TDA7052A single-ended output

[Mick Bailey]

I came across this schematic for a simple voltage controlled amplifier (VCA) using the TDA7052A - obsolete, but still readily available and cheap. The dynamic range is excellent, but I can't see how it works, being a bridged output IC. The spec sheet says it's capable of SE operation up to 100mA.

I'd be using it to drive a 100k Ohm load, though I'm unclear how it produces a complete waveform rather than just a positive or negative half when the load is taken from a single output.

[x-pro]

This all:

[Helmholtz]

Excuse my ignorance, but what made you think the internal output amplifiers would process only one half-wave each?
I guess each output sits at Vp/2?

[Mick Bailey]

A while back I breadboarded one of these ICs with SE output into a resistive load referenced to ground. Either output alone only produced half the waveform, whereas both in a regular bridged connection worked fine. This has got me thinking now whether I capacitor decoupled the output for SE - maybe I didn't and it affected the missing current limiter. I've just ordered a couple to experiment with - perhaps I can repeat what I did previously and come up with an explanation.

[dmeek]

The two outputs are both the full waveform but out of phase with each other.

Slightly off-topic:
A while back I had a Danelectro Honeytone to repair. It uses the TDA7052 and it was blown. I replaced it with a brand new TDA7052A and it immediately got very hot. Long story short,
the internals of the chip had been changed without clearly or even vaguely stating it in the datasheet. With the early version the input is reference to ground (pin 3 and 6).
The new version has a bias voltage on the input pin 2 so a capacitor is needed.

[Mick Bailey]

That's interesting - the spec sheet says no coupling capacitors are required but I've read quite a few articles saying there's a DC offset on the input. I'd seen something that later ICs also need the unused output balancing using a series R-C to ground if SE output is used to keep the current differential low. I'm thinking that for VCA use this wouldn't be a problem.

There's no schematic for the IC, just a conceptual block diagram. Given that the outputs are biased to half the supply voltage, if the output is not decoupled in SE mode I wonder if this would affect the reference voltage to the output amplifiers (effectively shorting out the reference if the resistance was low enough) and be a reason for the missing waveform half with my breadboarded circuit?

There are lots of these for sale and many are unbranded or have dubious looking markings. As a precaution I've ordered some original Phillips parts that hopefully comply with the spec sheet.

[Helmholtz]

I'd expect a DC load to cause trouble.

[dmeek]

I got my replacements from Digi-Key so I assume they're legit but who knows.
I still have a few left so I did a quick breadboard.
Here is what I get for DC voltages:

VCC =12V

pin 2 open, pin 3 grounded:
pin 2 = 2,5V
outputs = 6V

pin 2 open, pin 3 open:
pin 2 = 3.36V
pin 3 = 2.5V
outputs = .49V

Very strange.

[Helmholtz]

Why strange?

[dmeek]

I guess a wrong assumption on my part. It looked like an op-amp with + and - inputs.

[Helmholtz]

The application circuit in the datasheet shows input and output DC blocking caps. So I'd expect DCV there.

[The Dude]

Agree, especially given that it's designed for use with a single supply.

[Mick Bailey]

I'm not clear why the spec sheet says "TDA7052A/AT the DC volume control stage is integrated into the input stage so that no coupling capacitors are required...."

I think this is what confused me first time round when I'd experimented with this I.C.

My older Phillips parts turned up and I got chance to breadboard the VCA circuit. For processor compatibility I'm running it off 5v and the gain control is taken off the wiper of a 20k ohm trimmer - the control voltage comes from an envelope shaper. These examples I have also have DC on the inputs. The VCA works perfectly and I really like the way it fades to mute with a control voltage that starts at zero. It's especially good for audio since the control voltage acts exponentially on the output. I'm using a 100mV input signal and unity gain, so there's no discernable distortion.

There's one issue I need to track down in that if the CV is taken directly from an opamp buffer there's a heavy thump at the start and end of the envelope and a very distorted output (which is capacitor coupled). I'm seeing a DC transient on the output pin 5 that looks like one half of a low-frequency sine wave. This doesn't happen if the control is taken from a higher impedance source. I'm thinking now that this is maybe what I'd seen in my original breadboard experiment. I can't figure this out, since the wiper of the trimmer is at about 25%, or 5k ohm from ground. How would the source impedance of the CV be affecting it?

EDIT: I've now got this all working fine without the thumps/distortion. I'd omitted the 220uf cap on pin 1, as the supply was already well filtered and regulated. Installing this cap close to pin 1 solved the problem.

[Mick Bailey]

I've been experimenting with this IC and after my initial enthusiasm it became aparrent that using it with a typical envelope generator that charges /discharges a capacitor is flawed blue to the control pin having a log response. With anything other than fast envelopes the attack response results in effectively a delayed signal where there's a distinct gap before the note starts. Similarly, the note decay sounds as though it ends prematurely.

It looks like the way around this would be to use a linear envelope generator, but that negates the simplicity of this VCA.