HIP4080 based amp issues with DC offset (is it ?)

Okay, if i'm getting this right. Because this is full bridge class D amplifier, there is voltage on the - and + at the output terminals. And the sinewave swings in between 0v and rail voltage.
Both + and - speaker terminals sits at the middle of 0v and rail voltage. So if rail voltage is 60v, they will be at 30v.

Referenced to the main ground (GND), using 12v of B+input, there is +28v at +speaker terminal and +27,4v at the -speaker terminal. So there is ~ 0.6v offset.

Once I short pin 1 & 2 at the NE5532 there is:
37v at the +speaker terminal
19v at the -speaker terminal
So total of 18v DC offset between the + and - speaker terminals

I think the problem is that amp starts up with full rail voltage at + speaker terminal and sinks down to 28v, while the - speaker terminal starts at 0v and rises up to 27.4v. This causes a big issue if there is load at the speaker outputs and it takes like 15-20seconds to stabilize....of course this doesn't happen when I hook up a load, due to the fact this induces high current draw and crazy stuff at the output of the amplifier.

This equilibrium state at the speaker output + and - should happen faster, before even relay clicks in.
 
Last edited:
They are in order: 5-6-7
 

Attachments

  • 5.jpg
    5.jpg
    87.5 KB · Views: 113
  • 6.jpg
    6.jpg
    108.2 KB · Views: 113
  • 7.jpg
    7.jpg
    109.8 KB · Views: 111
Yes I do. I've seen test of this scope, reads pretty accurate to about a few Mhz.

Amps works just fine if i just wait 15-20seconds after relay clicks.

I have 4 pcs of IRF3710Z...this weekend i'm going to pull all of the B31N20D and replace them with the IRF3710Z (one per bank).

If this won't help, i'll order new hip4080 and replace it with ne5532 together
 
There are some inconsistencies. Previously, the duty cycle was 48%+. Here, it's 38%.

Using a multimeter and measuring the DC voltage from pin 4 of the op-amp to pin 6 and pin 7, when the voltage across the speaker terminals drops to (essentially) 0v, what is the voltage on those two pins?

I'm sorry to tell you but the scope isn't accurate, even at 65kHz. The waveform on pin 5 is going to be very nearly a perfectly shaped square wave. If the signal on the output FETs was as shown, the FETs would overheat and fail.

The difference in the waveforms from one bank to the other is due to an offset in voltage between the DC on pins 6 and 7.
 
DC offset goes as low as ~ 0.6v, does not drop to totally 0.
DMM for DC measurements, black probe (GND) on pin 4 of the OP amp red on pins 6 and 7.

NE5532:
pin 6 - 7v
pin 7 - 6.45v

HIP4080:
DMM black probe (GND) on pin 4 of the OP amp red on pins 6 and 7.
pin 6 - 5.9v
pin 7 - 6.05v

If I short pin 1 and 2 at the op amp ne5532 same measurements with (18v DC offset at the speaker terminals)

ne5532:
pin 6 - 7.2v
pin 7 - 6.9v

hip4080:
pin 6 - 5.9v
pin 7 - 6.45v
 
6/7 of the 5532 aren't important with pins 1/2 shorted.

Does the waveform on terminal 6 of the 4080 immediately go to 5.9 and settle there?

What's the DC voltage on the equivalent of (TeO) ZD1C?

If you remove TeO C20A and the 1/2 jumper, do you still have the initial offset problem on the output?
 
Does the waveform on terminal 6 of the 4080 immediately go to 5.9 and settle there?
- Yes, it does immediately
What's the DC voltage on the equivalent of (TeO) ZD1C?
- 11.35v (i'm not sure this has to be that high, this amp also uses lm393)
If you remove TeO C20A and the 1/2 jumper, do you still have the initial offset problem on the output?
- Yes sir, still the same.
 
Pin 7 from 393 goes through two 2k resistors to pin 4 of CD4013BE which datasheet says it's reset 1.
Also pin 1 from 393 is in direct short with the same pin 4 of the CD4013BE
 

Attachments

  • 83793037_3337845736230760_995387107826466816_n.jpg
    83793037_3337845736230760_995387107826466816_n.jpg
    132.9 KB · Views: 95
Last edited: