Building the Tiny Giant amp

[David]

I'm working on testing mine now, which is proving interesting since my guitars are in the shop and I'm
trying to improvise using a signal generator.  I had issues with the op-amp not being in all the way and
a faulty mute switch.  I tracked these down and I'm going to put a different switch in.

What I'm noticing is that the 10K is not acting quite as I expected.  It was marked 10K audio, so I don't
THINK it's linear, but it's not acting like it's logarithmic.  I get nothing for probably a third of the pot's travel,
then the volume climbs like a Saturn V taking off.

Any ideas?

I was wrong.  I fed it a bass instead of a Colpitts oscillator.  Holy cow, Houston!  We have an amp!  Taylor, this thing is INCREDIBLE!!!!!!!!!!!!!!!!!!!!!  I won't be able to use it this week because I don't have the pedalboard power takeoff yet.  It should be ready to go next month.  I CAN'T WAIT!!!!!!!!!!!!!!!!

[thedefog]

Hmm, that's quite strange. Generally, distortion means clipping. But I can't think of any normal way that something could clip where the amount of distortion does not vary with input volume.

Are you plugging your instrument straight into the TG? Is it a guitar or something else? If a guitar, is it a normal passive one?

This is plugging straight in with my guitar, or anything for that matter. Just got a dirty signal. It's like 1/10th of the signal is noisy/distortion and the rest is clean, which leads me to believe I've got a bad component somewhere. I'm gonna try different TL0x2's as well as some other duals I've got laying around first.

[Taylor]

This is plugging straight in with my guitar, or anything for that matter. Just got a dirty signal. It's like 1/10th of the signal is noisy/distortion and the rest is clean, which leads me to believe I've got a bad component somewhere. I'm gonna try different TL0x2's as well as some other duals I've got laying around first.

Have you checked that the supply voltage matches what it says in the PDF?

[thedefog]

This is plugging straight in with my guitar, or anything for that matter. Just got a dirty signal. It's like 1/10th of the signal is noisy/distortion and the rest is clean, which leads me to believe I've got a bad component somewhere. I'm gonna try different TL0x2's as well as some other duals I've got laying around first.

Have you checked that the supply voltage matches what it says in the PDF?

Yeah and they all check out to what they should. I will re-check later tonight though. Something may have moved when I put it into the cabinet.

[dylar]

Just finished building my Tiny Giant.  Works perfectly and is plenty loud.  Did it all for $45.
kit: $20
hi-fi speakers from Goodwill: $10
Sony VAIO power source from ebay: $10
misc. hardware: $5
This thing is great--I wanted to build one to play background percussion tracks while performing and it will do that nicely.
Thanks Taylor!

[Taylor]

Cool, post pics if you get a chance dylar.

[dylar]

yes, it looks like a set of speakers with a Radio Shack project box on top - because that's what it is.  Cost effectiveness was my main concern.  Not pretty but works great!

[PRR]

> I haven't had a chance to measure current draw yet, but I wouldn't be surprised if it was something like 3a or less in regular use.

Assuming 4 ohm load and about 12V regulated voltage to the TDA7240 chip: 2 Amps near-clipped sine-test-tone, 3 Amps maxxed-out utterly-clipped (LOUD!!).

On 8 ohms, 1-1.5A; on 16 less than 1A tops.

At sine-clipping in 4 ohms, the total heat in the box is 18V*2A= 36 Watts in, ~~20Watts delivered to speaker, 36W-20W= 16 Watts of heat in the box. (Similar when maxx-overdriven... more power in but even more power to speaker, about 18W heat in box.) This is a lot of power for these small boxes. They will get HOT. Too hot? Apparently not. If overheated, these chips will shut-down and re-start when cooler. If it craps-out under abuse, and chips are HOT, re-check chip thermal contact to case (flat, smooth, greased) then ponder a larger case or higher-impedance load (and less power out).

These calcs are worst-case. Many players won't come close to these numbers; it may be hard to approach 3A demand with actual fingers and speakers. I think typical heat is 5 to 10 Watts, the box will run very-warm not HOT.

Since you have a 4A-7A regulator and 4.7A cord-warts, there is at least another 12V 1A which could be stolen for pedals, direct or via a 9V regulator. That's ample for any pedal which does not need a dedicated wall-wart.

> the half of the TL072 that is creating the bias voltage

The unused side of TL072 is just UN-used. Not even "creating the bias"; no real reason to do that. Taylor connected it as follower and nailed the input to +12V, which just slams it into a "do nothing" state so it won't make trouble. (TL071 could be used, but is less common and no cheaper.)

Distortion? Measure voltages. There's a troubleshooting thread somewhere.

[David]

> I haven't had a chance to measure current draw yet, but I wouldn't be surprised if it was something like 3a or less in regular use.

Assuming 4 ohm load and about 12V regulated voltage to the TDA7240 chip: 2 Amps near-clipped sine-test-tone, 3 Amps maxxed-out utterly-clipped (LOUD!!).

On 8 ohms, 1-1.5A; on 16 less than 1A tops.

At sine-clipping in 4 ohms, the total heat in the box is 18V*2A= 36 Watts in, ~~20Watts delivered to speaker, 36W-20W= 16 Watts of heat in the box. (Similar when maxx-overdriven... more power in but even more power to speaker, about 18W heat in box.) This is a lot of power for these small boxes. They will get HOT. Too hot? Apparently not. If overheated, these chips will shut-down and re-start when cooler. If it craps-out under abuse, and chips are HOT, re-check chip thermal contact to case (flat, smooth, greased) then ponder a larger case or higher-impedance load (and less power out).

These calcs are worst-case. Many players won't come close to these numbers; it may be hard to approach 3A demand with actual fingers and speakers. I think typical heat is 5 to 10 Watts, the box will run very-warm not HOT.

Since you have a 4A-7A regulator and 4.7A cord-warts, there is at least another 12V 1A which could be stolen for pedals, direct or via a 9V regulator. That's ample for any pedal which does not need a dedicated wall-wart.

> the half of the TL072 that is creating the bias voltage

The unused side of TL072 is just UN-used. Not even "creating the bias"; no real reason to do that. Taylor connected it as follower and nailed the input to +12V, which just slams it into a "do nothing" state so it won't make trouble. (TL071 could be used, but is less common and no cheaper.)

Distortion? Measure voltages. There's a troubleshooting thread somewhere.

Sorry.  My mistake.  It looked like it was generating Vbias.  In any case, mine is working just fine.  Now I have to get the secondary power supply working.

[pruttelherrie]

Finished my Tiny Giant this weekend, it is indeed plenty loud. (Pics to follow later this week)
At first I thought I had some kind of overdrive going on when going past 9:00 on the (25k) volumepot, but that's only the case when I plug the guitar direct into the amp: going over 9:00 will not output more volume, only more distortion. But with a pedal in front I get waaaay more volume, although I still can't get over 9:00 with the volumepot. Haven't measured the levels yet, as far as I can see the TDA clips at that point?

Anyway, two things in the build pdf v2 caught my eye:

Before you power the TG up the first time
-Use your multimeter set to continuity test mode, to verify that neither terminal of the speaker jack is connected to ground.
-Use your multimeter set to continuity test mode, to verify that the heat sink tab of the LM338T is not connected to ground.
-Use your multimeter in voltage mode to test the voltage at pin 8 of the TL072. It should be approximately 11.6 volts.

That last point is not possible *before* powering up

Raise the 1k to: 1.2k 1.5k
the output voltage becomes: 13.75v 16.88v
the power becomes: 20w 24w

But wasn't 'stock' (1k) supposed to be 20W already?

All in all I'm very happy with this project, I'm pondering building a few more!

[Taylor]

Finished my Tiny Giant this weekend, it is indeed plenty loud. (Pics to follow later this week)
At first I thought I had some kind of overdrive going on when going past 9:00 on the (25k) volumepot, but that's only the case when I plug the guitar direct into the amp: going over 9:00 will not output more volume, only more distortion. But with a pedal in front I get waaaay more volume, although I still can't get over 9:00 with the volumepot. Haven't measured the levels yet, as far as I can see the TDA clips at that point?

That could be, it depends largely on what kind of instrument you're using. The amp is designed with some boost in the preamp section, to accommodate a variety of inputs, from weak to hot guitars, piezo-equipped acoustic instruments, etc. So for example my cello with piezo element is much quieter than my active bass - I can turn up the volume half way with the cello with no sign of clipping.

-Use your multimeter in voltage mode to test the voltage at pin 8 of the TL072. It should be approximately 11.6 volts.
That last point is not possible *before* powering up

You're quite right about that. Thanks for bringing that error to my attention.

Raise the 1k to: 1.2k 1.5k
the output voltage becomes: 13.75v 16.88v
the power becomes: 20w 24w
But wasn't 'stock' (1k) supposed to be 20W already?

The thing is that wattage ratings are - always, by everybody - an estimate or average or a shrug. In precision test applications, an amp must be measured at some incredibly tiny total harmonic distortion to find the "wattage" rating. For guitar, the standard is to measure at some much higher level of distortion and call that the wattage. So for example, the EHX 22 caliber is claimed to be a 22 watt amp, but I calculated at some point that it's not possible for it to be 22 watts except by a very lenient definition of watts.

Then you have DJ equipment and car stereos which are rated by a standard invented by some truly magical thinkers: this equipment is measured by the absolute biggest peak it can put out, with some monstrous amount of distortion. So you have car amps not much more powerful than the Tiny Giant, being rated at like 900 watts or something else ridiculous.

The amp chip in the Tiny Giant, the TDA7240A, is called by the datasheet, a "20 watt amplifier" - so that's why I've called it that in the literature. In reality, depending on supply voltage and the impedance of the speaker you plug into, it's a 12-24w amplifier. 20 watts is an average since I can't know how you'll hook it up beforehand.

Audio amp ratings aren't standardized, so there's no way to be objective about it, unfortunately. I do hope that the chart at least shows, in an understandable way, how to make it louder by changing the resistor.

Another thing to point out is that double wattage is not twice as loud, only 3db louder, which is on the edge of perceptibility. In other words, don't stress too much about watts. To get twice as much volume as this amp can do, you'd need 200 watts, and no resistor change will get that out of this design, so if something in the range of 15-24 watts is not enough, a 3886 amp would be the next step up.

[thedefog]

The ground wire from the input jack to the PCB had come a little loose and was the source of the distortion and 60hz hum in my build. I have my jack plate grounded, so the jack was grounding at the plate and not directly at the point on the PCB. As soon as I re-soldered it, it went away. I actually just redid my entire ground network to a single point on the jack plate and it quieted things up considerably. I should have done that originally, but never thought it would be an issue. I should have known better, having just built a tube amp recently. It's just good practice to always do that.

[Taylor]

Cool, glad you got it sorted. 

[Taylor]

Finished my Tiny Giant this weekend, it is indeed plenty loud. (Pics to follow later this week)
At first I thought I had some kind of overdrive going on when going past 9:00 on the (25k) volumepot, but that's only the case when I plug the guitar direct into the amp: going over 9:00 will not output more volume, only more distortion. But with a pedal in front I get waaaay more volume, although I still can't get over 9:00 with the volumepot. Haven't measured the levels yet, as far as I can see the TDA clips at that point?

pruttelherrie, can you give me some info on your guitar? What sort of pickups, etc.? I was testing the amp a bit today, and with a passive bass I was able to get it up to about halfway on the volume dial with absolutely no clipping, and this was quite loud.

When you turn it up to the point that it clips, is it ear-splittingly loud? Are you playing in some kind of practice space or just in your home? It could be fine, and you're just hitting the max volume, however if you're getting clipping at anything less than ear-splitting levels I wonder if there's an issue with your build, as in thedefog's ground situation.

[Poste]

I intend on putting this amp in a cab with a 8 ohm 12" celestion speaker, will this hook straight up.

Also how would I wirer up a superfly amp so that I can switch between the two in the one cab, would it be as easy as just using a 3pdt toggle with a couple of leds to show which is in use.

On a side note can the superfly circuit take 15V?

Thanks in advance

[pruttelherrie]

Finished my Tiny Giant this weekend, it is indeed plenty loud. (Pics to follow later this week)
At first I thought I had some kind of overdrive going on when going past 9:00 on the (25k) volumepot, but that's only the case when I plug the guitar direct into the amp: going over 9:00 will not output more volume, only more distortion. But with a pedal in front I get waaaay more volume, although I still can't get over 9:00 with the volumepot. Haven't measured the levels yet, as far as I can see the TDA clips at that point?

pruttelherrie, can you give me some info on your guitar? What sort of pickups, etc.? I was testing the amp a bit today, and with a passive bass I was able to get it up to about halfway on the volume dial with absolutely no clipping, and this was quite loud.

I tested with an Ibanez VBT-700, the short-lived Flying V model from early 2007 introduced together with the Xiphos but discontinued shortly after. It has DiMarzio D-Activators. When I plug it in clean, it clips at home levels. When I use a random distortion box in front, it gets ear-splittingly loud. Well, I didn't dare to test it at full volume at home. I will test it in the rehearsal room on sunday and report back. I suspect it's just the difference between clean and distorted, at clean signal you need much more headroom/voltage to get the same perceived volume.

When you turn it up to the point that it clips, is it ear-splittingly loud? Are you playing in some kind of practice space or just in your home? It could be fine, and you're just hitting the max volume, however if you're getting clipping at anything less than ear-splitting levels I wonder if there's an issue with your build, as in thedefog's ground situation.

Ok, I spent half an hour with the signal generator and the scope, and this is what I measured:
* Input voltage (HP laptop PSU): 19.3V
* LM338 output voltage: 11.55V (disregarding the error from the cheapo voltcraft multimeter)

Next I put the volumepot at max, so no divider between TL072 and TDA. Then I turned up the sine-generator until clipping occurred at the output terminal. Voltages were as follows:
* Output: approx. 20V peak-peak
* Input TL072: 60mV peak-peak
* Output TL072: 200mV peak-peak

Does this make sense?

Iwan

[crash415]

If I wanted to add an eq like the Tonemender or Mr EQ to this, where would I put it?  Right before the 1 microfarad cap before the volume?

[Taylor]

Since those are both active EQs, it doesn't really matter. I would just put it in front of the whole circuit, and treat it like you have an EQ pedal in front of your amp.

If the TGA had a clipping section, you'd want your EQ after it, but since it's a clean amp, EQ can go in front. I have been meaning to test out a mod that turns the preamp buffer into a TS-style clipper, just haven't gotten around to it yet.

[PRR]

> * Output: approx. 20V peak-peak
> * Input TL072: 60mV peak-peak
> * Output TL072: 200mV peak-peak
> Does this make sense?

Yes.

The TL072 is wired for gain a bit over 3; 200:60, check.

IIRC the power amp has gain of 50 per side, so 100 both-sides; 20V/200mV, check.

60mV p-p is 20mV RMS (if you can ignore meter error, I can use 2 * sq.rt.2 = "3"). 20mV is THE classic input sensitivity for ~~~1960 Fender guitar amps. (50mV early 1950s, 1mV for some modern iron-shredders.)

20mV is an extreme setting; many players in many situations will turn-down on axe or amp. 200mV seems to be a hot guitar-cord signal. 200mV RMS is 600mV p-p. Times preamp gain of 3 is 1.8V p-p,  below the TL072's 9V p-p limit. Taylor has designed so it takes 3V p-p 1V RMS input to overload the TL072.... such levels are rare, will overload the classic 1960s Fender's input, and if you have hotter signals you are probably looking for overload.

> ""Raise the 1k to: 1.2k 1.5k - the output voltage becomes: 13.75v 16.88v - the power becomes: 20w 24w""

The difference in "power" is, as Taylor says, moot.

> double wattage is not twice as loud, only 3db louder

At conversational levels, double power is not quite "one and a half times as loud". Above 80dB SPL (the only way to play guitar, even home alone) the 3dB increase is compensated by 1dB-2dB of inner-ear tense-up, so not even 1.4 times as loud. Barely audible. So whether it is 18W or 24W is pretty insignificant.

I do suspect that advanced users should trim the "12V" as high as it will go without exceeding 18V or getting voltage-drop on LOUD passages with low-Z speaker. Not so much for the "more Watts", but because this slightly reduces dissipation in the regulator and maybe total heat in the box.

[pruttelherrie]

> * Output: approx. 20V peak-peak
> * Input TL072: 60mV peak-peak
> * Output TL072: 200mV peak-peak
> Does this make sense?

Yes.

The TL072 is wired for gain a bit over 3; 200:60, check.

IIRC the power amp has gain of 50 per side, so 100 both-sides; 20V/200mV, check.

60mV p-p is 20mV RMS (if you can ignore meter error, I can use 2 * sq.rt.2 = "3"). 20mV is THE classic input sensitivity for ~~~1960 Fender guitar amps. (50mV early 1950s, 1mV for some modern iron-shredders.)

20mV is an extreme setting; many players in many situations will turn-down on axe or amp. 200mV seems to be a hot guitar-cord signal. 200mV RMS is 600mV p-p. Times preamp gain of 3 is 1.8V p-p,  below the TL072's 9V p-p limit. Taylor has designed so it takes 3V p-p 1V RMS input to overload the TL072.... such levels are rare, will overload the classic 1960s Fender's input, and if you have hotter signals you are probably looking for overload.

Ok, thanks for this explanation.

But now for the TDA: this one will overload way before the TL072 does. I have a feeling that the gain of the TL072 could to be turned down a bit, or I'll have to add a resistor to the volume pot to get more usable range. Or even make the preamp gain variable. A 'clip' led at the input of the TDA might be a handy feature in this case.

> ""Raise the 1k to: 1.2k 1.5k - the output voltage becomes: 13.75v 16.88v - the power becomes: 20w 24w""
I do suspect that advanced users should trim the "12V" as high as it will go without exceeding 18V or getting voltage-drop on LOUD passages with low-Z speaker. Not so much for the "more Watts", but because this slightly reduces dissipation in the regulator and maybe total heat in the box.

Yeah I was thinking that. In a simplistic way: All the current delivered to the speaker has to go through the regulator. So the less voltage drop, the less power it has to dissipate.

Above 80dB SPL (the only way to play guitar, even home alone)

Heheh this made my day