Wavebourn said:
Hi,
Are you suggesting using a variable resistor to control the gain of the op-amp, rather than attenuating its input or output? That will mean I have to attenuate the input to the lowest level I may want, and adjusting the feedback gain will also change distortion.
I was considering this. What do you other guys think of this idea?
I can't really put a pot before the buffer as it would have to handle 6 channels of balanced audio... I can't find a 12 channel quality pot!
Thanks for all your help so far.
Let's re-state the problem. You have a 6-channel, balanced voltage output DAC with an output impedance of 650 ohms. You want to control volume. Do you also need to convert the 6-channels all to single-ended to feed your preamp?
I would ignore the "likes to see 1K load" part.
You can't find a 12-channel pot but can you find a 6-channel one?
http://www.diyaudio.com/forums/showthread.php?postid=1159607#post1159607
I would ignore the "likes to see 1K load" part.
You can't find a 12-channel pot but can you find a 6-channel one?
http://www.diyaudio.com/forums/showthread.php?postid=1159607#post1159607
paulb said:
Hi thanks Paul,
You got everything correct except that the DAC output impedance is not 650R - that was the output impedance in conjunction with an RC filter and a transformer on the end of it. I don't know what the actual chips output Z is, as the datasheet doesn't say. It says it can drive 600R though.
Thanks for the link. I have a 24 position 6 layer (channel) switch that I could configure to switch resistors between the + / - phase. I am a little worried though. If you look at the datasheet for the lm4562 they suggest putting a resistor between the inverting and non-inverting inputs of the op-amp to increase its noise gain and make distortion measurable. I think this works by allowing some of the negative feedback to enter the non-inverting input. So if I used the method in your link, wont this also increase distortion dramatically as I lower the volume?
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
No. You wouldn't connect it to an op-amp this way.
You didn't answer my question; do you need to convert the balanced inputs to single-ended? If so, here are some ideas:
http://www.diyaudio.com/forums/showthread.php?postid=1564153#post1564153
You didn't answer my question; do you need to convert the balanced inputs to single-ended? If so, here are some ideas:
http://www.diyaudio.com/forums/showthread.php?postid=1564153#post1564153
paulb said:
Hi,
Yes the DAC has a differential output and my amps only have a single ended input so it needs to sum the DAC output.
Could you expand on exactly how that type of attenuator would be connected to an op-amp? A quick diagram would help a lot. Would it not be an issue because I would be connecting the op-amp output to the inverting input directly with no resistor for a gain of 1? So the distortion gain would still just be 1? Distortion Gain = 1 + (0 / whatever) is always going to equal 1.
Thanks.
Hi Tenson,
That part of the data sheet is only there to tell you how they managed to get measurements from such a quiet op amp. As Paul stated, you would not use the part that way.
With a current output DAC, the resistance determines how much output voltage is generated. You balance between getting a usable signal and clipping (voltage compliance).
You normally do not want to load a voltage output DAC down. This increases the output current and heats the chip up. You may end up with thermal gradients (a bad thing) or overheating (really bad). Normally the distortion will increase over what is nominal for the part. There is a difference between what you should do, and what you can do.
-Chris
That part of the data sheet is only there to tell you how they managed to get measurements from such a quiet op amp. As Paul stated, you would not use the part that way.
Okay. I was speaking in very general terms considering what is I guess a much older crop of DACs. Your DAC has a very low actual output impedance if it can drive a 600 line directly. A 10:1 ratio is commonly used at the minimum. Yours might be lower.
With a current output DAC, the resistance determines how much output voltage is generated. You balance between getting a usable signal and clipping (voltage compliance).
You normally do not want to load a voltage output DAC down. This increases the output current and heats the chip up. You may end up with thermal gradients (a bad thing) or overheating (really bad). Normally the distortion will increase over what is nominal for the part. There is a difference between what you should do, and what you can do.
-Chris
anatech said:
Hi Chris,
I realise that, but by adding a variable resistor between the + and - phase before the op-amp (as hinted at here - http://www.diyaudio.com/forums/showthread.php?postid=1159607#post1159607 )you would be doing a similar thing, wouldn't you? It may control the volume by leaking a certain amount to the opposite phase, but you will also leak some of the negative feedback to the non-inverting input.
However, as I realized, if I have a directly connected negative feedback loop for a gain of 1, the equation they gave for that distortion measuring trick always comes out as 1, because 0 divided by anything is always 0. On the other hand it might just be that the equation does not work when you have no feedback resistor.
That makes sense
Hi Tenson,
If you look up an "instrumentation amplifier", you will see that you can both combine your phases and set your output level with a single resistor for each channel. You will also prevent any loading to the DAC chip.
An "instrumentation amplifier" is a circuit configuration made up from basic op amps (and some resistors). You can also buy these already balanced in the form of a chip. This might be worth looking into for you. You can then drive a transformer with that if you really want to.
They also make fully differential op amps. I bought some, but have not used any yet.
-Chris
If you look up an "instrumentation amplifier", you will see that you can both combine your phases and set your output level with a single resistor for each channel. You will also prevent any loading to the DAC chip.
An "instrumentation amplifier" is a circuit configuration made up from basic op amps (and some resistors). You can also buy these already balanced in the form of a chip. This might be worth looking into for you. You can then drive a transformer with that if you really want to.
They also make fully differential op amps.
I bought some, but have not used any yet.-Chris
Hi,
I looked up instrumentation amplifier but I don't understand why I would need to buffer the inputs? Why can't I just do this?
I looked up instrumentation amplifier but I don't understand why I would need to buffer the inputs? Why can't I just do this?
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
Hi Tenson,
The goal of an instrumentation amplifier is to present an equal load to both phases, or inputs. Secondly, it buffers the DAC from loading changes and allows you to correct the balance. The circuit you presented does none of these things (unless the control swamps out the normal input impedance). If it were truly that simple, an instrumentation amplifier would not exist.
Good designs are only as complicated as they need to be. This does not mean that a simple circuit is better, although sometimes they are. This is not one of those times.
-Chris
Well, obviously you are changing the loading on the DAC outputs. Tell me, what happens at minimum volume?
The goal of an instrumentation amplifier is to present an equal load to both phases, or inputs. Secondly, it buffers the DAC from loading changes and allows you to correct the balance. The circuit you presented does none of these things (unless the control swamps out the normal input impedance). If it were truly that simple, an instrumentation amplifier would not exist.
Good designs are only as complicated as they need to be. This does not mean that a simple circuit is better, although sometimes they are. This is not one of those times.
-Chris
You need to convert balanced to single-ended, and you need to control volume. I kind of like the idea with the instrumentation amplifier, one resistor sets the gain and you have the switch. But it takes a lot of op-amps, 3 per channel.
To convert to single-ended, here is an example device:
http://focus.ti.com/lit/ds/symlink/ina2134.pdf
(you could wire an op-amp to do this as well but you need 4 precision resistors).
You would use a conventional volume control after it, or the differential volume control ahead of it.
Also see:
http://sound.westhost.com/project51.htm
http://sound.westhost.com/project87.htm
To convert to single-ended, here is an example device:
http://focus.ti.com/lit/ds/symlink/ina2134.pdf
(you could wire an op-amp to do this as well but you need 4 precision resistors).
You would use a conventional volume control after it, or the differential volume control ahead of it.
Also see:
http://sound.westhost.com/project51.htm
http://sound.westhost.com/project87.htm
The input of this circuit (does not matters if it is for mic) it comes from National app. notes and works just fine. You can see an example of two buffers (U1A, U1B) placed in the route of balanced signal before the subtractor (U2A). A single pot (VR5) can adjust the level symmetricaly. You can modify this input for line level signal - with less gain - and you can substitute ICs with better.
This is the output noise of whole circuit.
An externally hosted image should be here but it was not working when we last tested it.
This is the output noise of whole circuit.
An externally hosted image should be here but it was not working when we last tested it.
anatech said:
Hi,
At minimum volume I would think the DAC will see the load of the series resistors, which if they were 1K each is fine for it. The only problem I can see is that at minimum volume equal amounts of negative feedback would go to the inverting and non-inverting inputs. Is that the issue?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.