Building the Tiny Giant amp

[Lieuwe]

It's too bad your camera doesn't have macro focus - very difficult to see anything clearly in the photos.

Dear Walt, thank you for your reply. Here are some better pictures I just took. The back of the board looks a bit dirty from the resin in my solder. BTW, I use a 15v 4.5a laptop power supply.

It looks like your volume pot might be mis-wired.  On the PCB, the pads for the pot numbered 1-2-3 are connected to the yellow-red-orange wires. It looks like the red and orange wires are connected to the pot on terminals 1 and 2 (can't tell which is which from the photo).  The yellow wire should be on pin 1 (top terminal in the photo).

I know, but it's just wired in reverse, so that can't be the cause of this problem.

It's hard tell how things are connected from the photos.  As a general debugging technique, you should probably use your meter to check continuity for all the off-board wiring.  Is there continuity everywhere there is supposed to be?  Is there isolation everywhere else?  For example, with your standby switch off, make sure there is no continuity between the standby pads on the PCB.  With your speaker switch set to "internal" make sure the speaker pad on the PCB have continuity with the terminals on the speaker.  Make sure the speaker actually has some resistance between the terminal (about 3.2 ohms of DC resistance).

I did that before posting the problem here. Speaker measures 3.9 ohm. everything offboard seems to be isolated or continuous as it should be. Could it be that the problem is on the board? Is it my soldering, or can one of the components be broken?

update: I just did some more probing. The heatsinks of the LM338T and the TL072 are well isolated from eachother. However, there does seem to be continuity through the pcb. They read about 600 ohm apart. The heatsink of the TL072 alone reads continuity to ground, and the heatsink of the LM338T reads about 600 ohm to ground. Is this normal?

[Jdansti]

The heat sink = Vout. The resistance between the HS and ground should be the same as Vout to ground. You should have close to 0 ohms between the HS and Vout.



I'd use some isopropyl (rubbing) alcohol and a tooth brush or nylon bristle brush to clean the back of the PCB. You really can't see solder bridges with all of that flux on there.

[Lieuwe]

The heat sink = Vout. The resistance between the HS and ground should be the same as Vout to ground. You should have close to 0 ohms between the HS and Vout.

Resistance from heat sink to Vout = 0 ohms. From sink to Vin about 1k and from sink to ADJ about 100 ohms. That's normal, right?

I'd use some isopropyl (rubbing) alcohol and a tooth brush or nylon bristle brush to clean the back of the PCB. You really can't see solder bridges with all of that flux on there.

I cleaned the back, and can't find any bridges. Even resoldered some joints that might have been cold, but no change. I'm starting to doubt that I put on all the components correctly, going to check that now. What else can I do? Can I measure voltages at certain points or something?

Has anyone had this problem before (ultra low volume, but a working amp)?

[waltk]

It looks like your volume pot might be mis-wired.  On the PCB, the pads for the pot numbered 1-2-3 are connected to the yellow-red-orange wires. It looks like the red and orange wires are connected to the pot on terminals 1 and 2 (can't tell which is which from the photo).  The yellow wire should be on pin 1 (top terminal in the photo).

I know, but it's just wired in reverse, so that can't be the cause of this problem.

Correct, the volume pot will just work backwards until you rewire it.

update: I just did some more probing. The heatsinks of the LM338T and the TL072 are well isolated from each other. However, there does seem to be continuity through the pcb. They read about 600 ohm apart. The heatsink of the TL072 alone reads continuity to ground, and the heatsink of the LM338T reads about 600 ohm to ground. Is this normal?

Assuming your reference to "TL072" was just a typo, and you really meant "TDA7240"... The heatsink for the LM338 is internally connected to Vout, so there should be zero resistance between the heatsink and Vout.  The resistance between Vout (or heatsink) and ground should be the series resistance of the 1k and 120R resistors that set the voltage output - ~1,120 ohms.  Because the TDA720 heatsink is internally connected to ground, the resistance between them should also be ~1,120 ohms.  600 is too low, and means there is a short somewhere.

The gain of the TL072 is fixed by the ratio of the 220k and 100k resistors next to it.  The gain should be 3.2 (220k/100k + 1).  You have used a 1M resistor in place of the 220k resistor.  If your intent was to build a stock Tiny Giant, that's a mistake - but it doesn't explain why you have LESS volume.  Based on the resistors you actually have in the circuit, the gain of the input stage would be 11 (1M/100k + 1).  That is probably too much, and would cause clipping in the TDA7240 (which is not pretty).

At this point, you should probably try to figure out why there's only 600 ohms between the heatsink on the LM338 and ground.  You should probably do this without power applied as much as possible (to avoid risk of frying components).

You should have high resistance between pin 3 (input) on the TL072 and ground (like 3Mohm).  If you don't have a high resistance there, then your signal is being grounded somewhere before it gets to the opamp.

If you've corrected any faults that can be found without power applied, then try applying power again, and verify that pin 6 of the TDA7240 is getting power (about 11.6V).

Also, an audio probe might help determine where your signal is leaking off before it gets to the TDA7240. 

+1 on the idea to scrub the solder side of the PCB with alcohol and a brush.  It looks like there's not only rosin there, but a million flecks stray solder (maybe that's just an artifact of the photography though).


Good luck, let us know how it's going.

[Lieuwe]

Dear Walt, thanks again for your help.

It looks like your volume pot might be mis-wired.  On the PCB, the pads for the pot numbered 1-2-3 are connected to the yellow-red-orange wires. It looks like the red and orange wires are connected to the pot on terminals 1 and 2 (can't tell which is which from the photo).  The yellow wire should be on pin 1 (top terminal in the photo).

I know, but it's just wired in reverse, so that can't be the cause of this problem.

Correct, the volume pot will just work backwards until you rewire it.

Just rewired this.

update: I just did some more probing. The heatsinks of the LM338T and the TL072 are well isolated from each other. However, there does seem to be continuity through the pcb. They read about 600 ohm apart. The heatsink of the TL072 alone reads continuity to ground, and the heatsink of the LM338T reads about 600 ohm to ground. Is this normal?

Assuming your reference to "TL072" was just a typo, and you really meant "TDA7240"... The heatsink for the LM338 is internally connected to Vout, so there should be zero resistance between the heatsink and Vout.  The resistance between Vout (or heatsink) and ground should be the series resistance of the 1k and 120R resistors that set the voltage output - ~1,120 ohms.  Because the TDA720 heatsink is internally connected to ground, the resistance between them should also be ~1,120 ohms.  600 is too low, and means there is a short somewhere.

Yeah, sorry. I'm getting my parts numbers mixed up. To be clear: The resistance of the heatsink to ground in the LM338 = 1.12K. Heatsink to heatsink = 600 ohm.

The gain of the TL072 is fixed by the ratio of the 220k and 100k resistors next to it.  The gain should be 3.2 (220k/100k + 1).  You have used a 1M resistor in place of the 220k resistor.  If your intent was to build a stock Tiny Giant, that's a mistake - but it doesn't explain why you have LESS volume.  Based on the resistors you actually have in the circuit, the gain of the input stage would be 11 (1M/100k + 1).  That is probably too much, and would cause clipping in the TDA7240 (which is not pretty).

I see, don't know how that happened. Fixing it now.

At this point, you should probably try to figure out why there's only 600 ohms between the heatsink on the LM338 and ground.  You should probably do this without power applied as much as possible (to avoid risk of frying components).

You should have high resistance between pin 3 (input) on the TL072 and ground (like 3Mohm).  If you don't have a high resistance there, then your signal is being grounded somewhere before it gets to the opamp.

Resistance between pin 3 and ground is only 175K. Does that narrow down my search area?

If you've corrected any faults that can be found without power applied, then try applying power again, and verify that pin 6 of the TDA7240 is getting power (about 11.6V).

Also, an audio probe might help determine where your signal is leaking off before it gets to the TDA7240. 

+1 on the idea to scrub the solder side of the PCB with alcohol and a brush.  It looks like there's not only rosin there, but a million flecks stray solder (maybe that's just an artifact of the photography though).

That was just some glare from the camera, no real flecks. Just a lot of resin. I don't have an audio probe, but I should get one (I normally build guitars, lap steels and such, only did 3 BYOC builds before this, all went smoothly).

Good luck, let us know how it's going.

Thanks, will do.

[waltk]

The resistance of the heatsink to ground in the LM338 = 1.12K. Heatsink to heatsink = 600 ohm.

There's something really fishy about this.  The heatsink of the TDA7240 is internally connected to ground - so there should be zero resistance between the TDA7240 heatsink, pin 4 on the TDA7240 (ground), and any other ground point on the PCB.
In other words, heatsink on the LM338 to ground should give the same result as heatsink (lm338) to heatsink (TDA7240).

Maybe it's just my BSE (bovine spongiform encephalopathy) acting up, but I can't think of how you could get this result.

[Lieuwe]

The resistance of the heatsink to ground in the LM338 = 1.12K. Heatsink to heatsink = 600 ohm.

There's something really fishy about this.  The heatsink of the TDA7240 is internally connected to ground - so there should be zero resistance between the TDA7240 heatsink, pin 4 on the TDA7240 (ground), and any other ground point on the PCB.
In other words, heatsink on the LM338 to ground should give the same result as heatsink (lm338) to heatsink (TDA7240).

Maybe it's just my BSE (bovine spongiform encephalopathy) acting up, but I can't think of how you could get this result.

I've checked the schematic, and I agree, it makes no sense at all. I've checked all my joints, all looks fine. There's good isolation between the big heat sink and the parts.
Also, I just switched the 1meg resistor for a 220K, and now it does nothing at all. The PCB doesn't seem to like all my probing and resoldering, it's starting to crumble at places. I'm about to just give up and go back to building guitars.

[waltk]

 

I'm about to just give up and go back to building guitars.

Yeah, it's frustrating - but don't give up yet.  This little amp sounds great when it's done, and you're almost there.

Also, I just switched the 1meg resistor for a 220K, and now it does nothing at all.

Ok.  So switching that resistor has set the input gain where it should be.  It's not surprising that your output was reduced from a whisper to an inaudible whisper.  The good news it that even though your output was a whisper (before correcting the gain), there was some signal making it all the way through to your speaker.

The other good news is that all your other components appear to be in the right places. 

Is it possible that there's a short in the standby circuit? (in the switch or a solder bridge on the board)

[Lieuwe]

Well, I just discovered one silly mistake. I wired the LED so that it would switch on when the standby-leads would connect. So when the amp would switch to standby.
This still does not explain my problem: The amp produced whisper-like volumes while standby, and it does not work now when the leads are disconnected.

Yeah, it's frustrating - but don't give up yet.  This little amp sounds great when it's done, and you're almost there.

Not giving up yet, thanks again for the help.

Is it possible that there's a short in the standby circuit? (in the switch or a solder bridge on the board)

While disconnected, the standby leads have 5.2k resistance one way, and 4.7 the other way. Can this come from the 22uf cap?

[waltk]

While disconnected, the standby leads have 5.2k resistance one way, and 4.7 the other way. Can this come from the 22uf cap?

Nope, that doesn't sound right.  The resistance can't come from the 22uF cap (unless it's internally shorted).  What you're seeing is the internal resistance between pins 2 (standby) and 4 (ground) on the TDA7240.  The reason I say it doesn't sound right is that I measured a few chips, both in a working TG and just pins 2 and 4 on the naked chip.  All the chips I measured had between 55 and 60K ohms resistance.  I'm starting to think your TDA7240 chip is bad.  Try removing the 22uF cap connected to pin 2 (it's the one closest to the edge of the board), then measure the resistance between the standby leads again.

If the resistance is about 55-60K, then the 22uF cap is bad.
If the resistance is still about 5k, then the TDA7240 is likely bad.
The only other thing it could be is if there's a partial short to ground from the pin 2 trace on the board (and I don't really see any place that would be likely).

I suggested removing the 22uF cap because that's easier than removing the 7-pin TDA7240 - but if you determine that it IS bad, then you might as well take it out (and verify the suspiciously low resistance between pins 2 and 4).

Good luck - keep after it.

-Walt

[Lieuwe]

If the resistance is about 55-60K, then the 22uF cap is bad.

The cap was fine, so I removed the TDA7240. This was a real PITA, practically destroyed the PCB and the TDA7240 in the process (eyelets started coming loose, legs falling out). If the TDA wasn't broken before, it was now.

If the resistance is still about 5k, then the TDA7240 is likely bad.

luckily, the 2 and 4 pins were unharmed, so could be measured. And there it is: 6k resistance. So, problem found. However, there is a Dutch saying that roughly translates "operation a success, patient deceased". I hope that's not the case. I checked local electronics stores, and they don't sell the TDA7240. Even if I get a hold of one, will the PCB still work with the eyelets fallen out and all?

[waltk]

 

practically destroyed the PCB and the TDA7240 in the process (eyelets started coming loose, legs falling out). If the TDA wasn't broken before, it was now

Yes, desoldering a two-sided PCB with through-plated pads is a real PITA.  In this case, it might be possible to salvage.  The TDA7240 trace connections are all on the top (silk-screened) side of the board - except for pin 4 (ground) which is on the bottom.  You would have to be sure that all the traces on the top-side has a solid connection.  After all your trouble with it, though, it might be best to start over. 

The TDA7240 IS hard to find.  Maybe Taylor has some suggestions.  (Taylor, are you there? ...)

[Taylor]

Sorry guys, have not been able to check the forums as regularly lately. Lieuwe, if you contact me at the Music PCB site, I will set you up.

[papasteack]

Hi !
[! approximated english content !]
I just finished to solder the tga. It works near well. It seems the tda is prone to oscillation at near full volume. Without boxe, moving cables solve the problem. Happilly, stable in some position. Funny horrible sound when it oscillate, it is the first time i ear such a thing, and is very distinctive. It's the only bad point, it don't seems to be really stable. All is like the schematic, i don't have a oscilloscope...any idea to investigate more on this problem? Both tda & lm seems to get be really hot too. It runs at 11,60v, but didn't verified @ how much current. My speaker is a eminence governor 16 ohm.I'll try to see if current seems normal.
Thinking.... => Should it came from bad symetry of the two RC networks from tda to the speaker if values of resistors/capacitors are not the same...? Don't this network should be ajusted for speaker impedance, like zobels networks in other amps ?
As it happen near full volume and as where the volume controle is located , i say it's the tda...but maybe the lm...don't know how to resolve it problem/optimize my circuit to get it better stabilised.

As it didn't affect oscillation problem, I've added big tank caps before (3*2200) and after (2*4700) the regulator, with protective diodes as show in the lm datasheet. I've replaced the tl074 by a tle2072. Didn't tryed yet really at full volume playing because the only cab i have here now has a eminence 12" governor wich is really too much efficient (around 101db spl in cab).

The tga will be part of a bigger project. I've already posted about it, and it still evoluate.
At home, the governor use to be driven by a lm386 based amp wich is enough and never run at full volume...and the lm386 is only near 0,5 watt  .
I'll build a little tube ecl82(=6bm8=6f3p) amp (<2w) for this speaker to replace the lm386. I want to mix the tube amp with a mini sub with a active crossover for bass, i'll use the tga as a subwoofer amplifier.  For sub, i'll do a 4 ohm 6,5" tapped horn only 88db spl, but from 30 to 100hz in only 15 liters (with good group delay) with a 18sound 6nd430 (because high BL,low distortion and more H2 than H3, and dynamic lows to cross with the eminence). I've calculated i'll need 18 time the power to get the same sound level between the sub and the guitar speaker. (101-88=13db difference => (13db/3)^2=18 time more power needed). So 0,5watt i like at home for guitar cab => 0,5*18=9 watt for my relativly low efficience sub project. The tga will do the trick with not a lot of distortion
As i got a 16v 4,5A laptop psu, it should be enough to power both amp, and other effects. (i'll have to do a efficient boost converter, i'd like to try with a Ti tpa40210+ voltage doubler maybe to do something different from the ne555 and max1771)

Thanks Taylor for this funny tiny Giant project 
Cute pcb !

[PRR]

> prone to oscillation at near full volume. Without boxe, moving cables solve the problem.

All high-gain audio amplifiers WILL oscillate if input and output wires criss-cross and there is no box. Use a box.

> Both tda & lm seems to get be really hot too

If no box, what are you using for a heat-sink?

[jakobmagnusson]

Hi!
I ordered the Tinygiant and are doing some planning, I would appreciate some advice

1. I have a 18.5 V / 3.5 A PC charger, wll it do? Does the higher voltage ease the 4A  requirement?

2. I'm thinking of putting inside a small speaker cabinet (A TinyGiant combo   ), about 8 liters. The Tinygiant will be in a aluminum enclusure, inside the speaker, but with a cutout in the back, also with a heatsink at the back. The speaker will have bass ports (adding some ventialtion?). Will there be heating problems?

Thanks in advance,
Jakob

[waltk]

I have a 18.5 V / 3.5 A PC charger, wll it do?

Yes, that will be fine.

Does the higher voltage ease the 4A  requirement?

The 4A "requirement" is really a conservative suggestion.  You'll be fine with the 3.5A PS.  Having a higher-voltage PS, however, doesn't change the current required for the chip.  In fact, any excess input power has to be dissipated by the LM338.  For example, suppose the chip is drawing .3A at 11.6V - the voltage regulator has to dissipate 2.07 watts ( (18.5-11.6) * .3) as heat.  So a higher input voltage means that more heat will be generated in the LM338.

Will there be heating problems?

Probably not - sounds like a nice design.  Having decent heat sinks on the chips will be sufficient.  Just be careful to pay attention to the instructions about what can/can't be grounded.  The heatsink on the TDA7240 CAN be grounded (it's internally connected to ground); the heatsink on the LM338 CAN'T be grounded (it's internally connected to the +11.6V output of the chip).

Have fun, and good luck with your build.

[jakobmagnusson]

Hi,
Thanks Walt for your reply.
I'll post some pictures when it's ready, but it's my autumn/winter project so it'll take a few months time...
I'll be back...
/Jakob

[meffcio]

Ok guys, I'm kind of trying to marry the Tiny Giant power amp section with Marshall Lead 12 preamp section.
Here's what I've drawn:
http://www.mediafire.com/?4mfhuppg3wlo6xn
I'm, of course, very unsure of this design. Would you guys mind seeing it?

[EDIT]
S...t, forgot to mention I replaced the stock tone stack with a Fender type 2-knob tone stack.
Here's the original Lead 12 schematic:
http://music-electronics-forum.com/attachments/2552d1212538675-3005.gif
And the tone stack I used:
http://amps.zugster.net/images/articles/tonestacks/6G5.gif

[jakobmagnusson]

Hi,
I'm thinking of adding a few features to my TinyGiant.  My idea is to build a small combo with 6,5" speaker to be used when traveling. But I still would like the option of headphones, line in and also 12V (car/boat battery) power.

I would appreciate any comments or acknowledgment on below thinking

Headphones
Should be straight forward as suggested in reply 218:
http://music-electronics-forum.com/attachments/3932d1230346488-squier15.gif
I believe there are headphones outlets that has the possibility to disconnect the speaker.

12V
A car/boat battery is providing around 12V: can I connect it skipping the power regulator?
If the car/boat's engine is running the generator will raise the voltage up to 14,5 V, can it be directly connected after power regulator still?
Or should I in both cases connect it to the power regulator?

Line in
This was suggested in reply 119: http://i.imgur.com/Uill1.gif
Someone tried it but got a lot of noise. I'm thinking of connecting an MP3, Ipod or similar.
Is it a good solution? Alternative?

Tone stack
As suggested I'll use the tone stack, should not be a problem I think...

Many thanks for any reply on this.
Regards,
Jakob