Well, my first build is finished (yay!) and sounds pretty good, but I'm curious as to what an EF86 stage (in the context of a Vox AC30-ish amp) should look like....that is, what voltages are appropriate for this tube? Right now, I have 111v on Pin 1, 2v on Pin 3, and also 111v on Pin 6. I've heard that EF86s are super senitive/responsive to voltage changes, and that little changes make a big difference in how the circuit responds. For the sake of context, this stage is setup identically to a stock AC30/4 circuit, with 220k to the plate, 1M to the grid, a 2k2 and 25uf on the cathode, 8uf for a filter cap, and all this is sitting after a 22k dropping resistor coming from a Mercury Magnetics Woden clone tranny. I hope that gives enough context...
Oh, the tube is a used old stock Telefunken I grabbed off of Ebay. I have three others just like it, but in better shape. I figure I'll use the "more used" one first, while I tinker.
Wish I could help out Dave, but I don't know an EF86 from an F86. Ones a jet I know and the other could be just about darn near anything.
Dan
The Last of the World's Great Human Beings
Seek immediate medical attention if you suddenly go either deaf or blind.
If you put the Federal Government in charge of the Sahara Desert, in five years time there would be a shortage of sand.
Have a look at the tables on pp. 2&3 in the EF86 datasheet for component values at different supply voltages etc. There's even a table for using EF86 in triode mode which is quite interesting to try.
Have fun
Aleksander Niemand
------------------------
Life's a party but you get invited only once...
affiliation:TUBEWONDER AMPS Zagray!-review
Here's an Ampage thread I saved. It wavers on and off topic.
START NEW THREAD REPLY Top of Thread LIST
Read 315 times
From: Matt
Date: 1/4/2006 4:11 PM
Subject: EF86 - Reloaded
Hopefully the 'sequel' to my first EF86 trials will be better than the Reloaded sequel was to the original Matrix.
I tried the 'typical' EF86 circuit (220k plate, 1M/.1uF screen, 2k2/22uF cathode, 1M grid load, etc.) and it never sounded very good to me. Very little headroom and way too sqishy/chewy for me. The 'best' circuit I came up with (kg provided most of the details for this, I still have the post if anyone is interested) was this: 100k plate, 2k2/22uF cathode, the screen was connected to a voltage divider of 270k to B+ and 820K to ground and the .1uF screen cap was connected to the cathode (not ground). The grid load was 470k and the output was a .0047uF cap and 500k-A volume pot.
In any case, anybody have any other tricks/circuits to try with an EF86? My B+ at this stage is only ~350vdc, this will be the second channel of a 2-channel amp and will be primarily for cleaner tones. The EF86 will be the only gain stage before the PI and I was going to try the Route 66 tone stack. Any input appreciated. Thanks.
Matt
START NEW THREAD REPLY PREVIOUS LIST
Read 298 times
From: Matt
Date: 1/4/2006 4:15 PM
Subject: Re: EF86 - Reloaded
Just because it's good info, here's the post by ken:
i've messed with cascaded pentodes a bit.
one thing to keep in mind is that the character of the distortion will be more organic, a bit less compressed, with a bit less sizzle
to it, when using say one or two cascaded pentode stages vs. two or three cascaded triodes.
the reason boils down to the number of stages... every time you go through another stage you get a little sharper clipping of the waveform. one stage, massively overdriven, usually has a LOT of even harmonic distortion--looking at the waveform you'll see one half fairly accurately reproduced (usually the +ve half, when vg1 is going -ve), whereas the other half has been clipped hard (usually at the grid when vg1 is goive +ve and passing Vk, leading to grid clamping). now, futzing around with the quiescent point can alter these relationships--for example, using a low value of Rp and not dropping much voltage across it will cause the +ve plate half to be hard clipped at the b+ rail earlier--but generally you'll find the stage giving you one half of the waveform a lot cleaner than the other.
as a result, when you cascade two of these such stages, you end up with something resembling a square wave, since due to the polarity inversion each stage clips a different half. as you cascade your way through more of these stages, the wave becomes more and more squared off. as you can guess, using fewer pentode stages vs. more triode stages will lessen this effect.
the other thing about pentodes is that their operation at the extremes (like in guitar preamps) produces a little softer clipping
characteristic. this boils down to the screen grid drawing current during the peaks of cathode current. the Vp(min) will have a
round bottom instead of a flatline like in a triode. to some extent it is a form of compression, which can be very useful. cascade
two stages like this and you get more of it.
honestly, my tastes for the lead type of distortion come from a good hard clip, and GENERALLY i achieve this the good old marshall way--with a cathode follower. of course you can put a CF at the end of two cascaded pentode stages and see what it
will do--i think you might be pleased with the result.
a few hints:
a) use a fixed voltage divider from b+ to gnd to generate Vg2. i never liked the datasheet method of a simple series resistor from b+ and a cap to ground... any variation in tube characteristics and the Vg2 is all over the place, plus it tends to wander around as you drive it hard. a lower impedance divider will give you a bit less sag, and a bit less compression as the stages are
driven hard. a higher impedance will give a slightly spongier sound. bypass the divider with a cap. a smaller cap will give you a quicker recovery from the compression/sponginess, a larger will give a slower recovery. if you use cathode bias then tie the
bypass cap to the CATHODE, not to ground. this will help with preserving the transients.
b) the vg2 is what sets your vg1 range for the quiescent bias of the tube.. higher vg2 gives more -ve vg1 for the same Ik.
personally i like a large -ve bias on g1 so that clean headroom is preserved. this helps with the touch sensitivity.
c) keep Vp around the same voltage as Vg2. too much lower and Ig2 goes through the roof. higher doesn't seem to be a problem up to the voltage limit of the plate. Vp is of course set through the available b+ voltage, the Ik, and the value of Rp.
d) the value of Rp can be tweaked to adjust the gain through the stage. in pentodes it's handy to think of the gain as gm*Rp,
instead of mu*Vin, since the plate impedance is so high. clearly as Rp is made higher, the gain goes up. keep an eye to (c)
however since you can't arbitrarily adjust Rp without affecting other parameters.
e) as a corrolary to (d) the value of Rp also affects the value of output impedance.
f) as a corrolary to (e), with pentodes you can get away with much smaller coupling caps to net the same LF -3dB point, since
the output impedance is higher than in triodes.
that's all i can think of now. i say give it a shot... it's a different sound than triodes, and i kind of like the squishiness and pliability of the clipping characteristic.
ken
START NEW THREAD REPLY PREVIOUS LIST
Read 294 times
From: Matt
Date: 1/4/2006 4:18 PM
Subject: Re: EF86 - Reloaded
I forgot to mention that kg's post was in the context of cascading pentodes and getting distortion. In my case, the EF86 will be the only preamp tube and from there it will go straight into the long-tail PI. I would like a great clean tone but if I crank it I'd like it to overdrive the PI nicely.
START NEW THREAD REPLY PREVIOUS LIST
Read 285 times
From: MBSetzer @
Date: 1/4/2006 5:40 PM
Subject: Re: EF86 - Reloaded
Well, I use a 6186 which can be wired pin-compatible with a 6AU6.
It is very tasty as a solitary gain stage before a LTP or a SE power tube, but there will not be much overdrive except from a high-output guitar.
And if you put in a tone stack that will cause enough insertion loss to virtually eliminate distortion, I think a pentode input can be real good this way for a clean preamp topology.
To preserve the gain available from the pentode stage you really have to use high impedance components downstream from the plate output, for instance do not put a tone stack there but instead use a single carefully-selected coupling cap to go to the grid of the PI, and have that grid with no less than a 1Megohm grid load.
That way you have a baseline performance level to compare with after you put in a tone stack so you will know firsthand how much gain it costs you.
Also consider higher-mu power tubes like EL84 or EL34 since in operation they seem like they overdrive twice as easily as 6V6 or 6L6 types with equal levels of signal coming from the same preamp.
Mike
START NEW THREAD REPLY PREVIOUS LIST
Read 280 times
From: Matt
Date: 1/4/2006 7:14 PM
Subject: Re: EF86 - Reloaded
Thanks Mike. I can't get Duncan's site to work right now, are those pentodes? If so, have you tried running them with a voltage divider on the screen and the screen cap connected to the cathode?
I was concerned about having enough gain if I used a tone stack but that's what these little trials are for.
This amp is cathode-biased EL34s. The other channel will be similar to a Soldano Atomic 16 (or tweaked JCM800, same thing). There will be an FX loop but I want to get the base tones first. The channels will be in-phase with each other so who knows what I'll be able to get out of it (haha, that's what these little trials are for).
START NEW THREAD REPLY PREVIOUS LIST
Read 240 times
From: Ian Anderson @
Date: 1/5/2006 6:29 PM
Subject: Re: EF86 - Reloaded
has anyone here actually built a high-gain cascading EF86 circuit? - not necessarily multiple EF86s either, maybe an EF86 frontend with another two ECC83 gainstages after it.
... Ian
START NEW THREAD REPLY PREVIOUS LIST
Read 245 times
From: Rick Erickson @
Date: 1/5/2006 6:55 PM
Subject: Re: EF86 - Reloaded
maybe an EF86 frontend with another two ECC83 gainstages after it.
Does anyone have a source for EF86's that could work in a high gain situation? I have a hard enough time trying to find one that will work in a Matchless without being overly microphonic.
RE
START NEW THREAD REPLY PREVIOUS LIST
Read 241 times
From: MBSetzer @
Date: 1/5/2006 8:51 PM
Subject: Re: EF86 - Reloaded
The 6AU6 is the stereotypical small-signal pentode in 7-pin miniature format, with textbook construction having actual screen grids for g1, g2, & g3 plus a fully-enclosing plate structure. 6186 is pin-compatible with those 6AU6 designs having the cathode tied to the suppressor, since the 6186 has these internally tied a jumper often seen between pins 2 & 7 of the 6AU6 socket becomes moot. In that case (most of the time) its a plug-in substitute for 6AU6.
Schematically the 6186 would appear just like a 6AU6 or EF86, however it is not constructed very similarly. Somewhat like the way a beam tetrode is formally not supposed to use the same schematic symbol as a true pentode but usually does anyway.
Since there is no actual supressor grid in a 6186 either, it has what amounts to something like *beam-forming plates* instead not completely unlike power tetrodes. And there is not much to the plate structure of this RF-designed pentode which is why you can see all of this so well.
Well, I haven't tried the DC-biased screen, and usually tie the screen cap to ground, but with a bypassed cathode resistor this would be mostly equivalent to referencing the cathode anyway.
Keep in mind that the screen will be fluctuating with signal which is out-of-phase with the input grid and the cathode, so connecting the screen cap to an unbypassed cathode would be a form of feedback.
For ideal hifi pursuits a regulated screen power supply is preferred for pentodes, and the size (output capacity) and performance (internal impedance) of the power supply are major considerations which need to be made, dependant on the power handled by the tubes and specifically the milliamps of current needed by the screen(s) themselves under maximum conditions. It could be shown that having a voltage-divider screen bias and/or a typical or typical-to-oversized uF screen cap is less ideal than a regulated supply theoretically, but when dealing with the small signals of a guitar input tube these differences are probably smaller than you think. It will always be a balancing act between the main B+ screen-dropping resistor ohms versus having too little voltage on the screen at idle or during signal handling. Do not just put a 1megohm from B+ to screen and a 0.1uF cap from screen to ground and think you are done. That would really be where you are just getting started, for a head start in that direction though I would suggest a 220K plate resistor, a 560K screen resistor and a 1uF cap to ground, high voltage film cap. Goes well with a 4.7K cathode resistor and 2 or 3uF of cathode bypass, film caps please.
I would say it is highly advisable to use the clean output of a small pentode to then drive a 12AX7's cascaded triodes, and then not a bad idea to use a typical proven tone stack in between the triodes in a familiar way. The only problem as usual is selecting for low-microphonic pentodes, in addition I use high-temperature synthetic rubber damping material in the tube cover for the 6186. If the gain is high or the layout imperfect, also a DC heater for the pentode when there are too many stages following.
As usual on an input tube, you could also increase the plate resistor ohms or lower the cathode resistor ohms within reason to get some distortion out of the input tube rather than having it be a mostly clean first stage, but it could then turn out to be one of those amps with no real clean headroom.
Mike
START NEW THREAD REPLY PREVIOUS LIST
Read 231 times
From: Ian Anderson @
Date: 1/5/2006 10:40 PM
Subject: Re: EF86 - Reloaded
(snip) it could then turn out to be one of those amps with no real clean headroom.
you say that like its a bad thing
I know the trouble of getting non-microphonic EF86s. I had an AC30 I modded to have the old AC30/4 EF86 frontend and while the tone was the dogs bollocks, the EF86s would turn microphonic within a few weeks. I even tried primo NOS EF86s and they did the same. I reckon EF86s are realistically only useable in a head format due to their tendency to turn microphonic.
... Ian
START NEW THREAD REPLY PREVIOUS LIST
Read 187 times
From: Dai H. @
Date: 1/6/2006 4:05 PM
Subject: Re: EF86 - Reloaded
I know the trouble of getting non-microphonic EF86s. I had an AC30 I modded to have the old AC30/4 EF86 frontend and while the tone was the dogs bollocks, the EF86s would turn microphonic within a few weeks. I even tried primo NOS EF86s and they did the same. I reckon EF86s are realistically only useable in a head format due to their tendency to turn microphonic.
... Ian
how about if you use a 12AX7 (or half of one) for the input, then use the EF86 (there was some Selmer schematic like this?--but would you lose the "EF86 magic" though?). Other pentodes besides the EF86 and 6AU6 are the 5879 (9 pin mini--used in some gibsons), or how about the 6SJ7 (octal--if you got the space). E80F (and not EF80) is supposedly a EF86 sub. EF40 is a Rimlock base version of the EF86 (probably has the same problems w/microphony but perhaps you can find a big batch cheaper than EF86s)--but then you have to find a couple of the weird sockets even if you find the tubes.
START NEW THREAD REPLY PREVIOUS LIST
Read 179 times
From: MBSetzer @
Date: 1/6/2006 6:00 PM
Subject: Re: EF86 - Reloaded
I've got a couple of EF40's, but never knew such a socket existed.
8-pin miniature tubes with the locating key molded into the glass.
Where do you get sockets like this?
Mike
START NEW THREAD REPLY PREVIOUS LIST
Read 173 times
From: Dai H. @
Date: 1/6/2006 6:55 PM
Subject: Re: EF86 - Reloaded
Europe, or a local dealer (if you get lucky). Oh, I think I saw some new Chinese made ones, so maybe if you look around someone carries them. The Rimlock series tubes start w/"4" (ex. EF40, EZ40, etc.).
START NEW THREAD REPLY PREVIOUS LIST
Read 242 times
From: Chris/CMW amps @
Date: 1/5/2006 11:13 PM
Subject: Re: EF86 - Reloaded
Hi Rick
I've used TAD and Ruby EF86s with good/non microphonic results in such style amps. I have to say at least one time the socket was also mounted with springs/etc. to prevend chassis-rattles.
Hope this helps,
Chris
START NEW THREAD REPLY PREVIOUS LIST
Read 226 times
From: Carl Gigun @
Date: 1/6/2006 2:23 AM
Subject: Re: EF86 - Reloaded
I've played with EF86 and 6AU6 a little, not in high gain though. I just put an ef86 as the first stage in a hiwatt dr105 type circuit. That's like a jcm800 level of gain I guess, it goes ef86-vol-12ax7-tone-12ax7-12au7 PI-6v6 PA. It doesn't sound very good so far but it's not microphonic at least. The ef86 is a sylvania I believe from an old radio. IIRC it says it was made in russia. I'll have to try some of these tips once I get the rest of it set up better. I got the 6au6 (and sockets for them) from old TVs.
START NEW THREAD REPLY PREVIOUS LIST
Read 231 times
From: Rick Erickson @
Date: 1/6/2006 2:34 AM
Subject: Re: EF86 - Reloaded
Hi Chris,
Not familiar with TAD but I know that Ruby is just a distribution trade name for Magic Parts Co., so my question is - who manufactured the tubes for Ruby? The only EF86 listing I see in the MP's cataloge says "RUSSIA", which suggests either Sovtec or Svetlana. I just got some Sovtec's from New Sensor and they work O.K., but are still a bit microphonic if you crank the amp. I did notice that they were less likely to squeal if I removed the shield cover. I just ordered some generic # Svetlana "EF86" pentode tubes from another supplier and will see how they work. The price was very good. I'm hoping they are at least as good as the Sovtecs. I can't imagine using these tubes in the front end of a high gain circuit though.
RE
START NEW THREAD REPLY PREVIOUS LIST
Read 230 times
From: John Martin
Date: 1/6/2006 3:36 AM
Subject: Re: EF86 - Reloaded
As far as I know Svetlana is the only company currently making the EF86/6267. They make the ones New Sensor sells for sure. The last ones I bought had the "C" marking on them. I was lucky enough to buy a case of the Svetlana branded ones when they were on sale a few years ago and there have been very few rejects.
When I design a pentode stage for one of my amps, I follow guide lines from what I've gleaned from the hifi world. A good base line to tweak from is about 100v on the plate with the screen grid voltage at 30% - 40% of that. The screen grid cap is usually large enough so that it requires an electrolytic. If you keep the plate resistor at around 100 -150K the output impedance will be low enough so that coupling to the next stage will not be difficult. The output impedance rises rapidly with larger plate resistors and I think that is one reason many people have seen unsatisfactory results. This base line seems to work well for the 6AU6 and 6SJ7 too. There is a lot more tweaking involved with pentode gain stages.
jm
START NEW THREAD REPLY PREVIOUS LIST
Read 206 times
From: Matt
Date: 1/6/2006 2:09 PM
Subject: Re: EF86 - Reloaded
Thanks for the responses. Regarding the larger plate resistors, my limited experience agrees with John, I never can get a 220K plate resistor to sound very good to me.
I think part of my confusion has to do with the screen voltage. I've seen it all over the place. Drawing from my output pentode 'mindset', I usually associate a screen voltage as close to but not above the plate voltage as ideal and my limited trials bore this out.
John, so you like 100v on the plate and 30-40v on the screen? I'm pretty sure my voltages were higher than that, like 180v plate and 165v screen or in that ballpark (using my method above). Anyway, I've got some things to try out.
Thanks again.
PS - I think kg's suggestion of using a pentode input but then driving perhaps the direct-coupled gain stage-cathode follower (for the distortion) might be interesting. Almost all of my favorite distorted tones come from a circuit with a direct-coupled cathode follower at the end (aka some kind of Marshall or rodded Marshall circuit, like the Bogner XTC red channel).
PPS - my only EF86 was the Svetlana. I did have better luck not using a shield, the tube is internally shielded anyway.
START NEW THREAD REPLY PREVIOUS LIST
Read 190 times
From: Greg Simon
Date: 1/6/2006 5:39 PM
Subject: Re: EF86 - Reloaded
He was saying that the screen should be 30% to 40% of the plate voltage, not 30v to 40v.
The only new manufacture EF86 today that is available is the Svetlana factory made one, which sells in the US as Winged C, or relabelled as a Sovtek. They are very inconsistant, and prone to microphonics. They often have gain levels WAY higher than spec too. As far as I know, they don't make these any more....having stopped a year or two ago. JJ is supposed to be working on one but hasn't released it yet.
I used an EF86 input stage on a Silvertone 1484 I modded. The arrangement went like this:
EF86 > CF using 1/2 12AY7 > tone stack > volume control > gain stage using 1/2 12AY7 > stock Silvertone PI > fixed bias, no feedback 6L6GC power amp.
It sounds very nice, with cleans being better than your typical blackface Fender, and when cranked up, it will hang in Marshall JCM 800 territory. I put in a switch after the CF to switch a .003uf cap in and out to chop out some of the bass, or the tone gets flabby when the amp is cranked. Using a CF after the EF86 makes things easy since you can maximize the gain and not have to compromise. It has about 115v on the plate, and about 90v on the screen, using a 220k plate resistor, 1M3 screen resistor, 2k2 cathode resistor, and a 25uf cathode cap.
greg
START NEW THREAD REPLY PREVIOUS LIST
Read 171 times
From: Matt
Date: 1/6/2006 8:51 PM
Subject: Re: EF86 - Reloaded
Thanks Greg. In his example, the plate voltage was 100v so that's where I got the 30v-40v.
That sounds like an interesting circuit in your Silvertone.
START NEW THREAD REPLY PREVIOUS LIST
Read 169 times
From: John Martin
Date: 1/7/2006 12:29 AM
Subject: Re: EF86 - Reloaded
I use 100v as a starting point. The audiophile "ideals" don't always exactly fit guitar amp design, but I found most of the common small signal pentodes work well at lower plate voltages. I think I miss stated the screen voltage thing. I meant that they should be 30% - 40% less than the plate voltage. So with 100v on the plate it would be 60v - 70v on the screen. Its a lot more difficult to hit ideal values right on the nose like it is with triodes. Some tweaking from the "book" values is always in order for pentodes.
I hadn't heard that production on the EF86 had ceased. I think that the 6AU6 and 6SJ7 are still made in Russia as I have bought them in recent years.
START NEW THREAD REPLY PREVIOUS LIST
Read 164 times
From: BK
Date: 1/7/2006 1:09 AM
Subject: Re: EF86 - Reloaded
I tried an EF86 boosted plexi style circuit that squealed like a pig in heat. After several attempts at taming the EF86, I kicked the habit and adopted the cascode circuit with 12ax7's and ecc88's. WAY better sounding IMO.
I did have success with EF86's run real hot with a 47k anode load and a 820R cathode resistance. (I think the Vg was around -1 or -.5 with the svetlana EF86). Anyway, it was very organic in the lo gain EF86 -> 12ax7 -> CF -> tone stack.
BTW, the 6922/ECC88 is a perfect replacement/mod for the EF86 as the filament wiring is the same and you can use the screen filter as the grounded grid cap (with some other minor rewiring).
BK
START NEW THREAD REPLY PREVIOUS LIST
Read 167 times
From: Matt
Date: 1/7/2006 1:14 AM
Subject: Re: EF86 - Reloaded
Thanks for clarifying John. That makes more sense. Maybe I'll have more luck this time, the B+ is quite a bit lower in this amp.
FWIW - I just received my EF86 in the mail. It's labeled Sovtek but has a Cryllic C on it and says Made in Russia. The one I had previously was a Svetlana but I sold the amp that tube was in. I'm assuming they're the same tube.
START NEW THREAD REPLY PREVIOUS LIST
Read 164 times
From: Greg Simon
Date: 1/7/2006 8:48 AM
Subject: Re: EF86 - Reloaded
Yep, they are the same Matt. They're a good sounding tube when they work right, but very inconsistent, and probably has way more gain than spec or than an older one.
Thanks for the comments on the circuit I came up with. It worked well with the Silvertone so I'll try it in some other builds in the future to see how it works in them. It sure as hell sounds better than a stock Silvertone 1484!
Greg
START NEW THREAD REPLY PREVIOUS LIST
Read 172 times
From: Alf @
Date: 1/7/2006 10:58 AM
Subject: Re: EF86 - Reloaded
Hi B.K.
I'm very interested in your use of the ecc88 together with the 12ax7's. Could you explain a bit more about what you did exactly ? The ecc88 is a cascode tube and I never used one before in a circuit. What plate and cathode resistors did you use ? Do you happen to have schematic of it ?
I happen to have quite a few of those ecc88s but they are reported to be quite microphonic sometimes so did you experience any problems with that?
Thanks beforehand,
Alf
START NEW THREAD REPLY PREVIOUS LIST
Read 137 times
From: Matt
Date: 1/9/2006 9:26 PM
Subject: Re: EF86 - Reloaded
I did my first trial. I initially tried a 100K plate R, 2k7/22u cathode combo, voltage divider on the screen of 470k/220k, and the .047u screen cap connected to the cathode and it sounded pretty good. I used the Route 66 tone stack. Not as much clean headroom as I would like but it overdrives the PI nicely.
Voltages here were: B+ = 297, Vp=171, Vg2=166, Vk=4.
Then I increased the plate resistor to 220k and I thought it sounded even better. But the plate voltage went way down to 48v and the screen voltage only decreased a little bit to 160v. Cathode voltage and B+ didn't change. Is it a problem running the plate this much below the screen? When I calculated the currents, almost all of the cathode current appears to be going to through the plate resistor.
In any case, it sounds pretty darn good cranked. Very sustainy and wants to feedback nicely. Still doesn't get the best 'pristine' clean tone but all the 'in-between' tones are there all the way up to full volume. Nice and quiet, no microphonics or oscillations (that I can tell).
So is it OK to run the tube like this? Will I have screen problems?
PS - pins 2 and 7 are the two internal shield connections, I connected only pin 7 to ground, is that right?
START NEW THREAD REPLY PREVIOUS LIST
Read 108 times
From: Mike Conner @
Date: 1/17/2006 9:59 PM
Subject: Re: EF86 - Reloaded
Matt,
This is similar to my experience. I am using a 6SJ7GT.
I am running a 100K/220K split load plate resistor arrangement, with the 100K bypassed with a 100pF cap to boost highs. B+ node is about 310VDC and my plate voltage is only around 80 or 90 vdc. Don't remember screen voltage but it's probably at least double the plate voltage. The cathode resistor is 1K and bypassed with a 25uF.
I've tried a few other values/arrangments and dig this one. It sounds quite fantastic into a LTP pi and a pair of RCA 5V6GTs, and is plenty hot on output (hence the split load plate arrangement).
One basic question - what is the purpose of the screen cap and what's the effect of using smaller/larger?
START NEW THREAD REPLY PREVIOUS LIST
Read 99 times
From: Greg Simon
Date: 1/18/2006 3:23 AM
Subject: Re: EF86 - Reloaded
If I was you, I would experiment with running the screen lower than the plate and see how you like that. Aim for around 90v screen and 115v plate as a center value, and go up or down from there by changing your screen resistor value and your dropping resistor value. You get the most voltage output from the EF86 when the screen is 30% to 40% lower than the screen, and the most clean headroom too. When the screen is a higher voltage than the plate, you'll get less output out of the EF86, and less clean headroom too. If you aren't using a CF after the EF86, then the gain out depends also on what load you have after the EF86. Most people try to shoot for somewhere around 1M from what I've seen.
I think the screen cap is there to assure a clean DC voltage to the screen without a ripple problem. I used a .1uf on mine.
Greg
START NEW THREAD REPLY PREVIOUS LIST
Read 94 times
From: Matt
Date: 1/18/2006 1:51 PM
Subject: Re: EF86 - Reloaded
I still wasn't happy with the clean tone but I liked the OD tone, but I'm not really going for OD here. The other channel is a JCM800-type circuit and is sounding pretty good.
I ended up adding a series FX loop to the amp so instead of going straight into the PI, it now goes to a CF Send and then through the GS Return, then the PI.
After doing this I still wasn't happy, not enough headroom so I decreased the plate resistor to 150k giving me a Vp of 106v and a Vg2 of 155v. This sounded better. I'll try lowering the Vg2 to just below the Vp and see how that goes. When I had a Vp of 180v and a Vg2 of 165v it didn't all that great.
I too don't have enough sparkle even though I have a huge bright cap on the 1M volume control (820pF). However, when I connect my Chandler Tube Driver to that EF86 channel it sounds awesome. I haven't tweaked that Route 66 tone stack yet, I need to alter the voice a little bit.
I assumed the screen cap was to AC ground the screen and keep the signal from modulating the screen voltage. I would think any signal on the screen would be a form of NFB, so if you used a smaller cap you might get some local NFB in the low frequencies. But this is just a guess. I thought this was why kg suggested connecting the screen cap to the cathode (because the two are out-of-phase) when using the voltage divider on the screen and was the reason for the 'compression' difference between a low impedance voltage divider on the screen vs. a higher impedance divider.
START NEW THREAD REPLY PREVIOUS LIST
Read 84 times
From: Greg Simon
Date: 1/19/2006 2:36 AM
Subject: Re: EF86 - Reloaded
Most people seem to think somewhere around 90-120v on the plate and 30% to 40% less than that on the screen is the sweet spot for an EF86 from what I've heard. I think you had a too high voltage on the plate and screen before when you tried it with the plate voltage being higher than the screen.
On the setup I have on my brother's Silvertone, I have a LOT of highs and lows, with plenty of sparkle, using a 70pf bright cap on the volume control. Of course the arrangement of this one I did is different than what you're doing. This one used the EF86 > CF > tone stack (120pf treble cap, .1uf bass cap, .022uf mid cap, 100k slope resistor, 500k pots, no mid control) > volume control using a 500k pot > 12AY7 gain stage > cathodyne PI using the stock setup of the Silvertone 1484 with a 6CG7/6FQ7 tube > 6L6GC pair. I put in a switch after the CF to switch in a .004uf cap to chop some of the bass out when the amp is dimed. Without that, the bass was too flabby on higher volume stuff. On the EF86, I think I used a 220k plate resistor and a 1M3 screen resistor, with a supply of about 340v.
Greg
START NEW THREAD REPLY PREVIOUS LIST
Read 37 times
From: Matt
Date: 1/23/2006 12:50 PM
Subject: Re: EF86 - Reloaded
FWIW, here's the latest trial: B+ is 295vdc. Changed voltage divider on screen to 470K/470K which lowered the screen voltage to ~110vdc. Adjusted the plate resistor to 180K to get 140vdc plate voltage. Cathode is still 2K7/22u and is @ 2.7vdc. Screen cap is still .047u and is connected to the cathode. This sounds really good. Amazing how much the headroom increased and it still overdrives the PI very nicely. There's a nice, touch sensitive transition to OD as you turn it up. Also, before it seemed to compress when you hit the wound strings and the unwound strings seemed to stand out too much, now the lower notes are full and have punch. The Route 66 tone stack doesn't seem to need as much tweaking, if any.
START NEW THREAD REPLY PREVIOUS LIST
Read 13 times
From: Mike Conner @
Date: 1/23/2006 8:14 PM
Subject: Re: EF86 - Reloaded
I'm going to try this if I get a chance this weekend (and I haven't already soldered and resoldered the preamp to death!!)
START NEW THREAD REPLY PREVIOUS LIST
Read 10 times
From: MattR
Date: 1/23/2006 8:52 PM
Subject: Re: EF86 - Reloaded
Matt,
Is your voltage divider connected as such:
295v B+ - 470k - pin 1 - 470k - ground?
