Dick - is the unused IC5A input grounded or still attached to the circuit (a low impedance return to ground)? If neither, it could be oscillating causing your problem.
Quickshift - Pin 5 of which IC(s)? Sounds like you may have an issue with that pin's connection or you need a bypass cap on a gain setting resistor (C10, 11, 12, etc. try 100 to 220 pF).
If you are referring to IC3, check continuity between its pins 1 and 5. if none, your buffer may be oscillating. I know you checked the colder joints, but reheat them again and check. Thinking about it, check the continuity of whatever pin 5 you touch to get it quiet and the opposite point in the circuit.
Have you tried it in a shielded case?
Bob
Quickshift - Pin 5 of which IC(s)? Sounds like you may have an issue with that pin's connection or you need a bypass cap on a gain setting resistor (C10, 11, 12, etc. try 100 to 220 pF).
If you are referring to IC3, check continuity between its pins 1 and 5. if none, your buffer may be oscillating. I know you checked the colder joints, but reheat them again and check. Thinking about it, check the continuity of whatever pin 5 you touch to get it quiet and the opposite point in the circuit.
Have you tried it in a shielded case?
Bob
😀 😀 😀 😀 😀
I checked everything and then tried the bypass caps...
220pf caps bypassing R12 and R16 have totally silenced the white noise.
Thank you so much!!!!
YAY!
<< I'm chuffed to bits - can you tell 😀>>
I checked everything and then tried the bypass caps...
220pf caps bypassing R12 and R16 have totally silenced the white noise.
Thank you so much!!!!
YAY!
<< I'm chuffed to bits - can you tell 😀>>
Bob,
I haven't actually put IC5A outside the circuit (yet to come).
At this moment I use the output of IC5A (pin 1), while it's directly driven from IC5B (pin 6/7, R15 is shorted) while R20 is open.
My suspision is however that the oscillation is very much in the LP part. i filter at 3 kHz and the tweeter will not give the sort of sound level I get now.
Or can an oscillation in the HP part leak through to LP?
At another measurement I did tonight I got this reponse:
with the PC-signal to output of poweramp (off) and the filter open input, output at LP-filter, no connection to amp.
This looks very much like an oscillation. Can't reproduce it however.
I haven't actually put IC5A outside the circuit (yet to come).
At this moment I use the output of IC5A (pin 1), while it's directly driven from IC5B (pin 6/7, R15 is shorted) while R20 is open.
My suspision is however that the oscillation is very much in the LP part. i filter at 3 kHz and the tweeter will not give the sort of sound level I get now.
Or can an oscillation in the HP part leak through to LP?
At another measurement I did tonight I got this reponse:
with the PC-signal to output of poweramp (off) and the filter open input, output at LP-filter, no connection to amp.
This looks very much like an oscillation. Can't reproduce it however.
Attachments
I attach the schematic.
Both LP and HP are directly fed form IC3A.
LP section 1 is used to form a baffle step with a 2nd order filter and feed forward of the input signal to obtain -6dB levelling.
LP section 2 is severely changed mainly by using IC8A as a 2nd order LP filter:
Pin 6 and 7 are connected
Pin 6 is disconnected form R30
C22 is disconnected from R30 and pin 7
C22 is connected to pin 1
Pin 3 is discoonected from C22 and R30
R37 is filled by a capacitor (CR37)
an extra resistor is placed between C22/R30 and pin3/CR37
Took me quite a while to figure it out but it works (when not oscillating). I can't post the original schematic with the changes (haven't got a scanner).
Both LP and HP are directly fed form IC3A.
LP section 1 is used to form a baffle step with a 2nd order filter and feed forward of the input signal to obtain -6dB levelling.
LP section 2 is severely changed mainly by using IC8A as a 2nd order LP filter:
Pin 6 and 7 are connected
Pin 6 is disconnected form R30
C22 is disconnected from R30 and pin 7
C22 is connected to pin 1
Pin 3 is discoonected from C22 and R30
R37 is filled by a capacitor (CR37)
an extra resistor is placed between C22/R30 and pin3/CR37
Took me quite a while to figure it out but it works (when not oscillating). I can't post the original schematic with the changes (haven't got a scanner).
Attachments
I may have solved my problem.
A superfluous grounding wire.
Never knew this could lead to oscillation.
Listening is however still awkward, as I am always waiting for 'the moment'.
Thanks for all your suggestions.
A superfluous grounding wire.
Never knew this could lead to oscillation.
Listening is however still awkward, as I am always waiting for 'the moment'.
Thanks for all your suggestions.
dick van nierop said:
Dick,
Did you confirm your findings - and is the filter now working 100%?
\Jens
Hi Jens,
The problem hasn't occured since.
It's been playing for about 1 hour without problems.
So I suppose it is solved now.
I'm now building the 2nd board.
Thanks again.
By the way: a friend of mine has also wants to go active.
I bought 5 boards from 'Jim' and a lot of OPA's.
So I'll be designing again for him. Fun!
In order to be able to change resistors easily I plan to solder them not directly to the board but to some pins sticking out of the board.
A bit like your filter 2 but on the GB board.
Changing f and Q can to a large extend be done by changing resistors only, as you undoubtedly know.
By the way nr.2: in the filters I have designed so far it was not necessary to change gain and Q. I always have a gain of 1 and simply use the last opamp as output buffer. Gain can also be set with input or output buffer. Seems to work OK so far. I thought of designing a board with Eagle that would get rid of the buffers but am afraid of the time to master the program.
The problem hasn't occured since.
It's been playing for about 1 hour without problems.
So I suppose it is solved now.
I'm now building the 2nd board.
Thanks again.
By the way: a friend of mine has also wants to go active.
I bought 5 boards from 'Jim' and a lot of OPA's.
So I'll be designing again for him. Fun!
In order to be able to change resistors easily I plan to solder them not directly to the board but to some pins sticking out of the board.
A bit like your filter 2 but on the GB board.
Changing f and Q can to a large extend be done by changing resistors only, as you undoubtedly know.
By the way nr.2: in the filters I have designed so far it was not necessary to change gain and Q. I always have a gain of 1 and simply use the last opamp as output buffer. Gain can also be set with input or output buffer. Seems to work OK so far. I thought of designing a board with Eagle that would get rid of the buffers but am afraid of the time to master the program.
I’ve been following the LM4562 GB tread with great interest; apparently this opamp is a small miracle. Now I’m wondering if it will benefit my active filter build on this GB boards.
LM4562 GB: http://www.diyaudio.com/forums/showthread.php?threadid=91574
I’m using my active filter for a dipole system (a la Linkwitz), comprising 3 x 24dB LP/HP, 2 x shelving, 3 x notch and 4 X buffer, all in all 32 opamps. Currently I’m using OPA2134 (from the OPA2134 GB) – besides the filter is working perfect. Drivers are: Seas H1212 (alu tweeter), Seas 1224 (alu mid) and 2 x peerless SLS 12”.
– besides the filter is working perfect, thanks Jens/Bob.
As the replacement will be relatively expensive I’m looking for some advice beforehand. So please – will the replacement likely give a noteworthy improvement in sound quality?
LM4562 GB: http://www.diyaudio.com/forums/showthread.php?threadid=91574
I’m using my active filter for a dipole system (a la Linkwitz), comprising 3 x 24dB LP/HP, 2 x shelving, 3 x notch and 4 X buffer, all in all 32 opamps. Currently I’m using OPA2134 (from the OPA2134 GB) – besides the filter is working perfect. Drivers are: Seas H1212 (alu tweeter), Seas 1224 (alu mid) and 2 x peerless SLS 12”.
– besides the filter is working perfect, thanks Jens/Bob.
As the replacement will be relatively expensive I’m looking for some advice beforehand. So please – will the replacement likely give a noteworthy improvement in sound quality?
Sure hope it's good, I'll be using the LM4562 on my filters. I'm going for a simple 2-way setup with only 10 opamps, though, so the cost isn't quite the same.
That said, I think Dennis might do another GB run if the interest is there.
That said, I think Dennis might do another GB run if the interest is there.
Lars - judging from the community response regarding the LM4562 it's an excellent choice. Remember to consider the all-pass filter (delay), is a powerful attribute of the board.
Good luck with the project.
Good luck with the project.
Wanted to Buy!
Hi all,
I understand that Bob has no more of the active filter boards available.
Does anyone out there have any they would want to sell???
Be honest. If you bought them and know you will never get around to building them, please, please, please let me put them to good use.
Thanks very much,
Steve
Hi all,
I understand that Bob has no more of the active filter boards available.
Does anyone out there have any they would want to sell???
Be honest. If you bought them and know you will never get around to building them, please, please, please let me put them to good use.
Thanks very much,
Steve
Hi Steve,
I have some left. It would be cheaper to buy some from an american source, but if nobody want to sell then you can claim some from me. (cost+mail+paypal)
Greetings, Harry
I have some left. It would be cheaper to buy some from an american source, but if nobody want to sell then you can claim some from me. (cost+mail+paypal)
Greetings, Harry
Refering to my post 650.
Pleas - doesn’t anyone have some experiences with the effect of different opamps in active xo they want to share? The advice doesn’t need to be specifically on the LM4562 vs. the OPA 2132
Tanks
Pleas - doesn’t anyone have some experiences with the effect of different opamps in active xo they want to share? The advice doesn’t need to be specifically on the LM4562 vs. the OPA 2132
Tanks
Avi reported that he needed to pay close attention to PS bypassing with LM6172. The 6172 is a bit faster than the 4562 so you **may** have no issues on this board. Just be aware that these really fast chips might oscillate.
Thanks – I will be careful with those fast opamps.
Right now I’m looking into an alternative approach (not so expensive). I’m trying - by the help of Spice - to locate the most sensitive sections. I’m looking for high currence, hi freq feed back signal etc. E.g. my all-pass filters are candidates for better opamps, because of relative high current. The all-pass opampes are driving an 82nF through 681ohm resulting in approximately 1.3mA.
Any other suggestion, about what to look after?
Right now I’m looking into an alternative approach (not so expensive). I’m trying - by the help of Spice - to locate the most sensitive sections. I’m looking for high currence, hi freq feed back signal etc. E.g. my all-pass filters are candidates for better opamps, because of relative high current. The all-pass opampes are driving an 82nF through 681ohm resulting in approximately 1.3mA.
Any other suggestion, about what to look after?
Boards...
Is there anyone with a couple of these boards in spare they could be willing to sell..??
Lyra
Is there anyone with a couple of these boards in spare they could be willing to sell..??
Lyra
I am a bit puzzled by the balanced input.
My setup is R11 = R12 = R13 = R16 = 10K, C9 = C10 = 100pF. Tested with both LM6172IN and NE5532AP.
I hooked up the filter to my workshop computer soundcard. Played a 1000Hz sinus. Measured with an oscilloscope(got it when a local repair workshop got out of business, but my experience is very limited).
I did the following measurements:
1) directly connected to the computer soundcard. Adjusting the oscilloscope so that the sinus curve filled the whole height of the screen.
2) connected computers soundcard to In+, measuring at IC3A out. Now the sinus curve filled only half the height of the screen.
3) connected computers soundcard to In-, measuring at IC3A out. Now the sinus curve filled the whole height of the screen.
What I don't understand is why the curves have different magnitude.
I did the test on two different filters, one with LM6172 and one with NE5532 with the exact same results.
Any ideas?
My setup is R11 = R12 = R13 = R16 = 10K, C9 = C10 = 100pF. Tested with both LM6172IN and NE5532AP.
I hooked up the filter to my workshop computer soundcard. Played a 1000Hz sinus. Measured with an oscilloscope(got it when a local repair workshop got out of business, but my experience is very limited).
I did the following measurements:
1) directly connected to the computer soundcard. Adjusting the oscilloscope so that the sinus curve filled the whole height of the screen.
2) connected computers soundcard to In+, measuring at IC3A out. Now the sinus curve filled only half the height of the screen.
3) connected computers soundcard to In-, measuring at IC3A out. Now the sinus curve filled the whole height of the screen.
What I don't understand is why the curves have different magnitude.
I did the test on two different filters, one with LM6172 and one with NE5532 with the exact same results.
Any ideas?
- Status
- Not open for further replies.