• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Measuring Speaker and Headphone Outputs With Focusrite Scarlet 2i2 v3

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
I would like to measure the output of some tube amps using my 2i2. I am considering getting an oscilloscope too (<$500), but I don't really know if it will be useful.

I am wondering if I have a good dummy load circuit in mind, and how to calculate the values of the series resistors for the measurement parallel circuit.

I want to do a sine wave, multi-tone, etc. I'm pretty sure I won't have any trouble getting tests signals into the amps.

I was thinking about getting an oscilloscope with a budget of about $500 retail, so I would like advice on if that is likely to compliment my 2i2 in this application, and if so, which scope would be recommended.

I'll post pictures from my phone in a moment.

Thanks in advance for the help!
 
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
Here are the circuits i am evaluating.
 

Attachments

  • 20200902_174534.jpg
    20200902_174534.jpg
    947.4 KB · Views: 320
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
Here are the amps. Shuguang audio SG-845-7 SET amp and Darkvoice 336SE
 

Attachments

  • 20200902_174440.jpg
    20200902_174440.jpg
    1.2 MB · Views: 171
  • 20200902_174401.jpg
    20200902_174401.jpg
    3.2 MB · Views: 181
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
Sorry if my TS-M connector drawings are too masculine.

Each dummy load would be duplicated; for the speaker load, the whole thing; for the headphone load, the part right after the headphone output connector so it can Y to the two inputs on the 2i2.

I know the series resistors for the measurement circuit are going to be different for TS vs XLR because of how the 2i2 treats them differently, so I'd like to know if XLRs are going to be better, or if TS connectors are better going into the 2i2.
 

Blumlein 88

Grand Contributor
Forum Donor
Joined
Feb 23, 2016
Messages
20,524
Likes
37,057
Sorry if my TS-M connector drawings are too masculine.

Each dummy load would be duplicated; for the speaker load, the whole thing; for the headphone load, the part right after the headphone output connector so it can Y to the two inputs on the 2i2.

I know the series resistors for the measurement circuit are going to be different for TS vs XLR because of how the 2i2 treats them differently, so I'd like to know if XLRs are going to be better, or if TS connectors are better going into the 2i2.
You probably want to use the TRS into the 2i2 because the XLR goes into the microphone section. The Line section would handle higher voltages.

What you have drawn for speakers actually would not work to show you anything.

You could have your dummy load. In parallel with it you can have a resistor divider pair and tap it to get voltage that is low enough not to damage the 2i2 line inputs. For the low powered amps for instance you could parallel a series pair of 1 kohm and 100 ohm. Tap across the 100 ohm resistor with the inputs to the 2i2. Multiply by 11 to get the real value. You would be safe to amps of almost 60 watts with this setup. You would need to change the resistor pair for higher power items not to damage the 2i2.
 

P_M

Active Member
Joined
May 31, 2019
Messages
131
Likes
51
Location
San Francisco Bay Area, CA
@Helicopter
you need a specturm analyszer to test amp performance, not an oscilloscope. For some specific type of troubleshooting where you might want to see the waveform like for crossover distortion in AB or push-pull amps or flat topping of waves or ringing then an oscilloscope would come handy.

In any case, I recommend you install a basic test software like ARTA or REW and read the instructions that come with it. That will help you setup a simple jig for measurements. As for dummy speaker load I think amirm uses one, just search the forum.
 

GeorgeBynum

Active Member
Joined
Feb 14, 2020
Messages
121
Likes
103
Location
Greenville SC USA
@Helicopter
As for dummy speaker load I think amirm uses one, just search the forum.
Although not precision, I've used kitchen irons as dummy loads; a typical 1200 watt iron is 12 ohms in the USA (120V power). 2 of them in parallel give 6 ohms, a reasonable real load. NOTE use is intermittent; you DO NOT want the thermostat opening with the amp at full power.

I personally had access to scrapped returns years ago, and removed the thermostats on 2.

An electric stove element, commonly 4800W, will be 3 ohms in the USA.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,334
Likes
3,278
Location
.de
What you have drawn for speakers actually would not work to show you anything.

You could have your dummy load. In parallel with it you can have a resistor divider pair and tap it to get voltage that is low enough not to damage the 2i2 line inputs. For the low powered amps for instance you could parallel a series pair of 1 kohm and 100 ohm. Tap across the 100 ohm resistor with the inputs to the 2i2. Multiply by 11 to get the real value. You would be safe to amps of almost 60 watts with this setup. You would need to change the resistor pair for higher power items not to damage the 2i2.
Yes, the L-pad with two resistors is the way to go.
Actually, 4 - include a parallel of the same values between amplifier output ground and the "cold" line for balanced impedances.
1. out+ -> R1 -> ...
a) ... hot @ cable
+
b) .... R2 --> out-ground
2. out-ground --> R3 = R1||R2 --> cold @ cable

Note: Voltage equivalent to 60 W into 8 ohms will dissipate over 0.4 W in 1.1 kOhms. I'd go a tad higher in values in order to be able to use ordinary 1/4W metal film types (which in turn shouldn't be seeing more than half rated power), maybe a factor of 5-10. Do keep R1||R2 well below 1 kOhm if you can - make up R1 from two parallel or series if need be.
(P_total = V² / (R1 + R2) = P_amp_8ohm * 8 / (R1 + R2); P_R1 = P_total * R1/(R1+R2), P_R2 analogous.)

Note 2: You need a slightly different attenuator setup when measuring amplifiers with bridged (BTL) outputs, so you must always know whether the amplifier's negative output is ground referenced or not.

For a BTL output, you would need a double (balanced) L-pad:
1. out+ -> R1 -> ...
a) ... hot @ cable
+
b) .... R2 --> shield @ cable
2. out- -> R3=R1 -> ...
a) ... cold @ cable
+
b) .... R4=R2 --> shield @ cable

This should actually simplify to the non-BTL version if out- = out-ground connects to shield @ cable via the audio interface. So I guess this one could be used for both kinds of outputs.

Note 3: I don't actually see a problem with using the mic input assuming attenuation is increased accordingly (e.g. R1 = R3 = 15k || 15k, R2 = R4 = 470 ohm, should be good for max +9 dBu from a 200 Wpc output). It is not unusual to be seeing a line input being created internally by just putting some e.g. 22 kOhm resistors in series with the mic input, which in turn has a 2-3 kOhm input impedance and forms an L-pad this way. Our external L-pad with a couple hundred ohms of output impedance would be less noisy for one. If in doubt, evaluate both. The device specs may give some hints.
 
Last edited:
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
It does have the best specs with the mic input, so I can use that.

MICROPHONE INPUTS
Frequency Response 20Hz - 20kHz ± 0.1dB
Dynamic Range 111dB (A-weighted)
THD+N<0.0012%
Noise EIN -128dBu (A-weighted)
Maximum Input Level 9dBu (at minimum gain)
Gain Range 56dB
Impedance 3kΩ

I think that means about 2 volts to fry it with the input volume all the way down and 1 mV with the volume up. I'm not sure if that's right because I don't know if the 9bBu is referenced to 300Ω or 3kΩ, and I didn't check my math.

4 resistors... do you mean like this?

20201001_171941.jpg


Do I need another resistor on the bridge between +/-?

If not I could do 1 circuit like the schematic for each channel, with 4 ohm resistors (100W, irons, something like that) and then maybe pairs of 3 500 ohm metal films so the amp sees about 8 ohms and you get 3Kohm limiting current into the 2i2.

I have 4 100W 4 ohm resistors, so it would be great if the circuit without a resistor on the bridge would work, but I don't need to make it work with the stuff I have. I'm pretty sure there's a Parts Express order between me and this dummy load anyway.

Does the input of the 2i2 actually behave like a 3 Kohm resistor, drawing only half the current it would if you shorted the XLR after 3Kohms of resistors?

----

I found Amir's Darkvoice review

https://www.audiosciencereview.com/...ements-of-darkvoice-336se-headphone-amp.6549/

and his plans to review the Loxjie P20

https://www.audiosciencereview.com/...-be-reviewed-by-amir.4800/page-20#post-458293

Now I am more interested in measuring the Shuguang SG-845-7 with this load.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,334
Likes
3,278
Location
.de
I think that means about 2 volts to fry it with the input volume all the way down and 1 mV with the volume up. I'm not sure if that's right because I don't know if the 9bBu is referenced to 300Ω or 3kΩ, and I didn't check my math.
+9 dBu always is 2.18 Vrms whatever happens.

I would expect best distortion performance for a gain in the 20-30 dB range (it typically degrades near unity somewhat). So you could try R1 = 22 kOhm, R2 = 330 ohm or thereabouts.
4 resistors... do you mean like this?
Nyet.

Your "attenuator" makes no sense at all like that. (Homework: Explain why.)

You know how an XLR cable is normally wired up, right?
Hot --> pin 2
Cold --> pin 3
Shield --> pin 1

The attenuator is arguably best drawn with conductors in the order of:
Hot
Shield
Cold

Now try that with the pseudo-schematic given earlier, which I'll also paraphrase in plain German:
Output (+) connects to R1 in series. After R1, the Hot connection continues, and R2 branches off going to Shield.
Output (-) does much the same, except it's the Cold connection continuing this time.
Yes, the attenuator has two input connections and three output connections. Shield does not actually go all the way to the amplifier.

Your dummy load always just connects directly to the output, in parallel to the attenuator. It can be left off in testing as long as the amplifier does not require a load for stability (this is generally not an issue but can be in obscure low feedback tube designs).

You might consider making a small circuit board to not only hold the hold the attenuator but also connect the dummy load and facilitate 4/8 ohm switching. Something with sturdy wire terminals for input, male XLR jacks for output, and wires dangling off to connect to your dummy load(s).
 

somebodyelse

Major Contributor
Joined
Dec 5, 2018
Messages
3,682
Likes
2,962
An example with pictures showing a simple dummy load and attenuator with an RCA connection. this thread covers similar ground, but with a switched attenuator for easier use at different signal levels. The series of articles starting here is worth a look too - part 4 covers some of the interface options available to protect your soundcard, and part 6 has an example balanced attenuator that you could use in place of the single ended one in the first link.
 
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
+9 dBu always is 2.18 Vrms whatever happens.

I would expect best distortion performance for a gain in the 20-30 dB range (it typically degrades near unity somewhat). So you could try R1 = 22 kOhm, R2 = 330 ohm or thereabouts.

This confirms my understanding of dBu. Good point on optimizing attenuation at expected performance peak.

+9 dBu always is 2.18 Vrms whatever happens.
Nyet.

Your "attenuator" makes no sense at all like that. (Homework: Explain why.)

I didn't think it looked right. I was trying to understand where you would put 4 resistors. With no resistor on the link from + to -, you short the output to the 2i2 and I don't think you would actually divide the voltage. So you would just be attenuating current, shorting, and attenuating some more; the bigger problem is the short, which is why it wouldn't work at all.

You know how an XLR cable is normally wired up, right?
Hot --> pin 2
Cold --> pin 3
Shield --> pin 1

Yes.

The attenuator is arguably best drawn with conductors in the order of:
Hot
Shield
Cold

Now try that with the pseudo-schematic given earlier, which I'll also paraphrase in plain German:
Output (+) connects to R1 in series. After R1, the Hot connection continues, and R2 branches off going to Shield.
Output (-) does much the same, except it's the Cold connection continuing this time.
Yes, the attenuator has two input connections and three output connections. Shield does not actually go all the way to the amplifier.

Your dummy load always just connects directly to the output, in parallel to the attenuator. It can be left off in testing as long as the amplifier does not require a load for stability (this is generally not an issue but can be in obscure low feedback tube designs).

You might consider making a small circuit board to not only hold the hold the attenuator but also connect the dummy load and facilitate 4/8 ohm switching. Something with sturdy wire terminals for input, male XLR jacks for output, and wires dangling off to connect to your dummy load(s).

OK. Thank you.

An example with pictures showing a simple dummy load and attenuator with an RCA connection. this thread covers similar ground, but with a switched attenuator for easier use at different signal levels. The series of articles starting here is worth a look too - part 4 covers some of the interface options available to protect your soundcard, and part 6 has an example balanced attenuator that you could use in place of the single ended one in the first link.

Lots of good stuff here. Thank you.
 
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
4 resistors would work like this:

20201002_104244.jpg


Putting wires, lugs, or aligator clips for the 8 ohm resistor in this configuration should give versatility and allow the use of inexpensive metal film resistors for the 11K and 330 ohm resistors.

I think the result would be pretty similar with something like this:
20201002_104255.jpg


Although this would not give versatility, and all of the resistors would need to be high current.

Am I understanding the wiring correctly for XLRM Socket, with hot (+) tied to common ground?

I know the voltage divider doesn't need to be balanced as one resistor on one side of the circuit limits current to both sides, but I prefer this layout, and testing this morning with a DC load indicated it shouldn't change the outcome compared to a traditional voltage divider. Just add the values of the 4 or 11K resistors to get your R1 value of 8 or 22K.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,334
Likes
3,278
Location
.de
I meant 4 resistors plus dummy load resistors, in case that wasn't clear.
Output (+) connects to R1 in series. After R1, the Hot connection continues, and R2 branches off going to Shield.
Output (-) does much the same, except it's the Cold connection continuing this time.
Yes, the attenuator has two input connections and three output connections. Shield does not actually go all the way to the amplifier.
So there is two of R1 and R2 each for the hot and cold side, respectively.

EDIT: OK, screw it, here's a schematic:
spk2xlr.png
 
Last edited:
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
I meant 4 resistors plus dummy load resistors, in case that wasn't clear.

So there is two of R1 and R2 each for the hot and cold side, respectively.

EDIT: OK, screw it, here's a schematic:
View attachment 85927
Ahh. That makes perfect sense. Put the shield right in the middle of the circiut.
 

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,334
Likes
3,278
Location
.de
Ahh. That makes perfect sense. Put the shield right in the middle of the circiut.
I mean, you obviously wouldn't build it looking quite like that since the hot and cold wires will be twisted (as would the input wiring, generally), but it's good for illustration. The real thing has to be kind of folded up.
 
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
I mean, you obviously wouldn't build it looking quite like that since the hot and cold wires will be twisted (as would the input wiring, generally), but it's good for illustration. The real thing has to be kind of folded up.
Yeah. I just mean conceptually, and electrically. I was trying to figure out what to do with the shield other than nothing, like those crummy commercial products in the other thread where you posted your drawing. Your schematic is an elegant solution because the shield remains at the center of the AC wave through its entire positive, to zero to negative path. If something is not perfectly matched on one side due to tolerance, interference, the shield should help there.

In short, it is in the circuit, but it will not draw any current, which is important.
 
Last edited:

AnalogSteph

Major Contributor
Joined
Nov 6, 2018
Messages
3,334
Likes
3,278
Location
.de
In short, it is in the circuit, but it will not draw any current, which is important.
Exactly, since shield in general already connects to the amplifier's input side.

Incidentally, in case of a single-ended output (OUT - = SHIELD), the attenuator effectively simplifies to an unbalanced attenuator with an impedance balancing resistor in the cold lead.
 
OP
Helicopter

Helicopter

Major Contributor
Forum Donor
Joined
Aug 13, 2020
Messages
2,693
Likes
3,944
Location
Michigan
Finally made this attenuator box. I went 15kOhm/470Ohm for R1&R2/R3&R4. I made 2 channels with 3 SpeakOn connectors in parallel on each side so I can connect an amp, speaker or other load, and a multimeter if I want.

20210722_174329.jpg

20210722_174301.jpg


Next time I need to do D mounts in a project box I will use a vise... some of the big 15/16 holes got a little elongated.
 

Attachments

  • 20210720_112916.jpg
    20210720_112916.jpg
    163 KB · Views: 87
Top Bottom