Ruby Amp help, please

[glesconz]

Hi all! Im going to make a Ruby Amp and have knocked a veroboard layout in DIYLC. I read that you should keep the components as close to the chip as possible and hope I have acheived that. Also I read somewhere about keeping the grounds separate? Im not sure what is meant  by that. Is my grounding in this layout good or could it be better. Thanks for your input,
Cheers,
Glenn

[duck_arse]

if it were me, and I was worrying about grounds - I'd shift the pink [A2] and black [A4] wires to the top row, and link pin 4 from A3. this will bring the C2 bypass cap "closer" to the IC supply connections. the -ve spkr wire to A5, maybe?

[anotherjim]

keeping the grounds separate

Everything is so close already and the currents so low that it's hardly necessary.

But for reference...
Pick one spot for star ground point (a common distribution method). The input jack ground lug  is best. A metal case might be grounded via a metal jack socket for this job but it isn't best practise. The speaker jack if it's having one should be plastic body or insulated from the case.
It's usually best to break the circuit grounds into functional groups that connect on the board as closely as possible and then each given a unique connection to the common Star point.

One group is the preamp, that's the 2 resistors off the FET to one track and then one wire to the Star.

The next circuit is all of a lump since it's all done in the 386 - that's 4 connections on board and one wire to the jack Star. Since the supply cap needs a close connection to the chip, I'd add that making 5 connections in that group.
Speaker current has to return from whence it came, so the speaker jack negative goes close to 386 chip 0v power pin - making 6 board connections for ground.

The supply jack negative gets one wire to the Star. If there's going to be a battery snap, you want a TRS input jack and put the battery negative on the ring contact of the jack so it switches off when you unplug to save the battery.

If we want to "gold-plate things", a metal case should have a definite connection to the star using a solder/crimp tag to a bolt/star washer. If we go to that trouble, we can make that the star ground with several tags on the same bolt. The input jack ground then goes with pre-amp grounds on the board, something I like to do using screened cable to make a complete connection to the input jack from the board. That keeps the high impedance input nicely protected from internal interference and especially helps if we want to test the assembly outside of the box.

[pinkjimiphoton]

386 circuits are weird. the ruby is weird. i always end up with problems with these for some reason.
they are bone- simple circuits but can be kinda challenging.

you need to do all the decoupling mainly cuz of the amount of juice the fet is sucking down. a crappy jfet can "not work" in this circuit when everything else is spot on. uggh.

that said, jim says it all

[Paul Marossy]

386 circuits are weird. the ruby is weird. i always end up with problems with these for some reason.
they are bone- simple circuits but can be kinda challenging.

you need to do all the decoupling mainly cuz of the amount of juice the fet is sucking down. a crappy jfet can "not work" in this circuit when everything else is spot on. uggh.

that said, jim says it all

Few things based on my recent experience with my Little Gem MKII build from hell recently.

1 -- Use min. 10uF cap to ground on Pin 7 of LM386. Takes care of most possible noise issues due to power supply, etc.

2-- The newer LM386s seem to want an electrolytic cap right at the IC between Pins 4 & 6, I used 47uF. Without these all I got was noise, very loud low frequency oscillation that sounded like it could blow a speaker and loud farting noises, all while passing no signal at all.

The Little Gem MKII is a bridged design using a pair of LM386s, so hopefully you won't have these problems in the Little Gem/Ruby which only uses one LM386. My LM386s were National Semi's, I've heard JRC386s work better. I say build the Ruby per the schematic and if you have any problems like what I've described, then try those things listed above. You might or might not have same issues if using National Semi chips.

[pinkjimiphoton]

god. i just built one of my ruby doobie variants. never had an issue. couple parts, plug it all together and wail..

i gotta endorse what paul found with the 386's out there now. @#$%in garbage!!

all i can get outta my amp is oscillation and thats about it. i've built a shitload of these, the only diff is the 386 is a brand new one from tayda.

i'm gonna keep throwing caps at it til it plays nice, but, damn, dude...

[Perfboard Patcher]

Hi folks,

Just want to share my experience about dealing with "bad" LM386s.

I've cleaned up some of the notes concerning the LM386 I took while is was working on my guitar sustainers.

"The power chip I use is an LM386N without a number and without a brand name. I measured the power across an 8ohms load, it's a 1 Watt version of the LM386.  I have a dozen of them, they are all prone to oscillation. And I mean even oscillation with short wires from a 12 volts (DC, stabilized) net adapter or 9 volts battery to the IC, cap close to pins 6 and 4, both inputs earthed, pins 1 and 8 bypassed or not, using or not using the 47nF/10ohms snubber/Zobel circuit, and with or without load. There is a considerable signal think 400mVpeak 20Mhz(!) present at pin 5. The only thing that stopped the oscillation was to change the values of the snubber resistor: 2.2ohms instead of 10 ohms, I also changed the cap value to 470nF. That's not what the data sheet ordered!

Even though one cannot hear the 20 MHz signal one can hear that the sustainer is not running smoothly. Like a noise gate that is switching on and off. When examining this problem by oscilloscope it manifests itself as high frequency "shark fins" on top of the signal at certain signal levels."

Afterwards I found out that the best remedy against oscillation was to insert a small ferrite bead in the power supply line right before the cap. The ones I used are named in Dutch after their appearance, pig noses. There were only 3 or 4 turns of windings needed, I used the same enamelled copper wire as for my sustainer drivers, 0.10mm.

cheers,
Paul

[pinkjimiphoton]

i finally got one done last nite. its my own "ruby doobie" variant, basically instead of a jfet buffer front end, it uses a one stage fuzz <what's now the "photon phuzz"  on guitarpcb>.

i bought the enclosure as a kit on ali express a couple years ago and modded it slightly to suit my needs. it comes with the case, speaker, and a board for the vu meter and a chip amp, but i just chucked it in my junk drawer and used one of my own. it freekin screams. and is unruly. i diggit




















schematic if ya wanna mess with it



also... you may wanna try going with a far larger output cap. i used 1000uf @16v for mine, works better.. more bass = more volume. also will drive a larger cab better.