Building the Tiny Giant amp

[p_wats]

Hmm...after a few weeks of touring and actually using my Tiny Giant as my main amp at a couple small shows I've found something a bit concerning: everytime I use this amp while singing into a PA I get shocked. Not a huge spark or anything, but definitely noticeable with sweaty lips.

I assume this has something to do with a ground problem somewhere in my build. Anything jump out at you that would cause this?

I used a plastic jack for the speaker out, but a regular jack for the guitar input.

I just tested and I do get intermittent continuity between my LM388T and the enclosure---could that be it?

[Taylor]

Hmm.

You're sure it's only with this amp and not others in the same club? This happens a lot with all kinds of amps, when the amp is on a different power system from the PA. The grounds of each system will sit at different voltage potentials relative to each other, and when you touch your strings which are connected to the ground of the amp, then sing near the mic, the voltage between these 2 grounds goes through you. My friend actually was telling me earlier today that he experiences this at his new practice space, using his amp which is some old Crate tube amp.

So, I can't say without knowing more if this has anything to do with the TG at all. But, having the heat sink of the LM338t connected to ground is definitely a bad thing which you should fix in any case.

[p_wats]

I initially dismissed it as bad electrical at a club, but then it happened a 2 more clubs after that and in our jam space.

Tonight I managed to recreate the shock, then switch the Tiny Giant for a different amp (Roland Jazz Chorus) and the shock was gone.

As for the LM388T, I've installed the bushings/spacer, but still get intermittent continuity to the enclosure. Would there be obvious symptoms if this was connected to ground?

I also checked continuity between my plastic speaker jack and it too has continuity to the enclosure somehow. Are the grounds of the board connected to the ground running to the input sleeve?

[Taylor]

No part of the speaker jack should be connected to ground. This is a bridged amp, so neither lead to the speaker is connected to ground. I can't really think of how it could be unless you tied all your grounds together - are the lugs of the speaker jack connecting to anything else besides the two SKR pads on the board?

[p_wats]

No part of the speaker jack should be connected to ground. This is a bridged amp, so neither lead to the speaker is connected to ground. I can't really think of how it could be unless you tied all your grounds together - are the lugs of the speaker jack connecting to anything else besides the two SKR pads on the board?

I have nothing going to the lugs of the speaker jack other than the connections to the corresponding PCB spots.

I removed the board/jacks from the enclosure and tested for continuity on the board itself and I get continuity between one of the speaker connection points and the board ground.

Would these connections (speaker to ground, LM388T to ground) impede the amp from working, as it works absolutely fine until I sing in front of a mic (which is why I never noticed until now)?

[Taylor]

I'm pretty confused about the regulator. The tab is connected to Vout. In other words, if there really is continuity between the tab and ground, I would expect the amp chip and opamp to have no power and not work.

If one side of your speaker is being grounded, I imagine that means that the output of one of the amps inside the amp chip (it's technically two amps, the outputs of which are connected to either side of the speaker) is being shorted to ground. I'm not an expert on bridge amp fail states, but I imagine this means you'd only get half the power out of it, if no other problems.

[p_wats]

I'm pretty confused about the regulator. The tab is connected to Vout. In other words, if there really is continuity between the tab and ground, I would expect the amp chip and opamp to have no power and not work.

If one side of your speaker is being grounded, I imagine that means that the output of one of the amps inside the amp chip (it's technically two amps, the outputs of which are connected to either side of the speaker) is being shorted to ground. I'm not an expert on bridge amp fail states, but I imagine this means you'd only get half the power out of it, if no other problems.

Yeah, this is strange.

So I desoldered the offending speaker connection, but the pad on the board is still showing continuity to the ground pad. I don't really know what else to try at this point.

I mean... it seems to work fine, save for the fact that I can't use it with a PA...

[p_wats]

New discovery: even when not attached to the enclosure my LM388T connects to ground. That tells me there's something on the board connecting the two, right?

[StarGeezers]

P wats , are you using the 3 wire grounded power supply ....  Since going to that , I have NO problems ...

[p_wats]

P wats , are you using the 3 wire grounded power supply ....  Since going to that , I have NO problems ...

My power supply has three prongs, but it's from eBay, so who knows about it's quality. I could try another.

However, what worries me is the continuity between ground and the speaker, LM388T etc. I'm testing when it's not plugged in...if that matters.

[waltk]

Hmmm...

Continuity between the LM338 tab and ground?  As Taylor said, it's hard to understand how it would work at all if that were true.  The tab on the LM338 is internally connected to the 12V output.  Dumb question... are you really sure you are checking the continuity of the LM338 with ground? (The TDA7240 tab IS connected to ground, and looks very similar to the LM338 from the top side)

-Walt

[p_wats]

Hmmm...

Continuity between the LM338 tab and ground?  As Taylor said, it's hard to understand how it would work at all if that were true.  The tab on the LM338 is internally connected to the 12V output.  Dumb question... are you really sure you are checking the continuity of the LM338 with ground? (The TDA7240 tab IS connected to ground, and looks very similar to the LM338 from the top side)

-Walt

Oh I definitely made sure to check the right tab. BOTH tabs show continuity to ground. Is it possible that my DMM has a much higher sensitivity and the beeping is a false positive?

[waltk]

Paul,

Oh I definitely made sure to check the right tab. BOTH tabs show continuity to ground. Is it possible that my DMM has a much higher sensitivity and the beeping is a false positive?

DMMs that have a continuity test (beep) function only require some small non-zero resistance to consider the circuit complete - 50 ohms on my Fluke DMM.  That wouldn't make difference for you because the LM338 tab should measure off the scale (high) resistance to ground.  The TDA7240 tab should measure zero resistance to ground.

So based on your description (both tabs being connected to ground), if you measure the resistance between the two tabs (on the LM338 and the TDA7240), you are getting zero resistance between them?  If that's true, then I would expect your power supply to shut down (temporarily or permanently) as soon as you connect power to the board.  If the amp is working when you connect power, then the only explanation I can think of is that your DMM is bad.  What is the actual resistance measured between the tabs (disregarding the beeping)? Zero?

-Walt

[p_wats]

Paul,

Oh I definitely made sure to check the right tab. BOTH tabs show continuity to ground. Is it possible that my DMM has a much higher sensitivity and the beeping is a false positive?

DMMs that have a continuity test (beep) function only require some small non-zero resistance to consider the circuit complete - 50 ohms on my Fluke DMM.  That wouldn't make difference for you because the LM338 tab should measure off the scale (high) resistance to ground.  The TDA7240 tab should measure zero resistance to ground.

So based on your description (both tabs being connected to ground), if you measure the resistance between the two tabs (on the LM338 and the TDA7240), you are getting zero resistance between them?  If that's true, then I would expect your power supply to shut down (temporarily or permanently) as soon as you connect power to the board.  If the amp is working when you connect power, then the only explanation I can think of is that your DMM is bad.  What is the actual resistance measured between the tabs (disregarding the beeping)? Zero?

-Walt

Thanks for making me second guess my DMM. I don't think it's broken, but it would appear the threshold for continuity behaves funny.

For instance, the TDA7240 to ground beeps consistently and measures 0 resistance (it bounces around the .00-.06 range and back over and over) while the LM338 to ground beeps momentarily and then the resistance climps to infinity slowly to infinity. For some reason my DMM is slow on the uptake.

Also, the speaker connection to ground does the same thing--beeps briefly to indicate continuity, but then you can watch the resistance climb to .5K (that still seems low, though...doesn't it?).

I'm glad this seems to explain the mystery of continuity, but I'm still getting small shocks from the microphone when I use this amp with a PA, while others don't have the same effect.

Could it have something to do with my power supply? I have 19V 4A 3-pronged supply on which I soldered a new jack (to fit a regular DC pedal input). Should I try an insulated input jack?

I appreciate your help and patience with this build. I love this amp when it's not shocking me! Ha.

[StarGeezers]

  Paul , sounds like something is still not right !!!        Mine works SUPER , no shocks or any other problems ... I did however use  both isolators , supplied by Taylor , on the LM ,and both on the bolt inside and out ,  not wanting to take any chances ...  I get an Open reading on mine .. 

[p_wats]

  Paul , sounds like something is still not right !!!        Mine works SUPER , no shocks or any other problems ... I did however use  both isolators , supplied by Taylor , on the LM ,and both on the bolt inside and out ,  not wanting to take any chances ...  I get an Open reading on mine .. 

Do you use yours with a PA (ie, while singing into a mic)? As that's the only time it shocks me.

[slacker]

I think the reason you get shocked is because ground on the Tiny Giant isn't connected to earth at the wall socket, so it probably isn't at the same level as ground on the PA. So if you're touching ground on the TG via your guitar strings and touch ground on the pa, by touching the mic, you'll get a shock.

I did a test with mine by measuring the voltage between the sleeve of a guitar lead plugged into the TG and the sleeve of a guitar lead plugged into my Fender amp, where ground is connected to earth at the wall socket, and measured about 0.2 volts. I don't know if this is enough to give you a shock, I didn't try and find out.

I don't know if there's a solution to this or if it is even the problem, hopefully someone like Paul or R.G. will pop by.

[PRR]

> this amp while singing into a PA I get shocked

What Ian says. "ground on the Tiny Giant isn't connected to earth at the wall socket, so it probably isn't at the same level as ground on the PA."

It "floats".

This depends totally on the power supply wart you use. Computer power supplies don't ground the same as stage-amp supplies.

It is VERY possible there are two 0.05uFd caps from each side of the line to "system ground". On 120V line this gives 60V AC from amp chassis (and jacks and guitar strings) to ground.

The 0.05uFd value is a compromise between reducing (computer) RF EMI, and human safety. 0.05uFd at 60Hz is about 53K impedance. 60V across 53K is a bit over 1 milliAmp current into a short-to-ground, perhaps half that to ground through human skin.

0.5mA is a tingle, and quite shocking when un-expected, but is significantly below accepted safety thresholds (2mA to 50mA max leakage). 

> measuring the voltage between the sleeve ... and ... my Fender amp... measured about 0.2 volts.

Then your power supply either has extremely low leakage OR is somehow properly grounded all the way from wall to TG amp DC jack. 0.2V stray voltage is not uncommon on a "good ground". It may be 0.2V in the _Fender_'s input filtering.

> don't know if this is enough to give you a shock

Not unless you peel your skin and stick it right into a nerve. Put a dry finger on a 9V battery: no shock. Wet finger: maybe slight shock. Wire-up a 1.5V battery so both ends can touch skin close together. Very few persons will feel 1.5V. 0.2V is "nothing".

Not sure what you can do.

The "correct" answer is to have solid wire from wall-outlet ground pin to amplifier chassis. But that means hacking the power cord and adding a connection to go around the power-lump to the TG chassis.

You could keep buying $13 supplies until you happen on one with incredibly low stray leakage.

There is the old trick I used to use before modern grounding. Take any unused connector on the floating chassis. Run a cable to any unused connector another chassis which IS solid grounded. I took AUX OUT on the leaky film mixer to AUX IN on my PA mixer, then kept my AUX IN firmly down. The cable drained the leakage of the film mixer's 0.1uFd(!) caps across 2-pin plug. Signal cables are NOT rated for bet-your-butt grounding. Any REAL ground-fault would burn a hifi/guitar cable and leave you ungrounded. But sometimes you do what you do.


________________________________________________________________
> continuity between ground and the speaker

"Continuity" testers are for trailer light checks. Where you either have a solid metal path, or you don't.

Poking at semiconductors, there are all sorts of low-voltage and reverse-polarity "sneak paths" plus capacitors which confuse "continuity" tests.

> Is it possible that my DMM has a much higher sensitivity and the beeping is a false positive?

It is a poor tool for this work.

> beeps briefly to indicate continuity, but then you can watch the resistance climb to .5K

The "brief beep" is some large capacitor being charged by the "continuity test" voltage. Poke some caps and see.

The "0.5K" is a false reading; it is probably a 0.6 Volt threshold of some transistor junction. (Try some other diodes and see if they all read "half-K".)

[p_wats]

Wow! Thanks for all that info. Very helpful.

The option of hacking the power cord sounds like the most doable, but still an awkward bit of work. Otherwise I simply consider this a non-stage amp (which most people would anyway, but I play smaller venues with no drummer).

[StarGeezers]

 We just plug ours into the SAME power strip as the PA .... and use a foam mic cover... No shocks , no spit in the mic.  If we loan the mic out, it gets a different cover...