These capacitors will bring your DC gain down to unity.
If you removed these from the circuit, then with a 1 mV DC offset on your input signal, the amp would amplify it 32 times so you have a DC offset of 32 mV at your output.
So with these capacitors in circuit your gain would be 1 at DC, so 1 mV offset at the input would corolate to a 1 mV offset at your output.
If you removed these from the circuit, then with a 1 mV DC offset on your input signal, the amp would amplify it 32 times so you have a DC offset of 32 mV at your output.
So with these capacitors in circuit your gain would be 1 at DC, so 1 mV offset at the input would corolate to a 1 mV offset at your output.
Bassterror said:Thank you
So if I have an input cap, i do not need them, right?
It depends also, if you have large DC offset at the output put one in, it also may add sability to the amp so it doesn't oscilate.(Anyone care to verify?)
Simple addvice, try it with out and if you have problems add it later.
Hi,
the input cap you require is quite small, 1uF to 10uF, so can be Polypropylene or smaller and cheaper use Polyester.
The NFB cap needs to be a lot bigger and you have to use an electrolytic.
On that basis, I think the NFB cap will have more influence on the sound.
The NFB cap, as said before, reduces the amp DC gain to one (1). but the output offset will vary with temperature.
You may find that the offset will vary with ambient conditions, with signal output level, and with heatsink temperature. The designer may have made a good job of compensating for these and reduced offset drift to neglible proportions, but be careful to check the offset regularly in the early days.
If you fit an input DC blocking cap, remember to put in ground referencing rsistors on BOTH sides of the cap. The one on the RCA input side can be 1M0 to 2M2 to RCA ground. The one after the cap is your input impedance setting resistor and should match the upper leg of the NFB loop i.e. r1=r6.
I note that the PCB does not attempt to balance the impedances seen by the inverting and non-inverting inputs. This unbalance will exaggerate output offset drift.
the input cap you require is quite small, 1uF to 10uF, so can be Polypropylene or smaller and cheaper use Polyester.
The NFB cap needs to be a lot bigger and you have to use an electrolytic.
On that basis, I think the NFB cap will have more influence on the sound.
The NFB cap, as said before, reduces the amp DC gain to one (1). but the output offset will vary with temperature.
You may find that the offset will vary with ambient conditions, with signal output level, and with heatsink temperature. The designer may have made a good job of compensating for these and reduced offset drift to neglible proportions, but be careful to check the offset regularly in the early days.
If you fit an input DC blocking cap, remember to put in ground referencing rsistors on BOTH sides of the cap. The one on the RCA input side can be 1M0 to 2M2 to RCA ground. The one after the cap is your input impedance setting resistor and should match the upper leg of the NFB loop i.e. r1=r6.
I note that the PCB does not attempt to balance the impedances seen by the inverting and non-inverting inputs. This unbalance will exaggerate output offset drift.
I too worry about input offset current varying with temperature. Have you tried bypassing (connecting in parallel) with a "nice" cap (polypropylene, etc) C1/C2? At least the higher frequencies will get a nice shunt path. You could go nuts and blend in a DC Servo into the NFB. Then you could remove the shunt cap.
You could trickle in DC correction current into the NFB. The DC Servo circuit is just a gain circuit that is very LPF'd. You could wire your chip circuit's NFB shunt leg to ground (DC coupled) and take the output of your DC Servo circuit and, through a resistor, connect to the junction of your FB resistor and shunt leg. It will just make two paths for you. An analog path with a DC coupled shunt leg and a DC correction path where there is only gain from DC to, I don't know, .1 Hz. It looks like you already have an etched board, so it may be problematic at this stage.
Mrpong linkjed this on a different thread "About BPA200 schematic": http://www.national.com/appinfo/audio/files/BPA-200_Application_Note.pdf. There's an ap note about getting a DC servo in there.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.