matthias
Well-known member
when I measure the output of my converter at 0db it is 2,8V AC
it would be perfect if the signal is at 2V AC before the attenuator...
it would be perfect if the signal is at 2V AC before the attenuator...
new monitor controller - inspired by NewYorkDave's design
- Thread starter matthias
- Start date
Help Support GroupDIY Audio Forum:
Samuel Groner
Well-known member
2.8 V to 2 V is not even a 3 dB difference (and 2.8 V is not a high output level?). You shouldn't need to worry about that... And if you do: the 2k2 series resistors provide exactely the attenuation you asked for.
Samuel
Samuel
matthias
Well-known member
ah yes, you're right... 
then the input sensivity of my monitor speakers (Genelec 8050) is very sensitive... I already adjusted the input trim to its lowest setting which is just -6db
If I look at the output of my converters the meter shows -28 db at a normal listening level...
I think an attenuation of 6-12db would be in a useful range...
should I increase the value of the input resistors R1-R8 and leave the trimmer out.. ?
then the input sensivity of my monitor speakers (Genelec 8050) is very sensitive... I already adjusted the input trim to its lowest setting which is just -6db
If I look at the output of my converters the meter shows -28 db at a normal listening level...
I think an attenuation of 6-12db would be in a useful range...
should I increase the value of the input resistors R1-R8 and leave the trimmer out.. ?
Val_r
Well-known member
[quote author="matthias"]so... here's hopefully the final version of my monitor controller design....
monitor controller2008v3.pdf[/quote]
Go for version 4:
Outputs of OP1 and OP2 should be reversed, otherwise when you are listening in mono you are listening with reversed phase.
Look at NYD's design, which is correct.
I would leave the resistors on the non-inverting inputs of the mono-summing opamps, and add a small capacitance shunting feedback resistors R17/R24.
Consider the opportunity to go in floating mode: leave the gnd out from the outputs, disconnect the gnd from the C.T. of the 4-gang attenuator, and in the lowest poistion (lowest gain), phase and anti-phase connect together.
Also DC coupling can be ok, so I would remove C1 and C2, and, if you go for floating mode, remove R23 and R29.
Let us know how your build will sound. We are hopeful! :green:
Respect,
Val.
monitor controller2008v3.pdf[/quote]
Go for version 4:
Outputs of OP1 and OP2 should be reversed, otherwise when you are listening in mono you are listening with reversed phase.
Look at NYD's design, which is correct.
I would leave the resistors on the non-inverting inputs of the mono-summing opamps, and add a small capacitance shunting feedback resistors R17/R24.
Consider the opportunity to go in floating mode: leave the gnd out from the outputs, disconnect the gnd from the C.T. of the 4-gang attenuator, and in the lowest poistion (lowest gain), phase and anti-phase connect together.
Also DC coupling can be ok, so I would remove C1 and C2, and, if you go for floating mode, remove R23 and R29.
Let us know how your build will sound. We are hopeful! :green:
Respect,
Val.
matthias
Well-known member
thanks for your help, here's v4 :grin:
monitor controller v4
- phase issue at mono switch fixed
- added R54 and R55
my attenuator is already finished so i won't change that anymore,
but do you think floating mode would be a major improvement?
monitor controller v4
- phase issue at mono switch fixed
- added R54 and R55
my attenuator is already finished so i won't change that anymore,
but do you think floating mode would be a major improvement?
lolo-m
Well-known member
Did you checked it ? :grin:
matthias
Well-known member
sure.
Val_r
Well-known member
[quote author="matthias"]thanks for your help, here's v4 :grin:
monitor controller v4
- phase issue at mono switch fixed
- added R54 and R55
my attenuator is already finished so i won't change that anymore,
but do you think floating mode would be a major improvement?[/quote]
OK, but I still cannot see the utility to add the 4 trimmers....
If you are experiencing problems, why don't you lower the value of the feedback resistors (with a slight improve in noise performance)?
Going into floating mode can be useful, I would give it a try.
You can go DC coupled and remove C1, C2, R23, R29.
C8 and C9: I would use something like 100 or 150pF, well that's my experience with these type of opamps.
Do NOT forget to add 0.1uF bypass capacitors near the + and - supply rails of all of the opamps.
Respect,
Val.
monitor controller v4
- phase issue at mono switch fixed
- added R54 and R55
my attenuator is already finished so i won't change that anymore,
but do you think floating mode would be a major improvement?[/quote]
OK, but I still cannot see the utility to add the 4 trimmers....
If you are experiencing problems, why don't you lower the value of the feedback resistors (with a slight improve in noise performance)?
Going into floating mode can be useful, I would give it a try.
You can go DC coupled and remove C1, C2, R23, R29.
C8 and C9: I would use something like 100 or 150pF, well that's my experience with these type of opamps.
Do NOT forget to add 0.1uF bypass capacitors near the + and - supply rails of all of the opamps.
Respect,
Val.
Harpo
Well-known member
[quote author="Val_r"].., why don't you lower the value of the feedback resistors (with a slight improve in noise performance)?
Going into floating mode can be useful, I would give it a try.
You can go DC coupled and remove C1, C2, R23, R29.
C8 and C9: I would use something like 100 or 150pF, well that's my experience with these type of opamps.[/quote]
changing the feedback resistors R17 R24 would only handle the mono part. Keep it as is. Operating in floating mode (attenuator with no connection to gnd) would only work when not switched to mono as the summing stages have a reference to gnd.
I'd also keep C1/R23 C2/R29 to have the attenuator DC-free in case there is some offset comming from OP1 OP2. With these given values -3dB hpf is at 0,22Hz introducing a phase error larger than -1° only below 13Hz and -0,4° at 32Hz , so probably won't matter.
C8 C9 at 10pF is indeed a little low (-3dB lpf at 677kHz), 82pF (for 82kHz) or 68pF (for 100kHz) -3dB lpf might fit better. YMMV.
-Harpo
Going into floating mode can be useful, I would give it a try.
You can go DC coupled and remove C1, C2, R23, R29.
C8 and C9: I would use something like 100 or 150pF, well that's my experience with these type of opamps.[/quote]
changing the feedback resistors R17 R24 would only handle the mono part. Keep it as is. Operating in floating mode (attenuator with no connection to gnd) would only work when not switched to mono as the summing stages have a reference to gnd.
I'd also keep C1/R23 C2/R29 to have the attenuator DC-free in case there is some offset comming from OP1 OP2. With these given values -3dB hpf is at 0,22Hz introducing a phase error larger than -1° only below 13Hz and -0,4° at 32Hz , so probably won't matter.
C8 C9 at 10pF is indeed a little low (-3dB lpf at 677kHz), 82pF (for 82kHz) or 68pF (for 100kHz) -3dB lpf might fit better. YMMV.
-Harpo
matthias
Well-known member
thank you for the calculations.
I think the major layout won't change anymore, so I can go on and finish the pcb...
I think the major layout won't change anymore, so I can go on and finish the pcb...
Samuel Groner
Well-known member
You want similar small caps for R44, R52 etc. to improve stability.
R54 and R55 are pointless and just increase noise (is my feeling wrong that I'm repeating myself? :wink: ).
With respect to the PAD: I'd make R1/R2 etc. somewhat larger (say 3k3) and add a shunt resistor between them, about 1k (or a trim if you like). With the trimmers as shown you completely mess up your CMRR.
Personally I'd use different opamps as the OPA275 aren't very good examples of current IC technology, but I won't start another discussion on this.
Samuel
R54 and R55 are pointless and just increase noise (is my feeling wrong that I'm repeating myself? :wink: ).
With respect to the PAD: I'd make R1/R2 etc. somewhat larger (say 3k3) and add a shunt resistor between them, about 1k (or a trim if you like). With the trimmers as shown you completely mess up your CMRR.
Personally I'd use different opamps as the OPA275 aren't very good examples of current IC technology, but I won't start another discussion on this.
That's not true. As long as the signal is differential floating the connection works perfectly well. Ground referenced or not doesn't matter.
Samuel
Harpo
Well-known member
[quote author="Samuel Groner"]
Samuel[/quote]
Surely attenuation will work eigther way, but as I see it, signal would no longer be floating when referenced to gnd and varying this behaviour by switching to mono is probably not desired.
That's not true. As long as the signal is differential floating the connection works perfectly well. Ground referenced or not doesn't matter.
Samuel[/quote]
Surely attenuation will work eigther way, but as I see it, signal would no longer be floating when referenced to gnd and varying this behaviour by switching to mono is probably not desired.
Samuel Groner
Well-known member
OK, I see. The design is never truly floating due to the headphone amp buffers though. But I do basically agree with you that grounding is probably the better solution altough it might have lower CMRR.
Samuel
Samuel
solder_city
Well-known member
would r54/r55 still be 'pointless' if the output was dc coupled?
or maybe im misunderstanding what theyre for?
or maybe im misunderstanding what theyre for?
dagoose
Well-known member
I was wondering how you plan to do the headphone amp with the TPA6120?
I think there is just an SMD version of the opamp or am i missing something here? Is there a prepared board like this one for sale? http://www.pinkfishmedia.net/forum/showthread.php?t=7902&goto=nextoldest
But then without the trannies ofcourse..
Or do you have a different solution?
I'm also working on my monitorcontroller at this moment, i have done all the connectors on the backpanel (8x XLR + 2x 6.3 jack fits on a 1u panel :grin: ) and i soldered all resistors to the 4gang 24 stpes elma switch, loads of soldering but fun to do and looks cool too
The only real difference with the one i'm doing right now is that i going to do a subwoofer in/out.
The only thing that will happen is that i have an impedance change when i switch the sub in and out, anyone knows how i can make that not to happen?
I think there is just an SMD version of the opamp or am i missing something here? Is there a prepared board like this one for sale? http://www.pinkfishmedia.net/forum/showthread.php?t=7902&goto=nextoldest
But then without the trannies ofcourse..
Or do you have a different solution?
I'm also working on my monitorcontroller at this moment, i have done all the connectors on the backpanel (8x XLR + 2x 6.3 jack fits on a 1u panel :grin: ) and i soldered all resistors to the 4gang 24 stpes elma switch, loads of soldering but fun to do and looks cool too
The only real difference with the one i'm doing right now is that i going to do a subwoofer in/out.
The only thing that will happen is that i have an impedance change when i switch the sub in and out, anyone knows how i can make that not to happen?
matthias
Well-known member
yes, there is only a smd version. I will solder it directly on the solderside of the pcb...
I also use two headphone amps, one for each channel to reduce the heat dissipation of every ic.
here's a pic of the current layout...
Harpo
Well-known member
[quote author="matthias"]yes, there is only a smd version. I will solder it directly on the solderside of the pcb...
I also use two headphone amps, one for each channel to reduce the heat dissipation of every ic.
here's a pic of the current layout...
[/quote]
Hi Matthias,
some suggestions,
you may join TPAs pin 16 (Rin-) with pin 19 (Rout) and take pin 17 (Rin+) to gnd (or wired the same as for left side) if you don't use this part of the amp. Unconnected chips sometimes show strange behaviour. According to TI's datasheet (page 17), the 10R's could be placed closer to the TPAs output. Unfortunatly you can not get floating output by using both amps in the chip, feeding the other input with flipped signal for further heat reduction/double beef because of common gnd for standard headphone connectors.
I'd place some 22pF or 33pF across OP275s 47k feedback resistors close to the chip. (still don't get it, why you don't wire them as non inverting voltage followers) Rail decoupling at OP275s should be added and the TPA may like some additional 47uF or 100uF at it's rails. As always YMMV.
I also use two headphone amps, one for each channel to reduce the heat dissipation of every ic.
here's a pic of the current layout...
Hi Matthias,
some suggestions,
you may join TPAs pin 16 (Rin-) with pin 19 (Rout) and take pin 17 (Rin+) to gnd (or wired the same as for left side) if you don't use this part of the amp. Unconnected chips sometimes show strange behaviour. According to TI's datasheet (page 17), the 10R's could be placed closer to the TPAs output. Unfortunatly you can not get floating output by using both amps in the chip, feeding the other input with flipped signal for further heat reduction/double beef because of common gnd for standard headphone connectors.
I'd place some 22pF or 33pF across OP275s 47k feedback resistors close to the chip. (still don't get it, why you don't wire them as non inverting voltage followers) Rail decoupling at OP275s should be added and the TPA may like some additional 47uF or 100uF at it's rails. As always YMMV.
Samuel Groner
Well-known member
Never! That opamp uses a current feedback topology and will turn into an oscillator if you do this. You'll need a feedback resistor, about 2k seems appropriate. Basically a good idea to not leave things floating though.
Indeed, they should be placed closer.
One might argue if this chip is really the best solution anyway--I'd rather have used a decent buffer within the OP275 feedback loops. Simpler and likely higher performance as well.
I second both statements, as shown it is unlikely that these stages will become stable. It has all been said before in this thread...
Inverting will surely show lower distortion (though I'm not sure if this is the intention of the author to use inverting), at the cost of much higher noise (at least with these 47k resistors).
Samuel
matthias
Well-known member
the layout isn't finished yet... I only wanted to show how to implement the tpa on the backside of the pcb..
the tpa consists of two independent amplifiers, each amplifier also has its own ps inputs so I will leave one side disconnected.
I also think 47k is a bit too high... what do you think would be an appropriate value ? 15k ?
the tpa consists of two independent amplifiers, each amplifier also has its own ps inputs so I will leave one side disconnected.
I also think 47k is a bit too high... what do you think would be an appropriate value ? 15k ?
Samuel Groner
Well-known member
15k or perhaps 22k will be fine and reduces noise by more than 3 dB. You should include a small feedback resistor nonetheless.
Samuel
Samuel
