Getting back some gain
You're currently using the OPA627 as a unity gain follower. Its fairly easy to make up some of the lost gain (from snipping out the 100R in the T-network) by configuring this opamp to give you, for example, 20dB gain. Your attenuator circuit gives a maximum -6dB so at the very least it makes sense to run the OPA627 stage at +6dB to make up for that loss - I'd go for turning it into a X10 stage myself (18k feedback, 2k to ground).
If you want to keep the attenuator at the output of your OPA buffer, I think that's been the cause of your problem since it has an output impedance of 10k, and the input impedance of your LM3886 stage is only 1k. If you still have it between the OPA and the LM, it will only give you a gain of around 6dB from the power amp stage and plenty of instability as the 3886 doesn't like such low gains. So, instead of the 100R in the T-network, use a 1k in this position and your problem is solved.
You're currently using the OPA627 as a unity gain follower. Its fairly easy to make up some of the lost gain (from snipping out the 100R in the T-network) by configuring this opamp to give you, for example, 20dB gain. Your attenuator circuit gives a maximum -6dB so at the very least it makes sense to run the OPA627 stage at +6dB to make up for that loss - I'd go for turning it into a X10 stage myself (18k feedback, 2k to ground).
If you want to keep the attenuator at the output of your OPA buffer, I think that's been the cause of your problem since it has an output impedance of 10k, and the input impedance of your LM3886 stage is only 1k. If you still have it between the OPA and the LM, it will only give you a gain of around 6dB from the power amp stage and plenty of instability as the 3886 doesn't like such low gains. So, instead of the 100R in the T-network, use a 1k in this position and your problem is solved.
Thanks
Thanks a lot Abraxalito. I should have posted the attenuator circuit before,
then this situation would not have occured, I would have got the suggestion
of replacing 100R by 1K. I would not like to change the gain of
the buffer, also as I said I am reluctant to shift the attenuator at the
input of the buffer. So I will insert the t-network again by 1K and test the
circuit.
Regards
Roushon.
Thanks a lot Abraxalito. I should have posted the attenuator circuit before,
then this situation would not have occured, I would have got the suggestion
of replacing 100R by 1K. I would not like to change the gain of
the buffer, also as I said I am reluctant to shift the attenuator at the
input of the buffer. So I will insert the t-network again by 1K and test the
circuit.
Regards
Roushon.
all is well
Replaced the t-network by 1K. Now both the channel are silent. Although
initially some noise was appearing in the faulty channel when volume was
raised and voltage also was flactuating. Needed to do lots of isolation
of diferent power cables and speaker wires from ground bass to get rid of this
noise. Now everything is perfect.
Thanks a lot for all the help.
Regards
Roushon
Replaced the t-network by 1K. Now both the channel are silent. Although
initially some noise was appearing in the faulty channel when volume was
raised and voltage also was flactuating. Needed to do lots of isolation
of diferent power cables and speaker wires from ground bass to get rid of this
noise. Now everything is perfect.
Thanks a lot for all the help.
Regards
Roushon
Do I need to ungrade the amp?
So far so good. I was playing the amp on my Open Baffle speaker. But recently noticed
a problem. I was playing a cd (Belafonte) which was recorded in some concert. So there
was in between clapping of the audience. The music played fantastic, but the clapping
becomes very loud and irritating, I need to reduce the volume. I know that the CD is
of very good quality and the recording also. Where is the problem? Is it the 1K
resistor in the t-network feedback? Is this resistor changing the frequency
response of the amp?
As such the amp is silent, I hear a faint hum/noise when I put my ear very close to the
speaker.
The amp diagram is attached again for convenience.
Thanks and regards
Roushon.
So far so good. I was playing the amp on my Open Baffle speaker. But recently noticed
a problem. I was playing a cd (Belafonte) which was recorded in some concert. So there
was in between clapping of the audience. The music played fantastic, but the clapping
becomes very loud and irritating, I need to reduce the volume. I know that the CD is
of very good quality and the recording also. Where is the problem? Is it the 1K
resistor in the t-network feedback? Is this resistor changing the frequency
response of the amp?
As such the amp is silent, I hear a faint hum/noise when I put my ear very close to the
speaker.
The amp diagram is attached again for convenience.
Thanks and regards
Roushon.
Attachments
You might be getting RFI-induced sibilance. I've encountered that on many of my chip-amp circuits. The first thing to check is the grounding for the input cable. If the screen of your input cable connects on to the PCB, that's probably the problem. In my experience, the solution is to run a ground cable alongside the input signal cable connecting the input ground terminal (presumably on an RCA-phono input connector) to the star earth. Disconnect the screen of the input cable at the PCB.
If this does indeed solve the sibilance on applause it will have other beneficial effects on the sound stage too.
If this does indeed solve the sibilance on applause it will have other beneficial effects on the sound stage too.
Thanks Abraxalito.
This is probably correct, I do not remember. I was extremely careful regarding
grounding. I will open the amp and check again. Slight RF is there when power ON but
this completely disappears when volume raised and played for sometime.
I always remember your help when I was struggling with the amp beginning of this year.
Regards
Roushon.
So far so good...
Waking up this thread as I plan some modification in the
amplifier. So far I enjoyed the sound very much apart
from some minor issues.
Planing to open the amplifier in the coming summer
vacation if still I get some time then.
Firstly will see the RF issue. Just a two post and two years
back Abraxalito suspected that the screen of the input
cable is touching the pcb which is correct as I
see when studying through the pcb design and thought about
my wiring the amp in a clear mind.
Secondly I need a suggestion:
When I built this amp I did not have any speakers and
much knowledge about what I was doing. So built as
parallel of two inverted LM3886 design. Now that my
speakers (open baffle following Martin Kings design using
Eminence alpha 15a and Fostex FE103en) are 8Ohm
I see a parallel design is a waste. I would like to make
it a bridge. My first question is that is it worth taking
the trouble as I do not listen very loud music. I do not
know if making it bridge will have any effect on the
sound quality in medium to loud volume.
Finally there is a small problem with the attenuator
based volume control. The volume does not start from
zero. The attenuator circuit is attached in
post no. 59. Is there any solution to this problem?
Thanks and regards
Roushon
Waking up this thread as I plan some modification in the
amplifier. So far I enjoyed the sound very much apart
from some minor issues.
Planing to open the amplifier in the coming summer
vacation if still I get some time then.
Firstly will see the RF issue. Just a two post and two years
back Abraxalito suspected that the screen of the input
cable is touching the pcb which is correct as I
see when studying through the pcb design and thought about
my wiring the amp in a clear mind.
Secondly I need a suggestion:
When I built this amp I did not have any speakers and
much knowledge about what I was doing. So built as
parallel of two inverted LM3886 design. Now that my
speakers (open baffle following Martin Kings design using
Eminence alpha 15a and Fostex FE103en) are 8Ohm
I see a parallel design is a waste. I would like to make
it a bridge. My first question is that is it worth taking
the trouble as I do not listen very loud music. I do not
know if making it bridge will have any effect on the
sound quality in medium to loud volume.
Finally there is a small problem with the attenuator
based volume control. The volume does not start from
zero. The attenuator circuit is attached in
post no. 59. Is there any solution to this problem?
Thanks and regards
Roushon
Last edited:
Attenuator
Forgot to mention that I truncated the attenuator circuit
to make it 64 step by removing the stage 10K + 6R31.
If that is the reason for the problem I can add two resistors
10K and 6R31 at the output of the attenuator. Extending
the whole thing by one more stage will need lots of hard work
as I will have to change the logic circuit also which is switching
the relays.
Roushon
Forgot to mention that I truncated the attenuator circuit
to make it 64 step by removing the stage 10K + 6R31.
If that is the reason for the problem I can add two resistors
10K and 6R31 at the output of the attenuator. Extending
the whole thing by one more stage will need lots of hard work
as I will have to change the logic circuit also which is switching
the relays.
Roushon
opened!
Could not wait for the vacation. Just spent a day opening the amp and made
the following changes. Left it in its parallel form.
1. removed the 1 mfd input capacitors
2. moved the buffer in between the amp and the attenuator to avoid the
impedance mismatch
3. removed the 1K t-network resistors to move the gain down to 20. this
also solved the not-so-zero-volume start problem
4. solved the grounding problem: input cable screen sent directly to the
star ground
Now the advantages: zero RF, zero hum, sound seems to have more opened
up and filling the air in the room evenly.
But there is one problem with the attenuator which I am not able to solve.
When the volume moves from 15 to 16, 31 to 32, 47 to 48... there is a loud
TAK sound and when the volume is pulled down and crossed the above
stages. Do not know why...
Thinking of replacing the attenuator by traditional volume control. Any
suggestion regarding a very good quality dual 10K volume control will
be helpful..Also how does the motorized volume control work?
Thanks and regards
Roushon.
Could not wait for the vacation. Just spent a day opening the amp and made
the following changes. Left it in its parallel form.
1. removed the 1 mfd input capacitors
2. moved the buffer in between the amp and the attenuator to avoid the
impedance mismatch
3. removed the 1K t-network resistors to move the gain down to 20. this
also solved the not-so-zero-volume start problem
4. solved the grounding problem: input cable screen sent directly to the
star ground
Now the advantages: zero RF, zero hum, sound seems to have more opened
up and filling the air in the room evenly.
But there is one problem with the attenuator which I am not able to solve.
When the volume moves from 15 to 16, 31 to 32, 47 to 48... there is a loud
TAK sound and when the volume is pulled down and crossed the above
stages. Do not know why...
Thinking of replacing the attenuator by traditional volume control. Any
suggestion regarding a very good quality dual 10K volume control will
be helpful..Also how does the motorized volume control work?
Thanks and regards
Roushon.
Volume control pot
It seems I am talking to myself. Finally found enough information and
availability of the motorized volume control in India.
But there is an issue I really need help: what value pot as volume
control should be the best?
The present stepped attenuator attached is 10K. But I feel when the volume
is raised high the sound becomes irritating. Does the volume control resistance
has some effect on the sound?
Please post a comment.
Thanks and regards
Roushon.
It seems I am talking to myself. Finally found enough information and
availability of the motorized volume control in India.
But there is an issue I really need help: what value pot as volume
control should be the best?
The present stepped attenuator attached is 10K. But I feel when the volume
is raised high the sound becomes irritating. Does the volume control resistance
has some effect on the sound?
Please post a comment.
Thanks and regards
Roushon.
The pot loads the source.
Actually, the preceding stage sees the resistance of the pot parallelled with the input impedance. A 10k pot is advised when your source is capable of at least half that. A lot of sources do not like loads below 10k, so a 22k pot is kind of my minimum.
Lower impedances also result in lower current noise, so 22k is kind of my maximum as well.
thanks
I knew someone will offer a helping hand. Thanks Sangram! I will look for a
22K pot.
Thanks Andrew for pointing out a possible reason.
The PS is quite heavy: +-35 volt and 5amp on each rail
and there are about 15000mfd cap on each rail. Channels
have their independent PS starting from toroidal. Do not
know if there is any other reason which can cause the
problem.
Regards
Roushon
I knew someone will offer a helping hand. Thanks Sangram! I will look for a
22K pot.
Thanks Andrew for pointing out a possible reason.
The PS is quite heavy: +-35 volt and 5amp on each rail
and there are about 15000mfd cap on each rail. Channels
have their independent PS starting from toroidal. Do not
know if there is any other reason which can cause the
problem.
Regards
Roushon
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.