TDA7293 supply voltage bypass caps

[JBro]

TDA7293 supply voltage bypass caps

I have a guitar amp that I am fixing for myself from a manufacturer that does not make repair info public...no schematics, no parts. So i reverse engineered the power amp stage but am reluctant to post it in case i break the rules of the forum. So instead i will describe a circuit in the stage and ask for some advice on it.

The stage is based on a single TDA7293. The amp is rated 75 watts thru a single 12", 8 Ohm speaker.
It uses +48V for +Vs at pins 7 & 13
It uses -48V for -Vs at pins 8 & 15
Following Fig. 1 'Typical Application and Test Circuit' in the datasheet, the values for the supply voltage bypass caps are as follows:
C6, C8 - supply voltage bypass for both +Vs and -Vs = 47 uF
C7, C9 - supply voltage bypass for both +Vs and -Vs = ?? uF

C7, C9 are surface mount and appear to be beyond the range of my multimeter, which goes to 20uF AND i am not sure if i can get an accurate reading of the caps IN CIRCUIT (same as datasheet).

My question is, is it possible to guess at the appropriate value of C7, C9 based only on the values of +Vs,-Vs,C6,C8 and the values shown in the datasheet?
If so, how would i determine these values? Can anyone tell me what they should be and why?

Thanks,
JBro

[J M Fahey]

Feel free to post what *you* drew yourself.
For some quirk in the Law, circuits or schematics themselves (meaning what is connected to what and in what way) are not patented, but the actual *drawing*, be it a schematic, layout or PCB design is "a work of art" which is protected by law.
So often you can´t post a scan or photocopy or PDF sent to you by the manufacturers, which hold the rights to that drawing or image, but yours is yours and you are free to share.
Besides that, I can´t imagine how different it can be from the application suggested in the Datasheet.

[JBro]

OK...Thanx for the explanation on the rules.

There are 3 more boards..Power supply, DSP, User interface.(pre-amp on DSP board)
Jumper JB1 connects to the PS board.
The DSP has 10 wires jumping to it from the PS board.
There is a headphone out and aux in (both stereo) on the DSP board.

The amp seems to work because I get a good sound thru the Hphones (with guitar or mp3 thru aux). Just a really degenerate sig thru the speaker.
I have to admit, i replaced the 7293 once and had the amp working great.
Then, like an idiot, i sparked the heat sink tab of the 7293 on the chasis (GND).
I then replaced the 7293 and the amp did not work... I am hoping the second 7293 was just bad (got more on order). But just in case, while i am waiting for the new one, I wanted to get a jump on replacing the entire board if needed.

Also, i may want some advice on building a test circuit for the power amp plugged into JB1 and i'm not sure what exactly is going on there with all the apparent inputs.

Anyway here is the Power Amp PCB.

[teemuk]

I'm pretty sure you'll have some errors in that sketch but I'll throw in a guess: It's a Line 6 amp, right?

[JBro]

To hell with caution i suppose, seeing as how the moderators are not stopping me.
Yes, its a line 6 Spider IV 75. I have checked and re-checked the schematic i came up with against the actual pcb and this is the final outcome. I will give another check right now...*sigh*. Not too proud to admit i can make mistakes...already admitted to frying the first replacement chip stupidly. But, can you at least tell me why you think there are mistakes as that will help me narrow down where to look?
Thanks,
JBro

[J M Fahey]

Well, I think the problem lies on your sparking that second TDA7293.
It certainly was not bad when new.
To play it safe:
1) search here and build a lamp bulb current limiter, plug your amp there (and not straight into wall power) until you are certain it´s repaired, this will save you some $$ but more important avoid bad moments.
2) do not connect your speakers until all DC voltages look OK.
Specially you must not have DC on th speaker out
3) besides the power IC (or chipamp) itself, you may have fried some tracks in your PCB , molten some connector, stuff like that.
Check (and repair) everything.
When you have the new IC mounted, check that you have proper rail (power) voltages and that you do not have DC on the speaker out, only then connect the speaker and test sound, at low power.
Only after you can play properly, plug it straight into the wall.

[teemuk]

But, can you at least tell me why you think there are mistakes as that will help me narrow down where to look?

- For one, these things have two feedback loops: one will connect right from the chip amp's output (pin 14) to the inverting input, that's the main voltage feedback loop. The second loop will connect to non-inverting input after those two "current sampling" resistors hooked between chip's pin 14 and pin 1 on the connector, that's the current feedback loop.
- The non-inverting input should have somekind of a ground reference, so likely you'll find that circuit (consisting of two resistors and a cap wired in parallel) drawn between inverting and non-inverting inputs connecting to either ground or directly to chip amp's output pin (14) instead.
- Input signal hooks up to non-inverting input, the other circuit that at quick glance looks like another input is actually a resistor divider for the voltage feedback loop. It probably terminates to audio ground on the board it connects to.

You should get pretty far with that.

In addition, though Line 6 has a policy no to share service documentary the stuff can still be found with persistent search. You don't likely even have to find a specific schematic since the overall architecture of their chip -based power amps tends to be the same. I can try to dig up my archives because I know for certainty that I'll find a very, very similar Line 6 schematic of a power amp, with the exception of the chip being LM3886.

If so, how would i determine these values? Can anyone tell me what they should be and why?

You will commonly find them to be close to 100 nF. 47 nF - 150 nF ... it doesn't really matter, the important point is that they are in close vicinity to power supply pins, overcoming any inductance long wire or trace runs from main, larger filter caps might introduce.

[JBro]

Thanks guys,
Got the Limiter built. Revised the sketch (attached). Tested the circuit. No glowing bulb, no DC on the output. Supply voltage good. Good logic level on pins 9/10. Sound thru speaker is still weak and very distorted.

What else can I test? Thinking about prototyping a test circuit with the new chip amps rather then just going ahead and soldering one in and risking losing it...again. Wish these things fit in a breadboard! How do you guys prototype a circuit with these things?

Anyway, here is the schematic I came up with. All of the 20k resistors are in a smd. That, and a few other devices block my view of the board, so it was difficult to trace. Hope its right.

[J M Fahey]

There must be some mistake (s) in your schematic because pin 3 is the + input and can not be connected to the output (you show it connected through two parallel 20k resistors and a .043 ceramic cap.)
Pin 2, the - input yes , because the feedback loop goes there.
Sorry but can´t help without a proper schematic.
As a general rule, after checking that there are no broken traces, I´d inject clean audio signal straight into pin 3 , to know whether it comes already distorted or the chip itself is failing.

Worst case, (what I do in "unrepairable cases") build *another* TDA7293 amplifier, with its own mini-board, mount it "upside down" on the heat sink, so it does not even touch the old board, and connect it to power, ground, audio, speaker, etc.
Looks ugly, but saves the day.
Post some pictures showing the heatsink and power board area.

[teemuk]

Dug up some Line 6 stuff. Here's the TDA7293 section of Duoverb.

In Duoverb the +Audio and -Audio inputs are actually push-pull signals from a crude OpAmp phase splitter circuitry (simple inverting amp between feeding points of + and - audio inputs). In most Line 6 amps the -Audio input is simply grounded, though.

I don't know what you specifically have in your amp since I don't have the exact schematics for it.

The main architecture of the TDA7293 section is still similar enough to provide an example.

[JBro]

Thanx teemuk!
First, i will breadboard the power amp as i drew it and plug it into JB1's associated jumper.
If the amp works, i know i have a bad chip amp or some other component on the pcb.
If not, i'll have to dig deeper into the other boards.
I feel like i can make some progress now.

[J M Fahey]

Excellent idea.
If the "external" amp works, it will help you a lot either to troubleshoot the bad one or just plain replace it.

[JBro]

Ok,
The prototype works great. I only tested it for a very short period of time as it was sloppy and had no heat sink, I didn't want to push my luck. Also, it was somewhat of a hack as some caps were seriously underrated and most resistors were not perfect values. If anything, I suppose this proves the rest of the amp's circuitry is ok and the problem is somewhere on the power amp pcb.

I will order new, proper parts and build a new replacement power amp pcb. The original is too beat from multiple repair attempts.
Since the original pcb was a separate board mounted to the heat sink all by itself, I won't have to hack the power supply pcb or the heat sink.

After sufficiently testing the repaired amp I'll let you guys know how it all held up.

[J M Fahey]

Since the original pcb was a separate board mounted to the heat sink all by itself,

That´s great  :dbtu:
Post some pictures.

[JBro]

Here ya go.
Side-by-side of the test circuit and the original pcb... the green wire on the original is a repair.
Thanks for the encouragement J M!