Hey, I have built a preamp from the INA217 datasheet sample schematics (https://www.ti.com/lit/ds/symlink/ina217.pdf, section 8.2, attaching a screenshot to this post).
There is something I don't understand about C1,C2 capacitors.
Why are they so huge? From the hi-pass filter perspective it would make sense to have them 10 times smaller (like 4.7uf, given the following 2.2K resistor). From the phantom power filter perspective there is another 47uf capacitor near the phantom power switch.
The second part I don't understand about the same capacitors is the requirement: «Use non-polar capacitors if phantom power is to be
turned off». What will go wrong if I use polarized capacitors and will switch the phantom power off?
Next thing is that I want to modify that preamp to add an isolation transformer (I have a 10K:10K transformer). Using my mad photoshop skills I've painted the planned schematics (it's not marked on the screenshot, but the phantom power and XLR grounds are planned to be isolated from the grounds after the transformer).
So, the question about the same C1,C2: do I need them given the transformer (which is supposed to block DC from the power supply)? And the second question: is there something fishy with my modification?
Thanks.
There is something I don't understand about C1,C2 capacitors.
Why are they so huge? From the hi-pass filter perspective it would make sense to have them 10 times smaller (like 4.7uf, given the following 2.2K resistor). From the phantom power filter perspective there is another 47uf capacitor near the phantom power switch.
The second part I don't understand about the same capacitors is the requirement: «Use non-polar capacitors if phantom power is to be
turned off». What will go wrong if I use polarized capacitors and will switch the phantom power off?
Next thing is that I want to modify that preamp to add an isolation transformer (I have a 10K:10K transformer). Using my mad photoshop skills I've painted the planned schematics (it's not marked on the screenshot, but the phantom power and XLR grounds are planned to be isolated from the grounds after the transformer).
So, the question about the same C1,C2: do I need them given the transformer (which is supposed to block DC from the power supply)? And the second question: is there something fishy with my modification?
Thanks.
Attachments
Isolation. One of the usages is to connect the output of the guitar pedals chain (not just the mic).
The question is about why. I've only recently started to dig deeper into the circuits and my current level of knowledge doesn't answer the above questions about C1 and C2. Not just yes or no, but also why.
Could they be smaller than 47uF? Like 10uF (given the 10K resistor in front of the transformer)? One of the key questions is why are they 47uF in the first place?
They can be any size that you prefer, although this will affect the low frequency bandwidth.
As it is, the LF bandwidth of the input circuit is about 0.7Hz. This will vary inversely with the values
of the input capacitors. Using instead 4.7uF parts will give 7Hz LF bandwidth. However, the second
pair of capacitors dominates the LF bandwidth, since they give about 15Hz LF bandwidth.
Generally, electrolytic DC blocking capacitors (in series with the signal) are substantially oversized
in order to reduce their measured distortion. Nonpolar electrolytic capacitors have somewhat lower
inherent distortion due to their electrical symmetry, which reduces the even harmonics.
As it is, the LF bandwidth of the input circuit is about 0.7Hz. This will vary inversely with the values
of the input capacitors. Using instead 4.7uF parts will give 7Hz LF bandwidth. However, the second
pair of capacitors dominates the LF bandwidth, since they give about 15Hz LF bandwidth.
Generally, electrolytic DC blocking capacitors (in series with the signal) are substantially oversized
in order to reduce their measured distortion. Nonpolar electrolytic capacitors have somewhat lower
inherent distortion due to their electrical symmetry, which reduces the even harmonics.
Last edited:
Without the transformer, the polarized electrolytic capacitors could see reverse voltage,
depending on the input signal, which is not allowed.
With the transformer this does not happen, since the two series polarized capacitors then
form a single nonpolar capacitor, so nonpolar types would not be necessary in this case,
although they would be fine to use.
depending on the input signal, which is not allowed.
With the transformer this does not happen, since the two series polarized capacitors then
form a single nonpolar capacitor, so nonpolar types would not be necessary in this case,
although they would be fine to use.
Last edited:
I just cannot find a surface mount 47uF nonpolar capacitor with more than 50V rating.
So, if I take my schematics with the transformer, use polar 47uF before and polar 47uF after, it should work? Should I flip the capacitor after transformer (so that + goes to INA217 and - goes to the transformer)? And in that case, should the capacitors after the transformer also be 60V or some smaller voltage would do?
So, if I take my schematics with the transformer, use polar 47uF before and polar 47uF after, it should work? Should I flip the capacitor after transformer (so that + goes to INA217 and - goes to the transformer)? And in that case, should the capacitors after the transformer also be 60V or some smaller voltage would do?
When using the transformer, polar electrolytic capacitors will work both before and after the transformer.
Since the second pair of capacitors are electrically in series (like the first pair), they should be connected
like the first pair of capacitors. That is, both of the + terminals should be pointing in the same direction on
the schematic, though which direction does not matter in this case. A rating of 35VDC each would be enough.
Since the second pair of capacitors are electrically in series (like the first pair), they should be connected
like the first pair of capacitors. That is, both of the + terminals should be pointing in the same direction on
the schematic, though which direction does not matter in this case. A rating of 35VDC each would be enough.
Last edited:
Does it really have to be surface mount? If so, I did come across Nichicon UUN1H470MNQ1ZD.
Otherwise, I would go with the Nichicon Muse ES which are known to have excellent measured performance and are widely available.
I think I've got it. And I suppose I will need to make them even bigger (like 68uF or more) to get the same bandwidth as one 47uF does?
Thanks!
Well, not 100% necessarily, it would just be more convenient. Thanks for the tip.
One more thing: if I use nonpolar capacitors before the transformer and omit the ones after — should it be the same?
Capacitors are needed both before and after the transformer, regardless of whether they are polar or not.
The input and output circuits are isolated by the transformer. Each location requires capacitors because of
different reasons, since the circuits are different. A value of 47uF for all four capacitors will be fine.
The input and output circuits are isolated by the transformer. Each location requires capacitors because of
different reasons, since the circuits are different. A value of 47uF for all four capacitors will be fine.
Last edited:
You paid a lot (a whole dollar) for BIG input transistors in this chip. Why? For very lowest hiss.
But these are BJT devices. (Lo-low hiss JFET would be more costly.)
The Base Current in the source transducer AND the capacitors adds hiss. We want the input circuit very low impedance. As designed, below 100 Ohms.
10uFd will be 100r impedance at 160Hz. You actually plan 5uFd (two 10uFd series), so 320Hz. Hiss will rise below middle C (so you call it random rumble).
I have used systems this "bad". In fact many rooms rumble so bad that 1/f noise in the preamp is moot. However THAT Corp obviously wants to show their parts in best light.
Common mode DC is OK, differential isn't.. I guess it depends on how symmetrical the load current is on the inputs. It also depends on how tolerant of dc the transformer design is. (Mostly not very, I have learned this to my chagrin in more recent times than I would like to admit.)
If the transformer is internal you can get away with a single capacitor as long as no DC current flows through the primary.
Instead of mucking up the entire mic input performance with a 1:1 transformer that'll be a very bad compromise between mic and line requirements, why not use a passive DI box for the pedal chain? That would provide the required galvanic isolation.
50 kOhm in, 150-200 ohm out is not an unusual spec for one of these. Yes, they'll have the expected ~25 dB loss in voltage, but guitar pedals aren't exactly the last word in noise anyway and you have a super low-noise input, so what?
50 kOhm in, 150-200 ohm out is not an unusual spec for one of these. Yes, they'll have the expected ~25 dB loss in voltage, but guitar pedals aren't exactly the last word in noise anyway and you have a super low-noise input, so what?