I would like to kwon if resistors in ps rails of an opa2134 (which make a voltage drop of about 0.8V) can decrease dynamic and audio performance in general. I'm considering to remove these two resistors to eliminate these voltage drop and get the ps attached directly to the power pins of opa2134 (with 47µF ZA and 0.1µF decoupling). Do you think i'm right ? thx.
Resistors in the power supply legs generally provide some de-coupling of the power supply noise. De-coupling caps with or without the resistors can only help performance. The OP amp in question is already short circuit proof.
I would suppose, that the effects of the resistors would be weighted heavily by the load in which the op-amp is driving.
Feel free to experiment.
I would suppose, that the effects of the resistors would be weighted heavily by the load in which the op-amp is driving.
Feel free to experiment.
I am not an authority on this, but I would wager to say the Resistor combined with the caps emulate an inductor to some degree... an inductor is a capactive resistor, or a resistive capacitor if you want to look at it like that...
Now an inductor is great for class-A amps as it combats ripples and small fluctuations... trying to drive a near constant load even with no signal...
As for class AB and most of the designs you are likely to come up with, for that chip, a more dynamic PSU rail is probably more responsive.
BUT, is that worth increased ripple doing without?
Would you care to describe your PSU,... mains or battery driven etc... for a battery supply they would be pointless.
PS. This post belongs in the chipamp section...
Now an inductor is great for class-A amps as it combats ripples and small fluctuations... trying to drive a near constant load even with no signal...
As for class AB and most of the designs you are likely to come up with, for that chip, a more dynamic PSU rail is probably more responsive.
BUT, is that worth increased ripple doing without?
Would you care to describe your PSU,... mains or battery driven etc... for a battery supply they would be pointless.
PS. This post belongs in the chipamp section...
A step ahead, the resistor is between two caps, a 22µF ZA and a 47µF ZA so it's forming a CRC filter with 0.8V drop at R (100ohms) leads because of the 8mA requested by opa2134. In normal conditions (Z5500 dac sym output buffer) the I load is almost null, so there is no voltage variation in PS rails. But if you plug a head phone (low impedance 32ohms), the I load is not null anymore and PS rails swings.
Hi,
don't send this to Chipamps.
They have fewer resources over there.
Almost all opamps operate in ClassAB.
Some builders modify them to run in single ended ClassA, but that is a different solution.
The supply current on infinite load and zero signal is near enough constant.
As the load draws AC current the supply rails source or sink the extra current to feed the load.
If the output load were 10k and the peak output voltage were 10Vpk then the maximum output current would be 1mApk.
the supply rails would now have a varying current.
each would be Iq+-Ioutpk~=8+-1mA so one has rail currents varying from 7 to 9mA.
The variation is met by the last cap on the supply pins.
The average (about 8.35mA) is fed through the resistor.
Now, look at your headphone driving problem.
Let's set the load voltage to just 100mV and the load impedance to 32ohms.
Vpk=140mVpk and Ipk=4.4mA
Now the supply rails are varying around 8mA+-4.4mA. i.e.3.6mA to 12.4mA. Again the last cap tries to meet this varying load.
But look up the spec for a 2134, where does it say it is suitable for driving low impedance headphones? It's more sensible limit is driving 600ohm loads.
As an example, increase the output voltage to the headphones to 300mV. What are the limits of the supply rail currents?
What if a peak (transient) voltage of 800mVac (Vpk=1.13Vpk) is sent to 32ohms load?
don't send this to Chipamps.
They have fewer resources over there.
Almost all opamps operate in ClassAB.
Some builders modify them to run in single ended ClassA, but that is a different solution.
The supply current on infinite load and zero signal is near enough constant.
As the load draws AC current the supply rails source or sink the extra current to feed the load.
If the output load were 10k and the peak output voltage were 10Vpk then the maximum output current would be 1mApk.
the supply rails would now have a varying current.
each would be Iq+-Ioutpk~=8+-1mA so one has rail currents varying from 7 to 9mA.
The variation is met by the last cap on the supply pins.
The average (about 8.35mA) is fed through the resistor.
Now, look at your headphone driving problem.
Let's set the load voltage to just 100mV and the load impedance to 32ohms.
Vpk=140mVpk and Ipk=4.4mA
Now the supply rails are varying around 8mA+-4.4mA. i.e.3.6mA to 12.4mA. Again the last cap tries to meet this varying load.
But look up the spec for a 2134, where does it say it is suitable for driving low impedance headphones? It's more sensible limit is driving 600ohm loads.
As an example, increase the output voltage to the headphones to 300mV. What are the limits of the supply rail currents?
What if a peak (transient) voltage of 800mVac (Vpk=1.13Vpk) is sent to 32ohms load?
I make a few use of the head phone (a cheap sony) and there's no variation without it so no problem after all. The question was there is any sound degradation in normal condition (without headphone). But since there is no variation i think not so much and i will leave resistors in place. Thanks for your advices.
Hi Max,
what is the maximum voltage each 2134 is asked to deliver?
What is the minimum load impedance each 2134 is asked to feed?
What is the maximum peak current that each 2134 is asked to deliver?
Can the last cap supply this current with little voltage modulation of the chip supply pins?
Vripple=630000*Ipk/uF/f (Ipk in Amps, C in uF, f=modulation frequency) 100Hz @ 1mA > 130mV ripple. Have I got this formula correct? Can anyone confirm?
Now subtract the PSRR for that frequency.
Do the calculations for your loadings. Then decide if your 2134 can perform adequately.
what is the maximum voltage each 2134 is asked to deliver?
What is the minimum load impedance each 2134 is asked to feed?
What is the maximum peak current that each 2134 is asked to deliver?
Can the last cap supply this current with little voltage modulation of the chip supply pins?
Vripple=630000*Ipk/uF/f (Ipk in Amps, C in uF, f=modulation frequency) 100Hz @ 1mA > 130mV ripple. Have I got this formula correct? Can anyone confirm?
Now subtract the PSRR for that frequency.
Do the calculations for your loadings. Then decide if your 2134 can perform adequately.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.