Can a "Ruby" style amp be made sparkly?

[Badside]

I breadboarded a Ruby amp as per ROG's schematic, including the MPF102 buffer. It's loud enough for my needs (just noodling in my appartments) but I found the tone to be highly missing in high-end contents. I'm using Celestion G12T-75 speakers, which are pretty bright by themselves.

I tried to experiment with bright caps on the volume pot, which did seem to move the frequency contents more towards higher-mids, but the top-end still sounds as muffled.

Anyone has tricks to make the JRC386D sparkle a bit?

Also, I left pins 1 and 8 unconnected, as per the datasheet for 26dB gain (I need low volume, so no point in increasing gain). Do pins 1 and 8 have any effect on the frequency response?

Thanks
Yannick

[PRR]

> tone to be highly missing in high-end contents

LM386 (and a properly setup FET buffer) are dead-flat to the top of the guitar band; to well past 100KHz in the low-gain mode you have.

> Celestion G12T-75 ... pretty bright by themselves.

I wonder if they are really bright at all levels, or if they ring when driven HARD.

I also wonder about a person's hearing when they ask for more "sparkle". I quit doing live sound (and working in a printing press) when I noticed my sparkle was fading away.

Nevertheless, classic rock guitar amp does have considerable top-lift. A flat amp may be boring.

> Do pins 1 and 8 have any effect on the frequency response?

You have a clue. Gain is 26dB open, 46dB shorted. They suggest shorting with a large cap to avoid subsonic and DC trouble.

What if you had a part which was "open" at low frequency, "short" at high frequency? That's the general trend of a capacitor.

Does not have to be dead-short/open; what we are really doing is fighting the 150 and 1350 ohm resistors inside.

So a cap impedance which was near the internal 1350 ohms resistance at 1KHz would leave gain at 26dB for low frequencies, and give a slight rise at 1KHz. About 9KHz it would equal the internal 150 ohm resistor, and tend to 46dB gain at higher frequencies.

That's the right idea. But 20dB top-boost is probably way too much. You can use a stopper resistor to limit the amount of rise.

Counting on thumbs: 0.01uFd and 1K ohms acts at 17KHz, much too high, try something like 0.1uFd. If the stopper is lots bigger than 1350 ohms it won't do much, if it gets close to 150 ohms that's probably more boost than we should need.

[wavley]

Sparkly may be covered in this thread http://www.diystompboxes.com/smfforum/index.php?topic=79112.0

[Badside]

I wonder if they are really bright at all levels, or if they ring when driven HARD.

For sure they sound much much better when punished by my 50W tube amp!
But for the sake of comparison, I'm used to my usual "living room noodling" amp which is a 8W tube amp (6K6 power tubes in push-pull) that I can barely open, so we're talking the same very low volume, and the 8W amp has no tone stack, but it does have some high end emphasis in the preamp that I tried to replicate with a similarly configured JFET preamp feeding the JRC386D.

So using pins 1 and 8 for an extra gain boost in the high-frequencies eh? Pretty nifty! Hadn't thought of that.
I could put say a 2K pot in there and experiment, or leave it there for tweakability.

I'll try this and I'll report.

I was thinking too that maybe my 386 is a dud. I ordered 4 of them (they're so cheap) so I'll swap around and see.

Note also that I bypassed the 220uF output coupling cap with a 0.1uF film cap, didn't really do much difference (at least none that a 12" guitar speaker could render). I also tried removing the "zobel" and that wasn't it either.

Thanks

[Scruffie]

Isn't the idea of the FET Buffer in the front to add sparkle...? If you raise the 1M5 Value to something like 10M like on the input of the SHO, would that work? My electronics knowledge is too limited to know but I do know that cutting those 10Ms on the SHO down to 1M significantly decreases the sparkle and that the 1M5 is setting the input impedeance... or I could just be completely wrong.

[PRR]

> Isn't the idea of the FET Buffer in the front to add sparkle...?

Nearly.

The problem is that the 50K into the '386 sucks all the sparkle out of guitar pickup. They were designed for much higher loads. 470K is usually ample. Only a few pickups and ears will show a difference for >500K, but anybody can hear a 50K loading. Hence a buffer, and the JFET is cheap and cheerful.

Sure, raise the gate resistor to several Megs and see if it matters.

[Badside]

Isn't the idea of the FET Buffer in the front to add sparkle...? If you raise the 1M5 Value to something like 10M like on the input of the SHO, would that work? My electronics knowledge is too limited to know but I do know that cutting those 10Ms on the SHO down to 1M significantly decreases the sparkle and that the 1M5 is setting the input impedeance... or I could just be completely wrong.

All of my amps have 1M input impedance, so do all the pedals I've built. Doesn't seem to affect my guitar that much.
Not to mention the Les Paul has a mere 250K volume pot (came stock that way), so from 1M to 10M wouldn't make much difference on loading characteristics.

But yeah, I built it with the FET buffer to prevent the extra loading on the pickups, but it's still too muffled for me. The best I've had so far was with 2 JFET gain stage before it, cascaded, which gave me a nice crunchy tone that took pedal well, but I'm still looking for more "sparkle" (high-end). I'll try the small cap and pot between pins 1 and 8 and I'll report back. Might make a cool "presence" control.

[mac]

Add a 0.0047uf or higher from pin 1 to Gnd.

mac

[PRR]

DOH!! I completely missed a factor.

> they sound much much better when punished by my 50W tube amp!

These speakers are tube for tube amps, which typically have a high output impedance, low damping.

The LM386 has a low output impedance, high damping.

The speakers have a significant impedance rise above 1KHz. On the tube amp, output rises. On the transistor amp, output is flat. The designer intended the rise.

This is a bedroom amp? Then you can tolerate a little loss of maximum loudness. Put 22 ohms in series with the speaker. At this power level, a 1/4W part may be fine. Gain and output will be reduced, crank it up, what does it do for the sparkle?

There's ways to kill the amp's damping without the power loss but it gets tricky.....

[PRR]

.....not so tricky; bless the '386's odd architecture.

You do have to "float" the speaker jack shell off the system ground/common.

Use the "high-gain" connection (big cap 1-8). Add 1 ohm in bottom leg of speaker. Tie this point to the "-in" pin 2 of LM386. That may be too much, use 10K or 50K pot to trim to taste. The trim will affect gain as well as damping. Low apparent gain is least damping, allowing treble-rise to happen.

[soggybag]

While we're on the subject. Notice that the gain can be increased in several ways with the 386.

The app notes suggest a cap accross pins 1 and 8.

I've also noticed many designs just connect pins 1 and with out a cap.

I've also seen some designs that connect pins 1 and 8 to ground. Sometimes using a capacitor resistor network.

[PRR]

> many designs just connect pins 1 and {8} with out a cap.

That gives gain down to DC. It multiplies the input error by 200. I don't think this error is specified. It is probably less than 10mV, but let's pretend it could be. 10mV times gain of 200 is a 2V error at output. If we had a 6V supply, wanted about 3V quiescent DC level, a 2V shift would jam the output to the rail or to ground. Even at 9V supply it would reduce symmetric swing and output.

With the cap, DC gain is roughly the AC gain, 26dB or about 20. A 10mV input error is a 0.2V output shift, not a big deal even at lower voltage.

I bet 99.99% of recent production '386 have input error much smaller than 10mV. And that few of us really mind a little asymmetric clipping; there's boxes made to do that to color-up too-clean amps. So nothing wrong with it, except if you don't get full output, you can't complain to the maker. They did suggest a cap, even if only to cover their butt on the occasional half-baked chip.

> 0.0047uf or higher from pin 1 to Gnd

I'm foggy on that. It looks quasi-unstable. Of course a lot of good guitar effects are about unstable systems.

If I do get what it does, I do think "or higher" is needed to make a "sparkle" difference.

[mac]

> 0.0047uf or higher from pin 1 to Gnd
I'm foggy on that. It looks quasi-unstable. Of course a lot of good guitar effects are about unstable systems.
If I do get what it does, I do think "or higher" is needed to make a "sparkle" difference.

I did that for my portable mp3 player amplifier to add brightness. Somewhere in the members area iirc.

mac

[soggybag]

Let me see if I got this straight. Gain on the LM386 is set by the resistance between pins 1 and 8. Less resistance more gain.

A cap across these two pins would set the AC resistance. Seems you could create a treble boost with a pot in series with a smallish cap. Or add some tone shaping with a parallel cap and resistor.

[liquids]

I've not messed with a ruby....but a few thoughts...

You might be thinking that your typical guitar amp has some kind of 'flat' EQ.  Not so.   Most have a lot of EQ coloration, and their (typically passive) tone stacks have a lot of mid scoop to compensate for this and other factors....as passive guitar pickups, to the ear, are rather mid-rangey...

if the 386 is a dead flat power amp, it's not going to be a direct replacement for a standard amp, even if you send it a buffered signal like the jfet.

If you have the time, I'd work on integrating a tone stack of some kind to cut some mids and shape the frequency response...that will make it 'seem' more sparkly, for one. There's more than one way do this, and some are better than others, but it is a good starting point for expirimentation.

[Wild E]

.....not so tricky; bless the '386's odd architecture.

You do have to "float" the speaker jack shell off the system ground/common.

Use the "high-gain" connection (big cap 1-8). Add 1 ohm in bottom leg of speaker. Tie this point to the "-in" pin 2 of LM386. That may be too much, use 10K or 50K pot to trim to taste. The trim will affect gain as well as damping. Low apparent gain is least damping, allowing treble-rise to happen.

The Ruby feeds the signal to the inverting input. Would this make a difference to your suggestion?

[Badside]

Little update!

Just had some fun with the breadboard. The serie resistance is surprinsingly effective while remaining subtle.
I used a 100R part cause that's what I had on hand, but seeing my cab is 16ohm I figured a higher value would be a good thing.
Such a simple and clever workaround!

Tried the small cap across 1 and 8, .1uF made it into a harsh beat, .01uF was almost imperceptible. So I may put the .1uF with a pot to get some sort of presence control.

Didn't get the chance to try the current feedback method as I don't have any 1R resistor on hand, but I feel the serie resistance is good enough that the rest can be handled by some form of tone stack in the preamp. Right now a Thor is feeding it and it's a fun little practice amp!
Might just use a Big Muff type tone control.

As I mentioned, my other living room amp is a small 8W 6K6 amp I've built myself that has no tone control short of a "Cut" control that actually cuts highs.
There is only the typical Marshallesque high-mid emphasis in the preamp that I'm including the Thor (though the "cathode" bypass caps don't seem to work the same way as with Triodes)

Thanks

[arjespen]

I've had similar trouble with a ruby build. It seems too dark to me.
I tried the various suggestions listed here, but have a hard time finding a solution that sounds good.
The best results seemed to be the resistor + cap between pin 1 and 8, but I still couldn't get it to sound like I wanted.
Am I correct in thinking that varying the resistor value increases/decreases the boost, and the cap decides the frequency to boost?
The closest I got to something I like was with a 820R resistor and a 0.047µF cap.

Another problem I have, is that the amp gets quite farty when I put a pedal in front. Tried a timmy and a TS. Not good sound :-(

Is the ruby in general not too happy about pedals, or?

[liquids]

1) if you want it to sound like a guitar amp - you need a mid scoop.

2) These chips expect a low output impedance (buffered) signal.  Most pedals end with a volume control that makes them not compatible here.

Build the circuit with a dual op amp in front of it - op amp buffer --> passive marshall tone stack ---> op amp buffer-->ruby.  You'll like it.

If you prefer the blackface fender sound, try  op amp buffer --> 33k series resistor---> passive fender tone stack ---> op amp buffer--> ruby.  It will sparkle.

[Mark Hammer]

If a person wants a speaker cone that can produce audible treble content above, let's say 8khz, it is going to have to be made of remarkable materials and have a Nobel-Prize-for-physic-winning design if you expect the puny signal from a half-watt (max) power amp to do the work.  Remember that the upper mids and treble constitute significantly less than the half watt of output one might expect of a power chord on the wound strings.

Think back to your high school physics and consider just how much MORE energy would be required to make a cone accelerate fast enough from a dead stop to move enough air that glassiness would be audible.

Can you goose the content in the 1-2khz range and add more bite?  Sure. No problem. Expecting audible treble content above that from a 500mw, or even 1w amplifier feeding a speaker that is either a) small enough to not be underpowered but lacking much more than a 5khz bandwidth, b) possessing the bandwidth but with a big enough cone that it's going to take a lot more than 500mw to make it jump to, optimized for high frequencies (small and light) but tied to a passive frequency dividing network that sucks out a big chunk of the signal.....is expecting a little too much.

Having said that, clearly tiny low-power audio amp chips CAN produce sufficient power and bandwidth when pushing what they were designed for - earphones.  BUt then pretty much everyone looks incredibly strong when lifting up balsa-wood barbells, right?