This is a good idea. Not to permanently leave it like that but to rule out whether the power supply is the source of your noise problem.
If there is still noise, then you can look closer at the filter circuit.
I think Andrew is talking about these green ground loops and the other pic shows the close coupling method. Please correct me if I am wrong.
It's pretty funny that the schematic says to leave the ground un-earthed to avoid ground loops and then it shows a bunch of ground loops.
Attachments
YES !!!!
minimise the LOOP AREAS.
They are not ground loops, but they are loops.
Loop = aerial, receiving or transmitting interference.
Always arrange your wiring AND traces to minimise Loop Area.
There are also loops in the NFB components and their traces that need the same care in laying out on the PCB.
minimise the LOOP AREAS.
They are not ground loops, but they are loops.
Loop = aerial, receiving or transmitting interference.
Always arrange your wiring AND traces to minimise Loop Area.
There are also loops in the NFB components and their traces that need the same care in laying out on the PCB.
Last edited:
Ahhhh,
So connections haven't changed, the lines are just physically closer thus eliminating a coil and in turn, any interference.
Though this circuit is on vero board, I will keep this technique in mind.
Now that I understand this part - I would like to apologize for dismissing the suggestion to add buffer caps!
On the pcb of the diy XO, I realize that there are buffer caps already in place (see this pic on lower left of pcb).
Maybe this explains why XO runs clean and hpf doesn't - currently there are no caps subsequent to regulators of my psu, nor prior to hpf (same thing).
I'm told adding buffer to HPF (or subsequent to regs -same thing) should help.
So ... to do this must I buy another op-amp alongside additional caps?
Referring to most recent schematic above - there appears to be an additional op-amp on the left, 'buffer section' (and in schematic shown below).
It's not obvious if I can just use existing op-amp I have in place at heart of hpf circuit. (It's a 2134).
So connections haven't changed, the lines are just physically closer thus eliminating a coil and in turn, any interference.
Though this circuit is on vero board, I will keep this technique in mind.
Now that I understand this part - I would like to apologize for dismissing the suggestion to add buffer caps!
On the pcb of the diy XO, I realize that there are buffer caps already in place (see this pic on lower left of pcb).
Maybe this explains why XO runs clean and hpf doesn't - currently there are no caps subsequent to regulators of my psu, nor prior to hpf (same thing).
I'm told adding buffer to HPF (or subsequent to regs -same thing) should help.
So ... to do this must I buy another op-amp alongside additional caps?
Referring to most recent schematic above - there appears to be an additional op-amp on the left, 'buffer section' (and in schematic shown below).
It's not obvious if I can just use existing op-amp I have in place at heart of hpf circuit. (It's a 2134).
Attachments
Last edited:
that schematic shows the four decoupling caps connecting to the Power Ground.
Remember that the two HF decoupling connect from the Power PINS to each other. This creates the shortest loop.
The two electros connect to the power supply traces near the Power PINS and connect to each other. This creates the MF decoupling which has a slightly bigger loop.
Now connect the two junctions together with a very short wire/trace. This becomes the best place for an output Zobel if it is used.
It has the lowest impedance of any of the Power Ground connections, because it has the shortest cables and the smallest loops. It also benefits from any connection made here in that the two sets of decoupling are effectively in parallel for any current coming to this junction.
Remember that the two HF decoupling connect from the Power PINS to each other. This creates the shortest loop.
The two electros connect to the power supply traces near the Power PINS and connect to each other. This creates the MF decoupling which has a slightly bigger loop.
Now connect the two junctions together with a very short wire/trace. This becomes the best place for an output Zobel if it is used.
It has the lowest impedance of any of the Power Ground connections, because it has the shortest cables and the smallest loops. It also benefits from any connection made here in that the two sets of decoupling are effectively in parallel for any current coming to this junction.
Last edited:
Check the data sheet for your regulator and you'll probably see that they require a 10uF or larger capacitor
physically located right at the regulator input and output.
Alright chaps
As you can see C5-C8 'buffer' added as instructed.
@shreadhead perhaps a mistake in your diagram are polarized caps the right way up?
I've drawn them in the schematic as per your diagram but I think the lower one is up-side down?
Also, I realize HPF is actually dual polarity circuit, perhaps I'm wrong - please confirm.
I thought it was a single polarity at first. So I believe VE+ connects to +15v and VE- connects to -15v, subsequent to new capacitor arrangement, this right?
This and that power common (black triangle) should be isolated from commons in the HPF (white triangles)...
Thanks in advance
As you can see C5-C8 'buffer' added as instructed.
@shreadhead perhaps a mistake in your diagram are polarized caps the right way up?
I've drawn them in the schematic as per your diagram but I think the lower one is up-side down?
Also, I realize HPF is actually dual polarity circuit, perhaps I'm wrong - please confirm.
I thought it was a single polarity at first. So I believe VE+ connects to +15v and VE- connects to -15v, subsequent to new capacitor arrangement, this right?
This and that power common (black triangle) should be isolated from commons in the HPF (white triangles)...
Thanks in advance
Attachments
I've enclosed a pic of the PSU I use these days. You are right about the cap being out of polarity but it is a 100nF which is supposed to be a ceramic non-polarized anyway so it doesn't matter. If that was my fault I am sorry about that. From what I understand it is important for the 100nF and 330nF ceramics to be as close to the chip's pins as possible.
As far as whether you need a buffer on the input of your HPF, I read this on Rod Elliot's site pertaining to filters:
"It is very important to remember that inputs must always be from a low impedance source for any filter"
So probably you'll be ok without an input buffer but this simple buffer (see pic) will ensure that your filter will be fed with a low impedance source.
Also, Rod Elliot usually puts a 100 ohm resistor in series with the output of the last op amp in the circuit. About that he says:
"The 100 ohm output resistor isolates the opamp output from the capacitance of the output cable."
As far as whether you need a buffer on the input of your HPF, I read this on Rod Elliot's site pertaining to filters:
"It is very important to remember that inputs must always be from a low impedance source for any filter"
So probably you'll be ok without an input buffer but this simple buffer (see pic) will ensure that your filter will be fed with a low impedance source.
Also, Rod Elliot usually puts a 100 ohm resistor in series with the output of the last op amp in the circuit. About that he says:
"The 100 ohm output resistor isolates the opamp output from the capacitance of the output cable."
Attachments
Alright
Regarding low impedance source, the HPF will be fed from XO directly.
XO already has a buffer therefore I don't think I will need to add one to hpf.
I will try re-arranging the second 2200uf caps to after the regulator see if this makes a difference, I will try adding regular non-polarized ceramics subsequent to regs if not.
I'll keep this posted.
Thanks
Regarding low impedance source, the HPF will be fed from XO directly.
XO already has a buffer therefore I don't think I will need to add one to hpf.
I will try re-arranging the second 2200uf caps to after the regulator see if this makes a difference, I will try adding regular non-polarized ceramics subsequent to regs if not.
I'll keep this posted.
Thanks
The Low pass starts with a resistor into the active filter.
This resistance is made up from an actual resistor value plus the output impedance of the transmit circuit, be it a buffer or an amplifier.
Unfortunately the high pass starts with a capacitor and requires that the source impedance be a theoretical zero ohms.
I have asked two or three times how to correct the high pass filter to take account of the non zero source impedance. No one has been forthcoming.
Have you found a solution?
This resistance is made up from an actual resistor value plus the output impedance of the transmit circuit, be it a buffer or an amplifier.
Unfortunately the high pass starts with a capacitor and requires that the source impedance be a theoretical zero ohms.
I have asked two or three times how to correct the high pass filter to take account of the non zero source impedance. No one has been forthcoming.
Have you found a solution?
You've lost me.
Signal path will be
Signal Source > 3way crossover (lower band 0-87Hz sent out via respective jack) > into high pass jack > out hpf jack > amplifier.
Do you think think high impedance will be present XO output and enough to cause a problem?
I cannot confirm this myself currently as I don't have the tools.
I can however upload schematic of the XO for analysis.
Signal path will be
Signal Source > 3way crossover (lower band 0-87Hz sent out via respective jack) > into high pass jack > out hpf jack > amplifier.
Do you think think high impedance will be present XO output and enough to cause a problem?
I cannot confirm this myself currently as I don't have the tools.
I can however upload schematic of the XO for analysis.
Alright, here's the xo schematic. Yeh the wire will just be a immediate 10cm or so interconnect.
I see there is indeed resistors on xo's output, so i'll probably add 100ohm to hpf output aswell.
Not sure about why there is a second 11k resistor ( R5b) there. I wonder if this is also essential for hpf..
Thanks.
I see there is indeed resistors on xo's output, so i'll probably add 100ohm to hpf output aswell.
Not sure about why there is a second 11k resistor ( R5b) there. I wonder if this is also essential for hpf..
Thanks.
Attachments
Last edited:
Looks like high and low pass filters only. Yeah man, should all work out then as far as what I can figure. Wire it up! Remember Andrew's advice on the twisted pairs especially to your separate HPF. The two 11k's are just figuring using 1% R's to add to 22K. I guess they are hoping that it will get you closer to 22k than the next closest 1% R value which is 22.1K. Same story on the parallel caps C7&8.
This can't be right.
The treble needs a high pass.
That part is correct.
The bass needs a low pass
That part is correct.
The mid needs a band pass.
That looks wrong to me.
Remember that R19 requires Rout of A7 to be zero ohms.
similarly for R17 & R8.
If not zero ohms then the resistor values fitted must be reduced to account for the actual Rs of the preceding opamp.
I think they're making a bandpass by putting a low pass and high pass in series? I don't know how else it would work.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.