Hello! I introduce myself in this forum that I consider to be very interesting. My question is about the excessive gain of the EF86. I have built a VOX Ac15 schematic circuit and the clean channel has a very high gain and a lot of background noise amplified by this gain. Is it possible to reduce this? Is it normal for it to have a plate voltage of 120 Volts and an anode voltage of 95 Volts? Thank you
Attachments
It is supposed to be high gain. If the tube is noisy, replace it. Get several, they are often hissy.
'Plate' and 'anode' are the same electrode, at least in my language. (Plate is US and anode more UK.) "About 100V" is fine for a 300V supply. 1/3rd to 3/2rd of supply voltage is good. The screen (G2) voltage is very variable and not at all critical as long as the plate/anode sits in the right voltage range.
Thank you Excuse me, but I'm with a translator since my language is Spanish. I was referring to the screen grid, that is, the G2. I have already replaced the tube with a new one. Is it normal that with the input at 0 (pin 9 of the EF86 grounded) it still generates a lot of background noise when I increase the volume?
Pentodes are somewhat noisier than triodes. But noise may be partially due to high value input resistor (1 M). Try lesser value, something like 100 K, if your source allows. Gain can be reduced 2 times by replacing plate resistor with 100 K, and also reducing values of cathode and screen resistors. Check EF86 data sheet for correct values. You may also try E80F, which is a higher quality EF86.
You are right when you just look at the noise level with open input, but why would anyone want to listen to an open input? That 1 Mohm has less influence on the signal-to-noise ratio than 100 kohm would have when there is a source connected to the input. In any case, there is also hiss with the grid of the EF86 grounded.
Last edited:
Reducing the gain by taking out C5 not only brings the gain down but also makes it more linear.
You could even play with a resistor in series with C5 to tailor the gain to your hearts desire.
I'm sorry I can't make it more complex ...
Jan
Note that rafaelct now says:
In general, the full noise voltage of that 1Meg almost never happens because it is bypassed by guitar pickup (5k-200k) or shorting jack.
The AC15 is popular because it has a lot of gain, "plays easily", distorts on a whim.
Jan has a good solution if you actually want to get "less", but guitarists always want more.
Rafael, welcome to the forum!
You are exactly right. VOX made a terrible mistake with the original AC15. From the tube catalogs, they lifted a very high gain circuit that was designed to be used with a microhone, record-player pickup, or tape-recorder playback head. They stuck this in the front end of a guitar amplifier. This was a Very Bad Idea (TM).
The trouble is that guitar pickups put out a far bigger signal than microphones, record pickups, or tape heads. And guitar amplifiers are usually played at much higher loudness level (SPL). Worst of all, the AC-15 mounts the tubes and loudspeaker in the same box, so vibrations from the speaker can shake the electrodes in the EF86 input pentode.
The result was too much noise, too much distortion, too much microphony. VOX realized this very quickly. Within a few years the bad circuit was ripped out and replaced by lower-gain circuits using triode valves (not pentodes). VOX dropped the pentode front-end because it created too many problems for them.
VOX was smart, but other people weren't as smart. The AC15 and AC30 had made a name for themselves, and others were eager to grab some of the glory for themselves. So many other people made copies of these flawed amplifiers.
All of these "boutique" manufacturers and hobbyists made exactly the same mistake: they used the exact same problematic circuit, and (no surprise) experienced exactly the same problems with it. Even if the circuit behaves well when new, within a few weeks or months the pentode will become microphonic, and start to ring, screech, and howl. You replace the expensive pentode, and within a short time, have the same problems again.
The first intelligent discussion about this seems to have come from Merlin Blencowe, aka the "Valve Wizard". He wrote a book on designing tube guitar preamplifiers where he discusses this problem - and cures for it - in detail. A part of the pentode chapter is publicly available, free, on his website: https://www.valvewizard.co.uk/pentode.html
How do you fix your problems? The best solution is to (a) design for much less voltage gain from the pentode, and (b) to use the pentode NOT at the input stage, but later in the signal chain. I've done this successfully with some of my own tube guitar preamp designs.
You are somewhat stuck, because you've already built a fundamentally flawed design. It may not be easy for you to move the pentode to a later stage in the chain. So all you can do, is lower the gain of the pentode stage.
Read through Blencowe's web page first, so you understand how the pentode stage works. The circuit he describes on that page has half the voltage gain of the AC15 input stage, which is a step in the right direction. You can modify your input stage to match Blencowe's circuit by changing just a few parts.
If that doesn't solve the problems, you can put a trimpot between the output of the EF86 stage, and the input of the next stage, wired like a volume control. Set the trimpot to throw away some of the excess gain. Hiss/ noise/ microphony will also go down.
If you are able to follow Blencowe's design equations, you can make your own design, to lower voltage gain even more. Start by lowering the B+ voltage to the pentode a lot - that automatically lowers the voltage gain of the stage. Instead of designing for 250 volts B+, use a big series resistor and filter cap to drop the pentode stage B+ to, say, half that: 125 volts.
I suggest also halving the anode resistor (plate resistor). Instead of the 68k Blencowe suggests, try 33k.
These two changes will help a lot - pentode stage gain will fall a lot, which is what you want.
Of course you can't just change these two things - you will have to change all the other resistors to match, using the equations Blencowe provides.
You will find the voltage gain is now much lower. This is good, because it will reduce all your problems: less hiss, less microphony, less tendency to overdrive prematurely.
There is one big mistake in Blencowe's design equations: his formula for calculating the proper cathode bypass cap is quite badly wrong. I could go into the reason why, but it would be long and mathematical. I suggest that the best practical solution for you, is to try several different cap values here, until you find one that sounds right to your ears.
Good luck!
-Gnobuddy
You are making it sound as if the Vox engineers had no idea what they where doing, and I don't think that's fair. The designer of the AC15 was Dick Denney, a guitar player with electronics skills who decided to design his own amplifier. He brought his prototype to a company called JMI, and they were so impressed by the sound that they hired Denney to develop the amplifier into a commercial product. Denney used the EF86 because he loved the sound of it, and he must have been well aware of the microphonics problem. but the musician in him overruled the engineer. Until they developed the AC30, wich was so loud that microphonics caused too many problems with the EF86.
Maybe, maybe not. It's decades later, so who knows. But the smoking gun is that Denney - or whoever chose that input stage - did not actually design it in any way. The circuit was simply lifted, unchanged, from the back of a tube catalog, and then used it in a circuit where it performed poorly.
This is somewhat the equivalent of taking the handlebar off your kid's bicycle, and then bolting it to the steering column of a big-rig truck ("semi"). What worked elegantly and well on the bicycle, is not going to work very well on the 50,000 lb big rig.
The unknown actual designer of that EF86 circuit worked for the tube manufacturing company, and came up with a circuit that showcased the ability to squeeze a lot of voltage gain out of their shiny new product (small signal pentode). At that time in history, the ability to get that much voltage gain from one tube was a radically new development. Exciting new "high tech"!
The ability to get all that voltage gain from one tube might also have helped justify the higher price of a pentode, as one pentode stage could now replace two triode stages in cheap record-players.
The nameless engineer who developed the circuit certainly knew what he was doing. The quiescent operating point for the tube is a very bizarre one. Screen current is absurdly low, too low to read on an EF86 datasheet. The anode load is absurdly high. Anode current is absurdly low, once again, too low to read on the published datasheets.
This was not a "textbook" design. It is a very ingenious, "out of the box" one - IF used for its intended purpose, i.e., amplifying very small signals from a microphone, tape head, or record player pickup.
The transconductance of a pentode falls at low anode currents. Generally speaking, that lowers the voltage gain of the stage.
But the clever nameless engineer noticed that by using a much larger-than-normal anode load resistor, and a high B+ voltage, he could end up with even more voltage gain than you'd get at higher anode current. The high gain came at a cost - very high output impedance, over 200k compared to 40k for a typical 12AX7 triode stage. But if the input impedance of the next gain stage was high enough, it could work.
That engineer working for that tube company never intended his circuit to be used with the much higher output of a guitar pickup, and obviously, it was never tested at the extremely high SPL and vibration levels it was subjected to in VOX guitar amplifiers. He designed a very clever and elegant bicycle handlebar, but it was never intended to be bolted into a big rig.
I've heard Denney's name before, but have no idea what his actual engineering skills were. Did he actually know how to design a tube amp, or was he like most other guitar amp "designers" of that era: someone who knew just barely enough to wield a soldering-iron, read a schematic, and string together building block tube stages taken from tube catalogs and from competitor's designs? More a tinkerer with basic technician skills, than someone capable of actually engineering an appropriate circuit?
To me, the fact that he shoehorned the circuit into the front end of an AC15, unchanged, suggests the latter. If Denney had liked the sound of pentode distortion, but actually knew what he was doing, he would have redesigned the circuit to work better in the guitar amp. He didn't do this, probably because he had no idea how to design a pentode gain stage.
Remember, this was long before the Internet, long before pocket calculators. The pentode was a new device, and a complicated one to bias, with all these electrodes interacting with each other. Even if you knew the design equations, there were no pocket calculators or computers or spreadsheet software - you would have to know how to use a slide-rule, or a book of logarithm tables, to calculate the values of every component.
An actual electronics engineer would know how to do all that. A guitarist who enjoyed tinkering with amplifiers probably wouldn't.
Ultimately, all this is irrelevant today. The main thing is, what can we do to help out Rafaelct?
-Gnobuddy
Come on, the pentode was invented in 1926 and the EF86 is from 1953 (its predecessor EF40 from 1948, EF37A from 1942).
The OP needs to define "noisy": what kind of noise? hiss? hum? and under which circonstance? in his bedroom? in a pub? in a concert hall?
The AC15 is a gig-worthy amp, and it's perfectly normal for it to be a bit noisy at home. Even my Fender Champ hiss a bit when all is quiet, and I did run DC heater; crank it up, and all is fine.
If he really wants to do something, running the EF86 in triode mode is OK. It'll no longer be an AC15, but then again, I, personally, have no love for Vox, but that's only me. 🙂
Oh, and by the way, it's perfectly OK to lift circuits from the manufacturers' application notes.
The AC15 is a gig-worthy amp, and it's perfectly normal for it to be a bit noisy at home. Even my Fender Champ hiss a bit when all is quiet, and I did run DC heater; crank it up, and all is fine.
If he really wants to do something, running the EF86 in triode mode is OK. It'll no longer be an AC15, but then again, I, personally, have no love for Vox, but that's only me. 🙂
Oh, and by the way, it's perfectly OK to lift circuits from the manufacturers' application notes.
Last edited:
1926 is pretty much a red herring. Holst and Tellegen were probably the only two people on planet Earth who understood how pentodes work in 1926. They didn't even have a patent until 1934.
The AC15 was sold starting 1958, and its design had to have been finalized earlier than that. What percentage of everyday British guitarists do you think had pentode design skills in, say 1957?
Any way you look at it, pentode input stages in guitar amps were an evolutionary dead-end. The idea was repeatedly revived, usually by by small-scale boutique guitar amplifier manufacturers, even into the 1990s and 2000s. Yet, those designs never stayson the market for long (for instance, the 65Amps SoHo).
Online guitar forums seem to have plenty of posts from people who bought one of these boutique pentode-input guitar amplifiers, only to run into major headaches with microphony despite frequent (expensive) pentode replacement.
My opinion is that the VOX AC15 pentode front end was a bad idea in 1958, and is still a bad idea in 2022. You may not share that opinion. That's fine.
But I'd rather not see this thread wander off into useless arguments about history or opinon. The OP asked a technical question, and what would benefit him is technical answers.
-Gnobuddy
Yes, of course. That's why manufacturers provided them.
But it's silly to then try to use those circuits in applications where they don't work well. Don't bolt bicycle handlebars onto your big-rig.
-Gnobuddy
I have no opinion on that, as I hardly know anything about electric guitars.
Then why did you start them?
Jan came up with a good way to reduce gain and reduce distortion, Ketje with a good way to reduce gain and eliminate partition noise, and you with a good way to just reduce gain. I think the thread starter should try those and see if they do the trick.
Last edited:
I think Denney took the Mullard 10 watt amplifier as a starting point, and then made some changes. He got rid of the negative feedback loop, and made the EF86 stage even more gainy. All of this to make it sound good to his ears, wich is all that matters in the end. This is not an engineer trying to make a 'good working' guitar amp, but a guitar player trying to make a 'good sounding' one.
That's a problem, when discussing a circuit that works perfectly fine except in electric guitar applications.
I'm not just being difficult. Audio amplification for electric guitars really is a different world than "normal" audio amplifier design.
I learned this the hard way. I started in electronics by building a crystal radio at age 7 or 8, moved on to building TRF (tuned radio frequency) transistor radios by age 9, then reflex TRF radios, then superhet AM medium wave/shortwave radios. Then I got interested in Hi-Fi audio, and spent years learning how to design and build Hi-Fi audio amplifiers and preamps.
I didn't pick up a guitar until I was in my early twenties. And that's when I began to discover that everything I knew about "good" solid state amplification, completely failed me when it came to electric guitars. "Good" amplifiers sounded bad when you plugged a guitar into them. "Good" loudspeakers sound horrible when driven from a guitar amp. What's needed is a very specific type of "bad" amplifier, and an equally "bad" loudspeaker. Horrible for Hi-Fi, but good for e-guitar.
My turn to say "Oh, come on!"
I believe in facts, not hero-worship. The facts are that this circuit was problematic from the start, and VOX quickly abandoned it, probably because of too many complaints from customers and too many warranty repair claims.
Since that time, people found they could get all the overdrive they could ever want in a tube guitar amp, simply by adding more cascaded triode stages. For several decades now, there have been forms of music (metal, etc) that use amplifiers with very high gain and very high levels of distortion - far more distortion than Denney ever got from one EF86. Not one of these more modern high-gain tube guitar amplifiers uses pentodes in the preamplifer.
Agreed, and well worth a try.
In this circuit, the anode load is so much bigger than the cathode resistor, that I don't think Jan's suggestion will lower voltage gain sufficiently to help Rafael. But it's an easy mod to try. Takes only seconds with a soldering iron to lift one end of the cathode bypass cap.
Ketje's solution effectively turns the pentode into a triode. It makes perfect sense from a Hi-Fi audio engineer's point of view. It will certainly kill lots of gain.
But to most electric guitar players who have just shelled out extra money, and dealt with extra build complexity, because they believe in the almost mythical status of the VOX AC15, and the associated "pentode sound", this is not a reasonable suggestion. Turn that precious pentode into a mere triode? Horrors! Sacrilege! Blasphemy! 😀
Imagine if you were a bicycling enthusiast. You just bought yourself a $15,000 racing bicycle with carbon-fibre frame and rims, but you're finding the ride a bit harsh. How would you respond if I suggested you sell the bike and buy a fat, heavy, American pickup truck to replace it?
That's what telling a VOX AC15 enthusiast to triode-strap his EF86 is like. 🙂
All in good fun...please let's not start an argument about cycling, now!
Hope to hear back from Rafael.
-Gnobuddy
The EF86 stage in the Mullard 10 watt amplifier is a much more conventional one. Note that the anode resistor is about half the value used in the VOX EF86 stage, for instance.
I don't think Denney had anything to do with the design of that EF86 stage at all. He lifted it unchanged from application notes in the back of a tube catalog.
My memory is that I've seen the exact EF86 stage Denney used (all part values identical), in the back of one of the old tube catalogs. The tube manufacturer suggested that it could be used with a dynamic microphone, tape playback head, or record player "cartridge". Its claim to fame, was that it squeezed out higher voltage gain than other contemporary EF86 circuits did.
I'm pretty sure I've also seen that exact EF86 stage at the front end of a schematic for a budget-priced record player from the era. IIRC it was driving the output pentode, with no other gain stages in between. Just one EF86 stage, and one pentode output stage. This might have been in the same tube catalog.
One could speculate that Denney simply unplugged the pickup cartridge from one of these record players, and wired up his guitar cord instead. It wasn't uncommon for early electric guitar players to convert (tube) radios and record-players into impromptu electric guitar amplifiers.
(Note that the previous paragraph is purely speculation on my part. Plausible speculation, but I have no evidence to support it.)
I also think I've seen a very similar record player amplifier circuit that used two pentode output tubes in push-pull, with some sort of phase-splitter in between the EF86 and the output stage. This memory is more fuzzy, and I can't recall what type of phase splitter was used. That circuit was well on the way to a VOX AC15.
Exactly. In this case, he may have got a sound he liked, but it proved to be a bad design for an actual production amplifier.
If Denney started out plugging his e-guitar into a two-tube single-ended record player amplifier, it would have had much less output power, and much less overall voltage gain - and that would have made it a much less faulty design than the later VOX amps that evolved from it. It may have had too little output power to cause the microphony problems sufered by the AC15 and AC30.
I worked as an engineer in an audio company at one time, and I saw progressive "feature creep" pushed by sales managers ruin a perfectly reasonable amplifier designed by a co-worker, in much the same way. (Though this was a solid-state audio power amp, not a tube guitar amp.)
-Gnobuddy