No disagreement from me there. Certainly the behaviour of the two devices under heavy overdrive is extremely different.
But for smallish signals, well away from clipping / saturation / cutoff, the sprinkling of mostly second harmonic distortion from a JFET does seem to sound acceptably like a vacuum triode operated similarly away from its limits.
As a bonus, some back-of-the-envelope calculations I did suggest that (some) JFETs are probably the lowest-noise amplifying devices we have for interfacing directly with a traditional high-impedance guitar pickup. Just pick one with a low corner frequency for 1/f noise, and reasonably good noise specs above that. BJTs aren't really in the running at source impedances this high, and vacuum triodes are handicapped by the high cathode temperature (which raises thermal noise) as well as flicker noise.
To be fair, guitar-amp hiss hasn't been much of an issue for me personally, probably because my tastes run to relatively low SPL and relatively low gain.
-Gnobuddy
However it still needs to be finished. So far I bought a cheap 12 inch speaker that has a coil rub, and an amp box that was smaller than it looked in the advert. Grrr.
Remove the gasket, then carefully remove the dust cap, shim the voice coil (to center and to arrest it in it's correct vertical position), unglue the surround and the spider from the basket/frame, using MEK. The voice coil still sits firmly in it's position? Then glue first the spider and then the surround to the basket again.
If the speaker' values justifies your effort
.
Best regards!
If the speaker' values justifies your effort
.Best regards!
You have my commiserations. That stinks.
Since much of this thread has been about exploring the road not usually taken, I'll mention that I've been on the hunt for affordable speakers for small guitar amps for a while. After a lot of searching and at least one dead-end, I may have located another good candidate for experimentation.
A little background: in the early years of the moving coil loudspeaker, amplifier power was expensive and meagre, so the focus was on maximizing efficiency. That led to big cones (10" - 15") made of the thinnest possible paper for lightness. Voice coils were similarly light. There was no separate surround - just corrugate the edge of the paper cone, that'll do just fine. Magnet materials of the time were weak and expensive by today's standards, so the speaker had a relatively weak magnet.
Unfortunately, it turns out that approach has many sonic limitations. In the world of good-quality audio reproduction (Hi-Fi, for lack of a better term), the speaker quickly evolved into something different, and capable of producing much more accurate sound, at the expense of sheer efficiency. Surrounds and spiders got softer. Cones got smaller, stiffer and heavier. "Whizzer" cones were added to the centre to extend treble response. Later, separate tweeters and crossover networks became usual (maybe mid-ranges too.) Magnets and magnetic structures (pole-pieces) became massive. Voice coils became longer and heavier. The materials used in construction improved. All this added cost and complexity.
But there were a few shadowed areas of the loudspeaker world in which the early crude speaker barely evolved at all, for one reason or another. One of those areas was guitar speakers. They are still, for the most part, made the same way as they were in 1940: a single big light floppy cone with a stiff corrugated surround, and a singular focus on efficiency, at the expense of almost everything else. Electric guitars don't produce bass below 83 Hz, or treble above maybe 6 kHz, so the simple, single large speaker worked tolerably well.
(Some of today's guitar speakers do use better magnets, glues, and materials than their 1940s forebears. But the changes are slight compared to the evolution of Hi-Fi speakers in the same period.)
Because of all this, today's guitar speakers ought to be inexpensive: they are still made with very few of the improvements that gradually raised the cost of Hi-Fi drivers and speaker systems. But as all electric guitarists know, guitar speakers today are actually very expensive. The cost is completely unreasonable for what they actually are. Instead, the cost is driven by "image". If you believe the narcissists, image is everything, so you will pay dearly for it.
The next question is, were there perhaps other dim corners of the speaker manufacturing world in which ancient, primitive speaker designs continued unevolved? Maybe an unfashionable area, where pricing was driven by reality, not by a manufactured "image"?
It turns out there is at least one such area. The ceiling-mounted paging / public address speakers used in factories, warehouses, machine shops, dispatching stations, and other similarly unglamorous work places, are still built a lot like speakers from WW II, i.e, like electric guitar speakers. Big floppy paper cone, check. Stiff paper surround, check. Focus on efficiency, check. Budget prices reflecting product reality and not advertised image, check. (We're talking $10 or $20 instead of $100 or $200 for the famous-brand guitar speakers.)
So I began hunting for mass-production paging speakers like this. While they're built very much like vintage guitar speakers, there have been two changes: they are now made to an industry-standard 8" size, and many have a smaller "whizzer cone" mounted in the centre, to extend treble frequency response.
This means two things: One, if you want sheer ear-bleeding SPL, you probably need the big 12" $300 big-name guitar speaker. Two, if we use an 8" paging speaker for e-guitar, we might need to do something to limit the extended treble response, which we probably won't want for e-guitar.
This post is already too long, so I'll continue it later today.
-Gnobuddy
> sheer ear-bleeding SPL, you probably need the big 12"
No. 16 Eights in array will whup most single Twelves.
Arrays have real problems of balance and coherency. But they are LOUD.
Look for Marshall Full Stack. 8 Tens in two stacked boxes. It really is the same idea: Jim had access to inexpensive PA speakers.
No. 16 Eights in array will whup most single Twelves.
Arrays have real problems of balance and coherency. But they are LOUD.
Look for Marshall Full Stack. 8 Tens in two stacked boxes. It really is the same idea: Jim had access to inexpensive PA speakers.
> sheer ear-bleeding SPL, you probably need the big 12"
No. 16 Eights in array will whup most single Twelves.
Arrays have real problems of balance and coherency. But they are LOUD.
Look for Marshall Full Stack. 8 Tens in two stacked boxes. It really is the same idea: Jim had access to inexpensive PA speakers.
No. 16 Eights in array will whup most single Twelves.
Arrays have real problems of balance and coherency. But they are LOUD.
Look for Marshall Full Stack. 8 Tens in two stacked boxes. It really is the same idea: Jim had access to inexpensive PA speakers.
As far as I know, Jim Marshall used 12's by Celestion. It wasn't his own idea, but that of Pete Townshend and/or John Entwistle, wo wanted it as loud as possible, to get eight speakers into one cabinet for a total capability of 200 watts. Soon they found out that these were too heavy and a PITA to haul. So they have cut them in half and got the now famous 4 x 12 stack.
8 x 10's were/are sold as well. The first ones were introduced by Ampeg in 1969 and sold with their 300 watt SVT heads. These featured proprietary 32 ohms drivers, which are roumored to be made by Eminence especially for Ampeg.
Best regards!
8 x 10's were/are sold as well. The first ones were introduced by Ampeg in 1969 and sold with their 300 watt SVT heads. These featured proprietary 32 ohms drivers, which are roumored to be made by Eminence especially for Ampeg.
Best regards!
And from super-loud arrays of super-expensive 12" speakers, I'd like to go back to affordable 8" speakers, for low powered guitar amps of the sort discussed in this thread.
Okay, so stripped of marketing BS, a guitar speaker is basically a primitive speaker with a big floppy cone that beams treble like a searchlight, breaks up badly by 3 - 4 kHz, and has negligible sound output above maybe 5 - 6 kHz.
To make e-guitar sound good, our primitive speaker should preferably have a few other additional characteristics. The attached image shows the frequency response of a popular guitar speaker, along with some of its characteristics which I've noticed are frequently found in other guitar speaker frequency responses as well.
In the search for budget paging and public address speakers for e-guitar, then, we should look for at least some of these guitar-friendly characteristics. In particular, the slow and steady rise in SPL between roughly 300 Hz and 3 kHz, which is not a typical feature of paging speakers, but is ubiquitous in guitar-specific speakers.
Without this significant treble boost, guitars seem to sound dull and muffled, and valve amps sound less "valvey", probably because the treble boost brings up the level of the harmonics generated by the valve, effectively increasing the amount of low-order harmonic distortion arriving at the listener's ears.
Electric guitars make nasty harsh sounds at 6 kHz and higher, particularly when any distortion or overdrive is part of the sound. So the drastic treble cut above 3 - 4 kHz, produced by guitar speaker cone break-up, is also highly desirable. In conjunction with the slow treble boost from 300 Hz - 3 kHz, this built-in speaker EQ curve emphasises the low-order valve distortion some ears enjoy, and suppresses the high-order harshness they don't.
To sum up, our cheap P.A. speaker for e-guitar should, if possible, have at least some of the following characteristics:
1) A fundamental resonance not above 100 Hz, preferably between roughly 80 Hz and 100 Hz.
2) A poorly controlled fundamental resonance (high Qts, caused by a weak magnet or too-small sealed enclosure) if we want a little deep bass boost, like those famous 4x12 Marshall cabs ( Frequency Response of a Marshall 4x12 Cabinet ).
3) A slow and gradual rise in frequency response from around 300 Hz to around 3 - 4 kHz, if we want bright Fendery cleans.
4) An abrupt and drastic fall in high-frequency response above 3-4 kHz or so.
In my search, I found items 1) and 2) easy to find. Item 3) is considerably more difficult: a big treble boost like that does not sound good for actual paging and background music applications, so it's less likely to be found in speakers designed for those purposes.
Item 4) is also less likely to be found in a paging speaker, which is typically intended to be a "full range" speaker, i.e. to make at least some pretence of reproducing treble adequately for low-fi music applications.
-Gnobuddy
Okay, so stripped of marketing BS, a guitar speaker is basically a primitive speaker with a big floppy cone that beams treble like a searchlight, breaks up badly by 3 - 4 kHz, and has negligible sound output above maybe 5 - 6 kHz.
To make e-guitar sound good, our primitive speaker should preferably have a few other additional characteristics. The attached image shows the frequency response of a popular guitar speaker, along with some of its characteristics which I've noticed are frequently found in other guitar speaker frequency responses as well.
In the search for budget paging and public address speakers for e-guitar, then, we should look for at least some of these guitar-friendly characteristics. In particular, the slow and steady rise in SPL between roughly 300 Hz and 3 kHz, which is not a typical feature of paging speakers, but is ubiquitous in guitar-specific speakers.
Without this significant treble boost, guitars seem to sound dull and muffled, and valve amps sound less "valvey", probably because the treble boost brings up the level of the harmonics generated by the valve, effectively increasing the amount of low-order harmonic distortion arriving at the listener's ears.
Electric guitars make nasty harsh sounds at 6 kHz and higher, particularly when any distortion or overdrive is part of the sound. So the drastic treble cut above 3 - 4 kHz, produced by guitar speaker cone break-up, is also highly desirable. In conjunction with the slow treble boost from 300 Hz - 3 kHz, this built-in speaker EQ curve emphasises the low-order valve distortion some ears enjoy, and suppresses the high-order harshness they don't.
To sum up, our cheap P.A. speaker for e-guitar should, if possible, have at least some of the following characteristics:
1) A fundamental resonance not above 100 Hz, preferably between roughly 80 Hz and 100 Hz.
2) A poorly controlled fundamental resonance (high Qts, caused by a weak magnet or too-small sealed enclosure) if we want a little deep bass boost, like those famous 4x12 Marshall cabs ( Frequency Response of a Marshall 4x12 Cabinet ).
3) A slow and gradual rise in frequency response from around 300 Hz to around 3 - 4 kHz, if we want bright Fendery cleans.
4) An abrupt and drastic fall in high-frequency response above 3-4 kHz or so.
In my search, I found items 1) and 2) easy to find. Item 3) is considerably more difficult: a big treble boost like that does not sound good for actual paging and background music applications, so it's less likely to be found in speakers designed for those purposes.
Item 4) is also less likely to be found in a paging speaker, which is typically intended to be a "full range" speaker, i.e. to make at least some pretence of reproducing treble adequately for low-fi music applications.
-Gnobuddy
Attachments
A few years ago I found a $10 paging / public address speaker which turned out to have a 130 Hz fundamental resonance, and didn't have the built-in slow treble rise. With an equivalent treble boost built into the preamp, it sounded quite good for clean e-guitar, though. (Deep bass was a bit lacking, which might bother some, but it actually made the guitar "sit better in the mix" at jam sessions.)
Encouraged by that, I bought a second slightly more expensive full-range speaker, which turned out to be a bit of a dud. It wasn't awful, but was just a bit dull and characterless. Usable, but not thrilling. The problem seemed to be, once again, the lack of that steady rise in midrange response, from roughly 300 Hz to roughly 3 kHz.
Just recently, I found a budget speaker ( Misco JC8FD - Free Shipping Across Canada over $150 ) with the manufacturer's datasheet frequency response shown in the attached image. It seems to have almost all the characteristics I'm looking for for e-guitar use, except for one: there may be too much treble output above 5 kHz for distorted e-guitar.
If that does turn out to be an issue, I can think of a couple of ways to address that, though. One (suggested by J.M. Fahey in another thread) was to simply amputate the "whizzer cone" with an Exacto knife or curved nail-scissors. It would remain to be seen if this also removed the 3 kHz peak, which we do not want to lose.
Another possibility would be to add a small series inductor (I estimate around 200 uH - 300 uH) to roll off high treble. How well this works will depend on the guitar amp's output impedance - if it has a high output impedance, a series inductor won't do much.
A small parallel capacitor across the speaker terminals is another possibility, and could be used in conjunction with the small series inductor (basically making a simple crossover network of the sort normally used with woofers in Hi-Fi speaker systems.)
Finally, there is also the option of filtering out the high treble acoustically, for example, by using a sheet of open-cell urethane or similar foam between speaker and grille cloth. Simple, and easily tailored by varying the foam thickness.
I think this speaker looks like a very promising candidate, so I'm definitely going to get at least one of them to experiment with.
Incidentally, the speaker brand (MISCO) seems to be derived from "Minneapolis Speaker Company", so those of you in the USA can probably find this speaker in America at much lower cost than I'd have to pay here in BC.
-Gnobuddy
Encouraged by that, I bought a second slightly more expensive full-range speaker, which turned out to be a bit of a dud. It wasn't awful, but was just a bit dull and characterless. Usable, but not thrilling. The problem seemed to be, once again, the lack of that steady rise in midrange response, from roughly 300 Hz to roughly 3 kHz.
Just recently, I found a budget speaker ( Misco JC8FD - Free Shipping Across Canada over $150 ) with the manufacturer's datasheet frequency response shown in the attached image. It seems to have almost all the characteristics I'm looking for for e-guitar use, except for one: there may be too much treble output above 5 kHz for distorted e-guitar.
If that does turn out to be an issue, I can think of a couple of ways to address that, though. One (suggested by J.M. Fahey in another thread) was to simply amputate the "whizzer cone" with an Exacto knife or curved nail-scissors. It would remain to be seen if this also removed the 3 kHz peak, which we do not want to lose.
Another possibility would be to add a small series inductor (I estimate around 200 uH - 300 uH) to roll off high treble. How well this works will depend on the guitar amp's output impedance - if it has a high output impedance, a series inductor won't do much.
A small parallel capacitor across the speaker terminals is another possibility, and could be used in conjunction with the small series inductor (basically making a simple crossover network of the sort normally used with woofers in Hi-Fi speaker systems.)
Finally, there is also the option of filtering out the high treble acoustically, for example, by using a sheet of open-cell urethane or similar foam between speaker and grille cloth. Simple, and easily tailored by varying the foam thickness.
I think this speaker looks like a very promising candidate, so I'm definitely going to get at least one of them to experiment with.
Incidentally, the speaker brand (MISCO) seems to be derived from "Minneapolis Speaker Company", so those of you in the USA can probably find this speaker in America at much lower cost than I'd have to pay here in BC.
-Gnobuddy
Attachments
Hi, nice FR.
Regarding the foam trick, i recently try 1/2" foam between grill-cloth and baffle board with a 2.5/3" hole centered with the cone (the donut beam blocker?). It is supposed to widen the hf dispersion and to flatten on axis treble response. people have used 3/4 to 1" foam for that.
In my case, an old blues jr/stock speaker with 18w lite II inside, it works like a charm, a bit wider sweet spot, nicer sounding and no grill rattle. So i did it to my 5F2/emi1058 speaker with the same results.
8" speakers are less beamy than 12" but foam trick could do well.
I don't know why but conjonctive or zobel network never did it for me.
Regarding the foam trick, i recently try 1/2" foam between grill-cloth and baffle board with a 2.5/3" hole centered with the cone (the donut beam blocker?). It is supposed to widen the hf dispersion and to flatten on axis treble response. people have used 3/4 to 1" foam for that.
In my case, an old blues jr/stock speaker with 18w lite II inside, it works like a charm, a bit wider sweet spot, nicer sounding and no grill rattle. So i did it to my 5F2/emi1058 speaker with the same results.
8" speakers are less beamy than 12" but foam trick could do well.
I don't know why but conjonctive or zobel network never did it for me.
Thanks for your feedback!
I read about the foam donut trick on some other forum, where I think it was called the "Mitchell Doughnut". As you say, the purpose is to improve treble dispersion.
The "Mitchell Doughnut" works because that open-cell foam absorbs treble better than it absorbs bass. Cut into a doughnut, it reduces treble radiated by most of the cone, allowing only treble from the central 3" region to come through. In effect, the big 12" speaker turns into a 3" tweeter at high frequencies. A very clever idea!
With the little 8" Misco speaker, the problem I'm concerned with is something else: the treble response doesn't fall as sharply above the 3 kHz peak, compared to "real" guitar speakers. That might sound harsh, particularly for overdriven or distorted guitar sounds.
So what happens if we cover the entire Misco speaker with foam, without cutting the 3" hole in the middle? It should "turn down the treble", and we can control how much by choosing the thickness of the foam. I think this might be one way to tame excessive high frequency harshness, if that turns out to be a problem.
-Gnobuddy
In an
SOS interview with Charlie Watkins he makes the following comment describing the first PA he designed in the late '60s:
Charlie Watkins said:
Thank your for that link. I found it fascinating reading. I knew a little about WEM the sound-reinforcement company, and I'd heard about the Watkins Dominator, but that was all I knew. I had no idea who Charlie Watkins was, or just how much crucial technology he'd contributed to the popular music we all grew up with.
What an amazing man, and what a life story!
-Gnobuddy
You can try my other hack: mount the driver back-to-front.
It is serious ugly but significantly cuts treble from a wizzer equipped cone.
Can't remember the theory (my brain is suffering a bad case of Lord Lucas today) but solid patterns in front of the cone also work as filters. You could probably mount the baffle halfway back and use a carefully designed front baffle to tune the response the way you want.
But cutting the wizzer off might be easierThanks, that's one more idea to add to the list of things to try.
I recently found out that one of the studio tricks to tame a harsh electric guitar sound is to mic the rear of the speaker cab, and not the front.
I've tried two different budget-priced P.A. speakers in my guitar amps so far, both with whizzer cones. One of them actually sounds dull, even with the whizzer. Probably because it's missing that built-in treble boost from 300 Hz to 3 kHz.
-Gnobuddy
As a postscript to all this single valve agony: My take away experience was the simple FET front end gain/overdrive section. I have inserted that in both my single ended 6V6 Princeton style amps with some success.
Post about it here. 6AQ5 SE 4w guitar amp schematic and sound clips
Post about it here. 6AQ5 SE 4w guitar amp schematic and sound clips
As one of the few people here who is willing to mix and match solid-state and hollow-state devices in the same amp, you might find this interesting. Some years ago, a Russian engineer who goes by "KMG" figured out a way to make small MOSFETs have almost identical transfer functions to the triodes in a 12AX7. He then went on to build several guitar preamps and power amps using his ideas. There are soundclips on his website too, and I found them pretty impressive.
Here is one example: Fet version of the JCM800 using LND150 mosfets
And here (attached) is my LTSpice simulation of a single KMG MOSFET "triode", i.e. one LND150 MOSFET with the associated components that make it behave like a vacuum triode.
In the simulation, the upper green trace is the output (divided down with two resistors), the lower purple trace is the input. The fake "triode" has been pushed well into nonlinear operation, but the output signal shows none of the sharp corners (and harsh sound) we usually see from solid-state circuits driven to clipping.
As far as I understand it, this is what's going on in KMG's MOSFET-emulated "triode":
R2 is the "grid stopper".
R3, R4, and Schottky diode D1 emulate grid current flow which occurs in triodes on large positive signal swings, when the grid reaches zero or positive Vgk.
Diodes D2 - D6 (along with R3) lower the transconductance of the MOSFET to match the 1600 uA/V of the triodes in a 12AX7. Because dynamic resistance of a semiconductor diode increases as current through it drops, the diodes also cause the MOSFET to turn off slowly and progressively as gate voltage drops, like a vacuum triode. This shapes the positive peaks of output signals at the FET drain, gradually rounding them off instead of the usual sharp-cornered solid-state cutoff clipping.
R5 sets the DC and very low frequency gain, to match that of a similar vacuum triode stage. C1 ensures that the lower end of R5 is at AC signal ground.
R6/C2 sets the AC gain to match the frequency response of a vacuum triode. R6 emulates the internal "cathode resistance" of the triode, while C2 is the external cathode bypass cap.
The negative supply rail (-3.3V) is necessary to make the MOSFET match the triode's gain over the entire low frequency range, from DC up to a few kilohertz. Using the negative rail allows R5 to set both DC operating point and very low frequency gain, while R6/C2 set the AC gain at higher frequencies.
KMG comments that you might need 100 pF across R1 to accurately simulate HF rolloff from an equivalent half-12AX7 triode (above 15 khz or so.)
It's a lot of components to emulate one single triode, but they are small cheap components...and the whole thing can end up smaller than the real triode. None of this would matter if the results were poor, but this is far and away the best solid-state emulation of a vacuum triode I have ever come across.
-Gnobuddy
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.