I'm going to try the buffer with LPF in my GC monoblocks.
Easy I thought, I have the buffers already made (previously used for tape buffers in preamp) and a +/-15V PSU for them. All I have to do is put them i the GC between the inputy socket and the input resistor (I use an external attenuator).
But when I actually came to do it I wondered about the grounding arrangement. Should all the signal grounds of the buffer/filter go to their own star and be attached to the power ground star of either the buffer or GC (or both)?
To confuse matters a bit more, I have my signal ground star on the ground terminal of the input socket.
The best solution that I have come up with is to split the buffer/filter, treating the buffer as a separate unit (as it would be in the case of a preamp) and placing the filter between the input of the GC and the 10K input resistor as shown in this diagram.
.
Can anyone give any advice on this arrangement or a better solution please?
Easy I thought, I have the buffers already made (previously used for tape buffers in preamp) and a +/-15V PSU for them. All I have to do is put them i the GC between the inputy socket and the input resistor (I use an external attenuator).
But when I actually came to do it I wondered about the grounding arrangement. Should all the signal grounds of the buffer/filter go to their own star and be attached to the power ground star of either the buffer or GC (or both)?
To confuse matters a bit more, I have my signal ground star on the ground terminal of the input socket.
The best solution that I have come up with is to split the buffer/filter, treating the buffer as a separate unit (as it would be in the case of a preamp) and placing the filter between the input of the GC and the 10K input resistor as shown in this diagram.
.
An externally hosted image should be here but it was not working when we last tested it.
Can anyone give any advice on this arrangement or a better solution please?
Nuuk,
First, you don't need and input and output cap on the buffer.
I think you're using the OPA627, right?😉
You can even remove both caps and measure DC, you may have a surprize.
And I think you don't need that 680 ohm resistor...
As for the grounds, I think there will be no problem if you connect the ground from the buffer to your signal ground point.
A picture would help...
First, you don't need and input and output cap on the buffer.
I think you're using the OPA627, right?😉
You can even remove both caps and measure DC, you may have a surprize.
And I think you don't need that 680 ohm resistor...
As for the grounds, I think there will be no problem if you connect the ground from the buffer to your signal ground point.
A picture would help...
Of course he needs it - he's trying to make a low-pass filter! 😉carlosfm said:
I would leave in the 6u8 cap - I'd also try changing the op-amp decoupling so that there is only the one cap connected between the supply pins. Try it both ways, and check that no oscillations occur...
Cheers,
Mark 😉
Thanks guys,
Sorry Carlos, I don't have a digital camera here but you surely know what an OPA627 looks like 😉
Mark, the 6N8 must stay as it is the other component of the LPF.
I'm not sure what you mean about 'only one cap between the supply pins.
Each rail is decoupled by one cap which is connected to the power ground star. All the signal connections are connected to the signal ground star and there is then a link between the two stars.
The signal grounds of the buffer are also connected to the signal ground star in the GC as shown in the diagram.
I wondered about the second blocking cap but somebody else suggested leaving it in I'll leave it for now and take it out later.
Sorry Carlos, I don't have a digital camera here but you surely know what an OPA627 looks like 😉
Mark, the 6N8 must stay as it is the other component of the LPF.
I'm not sure what you mean about 'only one cap between the supply pins.
Each rail is decoupled by one cap which is connected to the power ground star. All the signal connections are connected to the signal ground star and there is then a link between the two stars.
The signal grounds of the buffer are also connected to the signal ground star in the GC as shown in the diagram.
I wondered about the second blocking cap but somebody else suggested leaving it in
I'll leave it for now and take it out later.Nuuk,
You don't need to bother with signal and power star grounds on an op-amp.
A board with just a buffer is so small there's not even space for that, and the current is so small you can join everything.
Take a close look at my tape buffer module on my preamp thread.
My advice is: don't fiddle with ground on the buffer, just connect the ground point of that module to the RCA signal ground star you have on your GC.
You don't need to bother with signal and power star grounds on an op-amp.
A board with just a buffer is so small there's not even space for that, and the current is so small you can join everything.
Take a close look at my tape buffer module on my preamp thread.
My advice is: don't fiddle with ground on the buffer, just connect the ground point of that module to the RCA signal ground star you have on your GC.
mhennessy said:
Oh, yes, right.🙄
I was in a hurry to have my lunch and didn't look at it properly.
Thanks Carlos. It seems to me the big question is - should the power grounds of the buffer and GC be connected, as they would if I connected the signal grounds of the buffer to the power grounds and then to the signal star of the GC.
Nuuk said:
Too many grounds here.
Forget the power ground of your GC.
Join the two grounds on your buffer module and connect one ground wire to the signal ground point of your GC (the RCA ground).
Nuuk said:
Err... I said 6u8.... Although I've taken a closer look at your schematic, and seen that the cap is actually labelled 6ufd - maybe 6uF or 6u would be less confusing. I too was eating my lunch 😉
Nuuk said:
Exactly as you've said. There are arguments for not decoupling to ground, especially in low-level circuits where you are trying to keep the ground as clean as possible. A single cap connected between pins 4 and 8 (for a dual opamp) or 4 and 7 (for a single) will decouple the opamp perfectly well. When using this scheme, the only place where there are caps to ground are back at the power supply, or perhaps where the power arrives at a PCB. If I'm still not making sense, take a look at the schematics of my preamp - you'll see how I did this.
Nuuk said:
I'm not sure if I'd connect the preamp decoupling caps to the power amp "power" ground. Although the differences are likely to be small with such a compact circuit, as Carlos has said...
Nuuk said:I wondered about the second blocking cap but somebody else suggested leaving it in
Have you still got a blocking cap in the GainClone circuit? If so, you don't need it. But it isn't clear from your diagram. The answer is that you should have at least 1 cap between the output of the buffer and the input to the LM3875. Although, if you're using a FET-input buffer, perhaps the DC offset from it will be tiny, and you might choose to DC-couple it straight to the LM3875. Of course, the tiny offset in the buffer will be amplified by the LM3875, causing a less tiny DC offset at the speaker terminals. I know that you were quite worried about such things in the recent past, hence my caution here.
Another thing - if you decide that you can omit all DC-blocking caps from buffer to LM3875, check the power on/off behaviour very carefully...
Cheers,
Mark 😉
PS - I'm also experimenting with a LPF - I see you've chosen ~35KHz - I'm trying 100KHz initially... I've chosen different resistor values so that the turnover frequency is much less dependent on source impedance - hence no need for a buffer...
mhennessy said:
Do that on a fast op-amp like the OPA627 and it will oscillate like hell.
You may even not know it's oscillating, but the sound quality will be hugely affected.
Put the bypassing caps to ground NEAR the supply pins of the op-amp back on and the magic returns.
I even noticed huge effects (and I was not alone) just by changing 2.2uf tantalum caps for 100uf electro / 0.1 ceramics.
And the 2.2uf caps were right next to the op-amp's supply pins, so it was not bad at all.
This was with OPA2132 op-amps on the preamp seccion of a Micromega integrated amp.
The guy was preferring the sound of the original NE5532s😡 .
Then I bypassed properly, my way, and he went crazy.
It ended with OPA2132s and the guy is very happy with the sound.
This was done with long listening tests and testing several op-amps.
Let me tell you that I always bypass with 0.1uf between + and - pins of the op-amp, under the circuit.
Thanks again chaps.
I'll start with the OSCON GC's decoupling the rails to the OPA627 and then add in the 0.1ufd across the supply pins later. There are umbilicals from the buffer PSU to the buffers about 300mm long.
Remember, this buffer has worked very well for a few years now.
Sorry about the confusion with the 6ufd Mark. I have removed the input blocking cap from the GC to make room for the filter so I'll keep that 6ufd on the output of the buffer for now.
So, I have now, the decoupling caps on the rails of the buffer connected to the ground wire to the buffer PSU. The signal grounds in the buffer connected to the star ground in the GC.
Nearly ready to try it but first the daily cycle ride 😉
BTW - I have found some 700 ohm carbon caps here so my filter is close to the 33Khz recommended originally by Joe R.
I'll start with the OSCON GC's decoupling the rails to the OPA627 and then add in the 0.1ufd across the supply pins later. There are umbilicals from the buffer PSU to the buffers about 300mm long.
Remember, this buffer has worked very well for a few years now.
Sorry about the confusion with the 6ufd Mark. I have removed the input blocking cap from the GC to make room for the filter so I'll keep that 6ufd on the output of the buffer for now.
So, I have now, the decoupling caps on the rails of the buffer connected to the ground wire to the buffer PSU. The signal grounds in the buffer connected to the star ground in the GC.
Nearly ready to try it but first the daily cycle ride 😉
BTW - I have found some 700 ohm carbon caps here so my filter is close to the 33Khz recommended originally by Joe R.
High-end bike system
Oh, Nuuk, you've just gave me an idea...
What about if I use this:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=17491
as an amp for my bike? 😉
Nuuk said:
Oh, Nuuk, you've just gave me an idea...
What about if I use this:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=17491
as an amp for my bike?
😉
Nuuk said:
Very nice Carlos but won't it depend on the frequency that you cycle?
Hey, not that simple, the dynamo would charge the batteries.😀 😱
Progress report
I have just connected the buffers/filters up as discussed and notice a couple of points that I would like to check.
On the output of the buffer, I get 2-3mV on each buffer. Should I be getting that DC after the blocking cap on the buffers?
The DC offset on the amps has gone up from around 0mV to 30mV on one channel and 20mV on the other.
I have never pressed 'play' on my CD player when measuring DC offset on the speaker terminals but when I do, the reading (digital) jumps around but only by about 20mV either side of what it is when the CD player is not playing.
So, it doesn't look too bad from DC offset point-of-view but I thought that I would just ask for thoughts before connecting the speakers!
BTW, I connected 270K resistors to ground on the output of the buffers (to discharge the cap).
I have just connected the buffers/filters up as discussed and notice a couple of points that I would like to check.
On the output of the buffer, I get 2-3mV on each buffer. Should I be getting that DC after the blocking cap on the buffers?
The DC offset on the amps has gone up from around 0mV to 30mV on one channel and 20mV on the other.
I have never pressed 'play' on my CD player when measuring DC offset on the speaker terminals but when I do, the reading (digital) jumps around but only by about 20mV either side of what it is when the CD player is not playing.
So, it doesn't look too bad from DC offset point-of-view but I thought that I would just ask for thoughts before connecting the speakers!
BTW, I connected 270K resistors to ground on the output of the buffers (to discharge the cap).
carlosfm said:
Good morning Carlos - I'm sorry, but I'm confused...
You seem to start this post by saying that my recommendation for a single cap between +/- supply pins is a bad idea...
You then talk about the effects of different sorts of caps (very interesting - I too have had similar experiences), but I'm not clear of the configuration you're using here.
You end the post by saying that you always use a single cap between +/-
I'm sorry to be confused, but it is early in the morning for me
Just to be clear, I was recommending a single cap between +/- pins next to every IC, and a pair of electrolytics or similar at the PSU or board edge. You think this is a bad plan?Cheers,
Mark 😉
Re: Progress report
Hi Nuuk,
If I understand you correctly, this DC will be coming from the LM3875 - it's caused by the operation of the LM3875 and it's input bias current - nothing to worry about, IMO
Not sure why this would be, but it's just possible that there is some DC leakage across the 6u cap between the buffer and LPF. Is it an electrolytic? If so, check the potential at each end of it, and possibly try reversing it...
Depending on the model and the way it works, this might be caused by the output muting circuit of the CD player or an internal change of DC level being transferred to the output via ac-coupling capacitors... Do you hear a pop when you press play? Portable CD players can be bad for this. It might even be something else entirely, but the DVM might be confusing the issue - what does it look like on a 'scope?
A good trick - when in doubt, try connecting a speaker to the output via a series resistor (say 10-20 ohms, 1/4 watt) - then you'll be able to hear any strange effects at a lower level - and if anything bad happens, the resistor will go open-circuit before any harm can be done to your speakers...
Not sure why you need the 270K resistor - the GC 10K input resistor will discharge the cap. Remember that the -ve input is a virtual earth!
Cheers,
Mark 😉
Hi Nuuk,
Nuuk said:
If I understand you correctly, this DC will be coming from the LM3875 - it's caused by the operation of the LM3875 and it's input bias current - nothing to worry about, IMO
Nuuk said:
Not sure why this would be, but it's just possible that there is some DC leakage across the 6u cap between the buffer and LPF. Is it an electrolytic? If so, check the potential at each end of it, and possibly try reversing it...
Nuuk said:
Depending on the model and the way it works, this might be caused by the output muting circuit of the CD player or an internal change of DC level being transferred to the output via ac-coupling capacitors... Do you hear a pop when you press play? Portable CD players can be bad for this. It might even be something else entirely, but the DVM might be confusing the issue - what does it look like on a 'scope?
Nuuk said:
A good trick - when in doubt, try connecting a speaker to the output via a series resistor (say 10-20 ohms, 1/4 watt) - then you'll be able to hear any strange effects at a lower level - and if anything bad happens, the resistor will go open-circuit before any harm can be done to your speakers...
Not sure why you need the 270K resistor - the GC 10K input resistor will discharge the cap. Remember that the -ve input is a virtual earth!
Cheers,
Mark 😉
mhennessy said:
Mark,
Or I didn't make myself clear, or it's early in the morning, choose one.😉
What I said is that besides bypassing from the op-amp's PSU pins to ground, I always put a cap from + to - pins TOO.
mhennessy said:
I'm sorry to be confused, but it is early in the morning for me
Well, here in Portugal it's now 9:52 Am, the same hour as in the UK, I think.😎
What happens is that we are going to bed too late every night...
To answer to your question, it's a bad plan.
A least you should have one cap from each op-amp's PSU pins to ground.
Your plan works with old and crappy op-amps.
I've even seen a preamp with no psu caps near the op-amps.
I followed the PSU lines and the caps were right after the regulators, tens of centimeters away (!).
That worked ok with the original and nasty JRC op-amps.
But put a good and fast one and it oscilates as a mad cow.
And it's not enough to bypass between + and -.
If the circuit doesn't have caps near the op-amps you can always bypass under the circuit.
Lack of progress report
Not a good report so far!
When I connected the speakers to one GC and played some music I got NOTHING! No buzz or hum, absolute silence.
On the other channel I got a buzzing noise but no music.
I checked the buffers and cannot find anything wrong. All connections are good and I have exactly +/- 15V DC on the power rails.
I changed one buffer by taking out the 6uF cap and putting the filter (680R/6N8) as close as I can to the output pin of the OPA627.
I made a test set up of a cassette player, a 100K pot, my old A60 amp and an old speaker, then connected the modified buffer to the amp but now all I get is a loud buzzing noise on the speaker and no music at all at any volume.
I have just rechecked the buffer and it seems OK, I can trace some AC from the input on the buffer to the output (it stops when the music stops). DC on the output is 0mV even with no cap there!
I tried the other buffer on the test rig but same thing - loud buzz but no music? It's very puzzling as I know the buffers/PSU work.
I have tried connecting the signal grounds in the buffer to the power grounds but it makes absolutely no difference.
BTW Carlos, removing the 270K to ground makes no difference either.
Not a good report so far!
When I connected the speakers to one GC and played some music I got NOTHING! No buzz or hum, absolute silence.
On the other channel I got a buzzing noise but no music.
I checked the buffers and cannot find anything wrong. All connections are good and I have exactly +/- 15V DC on the power rails.
I changed one buffer by taking out the 6uF cap and putting the filter (680R/6N8) as close as I can to the output pin of the OPA627.
I made a test set up of a cassette player, a 100K pot, my old A60 amp and an old speaker, then connected the modified buffer to the amp but now all I get is a loud buzzing noise on the speaker and no music at all at any volume.
I have just rechecked the buffer and it seems OK, I can trace some AC from the input on the buffer to the output (it stops when the music stops). DC on the output is 0mV even with no cap there!
I tried the other buffer on the test rig but same thing - loud buzz but no music? It's very puzzling as I know the buffers/PSU work.
I have tried connecting the signal grounds in the buffer to the power grounds but it makes absolutely no difference.
BTW Carlos, removing the 270K to ground makes no difference either.
- Status
- Not open for further replies.