Need advice on resistor values for a stereo line level to mono instrument level attenuator.
The stereo line level will come from the tape output RCA connectors of an Onkyo receiver. The mono level will go into a guitar recorder device.
Here's the circuit that I'm thinking about:
Would R1 and R2 = 5.6k, R3=1k be good values? Better resistor values?
The stereo line level will come from the tape output RCA connectors of an Onkyo receiver. The mono level will go into a guitar recorder device.
Here's the circuit that I'm thinking about:
Would R1 and R2 = 5.6k, R3=1k be good values? Better resistor values?
Last edited:
I would use 10k for R1 and R2.
For R3, the smaller the better, if it still gives enough output level.
The 1k for R3 would lose about 20dB of level.
For R3, the smaller the better, if it still gives enough output level.
The 1k for R3 would lose about 20dB of level.
If the output from the Onyyo is not separately buffered, that circuit will degrade the crosstalk between L & R channels for your stereo signal. Post the model number or link to a schematic and I can have a look at the tape out circuit.
20dB might not be enough attenuation, depending on the overload margin of the instrument input. The formula for attenuation in dB = 20 x log10((R1//R2)/R3). Note that the two output resistors R1 & R2 should be considered to be in parallel because the signals from L & R channels are being summed, adding 6dB when L = R. If R3 = 150, the attenuation in dB = 20 x log10(5000/150) = 30.5dB; R3 = 47Ω gives ~ -40db, R3 = 470 gives ~ -20dB. A smaller R3 will reduce the degradation in crosstalk; that means an attenuation that has the input control on the instrument amplifier at about ½ way is going to be around the optimum solution.
The better way to sum to mono is with an inverting opamp stage configured as a unity gain virtual earth mixer, followed by another unity gain inverting opamp to restore absolute phase, then attenuate the mono output to the level you require which will also attenuate the circuit noise.
20dB might not be enough attenuation, depending on the overload margin of the instrument input. The formula for attenuation in dB = 20 x log10((R1//R2)/R3). Note that the two output resistors R1 & R2 should be considered to be in parallel because the signals from L & R channels are being summed, adding 6dB when L = R. If R3 = 150, the attenuation in dB = 20 x log10(5000/150) = 30.5dB; R3 = 47Ω gives ~ -40db, R3 = 470 gives ~ -20dB. A smaller R3 will reduce the degradation in crosstalk; that means an attenuation that has the input control on the instrument amplifier at about ½ way is going to be around the optimum solution.
The better way to sum to mono is with an inverting opamp stage configured as a unity gain virtual earth mixer, followed by another unity gain inverting opamp to restore absolute phase, then attenuate the mono output to the level you require which will also attenuate the circuit noise.
Last edited:
Are your recommendations for mic level or instrument level?
My research so far indicates that the impedance of guitar outputs is high so should work into an impedance of about 1 meg ohms. So I assumed that R3 being 1K would be better than 500 or 100 ohms. Also that instrument level is about -30dB while mic level is about -50 to -60 dB.
The Onkyo receiver model number is TX-8011. I've searched the internet for a schematic without success.
My research so far indicates that the impedance of guitar outputs is high so should work into an impedance of about 1 meg ohms. So I assumed that the impedance of the guitar recorder that the circuit will be working into is about 1 meg ohm. With that, would R3 being 1K be better than 500 or 100 ohms? I need to knock the line level down to about -30 dB. What values of R1 and R2 would be appropriate?
Just need a simple way to sum to mono. Having OK audio quality will be fine. Pristine audio quality is not needed for this application but I will research the opamp method that you mentioned. It sounds interesting and something good to try to understand.
Thanks!
We are not talking low impedance voice microphones level by any means, so 20dB is way too much attenuation.
We are not talking straight Guitar pickups either, please read the OP 🙂
A common Tape out level is around 200mV, maybe 100mV in a weak one.
Standard Guitar preamps expect between 50mV and 200mV signals, suppose that "Guitar Recorder" will expect about same.
So ranges almost overlap, on average 2:1 ratio or so, or around 6dB, and not more; I would make all 3 resistors same value.
Personally would use all 22k ones, if anything to make it easier to drive by unknown Onkyo tape out (an old unwritten standard was to design for 10k loads, many preamps and Line/Tape/Aux outs mentioned this) and IF a factor, to minimize crosstalk as feared in post #3
Actual testing rules, but I guess those values will work fine.
If too loud (which I doubt), you can always lower R3 to 10k and even 4k7
We are not talking straight Guitar pickups either, please read the OP 🙂
A common Tape out level is around 200mV, maybe 100mV in a weak one.
Standard Guitar preamps expect between 50mV and 200mV signals, suppose that "Guitar Recorder" will expect about same.
So ranges almost overlap, on average 2:1 ratio or so, or around 6dB, and not more; I would make all 3 resistors same value.
Personally would use all 22k ones, if anything to make it easier to drive by unknown Onkyo tape out (an old unwritten standard was to design for 10k loads, many preamps and Line/Tape/Aux outs mentioned this) and IF a factor, to minimize crosstalk as feared in post #3
Actual testing rules, but I guess those values will work fine.
If too loud (which I doubt), you can always lower R3 to 10k and even 4k7
You need to quantify:
At present you're just asking how long is a piece of string.
- the source level
- the source impedance
- the target level
- the target impedance.
At present you're just asking how long is a piece of string.
Mike 7, Juan (JMFahey) manufactures guitar amps, so his guidance is good.
FWIW I just checked the input impedance of my two amps (a Vox and a Washburn) and both were 1MΩ. I measured the output of my Ibanez bass and my Stratocaster and both output ~50mV, whereas my two acoustic guitars (the Sigma has a Fishman preamp, the Greg Bennet 12-string is passive) output ~100+mV into 1MΩ, which is spot on what Juan said above. The 12-string has a very high output impedance and will not be happy driving anything much less than 1MΩ, however all of the other guitars with active outputs are probably fine with 100kΩ. But in any case the input impedance of your guitar amp has no bearing on the circuit you need.
Juan suggests trying 6dB of attenuation to start with. The only thing I would add is that CD players typically output 500mV, or about 6-12dB hotter than other source components. I agree that the minimum for R1 & R2 should be at least 10kΩ so as not to load down your sources. I'd probably still try R1 & 2 at 10kΩ and R3 at 1kΩ which will give ~20dB attenuation of Lch and Rch individually, equal to ~14dB attenuation of L & R added together. There is no exactly correct answer; just general principles to avoid problems.
A virtual earth opamp based mixer is a very simple device. If you are interested in exploring how an active mixer works have a read here: https://www.electronics-notes.com/a...amp/virtual-earth-mixer-summing-amplifier.php
FWIW I just checked the input impedance of my two amps (a Vox and a Washburn) and both were 1MΩ. I measured the output of my Ibanez bass and my Stratocaster and both output ~50mV, whereas my two acoustic guitars (the Sigma has a Fishman preamp, the Greg Bennet 12-string is passive) output ~100+mV into 1MΩ, which is spot on what Juan said above. The 12-string has a very high output impedance and will not be happy driving anything much less than 1MΩ, however all of the other guitars with active outputs are probably fine with 100kΩ. But in any case the input impedance of your guitar amp has no bearing on the circuit you need.
Juan suggests trying 6dB of attenuation to start with. The only thing I would add is that CD players typically output 500mV, or about 6-12dB hotter than other source components. I agree that the minimum for R1 & R2 should be at least 10kΩ so as not to load down your sources. I'd probably still try R1 & 2 at 10kΩ and R3 at 1kΩ which will give ~20dB attenuation of Lch and Rch individually, equal to ~14dB attenuation of L & R added together. There is no exactly correct answer; just general principles to avoid problems.
A virtual earth opamp based mixer is a very simple device. If you are interested in exploring how an active mixer works have a read here: https://www.electronics-notes.com/a...amp/virtual-earth-mixer-summing-amplifier.php
Really great info John. Thanks so much for taking the time to check the input impedance and voltage outputs of your instruments. I will start off with R1 & R2 as 10k and R3 at 1k. Will adjust R3 if necessary to get a good level for the equipment used. Will report results when the issue is settled.
Thanks again so much!
Mike 7
Thanks again so much!
Mike 7
This does not compute...the guitar recorder has a fixed noise floor, and you get the best S/N ratio by feeding in the strongest signal it can handle without clipping. This is one of the basic precepts of "gain staging" any audio chain to get the best S/N ratio from it.
In other words, you want the largest value of R3 that still keeps the recording device out of clipping. Not the smallest.
But I agree with Fahey...make all three resistors 10k to start with. Lower R3 only if the recording device is actually clipping on peaks.
-Gnobuddy
"For R3, the smaller the better, if it still gives enough output level."
The input noise may be dominated by Johnson noise. Because the level of Johnson noise is determined by the equivalent parallel resistance of source and load, a 1MΩ input is noisier terminated with a 100kΩ source than when terminated with a 1kΩ source impedance. Whether Johnson noise is what sets the noise floor in any given circuit is another question.
The service manual is available here: https://elektrotanya.com/onkyo_tx-8011_sm.pdf/download.html
Like most consumer receivers, the tape outputs are not separately buffered, but simply the input selected routed through a switch to the record output sockets, in this case an IC analogue switch Q302: TC9273N-010. Therefore the issue with crosstalk degradation is real.
Like most consumer receivers, the tape outputs are not separately buffered, but simply the input selected routed through a switch to the record output sockets, in this case an IC analogue switch Q302: TC9273N-010. Therefore the issue with crosstalk degradation is real.
Last edited:
Thank you for finding the schematic for the TX-8011. I downloaded it for future use in case it develops a problem in the future. Knock on wood. Looks like using 10k resistors for R1 & R2 will be a good thing.
If you meant to say "Reduce R1, R2, and R3 all to low but still practical values", then I quite agree.
Your earlier post read as though you were suggesting that only R3 should be reduced. That would increase signal attenuation, and worsen signal to noise ratio. Not a good plan.
I typed up a reply to this days ago, and thought I had posted it too. I guess I didn't click on the "Post reply" button properly.
-Gnobuddy
Of course you are correct, there is no point in using extra gain to make up excess attenuation.