1
\$\begingroup\$

I'm currently working on a buffered voltage divider, followed by a Push-Pull stage to be able to source/sink up to 500mA per rail.

The whole thing will is powered by 36VDC, then it's split by half through a 10k voltage divider, followed by a NE5532 as voltage follower to stabilize the new center. This thing then runs through a push-pull stage using a complementary NPN/PNP package which then feeds into two voltage regulators, 7815 and 7915 for a +/-15VDC around the voltage divider reference.

My problem now is: For some reason my push/pull package blows violently when I apply input voltage. The circuit works fine without it but the poor NE5532 is running really hot and is getting unstable under a bit of load.

Could somebody help me to find the reason for that behaviour? What am I missing here?

Link to dataheet of used NPN/PNP package: ZXTD4591E6 NPN/PNP SOT23

schematic

\$\endgroup\$
2
  • \$\begingroup\$ Did it work before you added the regulators? \$\endgroup\$
    – Miron
    Commented Mar 18, 2022 at 16:43
  • \$\begingroup\$ That's one thing I didn't try yet... I'll do that and report back! \$\endgroup\$
    – Khanorr
    Commented Mar 18, 2022 at 16:47

2 Answers 2

Reset to default
1
\$\begingroup\$

Your op-amp is sophisticated enough to protect itself. However, your transistors are certainly not and, quite likely, you are (or will be) running them beyond their safe operating area (SOA): -

enter image description here

For example, if the upper transistor is dropping 15 volts (and it's highly likely to be more) from the upper input supply (in order to regulate), then the peak DC current that the upper transistor can pass before exceeding its SOA is about 32 mA. I've shown that as a red dot on the graph above.

So, I just made up this example but you can do the math yourself. Look at the graph above and place your own numbers on it.

\$\endgroup\$
1
  • \$\begingroup\$ Your're absolutely right! Vce in my case will be hovering at around 18V (or -18V respectively). That is one thing I didn't realise (and didn't know that this is a thing to be honest..) while selecting components for this project. I think I got misdirected by the Ic continous collector current figure of 1A stated in the datasheet. That also explains the rather explosive deconstruction of the package. Thanks! \$\endgroup\$
    – Khanorr
    Commented Mar 18, 2022 at 17:43
1
\$\begingroup\$

In addition to exceeding SOA, the op amp will never be stable in this configuration with all the capacitive loading. You need to compensate the feedback loop for capacitive loading. As the op-amp oscillates, it drives AC signal from the power stage into a short (the capacitors!), and destroys the transistors with ease. The capacitors don’t like this either. And most importantly, look at the voltages! Use an oscilloscope and see what is happening. You can even single-trigger so that the destruction will be documented. Any digital scope nowadays will do that job.

You also don’t show any bulk capacitors, and the regulators need a bit more than 1uF capacitors. They’ll work with those but these capacitors miss the point. Those are fairly slow regulators and such tiny capacitors don’t do enough. Parallel them with at least 47uF or more.

\$\endgroup\$
4
  • \$\begingroup\$ I followed the recommended values in the datasheets for the caps in that case. You're right about the op amp not being stable, I have provisions on board to compensate for that, it's currently just jumpered to see what's going on. Without the NPN/PNP array it was surprinsgly stable(ish), only the negative regualtor was inducing oscillations, especially without load. \$\endgroup\$
    – Khanorr
    Commented Mar 18, 2022 at 17:49
  • 1
    \$\begingroup\$ @devnull Yes, I did think about that but they are very exotic components. This is (sooner or later) going to replace a now obsolete power supply element which is driving an analog device (audio switching/volume) and is designed around a pretty clean +/-15V supply. That's also the reason why I planned to use linear regulators instead of buck/boost stuff to minimize noise coupling into the audio paths. There's also an integrated amp that draws around 220mA under full sine load on each supply rail, so charge pumps and such are also not an option. \$\endgroup\$
    – Khanorr
    Commented Mar 18, 2022 at 18:11
  • \$\begingroup\$ @Khanorr If the power supply is going to be replaced, could the transformer also be replaced/augmented with one with a centre-tapped secondary (or two secondary windings)? \$\endgroup\$
    – Andrew Morton
    Commented Mar 18, 2022 at 18:53
  • \$\begingroup\$ As far as noise reduction: “linear” supplies are a bit of a lie. They are switching: that’s what the rectifier does. And you’ll still have the nasty 100/120Hz harmonic-rich switching noise unless you make the switching fast. Regular rectifiers suck for that. Use the fastest rectifiers you can get your hands on. They are a good upgrade on older lab supplies too. Make them much quieter. Why fast? So they switch while the current is as close to zero as possible. Or use mosfets for an active-diode bridge. Ie.: old designs aren’t necessarily all good. New parts can make them quite a bit better. \$\endgroup\$
    – Kuba hasn't forgotten Monica
    Commented Mar 18, 2022 at 20:49

Not the answer you're looking for? Browse other questions tagged or ask your own question.