I'm building a little buffer/preamp box for guitar with a cathode follower output since I have assumed that was good for driving cables. But while looking at valve microphone discussions , I find that valve mics use transformers (or ss circuits) to drive their cables rather than cathode followers because CF distort at high levels/frequencies when faced with cable capacitance. I realise I may be mixing chalk and cheese here and there's a lot going on I simply don't have the knowledge to understand... but my question is , is it ok to drive cables with say a 12au7 CF?
Yes, it is perfectly OK. You need to rememeber that, although a CF has a low output impedance which should minimise HF losses when driving a cable, its ability to drive current into a load (including the cable capacitance) depends on the quiescent current. Using a 12AX7 CF running at 1mA is not a good way to drive a cable. Use the 12AU7 you suggested and set its quiescent to about 10mA.
Cheers
Ian
Cheers
Ian
Hi i disagree about microphone aren't using CF: some alltime classics do: Sony C37 or C800, Akg c12a, some U47 tubes were transformed into cf in states...
The CF in this case doesn't distort or at least not more than 'typical' grounded cathode (there is way less gain than in grounded cathode circuits/ which is unity or less from the circuit side/ could have gain from the output transfo though). And the choice of a CF is exactly because of the inverse of your concern: if planed right ( tube is able to deliver current needed) a CF can drive cable capacitance for 10m without issue.
It'll all depend of how you engineer your circuit and it start by defining the environment it'll face.
Worst case scenario about signal for guitar will be something in the 1V rms with load input impedance in the vicinity of 1megohm. You don't need a lot of current ( but check to allow enough not to slew rate limit your stage).
The CF in this case doesn't distort or at least not more than 'typical' grounded cathode (there is way less gain than in grounded cathode circuits/ which is unity or less from the circuit side/ could have gain from the output transfo though). And the choice of a CF is exactly because of the inverse of your concern: if planed right ( tube is able to deliver current needed) a CF can drive cable capacitance for 10m without issue.
It'll all depend of how you engineer your circuit and it start by defining the environment it'll face.
Worst case scenario about signal for guitar will be something in the 1V rms with load input impedance in the vicinity of 1megohm. You don't need a lot of current ( but check to allow enough not to slew rate limit your stage).
As well as the resistive load, the CF needs to be able to drive current into the capacitance of the cable. To work out how much current we just use i = C.dV/dt where C is the cable capacitance and dV/dt is the slew rate required. If you want to drive 10V rms into 10m of cable with a capacitance of 100pf/m at up to 20KHz it works out you need 0.4mA of current to do it. This is why little 12AX7 CFs running at 1mA quiescent rapidly run out of current and start distorting. If you limit the output voltage to 1V rms maximum you should be OK but running a 12AU7 at a few mA has a much better headroom.
Cheers
Ian
Cheers
Ian
If you want MORE headroom then use a small pentode. EL84 or 6V6 CFs will drive anything (almost). And there are $3 TV video amp pentodes that will give you 50 ohm output impedance with 10 or 20 mA of current. Some with a free triode in the bottle. You wanna drive 300 feet of mic cable or 100 ohm load? Triodes available that will do it too, but not as common, plentiful and cheap in gm’s above 10,000.
There is also a risk of oscillation with capacitive load of a cathode follower. You must use a grid stopper resistor to prevent it, or a low value series resistor at the output. Known issue also with unity gain opamps driving a long cable.
A 12AU7 With 250V plate, -8.5V grid bias, and 10.5mA plate current (cathode current) . . .
Has a transconductance of 2,200 micro Mhos.
The cathode impedance is approximately 1/(2,200 x 10-6 Mhos) = 455 Ohms.
The output impedance of the cathode follower is 455 Ohms.
With -8.5V grid bias, a magnetic pickup of an electric guitar will be far less than 8.5 Volts peak.
There will be no grid current, and the much smaller signal will be in a linear range of the cathode follower.
Just my opinion
Your opinion may vary
Your Mileage May Vary
Has a transconductance of 2,200 micro Mhos.
The cathode impedance is approximately 1/(2,200 x 10-6 Mhos) = 455 Ohms.
The output impedance of the cathode follower is 455 Ohms.
With -8.5V grid bias, a magnetic pickup of an electric guitar will be far less than 8.5 Volts peak.
There will be no grid current, and the much smaller signal will be in a linear range of the cathode follower.
Just my opinion
Your opinion may vary
Your Mileage May Vary
Do add the grid stopper and build-out resistor - see Icsaszar's post (#6).
On a different thread, one diyAudio member showed scope captures demonstrating 10-volt peak-to-peak signals, straight out of the humbucker pickup in his electric guitar. Another member mentioned seeing around the same signal levels out of his Gretsch electric guitar.
Such large signals are not typical - vintage Fender guitar amps were designed for 25 mV input from the guitar - but if you strum a six-string chord hard, evidently 10 Vpp from your guitar is not impossible, either.
In practical terms, one of my Ibanez guitars (Artcore AS73, a Gibson 335 lookalike), with its stock Ibanez humbucker pickups, has a hot enough output to drive the input stage of a '65 Fender Princeton Reverb reissue all the way into audible distortion if you strum a chord.
The same guitar will also overdrive and distort(!) my Joyo compressor pedal, which is a clone of the old MXR Dyna Comp.
But the same guitar, plugged into a little JFET buffer I built, produces no audible clipping or distortion. The buffer has a JFET that biased up to +1.5 volts at the source (with the gate tied to ground). Bias voltage is not very different from the input stage (12AX7 triode) of the Princeton Reverb, where the cathode of the triode biases up to about the same +1.5 volts.
I think this is because grid current starts to flow in the triode for relatively small positive voltage swings at the input. There is already some grid current flow when Vgk is -0.9 volts, i.e. when the input has swung only 0.6V positive from its quiescent value of -1.5 volts.
I like tube guitar amps, and I think they usually sound much better than the vast majority of solid-state guitar amps. Still, driving a long cable is one job where I would pick an FET over a triode any day. Small, light, cheap, no heat, has a lower output impedance, has less noise...the FET is better in every way for this application.
That said, your triode buffer should work just fine. Add the grid stopper and build-out resistor to keep it from turning into an accidental radio transmitter, though! 🙂
-Gnobuddy
No one mentioned the White CF, which is the proper way to drive any kind of capacitance such as a cable
An ordinary CF only drives well when the signal is positive going.
The charge in the cap load depends on the CF resister for charging in the negative direction.
The White CF is the way out, it is active both +ve & -ve going.
An ordinary CF only drives well when the signal is positive going.
The charge in the cap load depends on the CF resister for charging in the negative direction.
The White CF is the way out, it is active both +ve & -ve going.
Or to use a current source in place of cathode to ground resistor and while at it a second cathode follower on the plate of the one alreday present driven by the circuit output. In other words what Allen Wright called a SLCF.
But i'm not sure the op is looking for a completly 'transparent' circuit? It's for a guitar after all (and as Gnobuddy pointed if looking for a 'cleaner' sound a jfet is easier to implement imho) this kind of circuits are 'classic' in guitar amps ( this is often used at fx send).
Anyway some food for thoughts (if needed)
http://www.valvewizard.co.uk/accf.html
http://www.valvewizard.co.uk/dccf.html
All good advice's from members. Grid stoppers are a must. Cathode stopper helped me in the past. Pentodes makes great tube wired as triode for cathode followers, even smaller ones ( Sony C-37 used a 6au6 triode wired).
But i'm not sure the op is looking for a completly 'transparent' circuit? It's for a guitar after all (and as Gnobuddy pointed if looking for a 'cleaner' sound a jfet is easier to implement imho) this kind of circuits are 'classic' in guitar amps ( this is often used at fx send).
Anyway some food for thoughts (if needed)
http://www.valvewizard.co.uk/accf.html
http://www.valvewizard.co.uk/dccf.html
All good advice's from members. Grid stoppers are a must. Cathode stopper helped me in the past. Pentodes makes great tube wired as triode for cathode followers, even smaller ones ( Sony C-37 used a 6au6 triode wired).
It seems to me . . .
When electric guitar strings are plucked very hard, most guitar amplifers become a fuzz box, without needing the external fuzz box.
Reminds me of Michael Hedges recordings of his Acoustic Guitar.
I always cringed and waited for the strings to break.
Could be almost as interesting as an arrestor cable breaking on a US Naval carrier.
When electric guitar strings are plucked very hard, most guitar amplifers become a fuzz box, without needing the external fuzz box.
Reminds me of Michael Hedges recordings of his Acoustic Guitar.
I always cringed and waited for the strings to break.
Could be almost as interesting as an arrestor cable breaking on a US Naval carrier.
Some COAX is 20 pF/foot. But you need to look at the cable spec.
RG8, RG9. RG58 & RG59 are more like 30 pF/Foot. But those are unlikely running across the floor from a Gibson.
I think Belden publish their data.
Has anyone out there ever heard of 'light foot"? Not the mayor of Chicago.
Or the famous Canadian folk singer. And song writer.
If we do the math it turns out that light travels One Foot in One Nanosecond.
So 8 inches in a piece of RG 58 COAX, That was good to know during my days in the research lab.
But doesn't mean much for our home audio.
RG8, RG9. RG58 & RG59 are more like 30 pF/Foot. But those are unlikely running across the floor from a Gibson.
I think Belden publish their data.
Has anyone out there ever heard of 'light foot"? Not the mayor of Chicago.
Or the famous Canadian folk singer. And song writer.
If we do the math it turns out that light travels One Foot in One Nanosecond.
So 8 inches in a piece of RG 58 COAX, That was good to know during my days in the research lab.
But doesn't mean much for our home audio.
MILES of cable? Or just a few feet?
For guitar specifically, you usually want to stay in hearing-range of the amplifier. And what may be a stricter limit: well within tolerable latency. The cable latency is insignificant but 1 foot of AIR is one milliSecond so 40 feet from speaker to ear is a LOT to play with.
Yes, you can drape your cable all around town and back. Or you can have a back-line amp 50 feet to your rear and a stage-monitor at your feet.
But "in general" you will never be driving even 40 feet (12m) of cable. So 1,000-1,500pFd.
Also guitar (especially e-guitar) does NOT go as high as a mike on keys. (What will we do for a sizzle-test when all our keys become phone-apps?) 6 KHz tops, really, while percussion goes an octave or more higher.
1,500pF at 6kHz is 18k Ohms reactance over a typical length guitar cord. Even if excessive boost puts 1.8V peak on it (which would make most g-amp inputs faint), that is only 0.1mA. A 1mA tube can do that clean enough for jazz.
OTOH many microphone situations ran thousands of feet of cable. Which is well under 1k at the top of the audio band. It used to be rare to have mike preamps on/near stage. (Today they do it with remote-controlled multiplex boxes, but that has not changed new or legacy mikes yet.)
I think this is what all guitarists expect now, and so that's what amp manufacturers give us.
But, decades ago, Leo Fender and Les Paul seemed to want something entirely different. Leo spent years refining his guitar amplifiers to try and make them sound cleaner and cleaner. His early amps were lower power and more prone to distort; he didn't like that, so output power kept going up, and he tinkered with the circuitry to try to keep the amps "clean" even at full volume.
The Fender Twin, for instance, was intended to stay clean (no obvious distortion) and still get more than loud enough.
Similarly, Les Paul always used a squeaky-clean tone. (Very unpleasant to my ears now, but perhaps people's expectations were different in the 1950s. And he had turned himself into a jazz player, and complex jazz chords sound best with very little amplifier distortion to obscure them.)
But there is always a guitarist for whom no amp is ever loud enough. Inevitably, some blues guitarist turned up a Fender Twin to ear-bleeding level, liked the overdriven sound with longer sustain that resulted, and so did it again. Same story with the Fender Bassman, intended to be a clean bass-guitar amp, but much more successful as an overdriven e-guitar amp.
-Gnobuddy
Gnobuddy,
Thanks for your views of guitar amplifier users!
I believe that is why there are three different subject areas on diyAudio . . .
Tubes / Valves: . . . Hi Fi and Stereo, done with tubes (yes, with some solid state parts / hybrids)
Instruments and Amps: . . . Guitar Amplifiers done with tubes, and with solid state
Solid State: . . . Hi Fi and Stereo, done with solid state
Thanks for your views of guitar amplifier users!
I believe that is why there are three different subject areas on diyAudio . . .
Tubes / Valves: . . . Hi Fi and Stereo, done with tubes (yes, with some solid state parts / hybrids)
Instruments and Amps: . . . Guitar Amplifiers done with tubes, and with solid state
Solid State: . . . Hi Fi and Stereo, done with solid state
Last edited:
In the late 1990s, I did some research work on using DSP to flatten the frequency response of a loudspeaker. (Routine today, but an idea ahead of its time a quarter-century ago, thought up by the owner of the company. A very smart man.)
I immediately ran across the problem of audio delay. At a 44.1 kHz sampling rate, you get one data sample roughly every 22.7 micro seconds. Both A/D and D/A converters need a data buffer to make sure that there are no dropouts. And if your total buffer size is, say, 256 samples, you now have nearly 6 milliseconds of audio delay! And that is not a particularly big buffer size at all.
Six milliseconds seemed like a lot to me. So I did a few informal tests to find out how much delay was acceptable to musicians. I took a guitar to work and tried it out, running the signal through a DSP processor set up to produce adjustable delay.
For me, 10 mS was okay. 15 mS started to feel weird, as though my guitar strings had grease on them and weren't responding to my fingers. 20 mS was akward. 50 mS was almost unplayable.
The funny thing is that 10 mS of electronic delay felt worse to me than having the speaker 10 feet away producing the same 10 mS of acoustic delay. (Guitar amp 10 feet away isn't unusual). I have no idea why there was a perceived difference. Probably something to do with hundreds of millions of evolution of the ear/brain system, which has made it very good at dealing with acoustic delays in an enclosed space (cave, room).
The faster you can play, the worse the delay problem. I'm not a particularly fast guitarist, and can only play around 12 notes per second, tops. We found a couple of drummers who could rattle out snare-drum rolls at much higher rates than that. They found 10 mS delay acceptable, but noticeable - the sound from the speaker came back well after their sticks had hit the drumskin. Luckily for them, the drums were loud enough that they didn't need a monitor, they just listened to the drums and ignored the speaker.
There have been guitarists who used 100-foot cables, who liked to wander out into the audience, far from guitar amp and monitor speakers, while still playing. I would bet those guys were playing slower forms of music, mebbe long bluesy notes.
I would also bet Yngwie and Vai and other lightning-fast shredders, stayed much closer to their guitar amps. Long acoustic delays would probably have bothered them.
-Gnobuddy
I made a LOT of terrible-sounding guitar amplifiers, spread over many years, before I finally realized that everything I knew about Hi-Fi, didn't work for e-guitar. Then I had to start over, trying to learn what made an e-guitar amp sound good.
The fact is that Hi-Fi amps and speakers sound awful with e-guitar, particularly with solid-body guitars.
And guitar amps and speakers sound awful for Hi-Fi. No doubt about that, either!
Our own Juan Fahey put it very well in a recent post, where he said guitar amplifiers are controlled distortion generators, and guitar speakers are final-stage audio equalizers. 🙂
-Gnobuddy
Gnobuddy,
Thanks for making me think again about delays . . .
Acoustic propagation in air, approximate, 1 foot / second (1 second)
Electric field propagation in air, approximate, 1 foot / Nano-second (0.000000001 second).
Electric field propagation in slow wires (good coax is faster), 1 foot / two Nano-second (0.000000002 second).
Worry about the delay of a 40 foot guitar wire? No
Worry about the acoustic timing of one performer on stage, and another performer with a guitar cable that is 40 feet out in the audience . . . Perhaps.
Your Velocity May Vary
Thanks for making me think again about delays . . .
Acoustic propagation in air, approximate, 1 foot / second (1 second)
Electric field propagation in air, approximate, 1 foot / Nano-second (0.000000001 second).
Electric field propagation in slow wires (good coax is faster), 1 foot / two Nano-second (0.000000002 second).
Worry about the delay of a 40 foot guitar wire? No
Worry about the acoustic timing of one performer on stage, and another performer with a guitar cable that is 40 feet out in the audience . . . Perhaps.
Your Velocity May Vary
I suspect this is one of the reasons why classical orchestras evolved to need a conductor. One person, centrally placed, to dictate timing to everyone - and to do it using light, not sound (i.e. waving his baton as a visual signal). Light travels at 3x10^8 metres/second rather than 3x10^2 metres/second for sound in air - so there's a million times less lag with a baton!
If there was no conductor, I think musicians on the left of a 50-foot wide stage would listen to players on the right, and lag behind them by 50 mS. Players on the left would hear the laggy sounds and respond 100 mS late. Et cetera.
I would imagine it could get very ugly, very quickly, with a big stage and a hundred musicians spread all over it!
It probably helps that classical orchestras did not improvise, but played pre-written parts that they had already rehearsed over and over hundreds of times. No need to listen to the other guys - just watch the conductor, read your sheet music, and play accordingly.
It seems the same methodology worked for the big-band swing era as well. The only change was that the conductor was now replaced by a bandleader.
Live performances of popular music in recent decades are done differently. Every (good) musician listens to everyone else in the band, so that they all stay in time and contribute to the performance. A lot of stuff is improvised, at least to some degree. It's not note-for-note identical to every other performance of the same song. I think all this makes it necessary for the musicians be able to hear each other without too much delay getting in the way.
Stage monitors placed close to each band member probably help a lot. No one is trying to listen to someone 50 feet away, but rather to the stage monitor 10 feet away.
-Gnobuddy