Hi Andrew,
Unfortunately, there is not an errata page on my website. It is something I should have done all along as I received feedback.
Cheers,
Bob
Hi Bob
I am looking at the gain clone circuit in Chapter 27 which improves on the basic LM3886 circuit by adding the opa2604 op-amps to reduce offset voltage and remove capacitor from signal path. I tried replacing the opa2604 by low voltage offset opa192's but the DC offset at output seems to get worse from 6mV to around 270mv. I tried changing capacitor from 1u to 10u and 10k/10k resistor pair in second stage with 1k/1k pairs too without much effect. The phase response also seems to be dergraded compared to circuit with opa2604. What am I missing ?
Thanks
DC
I am looking at the gain clone circuit in Chapter 27 which improves on the basic LM3886 circuit by adding the opa2604 op-amps to reduce offset voltage and remove capacitor from signal path. I tried replacing the opa2604 by low voltage offset opa192's but the DC offset at output seems to get worse from 6mV to around 270mv. I tried changing capacitor from 1u to 10u and 10k/10k resistor pair in second stage with 1k/1k pairs too without much effect. The phase response also seems to be dergraded compared to circuit with opa2604. What am I missing ?
Thanks
DC
Hi DC,
I don't have any experience with the OPA192, but there may be a chance that there is oscillation in the circuit. Just a hunch. It may have something to do with rail-to-rail outputs behaving a bit differently at HF. Did you put a scope on it?
More importantly, why did you want to change out the OPA2604 with the OPA192?
Cheers,
Bob
Hi Bob
I was just experimenting with the OPA192 since it seems to be designed specifically for an extremely low DC output offset. BTW this is just in Ltspice, i havent built the physical circuit yet but I hope to some day when I understand completely why each of the components are exactly that way.
DC
I was just experimenting with the OPA192 since it seems to be designed specifically for an extremely low DC output offset. BTW this is just in Ltspice, i havent built the physical circuit yet but I hope to some day when I understand completely why each of the components are exactly that way.
DC
Hi DC,
For most audio applications, including this one, dc offset and rail-rail swing are not priorities. Rather, sound quality is the priority. I don't know anything about the SQ of the OPA192. The device I specified is known to be well-regarded, so I would stick with that one.
Cheers,
Bob
Any dc offset less than 20mV at the output of a power amplifier should be certainly inaudible. Bear in mind that dc offset at the output of a power amplifier affects only the woofer, since the crossovers block any dc from getting to the mid or tweeter.
cheers,
Bob
Good point.
We could start with the fact that 20mV will cause only about 3mA to flow in an 8 ohm speaker. This corresponds to 60 micro-watts peak in a corresponding sinewave, or 30uW rms. This is 90dB below 1 watt. If the speaker sensitivity is 90dB, we get 0dB SPL ar 1 meter. We might not have to go any further than this realization.
Cheers,
Bob
This is something I've often wondered over.
Simply put, would say 10mv pk-pk of white noise applied to a speaker produce identical audio output from that speaker cone if that white noise could be superimposed onto a DC offset. Would there be any compression, any amplitude variation ? I have no idea... but I've often wondered.
I think the linearity becomes worse with higher excursions so yes, the signal superimposed on a DC would be distorted. How much - depends
Jan
If you displace the coil with DC, there comes a point where non-linearity and distortion become noticeable I guess. In the limiting case you would get the voice coil at the edge of the field or magnet, and that certainly will mess things up.
But certainly not with 20mV.
Jan
Since I have designed my speakers with a very long gap and short coil the problem that I see with a dc offset would actually be with an asymmetric loading of the suspension in both the surround and the spider. You would no longer be in a neutral position as far as restorative forces, this would be in my case the distortion cause. So the answer to this question becomes complex as you then introduce those forces plus any asymmetry in a common even hung or overhung voicecoil situation. The greater the dc offset the greater the problems become with any speaker due to these compounded elements.
I think the linearity becomes worse with higher excursions so yes, the signal superimposed on a DC would be distorted. How much - depends
Jan
That is my feeling also but as you say, 'by how much'.
Its easy to dismiss 20mv as being of no consequence and yet we chase down 0.000xxx distortion figures when it comes to circuitry. It wouldn't surprise me if a few 10's of millivolts caused more than 0.000xxx 'difference' in the output from the speaker... but how to prove the effect. We can't.