Tube Emulation & EQ

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
It is nearly finished.....well maybe. I need to wire up the transformers, and plug it in!
I'm glad you finally found it, and am looking forward to hearing how the eventual power-up goes!

For over a year now, I have been planning to make a guitar amp to gift to a friend. He is a senior citizen on a small fixed income, and dealing with a physical disability, so the amp needs to be reasonably small and light. I am on a tight budget myself, so the amp needs to be cheap.

I'd alread made some progress. I found an old thrift-store Sears "Hi-Fi" (anything but!) speaker for five bucks; it has a plastic grille with a Vox-like pattern, so it will become the enclosure. I also sourced a pair of more recent (80's?) boom-box speakers from the thift-store, and these donated their two woofers to the cause.

A few days ago I finished modifying the Sears enclosure to accept the pair of woofers. Then I hooked them up to an Ebay-sourced class-D power amp module, powered by a thrift-store 24V DC supply. The amp module is a pair of bridge-mode amps, one driving each 12 ohm speaker. I estimate something close to 20 watts RMS into each of the two speakers. 40 watts out to be ample for my friend.

I hooked up the class-D module to a Danelectro Fish-n-Chips guitar pedal - it's a 7-band graphic EQ, with a built-in clean preamp with adjustable gain up to 20 dB or so.

I plugged in my Yamaha electro-acoustic guitar, and it sounded pretty good through the combination of Fish-n-Chips preamp/EQ, class-D module, and boombox woofers mounted in the old Sears speaker enclosure. So far, so good.

Then I plugged in my old beater Epiphone LP Special II. And that's where things went seriously off the tracks: this electric guitar (solid body, pair of humbuckers) sounded absolutely awful. Harsh, ugly, and unpleasant.

It's not the guitar - plug it into any of my valve guitar amps, and it sounds fine. Not the greatest guitar in the world, but a perfectly usable instrument.

But into the solid-state preamp/class D power amp/boombox speakers? Awful. Completely unacceptable. No way I can give this to my friend, unless I can make it sound a lot better.

I've heard this awful harsh sound before, from every one of the solid-state DIY guitar amps I built in my first few years of learning to play the electric guitar, and then from a succession of commercial solid-stage electric guitar amps made by companies like Fender and Line 6. In fact, I came close to giving up playing the guitar entirely, because I sounded so awful, and I thought the problem was me.

That finally changed when I finally got my first valve amp. I was shocked that these primitive, obsolete amplification devices from the dawn of electronics made my guitars sound good, while every SS amp I'd tried made them sound awful.

Back to the present. My solid state gift-amp-in-progress sounds awful. Which leaves me wondering what exactly makes it sound so harsh and nasty, and what, if anything, I can do to remove at least most of the awfulness.

I've already found out that EQ alone doesn't do it (that's why I chose to use an EQ pedal as my temporary preamp.

I could try a speaker-emulation filter of some sort - a steep 3rd or 4th order low-pass, cutting everything above 4 kHz or so. It would go ahead of the class-D power amp module. Maybe that would remove the harshness? (But then, why did commercial SS guitar amps using real guitar-amp speakers still sound harsh??)

I also have various JFETS to experiment with in preamp designs. But I don't know if they can work magic on this thing.

Finally there is the option of using a $1 valve in the preamp. I'd have to come up with HV and heater power supplies, but I'll do it, if it's the only way.

The one thing I can't do is go all-valve. Too much weight, too much cost. So I have to do my best to make the class-D thing sound at least acceptably good.

Two big questions to attempt to solve. Where does the harshness come from? And how do I make it go away?

Hopefully I will have something worth reporting, as I look for a way forward with this thing.

-Gnobuddy
 
I think a lot of the sound we like out of guitar amps has to do with high frequency roll-offs.

I think it is more the way a 12" speaker will beam the highs.
I am hoping both of you are right, because filtering out high frequencies will be relatively easy to do, comparatively speaking.

There was a thread elsewhere on diyAudio where the subject of (guitar amp) loudspeaker frequency response came up. I grabbed the first frequency response I found on the net, drew a couple of straight lines on it, and then fiddled around with LTSpice to see if I could roughly duplicate the frequency response.

In the Hi-Fi world, speaker frequency responses are measured on-axis. In the world of Eminence guitar speakers? I have no idea, but if the published curve is on axis, as I suspect it is, it's already falling off dramatically (3rd or 4th order) above 3 kHz.

So if that's the on-axis response, imagine what the response looks like when you also point the speaker at your knees instead of your ears? Is it falling above 1 kHz? That's about a one-foot sound wavelength, small enough for a 12" speaker to beam noticeably. :eek:

In LTSpice, it took me an hour of fiddling and three virtual opamps (or one and a half dual opamps!) to roughly reproduce the Legend 1028 frequency response. In the attached schematic, the first opamp implements a crude first-order peak at around 3 kHz. The second one mimics the bass end of the speaker frequency response curve (2nd order high pass). And the third implements the high-frequency rolloff (2nd order low pass).

I'm hoping I can simplify the electronics with a little more tinkering. But, worst case, opamps are cheap.

Meantime, I have a couple of guitar pedals that are supposed to have speaker emulation built-in. Before I try to build my own, maybe I should try one of those pedals between a guitar and my nasty-sounding DIY amp to see if they improve things.

-Gnobuddy
 

Attachments

  • Eminence_Legend_1028_FR_001.png
    Eminence_Legend_1028_FR_001.png
    84.2 KB · Views: 899
  • Guitar_Speaker_Sim_011_No_Wien_Oct_14_2017.png
    Guitar_Speaker_Sim_011_No_Wien_Oct_14_2017.png
    23.4 KB · Views: 885
  • Guitar_Speaker_Sim_011c_Oct_14_2017.png
    Guitar_Speaker_Sim_011c_Oct_14_2017.png
    67.9 KB · Views: 907
That sounds like a worthy experiment.
And I haven't had a chance to try it yet - by the time I get back from work, it's too late to play guitar. Apartment living has its downsides!

It would be interesting to know what those speaker emulators are actually doing. Just EQ?
There is always superstition floating around when it comes to music gear, so there are probably some speaker emulators that contain blood from a pigeon's heart, moss taken from a gravestone at midnight, and the toe of a toad. :D

My old Super Champ XD has a crude speaker emulator built into the preamp ahead of the line-out jack. Fender published the official schematic. It's been a while since I looked at it, but as I recall, it was just a 3rd order low-pass filter implemented with an opamp or two.

I have one piece of Behringer gear, an active DI box that is supposed to include some sort of speaker/cab simulation. I think it's supposed to emulate a 4x12 cab stuffed with some sort of Celestions, which probably means it includes some comb-filtering because of interference between the four drivers in the cab. I've used this on a couple of home recordings, and I preferred the sound without the cab simulation...

One other data point: my most cherished guitar pedal, a Digitech Trio+, has two outputs, one for feeding to a guitar amp, the other for feeding to a P.A. system or flat-frequency powered speaker.

This latter output includes some sort of filtering to take the harshness out of the guitar when running into a full-bandwidth speaker, and it works quite well.

...a single-cap LPF with -3dB at about 5kHz. (That's what I remember, but perhaps it was lower, more like 3kHz.) That sounded pretty good to me.
I have tried a few similar experiments. To my ears, anything up to maybe 5kHz - 6 kHz is fine for clean tones, but 6 kHz sounds harsh with some types of overdrive.

I was quite surprised that the Legend 1028 rolls off as early as it does. Only three kHz? Geez!

That led me to the conclusion that a lot of the guitar amp's sound had to do with frequency response.
I agree, and I think this is true in every field of audio. Fifty to seventy years ago, everyone working in audio understood this, and that's why there was so much effort put into trying to attain 20 Hz - 20 kHz flat frequency response for Hi-Fi.

Since that time, Hi-Fi reached its technical and popularity peaks, the real engineers and researchers mostly left the field for good, and superstition and nonsense started to take over. so now we have people obsessed with WWII-era paper in oil capacitors and so on, things which actually don't matter at all to the sound.

The thing is, I've never seen much of anything published by actual researchers about guitar amps and guitar sound. Most of the real research is probably proprietary, locked up in the intellectual property of large corporations.

That leaves a lot of opinions, superstition, and a small mix of actual good information scattered around where you and I can find it on the Internet. It's not always easy to tell the good from the bad!

So, I took away that if you take a so-so amp, put a nice speaker on it (with a sound you like), and make sure the 'tone stack' you use has the kind of EQ/phase bumps and valleys you prefer, you can get that amp to sound pretty OK, at least as far as its clean sound is concerned.
And yet I've never heard a clean sound I like out of a solid-stage guitar amp, including various commercial ones (Fender, Line 6, etc)...

I'm convinced there is something more going on, whether it's the few percent of low-order harmonic distortion from triodes in the preamp, or something else.

One candidate: the typical 12AX7 triode gain stage is only biased to -1.5 V. From the datasheets, we know grid current in a 12AX7 is specified to reach 0.3 uA when the bias is -0.9 V. So if the guitar signal swings only 0.6 V positive (from -1.5V to -0.9V), grid current is already flowing in the input stage triode, and starting to soft-clip the positive peaks of the guitar signal.

As a rough estimate, a guitar with a 500k volume pot set to half-resistance has a source impedance a bit over 250k. 0.3 uA of grid current flow through this will drop 75 mV of voltage. The guitar had to put out a peak of 0.6 V - that's 600 mV - to make this happen. 75 mV is 12.5% of 600 mV.

So, input grid current flow is already squishing any 600 mV positive peaks from the guitar by about 12% before they ever make it through the first triode. Plenty of guitars can put out that much voltage, and that is definitely an audible amount of squishing, but I don't know exactly what it does for the sound of the guitar. Maybe it softens those harsh initial pick-attack transients just enough to make the guitar sound better, smoother, more musical?

I have never seen any discussion of this issue anywhere on the Internet or in print, though I brought it up once before, I think on this very forum. People have certainly thought about grid current flow in output valves during overdrive, and even in overdriven preamp valves in high-gain amps. But if anyone's thought about the possibility of grid current flow being an issue at the very input of a typical 12AX7 guitar preamp, they've kept quiet about it.

-Gnobuddy
 
A few vids on mic placement out there.

YouTube
It's amazing how much of a difference it makes to go off-axis by just a few inches.

Here is my first attempt at a very simplified speaker emulation circuit (simulation only, not built yet). This time, I made no attempt to duplicate the low-frequency end of the speaker response (my little boom-box woofers already have their own bass roll-off). So I tried to duplicate only the roughly 3 kHz peak and rapid fall-off above it.

I grew up with the wonderful BC147/148/149 transistor family, so I used one of their descendants in my circuit. It looks like one single transistor will do the job, though it must be fed from a low impedance, not direct from the guitar.

-Gnobuddy
 

Attachments

  • Guitar_Speaker_Sim_100_one_transistor_Oct_27_2017.png
    Guitar_Speaker_Sim_100_one_transistor_Oct_27_2017.png
    72.4 KB · Views: 856
I had a little time to spend tinkering with my budget hacked-together DIY guitar amp today. As before, I used a Danelectro Fish-n-Chips 7-band graphic EQ pedal as the preamp, mainly because it lets me quickly and easily tinker with voicing and gain.

In an earlier post I said that EQ alone wasn't fixing the nasty harsh sound I was hearing. I was wrong - the problem was that I wasn't going low enough in frequency with the EQ! I had tinkered with the 6.4 kHz, 3.2 kHz, and even 1.6 kHz sliders on the pedal, and none of them fixed the harshness. I didn't go lower than 1.6 kHz, because it seemed too low to bother with - surely the problem didn't go that low?

Well, it does! Today I found out that the 800 Hz band is the key to improving this particular amps sound. With the 800 Hz slider dialed well down, the nasty harsh sound disappears. I can bring the 1.6 kHz and 3.2 kHz sliders back up, and the amp sounds good, as long as 800 Hz stays notched out. But bringing up 6.4 kHz starts to sound scratchy - too much of a good thing.

Overall, the amp sounds best with a notch at 800 Hz, a gently rising bass response below that down to 100 Hz, and treble flat or slightly peaking at about 3.2 kHz, and the 6.4 kHz band rolled right down as low as it will go. The 3.2 kHz peak is pretty well in line with the guitar speaker frequency responses we've been looking at on this thread.

The speaker enclosure I'm using has a height of 463 mm (standing upright, this is the longest internal dimension between two parallel walls). It's about 19 degrees C in the house now, so the speed of sound should be around 342.4 m/S. ( Temperature and the Speed of Sound ). A wavelength of 463 mm calculates to a frequency of about 740 Hz. My EQ pedal has only 7 bands, each band covering roughly an octave, so 740 Hz may be close enough to 800 Hz to be what I'm hearing.

I might try stuffing the box with rags to see if that changes anything, but right now, I don't know if I'm dealing with a peak caused by acoustic reflections in the box, or a peak in the frequency response of the boombox woofers I'm using. Either way, it seems that inserting a notch at 800 Hz is a good way to remove the problem.

Perhaps a two-knob Fender tone control circuit (like the one in a Princeton reverb, with a fixed resistor for mids) would be a good choice for this amp. I just have to work out component values to put that mid-scoop at 800 Hz. Alternatively, I might put a twin-tee notch filter tuned to 800 Hz somewhere in the circuit.

And once again, I am thankful to Danelectro for their wonderful little Fish-n-Chips EQ pedal. Without the ability to quickly dial in different frequency responses, I doubt I would have found the source of the harsh sound from this amp, and certainly, not this quickly or easily.

-Gnobuddy
 
Can you try playing the "nasty" amp through a real guitar speaker?
I will try that today and report back. The only real guitar speaker I have at the moment is in my Princeton Reverb reissue, but it is wired with the usual 1/4" mono plug, so all I have to do is put a 1/4" mono jack on the class D amp in the "nasty" amp.

The one-octave dip centered on 800 Hz does a remarkably good job of de-nastifying the nasty amp, by the way. It goes from nasty to just the usual blah sterile-clean solid state guitar amp sound. A huge improvement.

So step one will be to implement a similar notch without using the Danelectro EQ pedal. I did a little tinkering with a twin-tee filter in LTSpice last night that looks promising, but it will take some more tinkering on the bench before it sounds right, I'm guessing.

With that done, I'm hoping I can improve on the "too clean" sound with a JFET or two in the preamp.

-Gnobuddy
 
Can you try playing the "nasty" amp through a real guitar speaker?
I did try this today. My guitar went into the Fish-n-Chips graphic EQ, the output from that drove the class D power amp module, and that in turn drove the stock speaker in my Princeton Reverb reissue, still mounted in its stock combo cabinet.

To start with, I left the EQ curve alone (as dialed in yesterday for the little boombox woofers). Through the Princeton Reverb speaker and cab, there was noticably more bass, but the overall sound quality wasn't all that different.

The next step was to bypass the EQ curve. The guitar sounded worse, with some harshness, typical of every solid-state guitar amp I've ever heard. But not as harsh as when using the boom-box speakers.

Step three was to set the Fish-n-Chips to a flat response (all EQ sliders centered), play the guitar through it into the Princeton Reverb speaker, and then, by ear, dial in the EQ curve that sounded best to me. (We're talking entirely clean tone here.)

Surprise - when I was done, the EQ curve was not all that different from what I'd dialed in for the two boombox woofers in their smaller enclosure, except for less bass boost. I still had a notch at 800 Hz, still had some bass boost, and still had the 6.4 kHz slider dialed all the way down.

I took a photo of the Fish-n-Chips (attached), and then sat down and measured its frequency response the old-fashioned way (function generator, Fish-n-Chips, oscilloscope). I have the data in numerical form in a Libre Office spreadsheet, if anyone wants them, but for now, I'll just post the frequency response plot I generated from the spreadsheet.

Evidently the 6.4 kHz slider does interfere with the 3.2 kHz response, as the frequency response I measured actually peaks at around 2 kHz, rather than 3.2 kHz.

So my next task is to try and generate something close to this frequency response from some simple analogue circuitry, which should give my guitar amp a good clean base tone on top of which I can experiment with adding harmonic distortion (JFETS), tone controls, et cetera.

There are lots of caveats - obviously, this frequency response was tuned for my ear, and maybe also for the two guitars I tried today (both twin-humbucker models).

Nevertheless, I think this method of voicing a guitar amp is a promising one, and I think it might save a lot of time, and produce better final results, than the traditional "tweak one capacitor at a time" method.

-Gnobuddy
 

Attachments

  • Fish_n_Chips_close_up_726px.png
    Fish_n_Chips_close_up_726px.png
    379.3 KB · Views: 822
  • Optimum_Preamp_Frequency_Response_plot.png
    Optimum_Preamp_Frequency_Response_plot.png
    30.5 KB · Views: 224
Jolly good show and all that..

Nevertheless, I think this method of voicing a guitar amp is a promising one, and I think it might save a lot of time, and produce better final results, than the traditional "tweak one capacitor at a time" method.
Filing this one away for use later.

I'm afraid my cheap Behringer EQ doesn't have a fancy name (oh the shame!) but I'm sure it'll do. :D
 
I'm afraid my cheap Behringer EQ doesn't have a fancy name (oh the shame!) but I'm sure it'll do. :D
In North America, admitting to owning a Danelectro pedal is almost as bad for your popularity as admitting you have a "social disease". :D

The Fish-n-Chips has a plastic enclosure crudely shaped to look vaguely like a fish, and a plastic stomp switch that responds to your toe with about the same precision as an overcooked potato. And it's cheap.

In the land of consumerism and baseball, that's three strikes against the Fish-n-Chips. Horrors!

I don't love the plastic enclosure or soggy plastic footswitch, but the electronics does exactly what it's supposed to do, and does it very well. I've found it to be an incredible tool for fixing or tweaking my guitar sound. :up:

-Gnobuddy
 
So my next task is to try and generate something close to this frequency response from some simple analogue circuitry
I did some tinkering with LTSpice, and a combination of a mistuned twin-tee notch filter, and a modified second-order lowpass filter, comes pretty close to the frequency response I'm targeting.

Screenshot attached.

-Gnobuddy
 

Attachments

  • Optimum_Preamp_Frequency_Response_plot_with_simulated_preamp_schematic_020.png
    Optimum_Preamp_Frequency_Response_plot_with_simulated_preamp_schematic_020.png
    149.7 KB · Views: 426
And now for the next interesting thing to share. I stumbled across this old thread yesterday: 2BJTE - 2 BJT triode emulator

The thread was started by yet another guy trying to make bipolar junction transistors sound good in a guitar distortion circuit. It's been done a thousand times, and it never sounds as good as a proper "toob" amp, right?

But this thread includes a link to a sound clip, and hearing it made me sit up and take notice. Here's the link: http://www.diale.org/mp3/v1.mp3

Now, to me, that is some pretty good-sounding rock guitar distortion. I have never heard anything that sounds as good as that come out of a BJT-only circuit before, so I am more than a little intrigued. Does this circuit really do what no other BJT circuit has ever managed to pull off - sound really "valvey" while driven deep into classic-rock guitar territory?

I threw the circuit together in LTSpice and it produces visually very similar waveforms to the ones posted in the thread.

I have also done a little fiddling and tweaking of component values in the simulator. One thing I can tell you is that the rounded corners of the negative half-cycles are very dependent on the bootstrap circuit - the values of the 500 ohm resistor (!) and 100 uF capacitor.

That is only a 50 millisecond RC time constant, so the capacitor discharges appreciably during each half-cycle at the low-frequency end of the guitars range (83 Hz), and this seems to cause the rounded corners of the negative half-cycles.

I can also tell you that fiddling with the 1k (!) emitter load of the second transistor changes the output waveforms drastically (harder corners, unlikely to sound good). Much though I would rather not operate a BC109 at 9 or 10 milliamps of collector current, that seems to be the magic spot for this particular circuit.

Note also the size of the input signal - this circuit is already beginning to distort with an input of 15 mV peak value, and is well and truly deep into clipping at 50 mV peak. There are going to be no cleans with this circuit and a normal guitar, not unless you attenuate the heck out of the signal before it gets to the first BJT.

This isn't necessarily a problem, though; you can get decent cleans out of JFET or two, and this two-BJT circuit could comprise a separate, parallel "dirt channel", with a switch or relay to changeover.

This is all hinged around one thing - how good that clip sounds. There is one big question mark in my head, though. The person who posted that sound clip says he used a BOSS OD3 overdrive pedal as a buffer/clean preamp to drive this two-BJT "tube emulator" circuit. He says the OD3 was set clean - but was it really?

The OD3 is an unusually valvey-sounding OD pedal that sounds very good in the right hands, and works its magic with a succession of slightly overdriven JFET stages, including a differential stage or two that mimics push-pull valves. It wouldn't be too far out of line to describe it as an analogue valve amp emulator built with JFETs. You can find plenty of good demos on You Tube if you want.

So the question in my head is, was it really the OD3 that produced the nice rock guitar timbres in that MP3 clip? Or was the OD3 really set clean, and did all that rich-sounding distortion really did come from that weird two-BJT/ weirdly bootstrapped/ weirdly mis-biased circuit?

The obvious thing to do would be to breadboard it and try it out, but I want to stay focused on my current project, building a preamp to take the "nasty" out of the little solid-state guitar amp I'm building to give to a friend, hopefully, this Christmas. Not to mention, I have no BC109s or BC108s around. I can't even find the 2N2222s I was sure I had somewhere. The only small-signal NPN BJTs I have are lovely BC550Cs, which have far higher beta than the BC109, and so may not sound the same in this rather oddball circuit.

But if any of you is interested enough to breadboard and try this out, I would love to hear about your results. (And if nobody else is interested, I will definitely build this and try it out once I'm at a good stopping-place with my current project.)

-Gnobuddy
 
Last thing to share - this (attached) is what I've been working on. This is a solid-state preamp with a JFET input stage / buffer / triode emulator, and a frequency response shaped to match the one I got from a 7-band graphic EQ set to product the best clean tone (to my ears, obviously).

To recap, I mounted a pair of small 6.5" woofers taken from a thrift-store pair of boombox speakers in an 8"x10"x19" MDF enclosure. This sounded absolutely awful with a solid-body electric guitar played through it using a solid-state preamp and power amp, an experience many of us may have shared if we tried playing an electric guitar through anything other than a real guitar speaker.

So far, so futile. The interesting part came when I found out that I could dial out virtually all the "nasty" sound using a graphic EQ pedal, and lots of trial and error.

This is a very interesting result to me, because real guitar speakers are expensive, and the cheap good-sounding alternative speakers from vintage Hi-Fi or vintage organs seem to be very hard to come by (except for Printer2, who gets them regularly from his tooth fairy. :) ) Also, real guitar speakers are physically large, and the light cone and tight cone suspension demand physically large speaker enclosures as well.

But, because smaller and more modern speaker designs always seem to sound absolutely nasty when you play a solid-body electric guitar through them, there was no acceptable alternative.

But if careful EQ can take out the nasty - which is what my ears are telling me - then a whole world of speaker possibilities opens up to us DIY guitar amp builder/tinkerers.

So I designed a circuit to emulate the frequency response I had dialed in with my EQ pedal, and it is going to be part of the preamp. If all goes will, it will work the same magic as the EQ pedal, and the pair of 6.5" woofers will sound good instead of nasty.

JFETs have such wide parameter spreads that it takes special circuit design techniques to control them. In a design intended to go into mass production, there is little alternative. But I'm building a one-off for a friend, so I chose a route that would allow me to use a simpler circuit design: I decided to simply measure one specific JFET from my junk-box, and design the biasing circuit for it using the specific parameters of that exact JFET. Throw in another JFET and it may not bias up properly, but as long as I use this particular JFET, it will work perfectly.

So yesterday I picked out and measured an MPF102 N-channel JFET from my box of goodies (Vp was 1.96 volts, Idss was 4.66 mA). With those all-important parameters known, I used the usual JFET equation and calculated that it would take a 1k source resistor to set Is at 1 mA. On the breadboard, it actually took 1.2k to get to the desired 1 mA.

Today I transferred the input JFET stage from the solderless breadboard to one of those general-purpose PC boards (protoboard), and added on the first BJT (buffer) stage. Powered up with 18V DC (from a pair of 9V flat batteries), everything biased up properly. So far, so good, then I ran out of time, and had to start on some of my Sunday chores.

I also realized I only have the 1nF caps I need for the twin-tee filter in a 10% tolerance. That's not going to cut it - one decibel corresponds to roughly a 12% change, and with three caps in the filter, using 10% components might change the frequency response by up to 3 dB.

That's too much, so I need to get some 1% or 2% film caps for the twin tee, and maybe also for the low-pass "speaker emulator" that comes next in the signal chain.

It will be slow going during the week, but I expect to make some more progress next weekend, if not sooner.

-Gnobuddy
 

Attachments

  • JFET_pre_w_EQ.png
    JFET_pre_w_EQ.png
    59.7 KB · Views: 478
...too clean and lifeless...plugged his ES335 into that amp...that combination did sound good
I have had the same experience several times. My Yamaha electro-acoustic guitar, or my Korean ES 335-like thin semi-hollow guitar, will sound fairly good plugged into just about any sort of amp, including keyboard amps, powered P.A. monitors, or home Hi-Fi. The guitars sound good on their own, and as long as the amp doesn't totally wreck the guitars own sound, the combination will sound good.

Things are very different with my solid-body guitars, though. All of them sound pretty nasty played straight into a flat-response amp and speaker. These guitars intrinsically sound nasty, and need the magic of "toobs" to make them sound good. But, paired up with the right "toob" amp, they can sound delicious.

A year or so ago, I stumbled across an undergraduate research project that attempted to mathematically model a typical guitar pickup / string combination. The big takeaway was that the magnetic field falls away with distance from the polepiece in an extremely non-linear fashion - and when the guitar string moves through this nonlinear field, it generates a quite distorted waveform, straight out of the pickup. This is where a good part of the harshness of a solid-body electric guitar comes from, from the nature of the magnetic pickup itself.

Evidently having a hollow guitar body that vibrates with its own carefully tuned response can take quite a lot of the harshness out of the signal from the bare guitar itself.

-Gnobuddy
 
<snip>6V6's were born to sing the blues.
<snip>the 6V6 has about half the emission capability of the 6BQ5 (EL84).
<snip>
The 6V6 has less than 3 watts of heat in it's tiny little cathode.
What happens if we take a valve that is more generously endowed in the heater power department, and then run the heater at reduced voltage?

The work function of thoriated tungsten is somewhere between 2 eV and 5 eV, while (kT) for electrons at typical cathode temperature of 1050 K is only about 90 meV. This means only a tiny fraction of the statistically most energetic electrons will ever manage to burst out of the metal cathode into the unwelcoming vacuum beyond. If the cathode is run a little cooler, we have raised the bar, and only an even tinier fraction of electrons will manage to escape into the vacuum. So we should be able to control the amount of electrons emitted from the cathode via cathode temperature, it would seem.

So can we turn an enthusiastic EL84 into a reluctant 6V6, just by cooling off it's cathode a bit?

I remember seeing a graph somewhere that showed valve lifetime plunging dramatically when the heater was operated at lower-than-normal voltage, though. I don't think the reason for the loss of longevity was explained. If correct, that might put a damper on the prospect of making cheap 6V6-alikes out of other, lower cost, valves.

Thoughts?

-Gnobuddy
 
For your listening enjoyment (ok viewing, that just sounds better).
Wow. That is just an amazing body of work. My hat is off to "KMG", whomever he is.

I know there are some things you can't hear until you're standing next to the loudspeaker, but for what they are, some of those clips sound really, really good.

I usually find that solid-state amps fail most obviously when it comes to what we guitar players call "clean tones", which aren't clean at all, but have several percent THD. But the Bogner Ecstacy cleans being imitated by an LND150 (in this thread: Fet version of the JCM800 ) sound really good.

Here is what KMG had to say, including the link to the sound clip:
KMG said:
First was Bogner Ecstazy preamp. Even without grid current emulation it gives results close to tube sound.
Fet version:
http://milas.spb.ru/~kmg/files/projects/xtcfet/pedal/samples/denn/kmg_xtc_mix.mp3

-Gnobuddy
 
Notch filter and lowpass (aka "nastiness remover") built and ready for audio testing (see pic).

I managed to cram my entire "de-nastify" filter (four active devices plus support circuitry) into one half of that little protoboard, so there is still some real estate left for whatever comes next. Direct-coupling the whole thing helped a lot, by cutting down the number of capacitors, and simplifying biasing of the 2nd, 3rd, and 4th active devices.

The board passed basic DC checks late last night, just before I went to bed, with the output sitting at right around 9V when powered by 18V DC.

The two BJT emitter followers each drop the usual 0.7 volts DC or so, but the MPF 102 source follower brings the DC voltage back up a couple of volts, so in the end, the output is at about the same DC voltage as the drain of the input MPF 102.

I'm hoping the input stage (MPF 102) will add at least a small sprinkling of the good kind of harmonic distortion. There is a 1.2k source resistor, chosen to provide the right DC operating point. For triode emulation (i.e. three-halves law) when biased exactly halfway between cutoff and saturation, the source resistor should be 353 ohms according to Dimitri Danyuk's paper ( Triode Emulator by Dimitri Danyuk | Field Effect Transistor | Amplifier ).

I have the option of AC coupling a second resistor in parallel (510 ohms or 470 ohms) with the 1.2k source resistor, so that the parallel combination is about 350 ohms to the AC signal.

However, real vacuum triodes don't exactly follow the three-halves law even when the anode is held at constant voltage, and they very definitely don't follow it when a non-zero anode resistor is connected between anode and B+. The presence of an anode resistor causes internal negative feedback, which straightens out the transfer characteristic a little bit, presumably reducing that exponent to something less than (3/2).

So I'm not going to worry about the exact AC value of that source resistor until I've at least had a chance to hear how it sounds, as-is, first.

If the notch filter and low-pass do their job, and "de-nastify" my amp, then I'll move on, and try to get some good clean tones, maybe using KMG's ideas and an LND 150.

If I get as far as having good clean tone, after that comes the question of tone controls. This one gives me a headache, as I have yet to find a tone control circuit that I really like for use with guitar.

The friend I'm building this for really only ever uses clean tones, so that is probably it as far as he's concerned. But I do want to investigate that two-BJT bootstrap distortion circuit (post #2108), to see if it really sounds as good as the posted MP3 clip.

-Gnobuddy
 

Attachments

  • IMG_1059.JPG
    IMG_1059.JPG
    769 KB · Views: 388
I got home late, but managed to get a few minutes at the work-bench. I hooked up a function generator to the input of the "de-nastifying filter", and a 'scope to the output, and powered the thing up.

I didn't have time for multiple accurate measurements, but a few quick twists of the frequency knob on the function generator showed that (a)there is indeed a notch in the frequency response, somewhere near 800 Hz, and (b) there is a slow rise in bass response below that, and (c) there is a slight peak around 2 kHz, followed by a fairly rapid fall-off.

So it looks like the circuit is more or less working as designed. Hopefully I'll have a chance to check it out properly soon, and also play some guitar through it.

I also found a few minutes to throw together KMGs first LND 150 tube emulation circuit from the thread that Printer2 linked. Here is a screenshot, and the LTSpice schematic file, if anyone wants to tinker with it. Oh yeah, also the LND150 model (LND150.txt).

-Gnobuddy
 

Attachments

  • KMG_MOSFET_Triode_Emulator_001_Nov_09_2017.png
    KMG_MOSFET_Triode_Emulator_001_Nov_09_2017.png
    82 KB · Views: 432
  • LND150.txt
    244 bytes · Views: 101
  • KMG_MOSFET_Triode_Emulator_001_Nov_09_2017.asc
    2.2 KB · Views: 90
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.