Another Mini Amp Thread (Expanding on the Ruby)

[Scruffie]

Okay, so I just built my first Ruby amp... all very nice and sounds good but I was thinking I might want a bit more control and clean head room for a portable amp so i've expanded it with the following things in mind and just wondered if there's anythings I should add or remove etc. I'm hoping to get it all into a 1590BB with a tiny cab fixed to the underside with a 5" Speaker in it but i'll want an output for a cab incase I want to plug into something larger without removing the internal speaker.

Specs:
MAX1044 - 9v to 18v

Dual Op-Amp -
1/2 Op-Amp Input Buffer
1/2 Switchable DOD 250 for boost

Swithable Baxandall Tone Stack

W/
Headphone Out
Switched Cab Out

And I've drawn up a preliminary schematic... I think it's right.


Any Comments or Thoughts on it?

[Derringer]

I'd put the buffer before the tonestack

I'd put a switch on the diodes too ... or a variable resistor between the diodes and ground

[Scruffie]

I'd put the buffer before the tonestack

I'd put a switch on the diodes too ... or a variable resistor between the diodes and ground

Well the J-Fet is acting like a pre-amp for gain recovery before the tone stack and then there's the Op-Amp buffer after it, any reason you'd stick it at the front?

As for the switch, i'm trying to keep the amount of knobs and switches down at the moment there's already 5 knobs (I missed the volume off from the output of the DOD 250/ Input of the 386) and 2 switches (3 if you include a power siwtch) which will cover the side of a 1590BB.

(Although I think I might loose the ruby gain knob and make it just clean to get rid of one knob and just use the DOD 250 for if I want more gain)

[Derringer]

everything I've ever read about tonestacks in amp and pedal designs points towards having a low impedance source to drive a tonestack more efficiently

The jfet will boost as you've said, but you have a high impedance coming out of it. Most (not all ... but most) designs i've seen have a buffer driving the tonestack and then being fed into another gain stage or splitter stage to then be amplified.

regarding the diodes, me personally I'd rather not have the diodes there since you're talking about more clean headroom ... why clip the signal with a pair of diodes? But, it's your ear, not mine. I built a superfuzz where I have a 10K trim between the diodes and ground and I really like how it lets me dial in the amount of clip. You could just use a trim like that too instead of a pot ... set and forget.

here's the schem http://www.diystompboxes.com/smfforum/index.php?topic=77512.0

[Scruffie]

Ah okay... well if I build it in sections on the breadboard it wont be hard to swap where the input comes from so i'll try it with the buffer first and with the JFet first and see what's prefferable, if there's no difference i'll probably go with your way.

As for the diodes that's why the DOD 250 is switchable, without the DOD 250 the diodes wont be there and the amp will be clean, but when I want distortion then i'll probably want some clipping, but maybe not i'll try it with & without, a trimmer could be an option though, I realise i've written 1N4148s aswell but i'll probably use 3mm Red LEDs as I prefer them for a crunchier sound)

[davidallancole]

Duncans tone stack calculator shows the Baxandall tone stack with a source impedance of 38k.  I think it will be fine without the buffer.

[Derringer]

completely missed you had a switchable bypass around the dod 250 part ..

[Scruffie]

completely missed you had a switchable bypass around the dod 250 part ..

Ahh  yeah, it's there as a second channel almost and to use half an op amp up (could go single but why not go double)

Duncans tone stack calculator shows the Baxandall tone stack with a source impedance of 38k.  I think it will be fine without the buffer.

Ah alright then, well i'll keep the Op-Amp buffer there anyway so basically it's like running a Fetzer valve pedal with an added tone stack, into the amp.

[PRR]

Will any '386 take 18V?

Will a MAX1044 deliver the ~~300mA that an 18V amp driving 8 ohms will suck?

Even 9V 8 ohms is a heavy load for most 9V batteries.

I have other reservations but if you work on bredbord before you stuff a too-small box most of them will be resolved.

[Scruffie]

No has to be the -N4 '386 to take 18V the -3,2 or 1 only go up to 12V and also produce less wattage the lower the number.

Didn't think it would take 300mA to drive a 6" speaker (i'm actually using a 30 ohm speaker) so it might have to be a LT1054 Charge Pump to get enough (not even sure that supplies 300mA... that may have stumped the whole thing)

Perhaps i'll use Mark Hammers old favourite of an octet of AA Batteries to deliver enough mA (12v will have to do I suppose and then I can just use any old '386, it'll mainly be plugged in anyway, I just want the option of battery power) I realise a 9V battery in a Ruby dies after about 2-3 hours anyway.

It's a small box.. but with a small cab attached too, some of the extra jacks for cab & headphone out, maybe even the battery can be spread into it so there should be room (people have gotten tube amps into smaller boxes... i'm sure it'll go).

[PRR]

MAX1044 is rated 10mA. No-go.

MAX660 and LT1054 are rated 100mA.

> Didn't think it would take 300mA to drive a 6" speaker

Speaker diameter does not matter.

As a rough approximation: a single totem-pole audio power stage making full Sine power loads its power supply like a resistor 6 times the speaker impedance. So 8 ohm speaker, the power supply "feels" a 50 ohm load. 18V/50r is 0.36 Amps.

(The exact answer has 1.414.. in it, several times, but also the amplifier losses. So not exactly "6", but close enough to select yes/no/maybe.)

BS check: Read LM386 datasheet page 4, dissipation and output for various voltages. Take 9V into 8 ohms: it can put 0.75 Watts into the load and 0.5 Watts into itself, total power demand is 1.25 Watts. At 9V that would be 1.25W/9V= 0.139 Amps. And 9V at 0.129A is 65 ohms. A little higher than "6*load" because LM386 has a lot of voltage loss (0.75W out is 7V peak to peak, 2V wasted.)
   
> I realise a 9V battery in a Ruby dies after about 2-3 hours

Further confirmation that pulling a speaker is hard work.
   
> i'm actually using a 30 ohm speaker

Ah, then 6*30 is 180 ohms, and 18V across 180 is 0.1 Amps, 100mA. So the beefier boost-chips may work.

Just for curiosity: the '386 datasheet graphs do not show 18V in 30 ohms, but has 16V (close) in 16 ohms (about half your plan). The power in 30 will be about half what they show in 16 ohms, just under One Watt. Oh! Page 2 specs 16V supply in 32 ohm load: 1,000mW typical. In a good speaker, that will make some sound.

OTOH, an 8 ohm speaker on 9V through a '386 will also make nearly a Watt, withOUT the complication and losses of conversion. And the selection of 8 ohm speakers is much larger. (And you did mention external speaker, which are rarely 30 ohm.)

You could still use a '1044 for the small stuff, but you have to think: the clipper only needs >0.6V signal to do its thing, the '386 input needs even less, your guitar signal is less.... why boost-up to 18V allowing 6V signals when you only need less than a volt of signal?

BTW: The classic 9V battery is designed for 0.1W pocket radios. A lowest-price (I just got two for 97 cents) 9V will sag badly trying to pull a near-Watt load. But improved chemistries (and higher cost) 9V batts may serve OK. I guess it depends how much fun you can have in "2-3 hours".

> octet of AA Batteries

When you work it out, and find a good price, they are shockingly cheaper per watt-hour than 9V batts.

But at under 16 ohms and over 10V, the LM386's ~~0.4A peak current limits output power. A better choice becomes the LM383.... ah, out of production. There were several single-channel "7 Watt" 5-leg chip amps which did very well on 12V.

[Scruffie]

Aha, I didn't know the math or follow the datasheet that much but that makes alot of sense now you've written it out for me, thanks for clearing that up (my knowledge of amplifier design is pretty low, I assumed an LM386 was pretty tollerant of design)

This all being the case... I think i'll drop the DOD 250 & Opamp buffer, stick a J-Fet buffer following the tone stack and just work out a better gain control for the 386 and power it all from 12V (8 AAs) it'll fit in a box alot easier that way too... or build The LM380 amp mark hammer posted.

That or look at building one of Rick 'Cyborg from outerspace'  Frequency Centrals Sub-Mini Tube amps... perhaps the better idea.

[Brymus]

I have ran a 386-LM4 at 18V into 16ohms with good results.
In fact my friends were blown away at how loud and good it sounded,I used alot of AAs to power it with good results.
When upping the gain across pins 1-8 it does get hot,I smoked a couple in my testing.
Using it at a gain of 20 it works pretty well,also adding one of those glue on heat sinks helps.
But for real abuse get the SIP package and attach an oversized heat sink and abuse away,you will be amazed at how much the puny 386 can handle.
I ran two in BTL mode at 18V with heat sinks and could play with a drummer when using a 99Db SPL cab.
Everyone thought it was my EVJ I was using until I would show them the cab under it was in fact wired to my breadboard.

Scruffie I have a huge folder full of 386 designs (I was obsessed with that chip for months) if you send me an E-mail I will
send it to you.
I cant really host it becuase only 3-4 designs in it are my own,the rest from all over the net (ROG,here,there,you get the idea)
I made a really nice IMO amp using the 386 for a pre,found here >http://www.diystompboxes.com/smfforum/index.php?topic=78598.msg648144#msg648144
Which could be used with out the TDA chip amp after it,just swap in the LM4s in BTL mode

[Scruffie]

Cheers Brymus, PM'd