Hi,
I was hoping someone can tell me the function of rp1 in the attached schematic
I see the tlo74 op amp acts as the servo intergrator, but on the board the rp1 trimpot is maked as dc servo?
Another reason I ask is that dc offset jumps around between -2 mv and -8.5 mv dc on both channels. Is this normal for a dc servo set up. I am used to seeing a stable reading on capacitor coupled amplifiers
I have set dc bias as per manufactures instructions
Cheers
Chilli
I was hoping someone can tell me the function of rp1 in the attached schematic
I see the tlo74 op amp acts as the servo intergrator, but on the board the rp1 trimpot is maked as dc servo?
Another reason I ask is that dc offset jumps around between -2 mv and -8.5 mv dc on both channels. Is this normal for a dc servo set up. I am used to seeing a stable reading on capacitor coupled amplifiers
I have set dc bias as per manufactures instructions
Cheers
Chilli
Short answer: I would not worry about <10mV offset, though "jumps" is odd.
Long answer:
> the rp1 trimpot is maked as dc servo?
This sets the tail-current.
Some designers use a thing around VT5 VT6 which makes exact tail current non-critical; here they didn't. Ideally R7+RP1 should equal R6. In practice it has to be a little off to null device tolerances.
Except the servo will "correct" your attempt to trim RP1.
First check the voltages on the servo amp's pins 8 and 14. Both should be few-Volts at most, NOT 13V (slammed against their power supply). If slammed, suggest a bad joint or part failure and doubt caps C21 and C20 first.
I do NOT know the circuit or philosophy or manufacturer's instructions. But I would think that if "XS3" is easy to jumper-across, I would jump it. Now the servo is disabled. Now you can trim RP1 for lowest raw offset. Then un-jump XS3 to let the servo work.
This is still odd because the TL074 is not a low-offset part so several-mV offsets are likely.
Long answer:
> the rp1 trimpot is maked as dc servo?
This sets the tail-current.
Some designers use a thing around VT5 VT6 which makes exact tail current non-critical; here they didn't. Ideally R7+RP1 should equal R6. In practice it has to be a little off to null device tolerances.
Except the servo will "correct" your attempt to trim RP1.
First check the voltages on the servo amp's pins 8 and 14. Both should be few-Volts at most, NOT 13V (slammed against their power supply). If slammed, suggest a bad joint or part failure and doubt caps C21 and C20 first.
I do NOT know the circuit or philosophy or manufacturer's instructions. But I would think that if "XS3" is easy to jumper-across, I would jump it. Now the servo is disabled. Now you can trim RP1 for lowest raw offset. Then un-jump XS3 to let the servo work.
This is still odd because the TL074 is not a low-offset part so several-mV offsets are likely.
I measured the voltage on pins 8 and 14. Both were around .9vdc
I was able to do this easily , there are two test pins marked XS3.
Anyway it did disable the servo and I was able to get low and stable dc offset just using rp1 !
So why add the servo if there is no benefit of removing coupling capacitors.
Cheers
chilli
> why add the servo if there is no benefit of removing coupling capacitors.
The designer omitted the usual coupling cap at the bottom of the NFB network, series to R22 1K. Input pair offset is amplified same as audio, about 47K/1K or 48 times. Rough-assuming 5mV input offset, this is 250mV offset at the speaker. This is high.
The servo wants to reduce this to the offset of a TL074, nominally 3mV though 15mV over full temperature range. This is plenty low.
The "servo" (not) pot RP1 is an odd design detail. We would like the input pair to work at dead-equal currents for lowest THD (see Self's books). RP1 sets the both-sides tail current. If this is double the current in VT6, then equality happens. There are several ways to do this. A manual trim is one way.
I mis-wrote:
me> Ideally R7+RP1 should equal R6.
Actually (obviously) R7+RP1 should be *half* of R6. If all devices were equal and all Base currents negligible, this would force 2=1+1 current equality. Because parts are not perfect, they gave a +/-10% manual trim. Because the trim will also drift over time (and input abuse) and may not be dead-on zero voltage offset, they use the servo to trim the output error.
The designer omitted the usual coupling cap at the bottom of the NFB network, series to R22 1K. Input pair offset is amplified same as audio, about 47K/1K or 48 times. Rough-assuming 5mV input offset, this is 250mV offset at the speaker. This is high.
The servo wants to reduce this to the offset of a TL074, nominally 3mV though 15mV over full temperature range. This is plenty low.
The "servo" (not) pot RP1 is an odd design detail. We would like the input pair to work at dead-equal currents for lowest THD (see Self's books). RP1 sets the both-sides tail current. If this is double the current in VT6, then equality happens. There are several ways to do this. A manual trim is one way.
I mis-wrote:
me> Ideally R7+RP1 should equal R6.
Actually (obviously) R7+RP1 should be *half* of R6. If all devices were equal and all Base currents negligible, this would force 2=1+1 current equality. Because parts are not perfect, they gave a +/-10% manual trim. Because the trim will also drift over time (and input abuse) and may not be dead-on zero voltage offset, they use the servo to trim the output error.
Hi,
I feel, that this jumping offset resulted by the some LF oscillation. I see triple filtering in the offset. One of them are set around 5Hz, another to 0.7Hz, and the last to 7Hz. So I guess that, there is not enough phase margin.
Sajti
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.