor, per tubenit's schem; the stack sounds great but could use a touch more midrange.
It was hard to put the guitar down, it's a pretty nice feeling when that happens.

[ncusack]

Glad to here you've almost got things where you want them. Let me know how the tonestack sounds with the 33k mid since thats going to be my starting point I think.

[firemedic]

Go ahead & build it, the midrange resistor is about the easiest thing to change out. 33k is a pretty good bet IMO.
Hopefully you won't have to drop your B+ voltage like I did, I still feel weird about that big 220R even though it's most likely fine.
I'm going to see how it sounds w/ 5751 & 12AY7, just for jollies. It may fix the minor distortion problem until I open it up again.

[ncusack]

Yeah I need to make an order for parts before I can really get rolling so it really doesnt hurt to get my ducks in a row. I plan to throw in a few extra resistor and cap values to have something to experiment with when I do order. I think I have a good handle on my B+rail for a starting point but thats where a lot of the extra resistors will come in handy.

Speaking of the B+ rail im curious why both you and tubenit decided to feed the 5879 CF stage from the last filter stage on the rail instead of the one before it? Was that just because you didnt feel like switching all the dropping resistors around or did you want to use the smoothest DC for the CF? Im in the process of working out some of the different B+ possabilities now and I can get pretty caparable plate voltages across the board flipping the B+ feeds and using different dropping values. I'm just not sure if there is a benefit to using that final stage of filtering on the 5879 CF or not.

[firemedic]

The pentodes like lower voltage than 12A_7s, at least according to the datasheets. Especially the screens. That is, to get the "pentode sound" according to Merlin. So far my in my builds this has proven to be true.

[ncusack]

Yeah I've noticed that and from what i've read about the 5879 so far around 150 on the plate and like 75 on the screen is the sweet spot. My setup will be a custom build so i dont need to worry about working around a prebuilt circuit. Right now using the Duncan amps PSU simulator I can get voltages that look like this.
           


This is using a 32mA current to the 6V6 plates 3mA to the screen. The 5879 CF section is assuming a 5.6ma draw and the first preamp stage the usual 1.2mA for the 12ax7. Assuming a 2.2mA current to the plate of the 5879 and a 68k resistor would put 145V on the plate and a 0.44mA screen draw across 470k puts 75V on the screen. Im not 100% on the CF triode but I estimated a 2mA draw. The firstat satage would use an 82k putting 194V on its plate or 100k for 173V. The 100k 173V is reallt close to the old champ rated voltage of 150V.

Now all of this is speculation but I think my numbers make sense. Im really just hoping someone will cut me off im way off the mark.

[ncusack]

Ok so heres a quick Hoffman style layout for what im planning. This should be a simple layout to modify values. Do me a favor and check if I missed anything.

[firemedic]

A couple things-

I would get some smaller value caps to experiment for the treble bleed snubber across the 5879 plate resistor- esp without a bypass cap on the V2 grid you might kill your treble. This is not a High Gain design & you may not need to bleed off the highs at all.

The 5879 datasheet suggests connecting p6 & p2 to ground for additional internal shielding, probably not much difference.

That's all I see at 1st pass, have you ordered the parts? Keep in mind, you are almost certainly going to have to tweak some values to get this sounding like you want it....

[ncusack]

No parts have been ordered yet. What sections of the circuit did you find yourself tweaking with the most and ill graba few values above and below when i place the order. Thanks for the pin 2&6 info for the 5879 I hadnt checked the datasheet yet but was curious what was done with those pins. I noticed you had a treble bleed on the 5879 grid on your build so I intended to grab a few values for the grid resistor. If you can make value suggestions I'd greatly appreciate it.

[jojokeo]

I can't see the schem (no jschm program loaded) on my fancy electronic device but you have power going to v1 pin 6 but the signal isn't going anywhere? Wait I see, it appears you need a wire connecting tag board 22n/630v to 82k/2w. You may only need a 2n2 cap instead here.
The grid resistor 68k is pretty high, consider 22k, 10k, or you may not need anything here? Anything less the full up on gain pot gives you a grid stopper value.
Interesting descision on panel/pot layout order? It seems you are missing a filter stage. With a single ended amp, you'd be wise to follow prior examples using 4 filter caps w/ a resevoir>resistor>then 2nd cap for node A>resistor>3rd cap/node B>resistor>4th cap/node C, etc... check other single ended amp's power sections to see what I'm talking about.
For more values to sub, you may want something different than the 820 on 5879 like 1k or 1.2k? You may want different values on 5879 like 56k/270k or 220k combo (the higher here the more headroom available for pentode). Get a few small caps for treble bleeders on your vol pot - like 100pf, 250pf, and 500pf values to tailor more clarity on low settings, especially if using humbucking pu's.

Hope this helps,
Keo

[firemedic]

Oh, yeah, I really like a choke on my PS's, BTW. I used a small fender choke from Doug on mine. Nice quiet build, for a SE.

I have procured Merlin's preamp book, thus I have this idea that (among other things I only half understand) the 5879 grid can have a fairly high resistance on it before affecting the tone too much. I tried a bypassed 220k which WAS too much, then settled on a 56k bypassed with a 390p cap.

Note that it is a bypassed resistor (high-pass) rather than bled to ground or B+ (low-pass) Keo is probably right but I'm happy with the sound.

As far as tweaking, I tweaked EVERYTHING on this build. Resistors are cheap, get a bunch of different values & don't forget 1 & 2 watters for your plates/PS. Plus various 10Ws for your power tube bias. Ditto caps. Your calculations are awesome, & you're likely right on the money with yer values but your ear is not a calculator.

Find a local electronics supply store besides Radio Snack (good for fuses, jacks & solder in a pinch), you will be going there a few times for common value stuff you forgot to order; xicon caps are real common & work fine. They usually carry silver mica caps too. Mallorys, ODs, not so much. You don't want to be stuck, one little resistor away from firing up your creation.

I agree w/ Keo that your layout is unusual, consider giving the 12AX7's filter cap a little more room. Also take another look at the bass pot connections, somehow that doesn't quite look right.

PS I'm real flattered that you chose my topology -really tubenit's- on this project, I'm new at this too.

[ncusack]

So you guys are suggesting I do a PSU setup like this?

Cap - Choke - Cap to OT - Resistor - Cap to Screen - Resistor - Cap to 5879 CF - Resistor - Cap to 12ax7

I can get my hands on a hammond small fender choke replacement pretty easy so Maybey Ill adjust for that. As far as the layout goes how much would you say a hoffman style layout would really make in a single ended amp like this?

[jojokeo]

Cap - Choke - Cap to OT - Resistor - Cap to Screen - Resistor - Cap to 5879 CF - Resistor - Cap to 12ax7

Either a choke or power resistor - not a real big deal in a small amp but it's better to not follow Fender's Champ style power rail by going stright to the OT w/ the first resevior cap node on a single ended design.

As far as the layout goes how much would you say a hoffman style layout would really make in a single ended amp like this?

How much what? "Difference" would it make? As compared to what? This needs clarifying.

BTW, what will make a big difference as far as noise/humm is concerned is where & how you run and center tap your heater wires, your grounding plan, and lead dress. I don't think these were mentioned?

[ncusack]

In terms of "Difference" I meant compared to a typical non hoffman design.

Grounding I intend to split the Power section grounds and preamp grounds to seperate chassis points. For example the resevoir cap the OT cap and screens filter cap would be grounded at one point along with the power tube cathode and the other grounds would be somewhere on the other side of the chassis. The HT center tap would join at the power section ground and the heaters would be atrificially centertapped with 100R resistors at the positive side of the powertube cathode to provide a DC level. The pre grounds I would likely join somewhere near the input and the power ground would be near the power transformer.

[jojokeo]

sounds like you'll be fine

[firemedic]

Ditto. Plug in your iron, let's see what happens.....

[ncusack]

Hey Firemedic how would you describe the tone of your 5879 SE build? Is it a Chimey distortion or more creamy sounding?

[ncusack]

Ok so I think I have the layout set up as clean as I can make it let me know what you think. I didnt draw in a few basic connections but I believe the meat nad potatoes are all there. If this one looks cool Ill fire an order off monday for parts. Thanks again for all the help and inspiration on this one. 

[jojokeo]

At a quick glance it all looks good except remove the 3k3 grid stopper to the 6v6 from the board and attach it directly to the 6v6 pin socket w/ a lead back to the board.

As for chime, I have a 5879 pp using 6bm8s and it's very chimey w/ Weber blue speaker. I noticed that you have a 33k mid resistor. This will give you more mids and help to get more smooth and full tone. Smooth and creamy seems to have a direct relationship w/ the signal containing a lot of mids and high mids - aka Tweed-type of tone. A certain type of speaker as well as humbuck pu's also helps too if that's your goal for "smooth & creamy".

[John]

I've got a question for my own interest. I notice on the 5879 you have 2 and 6 connected externally, and then going to ground. Is that the way it should be done? If so, I've made (one more) mistake with mine.

Thanks,
John

[jojokeo]

no, you're fine - these are "no connection" see data sheet