Building the Tiny Giant amp

[Clarke]

Hey Taylor,

I put together my TG yesterday. Hooked it up this morning and the LM338 (literally) exploded immediately.

Exact thing just happened to mine, as soon as  i plugged it in. going to try re flowing the joints.

I didn't realize that I had the polarity reversed on the power jack. I wired the power supply plug to be center positive and didn't make the connection that it's the opposite of how pedal PS are wired.

Luckily I didn't blow the replacement chip--it just got fairly hot while I tested voltage and made the realization.

At least it's working now! Lesson learned...

I Guess I should try switching mine too?

[garyh]

How long does it take to get a pcb from you Taylor?

[corneliusduff]

So I just finished building mine. Sounds great.  I love the overdrive.  Though I'm concerned because the TL072's pin 8 still isn't reading 11 volts.  Need I worry? 

[Taylor]

What voltage are you getting there?

[corneliusduff]

What voltage are you getting there?

Most of the time, 0.  After playing through it for the first time, it started to read something but it shot down afterwards.   I wanna say it was 9v but it was probably 1 or -1.  It was hard to tell because it was so sporadic.   Though I'm a bit of a noob at this so I could be reading it wrong but all the other options I've tried didnt yield anything either.  Multimeter set at DCV, 20V would pick up on it right? And you say check this before powering up so does that mean i should have done this before I start playing through it or before I plug in the power, as I don't know where 11v would come from without some power.  The amp sounds fine, I just don't know if I'm danger of blowing it up, where to start looking for the the problem or what it even really means. 

[Taylor]

Right, you definitely need to have the power on to read voltage. Sorry if this is a rude question, but you know that to read voltage you have to touch the black lead to ground simultaneously right? You probably do, just checking.

[corneliusduff]

Right, you definitely need to have the power on to read voltage. Sorry if this is a rude question, but you know that to read voltage you have to touch the black lead to ground simultaneously right? You probably do, just checking.

Haha, not a rude question at all because I'm all new to this.  At first I thought the enclosure was synonymous with ground due to the plastic jack and the lm338t, so it took me a minute but then I figured you meant the ground from the power supply and it read fine! Thanks! 

[bcalder]

Hi all,

I'm close to getting this guy in its enclosure. I had it out to test it, and it's motorboating like crazy, with the guitar volume up or down.

A couple of possibilities that crossed my mind were:
-     the power supply is too close to the amp - it won't be in the enclosure
-     the volume, standby and input wires are too long & are picking up RF

Any additional thoughts on what it might be? Thanks in advance.



Uploaded with ImageShack.us

[bcalder]

Seems like it was the lack of good grounding. After a good night's sleep, I arranged things better, and it was fine. What fun it is!

And boy, that little sucker gets WARM for sure.

[Heemis]

Here's my build!  I used the carcass of a dead yamaha VX15 amp, gutted it and re-used as much of the original as possible.  Mods include a gain control (goes from about 2x - 7x), a speaker out which kills the internal speaker, and a 3 band tone stack as suggested by PRR in a previous post in this thread.  I love it, perfect clean volume for bedroom playing and prototyping... can't wait to plug it into my 2x15 cab at the space though!

One quick worry though... the first time I powered it up, I hadn't noticed a solder jump between 2 resistors... the chips immediately got very hot, and I was measuring voltage on pin 8 of the TL072 fluctuating between 8-10 volts.  I didn't leave it plugged in that was for long, and I got it fixed up pretty quick... just wondering if permanent damage might have been done to the regulator or amp chips?

Thanks for a great project Taylor!

Here are some pictures:

[Taylor]

Tough to say, not knowing which resistors were bridged, but if everything's working right now, you're probably fine. Nice build!

[therecordingart]

I have a 15v 3A power supply. You think it'll do?

[papasteack]

Hi,

Sorry for my english !
To answer Therecordinart, max volt drop of the lm338 is less than 3V. So 15-3=12v. 12v>11,47v. So 15V seems very good if you keep the adjustement resistor of the regulator as given by taylor schematic. The regulator won't get hot ! The current of 3A seems enough, because it as been said on this forum that the functionning current is around 1/1,5A. Maybe a bit low for peaks, but smps recharges really fast their output capacitors.


I've buy a TGA from Taylor. I've got a IBM 16v 4,5A laptop power supply. I got it to make a evolutive amplifier, from simple amp, to multi effect bi-amp cab by adding modules. At the end, the TGA would be use to amplify low frequency after a low pass filter and a sub cut, (and other effects) a with a 2x10" isobaric cab of fane 10-125 (seems a cheap and good solution for lows). But at the beginning of the project, it will be use only with a eminence governor 16ohm.
As it's a big project, I take me time. I like the concepting aspect. So, I don't have yet soldered stuffs, as i would like to try to understand the circuit, and maybe try to make some improvement.
So, since near one month, i interest myself in electronic. I've make a lot of search on forums to make my own opinion about audio power supply. Here is what i've understood with my poor electronic basis :

_First, it seems using a laptop supply is a very good idea to not use transformer. As it operate at very high frequency, output capacitors of the laptop power supply ciruit is recharged at very high frequency, so we don't need to have big reservoir capacitor on power part of the circuit (all before the regulator). The problem should be that it probably emit EMI so that we have to

_Regulators provide a easy way to get a very low impedance output, and this is a very good things for audio amplifiers. But it seems that it need time to detect voltage drop to correct it, wheareas capacitor reservoirs don't needs as they're precharged (but have higher impedance).

So i've got some question. It seems that the regulation could be improved by different ways, and i ask your help about their real utility:
_ adding a capacitor on adjustement pin, as the lm338 documentation suggest seems to improve transient response, noise (tnt-audio site explain it http://www.tnt-audio.com/clinica/regulators_noise2_e.html: Noise is reduced tenfold with about 20dB on lm317 or lm337 regulator) and lower the impedance. But it would need as the datasheet explain, a protector diode.

_using the regulator only as a output capacitor charger (for the 470uF capacitor on the circuit) : so capacitors after it would'nt need to support high voltage, and would cost less. Studying capacitors datasheets, it seems that lower voltage capacitors have lower esr...the idea is to putt for example a 0,1R in serie after the regulator, so that the tda current went first from capacitor, wich react faster than the regulator, as it is precharged, and don't need signal error to react to more current demand from the tda. The resistor would protect the regulator against capacitor inductance (i don't really understand this...). But in this way of supplying the tda, to get low impedance at low frequency, i estimate that it is needed to use 4 x 3300 uF panasonic FM capacitors wich have very good esr (0,013@100khz so maybe around 0,05@100hz=>0,0125 for four, near the same impedance of the lm338 without capacitor on adjustement pin) to get good current feeding at low frequency. Moreover, it make a RC filter after the regulator  low pass filter below  (i don't remember at wich exact frequency i had calculated).
But the problem with this solution is for peak demand of the tda : the use of a 0,1R means  proportional volt drop to voltage demand of the TDA. So maybe we would loose a bit dynamic ? Another problem is that the use of big capacitor may require a protection diode for the regulator...and i have understood that diodes increase noise.
The use of the resistance is to prevent the regulator to became non-stable with big capacitors inductance...

I know there is also "snubber" tricks on the web with the lm338 for audio, but it put another resistor on the output reservoir capacitor and reduce so the impedance of capacitor; it don't seems at all a good idea.

To resume, my idea is to use the regulator as a stabilisator so that power came from a big capacitive reservoir with very low impedance as we can use low voltage capacitors, wich are moreover, cheaper. Tweakings would add :
_a 10uf capacitor on adjustement pin (for noise) and a 1n4002 diode protection.
_a 0,1R 10 watt +4*3300uF 16V very low esr capacitors (chosen for the lower esr found in capacitors datasheets and limiting their cost to 8 euros for four).

Total tweaking cost would be less 10 euros, and seems interesting. Maybe other cap could be better, or cheaper, i didn't read all datasheets.

This is all hypothesis based on what i've searched on web, i know it's a esoteric approch of electronic,  but interesting.
Does it seems you coherent ? Do you see better / cheaper improvement ? (as providing symetric and powerful supply to the aop maybe...i know it would improve, but i've not already worked on the question)

Thanks,

Damien

[PRR]

> Noise is reduced tenfold

But can we hear the noise now? The simple regulator does have good ripple reduction. The car-amp power chip has very good power supply rejection (a car's electrical system can be VERY noisy).

In most DIY builds, ground layout and wires too close to each other will leak more power supply noise than the chips.

> very good esr ... 0,0125

The elephant in the room is the 3 or 6 ohms of speaker resistance. We want to make other resistive losses "small" compared to 3 ohms. Smaller than 1 ohm. But I would not go to a lot of trouble to get 100 times smaller.

[papasteack]

 
I didn't have any other reference than what i read, so i have mixed lot of different things  . Thanks for your knowledge !

Ok, so I don't add capacitor on adjustement pin.
I'll add the diode protection and two 6800uf standard caps i have in stock  for lows.
Is it useful to add a resistor on the output of the regulator for a story of inductance relation?

Damien

[PRR]

I think it is a very fine design just the way it is.

[papasteack]

Thanks again for your experience.  I've seen in the forum you make really good advices.
I'll just do as the original circuit. As i said, i (re)discover electronic since few month.
There's so many thread about audio circuit improvements, it seems there is always a lot of things to tweak to make better.

[Taylor]

I think it is a very fine design just the way it is.

Thanks Paul. 

Papasteack, I'm glad that the project has given you a number of ideas. Let us know how your build goes.

[papasteack]

As it seems to interest...let's saying more about it...
With 11 different tweaked modules in a Bi-amp home combo, it will be a loooooooooooooooong funny step by step project :
_choosing modules/designing...yet difficult !! I'm still at this step since 2 month ^^
_making the box + TGA to drive a already made 12" bass reflex eminence governor 16 cab
_adding Bloviator (madbean bbe soundstomp clone) + Cupcake (madbean orange squeezer comp)
_adding rebote delay 2.5 (tonepad)
_adding 2x split_blend (runoffgroove) + sabertooth (madbean fuzz)
_adding 10 band tl074 eq (with virtual ground for symetric power cicuit used for the subsaver too)
_making the combo enclosure :  2*10" isobaric fane 10-125 + 12" open back governor 16 ohm (<80l for all) => to be usable either for bass or guitar without compromise
_adding lm386 amp with hexfet preamp (french "mini-mos amp")
_adding 2 way active cross over (sound.westhost) + subsaver (subwoofer-builder)
_adding a distortion (diefet)


Here is the first try of design...I don't draw very good, so i did as i can. The idea is to mix CLEAN/DIRT sound. The DIRT beeing itseft a mix between fuzz/disto sound, with or without a high pass for bass use. The big EQ position is selectable : before mixing channels, off, or only on clean channel. Rebote delay, cupcake and bloviator are global effect so they came after in the chain. Output effect are the subcut, and crossover. Tiny giant is for used for low when crossover is selected. The high are amplified with "mini-mos", wich has a hexfet preamp (so only "high" benefit of preamp when crossover is activated, bass are intented to be more conserved). Rebote delay would include infinit mod. EQ, fuzz and disto potentiometers are 9mm models. Position of miniswitch will probably have to be adjusted as i didn't found their datasheets. Crossover frequencies will have to be adjusted. With 10 potentiometer, this is the EQ circuit wich will cost the more (sorry for my english, i write as i can ^^ )
The inside of the box, where cable would pass have been designed yet too making attention to parasite effects between them.

The only missing thing would be a foot pedal sampler/looper...but i think i would buy it if i want one...

[arma61]

I don't draw very good, so i did as i can.

You sure, looks like you "can" very good, I'd say 



well done, I'm working on "something like that", more eeeeeasy.... just an overdrive and chorus, and a cabsim.... couple of months and I'm done