Hi,
I'm interested in building a balanced line level splitter for my home recording studio. I need it so I can monitor the input signal without the problem of computer-related latency. I've come across a couple of designs - one very simple and another more complex. I'm unclear as to the benefits of the more complex design. I was wondering if someone here could explain the additional features. The two splitters are:
Paul In The Lab: Buffered Y Splitter Stripboard Veroboard
The one described in the Jensen AS041 design note (you need to register to see this one - sorry).
Also, the device will be in the recording path so I was wondering how much difference there would be between different opamps - the simple splitter uses a TL072. How would this compare with something like an OPA1662? Would there be any noticeable difference?
Cheers,
Chris
I'm interested in building a balanced line level splitter for my home recording studio. I need it so I can monitor the input signal without the problem of computer-related latency. I've come across a couple of designs - one very simple and another more complex. I'm unclear as to the benefits of the more complex design. I was wondering if someone here could explain the additional features. The two splitters are:
Paul In The Lab: Buffered Y Splitter Stripboard Veroboard
The one described in the Jensen AS041 design note (you need to register to see this one - sorry).
Also, the device will be in the recording path so I was wondering how much difference there would be between different opamps - the simple splitter uses a TL072. How would this compare with something like an OPA1662? Would there be any noticeable difference?
Cheers,
Chris
Last edited:
#1 is admirably simple, I don't think you could do it with any component less.
I see a number of issues with it though:
You may have difficulty procuring 10 megohm resistors, plus that section of the circuit is likely to be extremely sensitive to any kind of contamination. I mean c'mon, 10 megohms is like what, a typical multimeter input?
A TL072 at unity gain can use an output isolation resistor of as large as 470 ohms to fend off capacitive loading such as provided by longer cables. Really wimpy output stage. On a +9V supply, distortion is only going to be moderately good either, due to common-mode distortion effects. OPA1642 and similar should fare much better.
There is no supply bypassing at all. TL072s aren't fussy but that's a bit extreme.
Before going into the details of circuits, you should really specify what you need though. "Home studio" sounds like you might in fact require a balanced splitter rather than this unbalanced one. Besides, some audio interfaces already implement "zero latency monitoring".
I see a number of issues with it though:
You may have difficulty procuring 10 megohm resistors, plus that section of the circuit is likely to be extremely sensitive to any kind of contamination. I mean c'mon, 10 megohms is like what, a typical multimeter input?
A TL072 at unity gain can use an output isolation resistor of as large as 470 ohms to fend off capacitive loading such as provided by longer cables. Really wimpy output stage. On a +9V supply, distortion is only going to be moderately good either, due to common-mode distortion effects. OPA1642 and similar should fare much better.
There is no supply bypassing at all. TL072s aren't fussy but that's a bit extreme.
Before going into the details of circuits, you should really specify what you need though. "Home studio" sounds like you might in fact require a balanced splitter rather than this unbalanced one. Besides, some audio interfaces already implement "zero latency monitoring".
Hi, yes I mentioned that I wanted the splitter to handle a balanced signal. I was thinking that I could use two of the simple splitters to handle the two parts of the balanced signal (but maybe that's a bit naive?).
Yes, my interface has zero latency monitoring but there are two problems. I want to put some "comfort" effects on the headphone monitoring using a small mixing desk and some outboard effects. Firstly, I can't apply reverb to the input (zero-delayed) signal without it getting on the backing too. Also, the zero-delay output is for phones and I was under the impression that this won't work with a line-level effects unit like a compressor. The compressor would have to be inserted pre-mixer as the mixer doesn't have inserts.
Thanks for the explanation though.
That would probably work alright, even if purists would probably prefer a "native" differential part like OPA6132.
For practical CMRR concerns it is important that CM input impedance be kept as high as possible. CMRR would also take care of PSRR in a balanced case. Instead of the problematic 10M resistors, I'd suggest something more like a 100k/(100k||1µF) filtered supply voltage divider followed by a 1 Megohm series resistor or somesuch for input bias. (And probably something more akin to an OPA1652 rather than a TL072, with 100 nF + 10 µF across supply, and a 100-220 ohm output series resistor, maybe 120 or so.) A bipolar input part may require going lower.
What are you planning to use for a power supply? While a 9V block single supply circuit should be able to take medium levels, something like a mains-powered +/-15 V split supply would obviously be nicer. Then your voltage divider would no longer be required either.
BTW, I have so far assumed line-level signals.
Thanks for the info but you've already surpassed my level of electronics knowledge!
What I was hoping was that someone could explain the additional features in the Jensen splitter particularly wrt what they add compared to the simple version so then I can work out if I need them.
Another reason I want a splitter is so that I can record (balanced, line-level signal) through effects but also record a dry version just in case.
9V battery limits your maximum signal to about 3Vac unbalanced. That's just +11dBu. that's very low for professional levels.
If you maintain balanced impedance and maintain equal opposite phase throughout that will give you an extra 6dB of signal. i.e. +17dBu. Much closer to the +20dBu to +24dBu levels generally allowed for.
If you maintain balanced impedance and maintain equal opposite phase throughout that will give you an extra 6dB of signal. i.e. +17dBu. Much closer to the +20dBu to +24dBu levels generally allowed for.
What is the model of equipment that is the output (I presume some mixer), and what are the two devices you want this signal to go into (it appears one is a computer interface, the other is a headphone amp or another mixer used to drive headphones)? The manuals for these probably have specs for input and output impedances.
Virtually all equipment for the last several decades uses voltage drive (low impedance output, high impedance input) for line level signals, and an output can usually drive over ten inputs before there's any problem such as loading the output. So doing this could be as simple and easy as making a balanced Y cable.
I recall a basic rule that (in voltage transfer equipment) the input impedance "seen" by an output should be at least 10 times the output impedance (for a signal drop of less than some fraction of a dB). Output impedance might be 100 or 600 ohms. Input impedance of one device might be 10k, and the other 47k. The total impedance the output will see driving the inputs in parallel will be 1 / (1/10 + 1/47) = 8.2k, surely well over 10 times the driving device's output impedance.
Hi, yes, you're basically right. Output is from the infamous $5 mic preamp:
Open Chat - Microphone preamp vero layout
I want to split this and send it to an Edirol UA-101 (sometimes via a DBX166XS) for recording and also to a Behringer MX1602A mixer for monitoring.
I've had a look through the manuals for both pieces of gear and can't find any mention of their input impedence.
- Status
- Not open for further replies.