TDA2030 DC at output

[Rob Strand]

And.... that's the only one comes to mind (outside obsessed one-offs).

There's always one!

(Come to think of it I have a feeling Electronics Australia had an amplifier project which use a regulator.  It was a similar design to the Dynaco with 2N3055's.   Maybe around 1970.)

[PRR]

> peak current is a bit higher

Yes, but if it will stand >1,000uFd across the rails the raw supply only feels 1/2.828 of that.

This supply should be OK sine-wave down to 4 Ohms. (Guitar may touch square-wave operation, true.)

EDIT- dropped a zero

[Rob Strand]

Yes, but if it will stand >1,00uFd across the rails the raw supply only feels 1/2.828 of that

Yes buffering with a cap will help drop the peak.  Cap buffering relies on the impedance of the leads between the PSU and the added cap.  The lower the lead impedance the larger the cap required but a lower voltage dip.  (I suspect the cap needs to be big, like 1000uF).

I doubt there will be problems.  It's only if the PSU is one of those crazy ultra-sensitive ones.  Most aren't *that* sensitive.   If need be there's a few back-up plans.

[PRR]

> big, like 1000uF

That's what my brain typed, a good value for speaker power amps.

I see my fingers didn't hear so good, and have corrected my post.

I'm not sure how lead impedance matters, except in exceptional cases. 1,000uFd will supply the one positive peak in an audio wave with only small sag. In one sense if the lead impedance is real low, we hardly need this cap. If the power supply is cutting-out for a slight over-current, significant lead impedance hides the overcurrent from the power supply (but ensures sag on sustained passages).

This all assumes the supply tolerates 1,000uFd on its output. Some switchers will flip-out. Lead impedance may then be vital. 1 or 2 Ohms in this line may mask the effect with only slight effect on speech/music performance.

[MrStab]

hi guys,

i didn't get round to replying at the time, but i eventually consolidated everyone's advice and rebuilt the pre & power-amp for the talk box using a higher supply voltage.

the friend it's for seems happy with it. originally the plan was to send the modulated signal from the mic back into the guitar chain, but as he just needed it to record for now, i only built the pre/power amp section. he needed it asap and didn't consider the switching noise to be significant, so he's taken it away to try.

i used a generic 24V, 1.75A SMPS for laptops. i think the switching noise is partly from the input wire being close to the power line at one point, as a low impedance input was much less susceptible. i've offered to take it back for tweaks if needed, so i could hopefully resolve that. apparently the noise isn't noticeable until i point it out right next to the speaker.

what fixed most of the clipping was regulating the pre-amp to 15V. the pre is now just a buffer with spare half-op-amp used for buffered (overkill) Vref and a volume pot. The TDA2030 gain is at 22. It only slightly distorts on hard strums from Invader pickups, but it's really clean with a slight attenuation or cooler input. Bypass works more or less like the Rocktron Banshee (i think), acting kinda like an amp selector by grounding either path. iirc the heatsink is rated either 7 or 9 C/W, and fits snugly whilst keeping things cool.

atm the tube is too long, like 3-4 metres or something, but it still gives more than enough volume to do, er, talkboxing, so it'll be better when he cuts it down a bit. i have a quick video on my phone of me hitting strings without a tube connected, i'll try to upload it at some point.

thanks again for all the pointers!

[Rob Strand]

Good to see it all worked out.

BTW, thanks for the report, it's nice to hear how things turn out.

As for the noise often this stuff comes down to specifics.   It could be the proximity of the signal input to the PSU input.  Another common problem is if you follow the wire/track from the signal input ground all the way from the input socket to the input of TDA2030, if the power supply ground wire connects anywhere along  this wire then it is a potential source of noise.  The PSU needs it's own ground path to the TDA2030.  You have like a quite ground and a noisy ground.  Also if the PSU bypass cap connects to the quiet ground track you also get issues. It needs to be on the noisy ground track too.
The last point, which can be hard to see at first, is when you have grounded input and output sockets.  What this does is connect the quite and noisy grounds together through the chassis. You can get noiuse even if you have done the right thing with the grounds on the PCB!

[MrStab]

the TDA2030 section, the speaker and the power filter+preamp ground all go straight to the PSU negative separately, with a connection from there to chassis star ground, and the jack and switch grounds go straight to the star ground. so i think i might've got most of the isolation fairly okay, apart from allowing the filter caps and pre-amp to be on the same wire?

ie. the filter caps, an RF LPF, an opamp ground, Vref divider ground and Vref bypass cap all go on the same wire. sounds bad just typing it out! if i get it back i'll try splitting those. would everything in that list, minus the power filter caps, be better going straight to the PSU or chassis ground? none of it is (directly) for reference, just dumping crap from the opamp and filters, so PSU i'd imagine

[Rob Strand]

, Vref divider ground and Vref bypass cap all go on the same wire. sounds bad just typing it out!

The Vref stuff can be on the "quiet ground", in fact it is probably better there.

if i get it back i'll try splitting those. would everything in that list, minus the power filter caps, be better going straight to the PSU or chassis ground?

Any large PSU caps should connect to the "noisy ground".

[MrStab]

ok, i thought as much. thanks for that!

here's the clip i was on about:

the guy hasn't used it "properly" yet as he ran out of time when recording.