US-144 Calibration Question...

[etc6849]

I just bought the US-144 and the ECM-8000 mic just to use the wonderful REW software on my xp based laptop :-) I'm a newbie, so please excuse any of my ignorance. I'm a pretty technical minded engineer though.

I had a few questions about calibrating the US-144 to determine if I need to order any additional cables:

1. I plan to use an XLR mic cable from the ECM8000 to the XLR input of the US-144. Will calibration through the left or right 1/8" analog input/output channels be good enough for the soundcard calibration? Obviously, there is some mic pre-amp circuitry that would be unaccounted for in the calibration right?

To account for the mic pre-amp circuitry in the US-144, should I use the soundcards analog output and feed this through the XLR input on the US-144 with phantom power off? Do you think this would increase the accuracy of the measurements or would it be a waste of money to buy a XLR to rca cable just for calibrating this way?

2. Does REW or others support response plots using an SPDIF output? If so, how would you calibrate this type of set up? For this I would use the SPDIF out on the US-144 or SPDIF on a soundcard combined with the XLR input on the US-144.

3. What about running a test sweep through 6 analog channels of a ASUS Xonar card and using the US-144 as the mic input with one microphone at my main listening position? Is this possible or a good idea? I don't have a mixer or anything to provide phantom power is why I'm asking.

I'm interested in this type of all channel test to tell what the response might be if I'm using all channel stereo on my receiver versus response for 2 channel stereo. I'm assuming it's easy for the software to accurately combine six individual response plots together, is this how a measurement like this is done?

Thanks in advance

[Wayne A. Pflughaupt]

etc6849 wrote:

1. I plan to use an XLR mic cable from the ECM8000 to the XLR input of the US-144. Will calibration through the left or right 1/8" analog input/output channels be good enough for the soundcard calibration? Obviously, there is some mic pre-amp circuitry that would be unaccounted for in the calibration right?

Yes. Our general recommendation has typically been to calibrate the regular line in/out and hope that the mic pre amp is flat. However, I personally don’t see why the calibration routine can’t be run from the mic in to the line out. In fact, it seems to me like it needs to be. Maybe John can shed some light.

To account for the mic pre-amp circuitry in the US-144, should I use the soundcards analog output and feed this through the XLR input on the US-144 with phantom power off? Do you think this would increase the accuracy of the measurements or would it be a waste of money to buy a XLR to rca cable just for calibrating this way?

Yes, turn off the phantom power. As to whether or not it would be a waste of money for the cable, maybe you can be the first to give it a go!

2. Does REW or others support response plots using an SPDIF output? If so, how would you calibrate this type of set up? For this I would use the SPDIF out on the US-144 or SPDIF on a soundcard combined with the XLR input on the US-144.

I’ll defer that one to someone else more knowledgable.

3. I'm interested in this type of all channel test to tell what the response might be if I'm using all channel stereo on my receiver versus response for 2 channel stereo. I'm assuming it's easy for the software to accurately combine six individual response plots together, is this how a measurement like this is done?

Typically you want to measure one speaker at a time, with all processing turned off. But if you’re just looking to see what kind or response you’re getting with all channels running, just split the soundcard’s output to an L/R input, and then turn on the receiver’s all-channel stereo function.

Regards,
Wayne

[etc6849]

Thanks Wayne. I'm new to this stuff, but I'll give your suggestions to numbers 1 and 3 soon. I really like your idea for number 3. I'm also looking forward to doing tests with audyssey multiEQ on/off and see if audyssey really corrects for my room or not.

I'm still not sure if the US-144 limits the input frequency range for the mic inputs, but I know the FAQ's said something about never using a mic input on a laptop with REW. I'd think the XLR inputs on the tascam US-144 are much better mic input with fully range. I've also read a lot of people on avsforum are using this type of set up for portability. I couldn't find any specs for the US-144 mic XLR inputs other than these:

Audio specifications
Nominal input levels
MIC IN L and R (XLR balanced) –58 dBu (TRIM=max) to –14 dBu
(TRIM=min)
MIC/LINE-GUITAR L and R in
MIC/LINE position (1/4” jack, unbalanced)
–40 dBu (TRIM=max) to +4 dBu
(TRIM=min)
MIC/LINE-GUITAR R in GUITAR
position (1/4” jack, unbalanced)
–51 dBu (TRIM=max) to –7 dBu
(TRIM=min)
Maximum input levels
MIC L and R (XLR balanced) +2 dBu (TRIM=min)
MIC/LINE-GUITAR L and R in
MIC/LINE position (1/4” jack, unbalanced)
+20 dBu (TRIM=min)
MIC/LINE-GUITAR R in GUITAR
position (1/4” jack, unbalanced)
+9 dBu (TRIM=min)
Input impedance
MIC IN L and R (XLR balanced) 2.4 kΩ
MIC/LINE-GUITAR L and R in
MIC/LINE position (1/4” jack, unbalanced)
10 kΩ
MIC/LINE-GUITAR R in GUITAR
position (1/4” jack, unbalanced)
1 MΩ

[JohnM]

Wayne A. Pflughaupt wrote:

Our general recommendation has typically been to calibrate the regular line in/out and hope that the mic pre amp is flat. However, I personally don’t see why the calibration routine can’t be run from the mic in to the line out. In fact, it seems to me like it needs to be. Maybe John can shed some light.

There are 2 difficulties with calibrating a mic input. The first is level, there would need to be substantial attenuation between the line output and the mic input to prevent massive over-driving of the input. The second is the difference in impedances of the sources, i.e. the low impedance line output versus a high impedance mic, which may affect the frequency response. It is possible to make a cal measurement with an attenuator in front of the mic input (that's how I determined the grim response of the laptop mic input I tested) but it is probably only useful to see if the input has gross roll-offs at low and/or high frequencies.

Does REW or others support response plots using an SPDIF output? If so, how would you calibrate this type of set up? For this I would use the SPDIF out on the US-144 or SPDIF on a soundcard combined with the XLR input on the US-144.

We don't generally recommend using an SPDIF output because of the calibration problem. One possibility is to loop back from an analog output of the receiver, but that can be complicated by the effects of crossover filtering - would need to ensure the output being used was configured for a full range speaker. Easiest is to just use an analog out from the soundcard.

[Wayne A. Pflughaupt]

JohnM wrote:

There are 2 difficulties with calibrating a mic input. The first is level, there would need to be substantial attenuation between the line output and the mic input to prevent massive over-driving of the input. The second is the difference in impedances of the sources, i.e. the low impedance line output versus a high impedance mic, which may affect the frequency response. It is possible to make a cal measurement with an attenuator in front of the mic input (that's how I determined the grim response of the laptop mic input I tested) but it is probably only useful to see if the input has gross roll-offs at low and/or high frequencies.

How 'bout if the soundcard (or USB interface, as I've seen some call them) had a balanced output? Could you do a mic pre-amp calibration then?

Regards,
Wayne

[JohnM]

Still have the same level problem Wayne, the mic input will have anywhere between 20 and 40dB of gain so the line output would need a corresponding amount of attenuation.

[etc6849]

JohnM wrote:

Still have the same level problem Wayne, the mic input will have anywhere between 20 and 40dB of gain so the line output would need a corresponding amount of attenuation.

Thanks John and Wayne for the interesting discussion.

Looking at the relevant specs of the US-144 I see:
Nominal input levels
MIC IN L and R (XLR balanced) –58 dBu (TRIM=max) to –14 dBu
(TRIM=min)
Maximum input levels
MIC L and R (XLR balanced) +2 dBu (TRIM=min)
Input impedance
MIC IN L and R (XLR balanced) 2.4 kΩ

and

Nominal output level
LINE OUT (RCA unbalanced) –10 dBV
Maximum output level
LINE OUT (RCA unbalanced) +6 dBV
Output impedance
LINE OUT (RCA unbalanced) 100 Ω

I think it's easy to mathematically show that I can/can't run a line from the line out to the mic input without clipping the mic input using the data above and the definitions of dBu and dBV (not the same as dBv). Once I figure out how to convert dBu to dBV I would know whether or not connecting them is a good idea or not. Using definitions found on wikipedia, it looks easy to do this. My thought is if the maximum dBV output from the US-144 is less than the mic maximum input (+2dBu), it's ok to do this, but may cause clipping. I'm guessing the best check would be to compare the nominal values for each. At any rate, once I do this calculation, I would know how much attenuation I would need.

[brucek]

Well, it shows the maximum output line-out level to be ~ 2 voltsRMS (+6dBV), and the maximum input mic level to be ~ 1 voltRMS (+2dBu).

By those max values you would need to pad the input by 1/2, but I would go further than that (at least -25dB - you could require down to -40dB).

brucek

[etc6849]

Sorry I can't post the link (I don't think newbies are allowed to), but if you go on analog devices website they have a calculator that converts dBu to dBV. Converting the nominal output of -10dBV to dBu, I get: -7.782dBu. So since the nominal mic input with trim all the way down is -14dBu, I would need atleast -6.218dBu of attenuation (minimum). I would think you'd want to get in the middle of the -58dBu to -14dBu range anways, so add another 30dBu of attenuation and you get pretty close the attenuation John suggests.

Interesting though, -7.782dBu is within the max mic input of +2dBu, so I don't think it'll hurt anything to try it once, but what is the "nominal" volume level for the output? Where should my output volume be when I run these tests? any thoughts? My onkyo receiver says the rec out is 200mV@470 ohms, if I convert this to dBu I get -11.76 dBu so maybe this is the better way to go (assumes onkyo specs are in volts rms and a sinewave).

etc6849 wrote:

Thanks John and Wayne for the interesting discussion.

Looking at the relevant specs of the US-144 I see:
Nominal input levels
MIC IN L and R (XLR balanced) –58 dBu (TRIM=max) to –14 dBu
(TRIM=min)
Maximum input levels
MIC L and R (XLR balanced) +2 dBu (TRIM=min)
Input impedance
MIC IN L and R (XLR balanced) 2.4 kΩ

and

Nominal output level
LINE OUT (RCA unbalanced) –10 dBV
Maximum output level
LINE OUT (RCA unbalanced) +6 dBV
Output impedance
LINE OUT (RCA unbalanced) 100 Ω

I think it's easy to mathematically show that I can/can't run a line from the line out to the mic input without clipping the mic input using the data above and the definitions of dBu and dBV (not the same as dBv). Once I figure out how to convert dBu to dBV I would know whether or not connecting them is a good idea or not. Using definitions found on wikipedia, it looks easy to do this. My thought is if the maximum dBV output from the US-144 is less than the mic maximum input (+2dBu), it's ok to do this, but may cause clipping. I'm guessing the best check would be to compare the nominal values for each. At any rate, once I do this calculation, I would know how much attenuation I would need.

[etc6849]

brucek wrote:

Well, it shows the maximum output line-out level to be ~ 2 voltsRMS (+6dBV), and the maximum input mic level to be ~ 1 voltRMS (+2dBu).

You would need to pad the input by 1/2, but I would go further than that.

brucek

My Onkyo 905 receiver says this for rec out:
Output Level and
Impedance 200 mV/ 470
Ω
(REC OUT)

Following the same lines, is it ok to use this as the output to feed to the mic? I believe this would be a lot better choice over the soundcards output. .2vrms < 1 vrms, but .2vrms is not less than -14dBu~.1546 v rms and -14dBu is the nominal value.

[brucek]

My Onkyo 905 receiver says this for rec out:.....................Following the same lines, is it ok to use this as the output to feed to the mic?

You're not feeding the rec out to the mic input. The need for the pad is for when you do the one-time soundcard calibration with a short cable from line-out to mic-in of the soundcard. (personally, I think it's not needed and you can simply do a line-out to line-in cal and it would be fine, but that's another story).

As I indicated above, the max out of the line-out is a couple volts. That fairly standard. The nominal line-output is about 300mvolts (-10dBV).
A mic in is very sensitive. It shows a max input level of 1 volt (+2dBu), but that's quite high and would normally be within the ranges as shown from 150mvolts (-14dBu) down to 1mvolt (-58dBu).
So if you needed a pad to go from 300mvolts down to 1mvolt, it would require a ~-50dB down. If you wanted to go from 300mvolts down to 150mvolts, it would require a ~-6dB down.

So, as I calculated on the back of a napkin before, you probably would get away with about a -25dB pad. (I can tell you the resistors you would need)

brucek

[Wayne A. Pflughaupt]

JohnM wrote:

Still have the same level problem Wayne, the mic input will have anywhere between 20 and 40dB of gain so the line output would need a corresponding amount of attenuation.

Like this? (Balanced output)

Regards,
Wayne

[brucek]

Like this?

Twitchy is the first word that comes to mind.....

brucek

[etc6849]

brucek wrote:

You're not feeding the rec out to the mic input. The need for the pad is for when you do the one-time soundcard calibration with a short cable from line-out to mic-in of the soundcard. (personally, I think it's not needed and you can simply do a line-out to line-in cal and it would be fine, but that's another story).

As I indicated above, the max out of the line-out is a couple volts. That fairly standard. The nominal line-output is about 300mvolts (-10dBV).
A mic in is very sensitive. It shows a max input level of 1 volt (+2dBu), but that's quite high and would normally be within the ranges as shown from 150mvolts (-14dBu) down to 1mvolt (-58dBu).
So if you needed a pad to go from 300mvolts down to 1mvolt, it would require a ~-50dB down. If you wanted to go from 300mvolts down to 150mvolts, it would require a ~-6dB down.

So, as I calculated on the back of a napkin before, you probably would get away with about a -25dB pad. (I can tell you the resistors you would need)

brucek

Thanks Bruce you have a great idea! What I was getting at is I could run the lineout from the soundcard into the receiver, then use the rec out to go into the mic input, then calibrate. However, I would still be over the max end of the nominal range for the mic input.

I think I'll follow your suggestion of a voltage divider and aim for an input of about -22 dBu (about 61.5 mVrms) which is the middle of the mics nominal range (or should I be using 1mVrms??). As you state, I will assume a 300mVrms from the output. When I do this, I get these values: 1kOhm and 3.9kOhm (using standard resistor values). The 1kohm will go between the mic signal and ground. Any thoughts? What values were you thinking for the voltage divider? Thanks again.

[Wayne A. Pflughaupt]

brucek wrote:

The need for the pad is for when you do the one-time soundcard calibration with a short cable from line-out to mic-in of the soundcard. (personally, I think it's not needed and you can simply do a line-out to line-in cal and it would be fine, but that's another story).

I dunno, I just have a hard time believing and trusting that we can expect linear response from a budget product. Am I behind the times? I mean, I know that technology "trickles down," but has it trickled down all the way to the bottom end?

Regards,
Wayne

[Wayne A. Pflughaupt]

brucek wrote:

Twitchy is the first word that comes to mind.....

I have no idea what that means, at least as it relates to a level attenuator. Can you help me out?

Regards,
Wayne

[brucek]

When I do this, I get these values: 1kOhm and 3.9kOhm (using standard resistor values). The 1kohm will go between the mic signal and ground. Any thoughts? What values were you thinking for the voltage divider?

Well, the 3.9K is fine (between the RCA output and mic input), but the 1K to ground is too high. This becomes your output impedance and would not work.
I suggest about a -32dB attenuation, which would be 3.9K (between the line-out and mic-in) and then a 100 ohm to ground (on the mic side of course). This would be about a -32dB pad (fairly typical for a mic pad), and would offer 100 ohm output impedance.

What I was getting at is I could run the lineout from the soundcard into the receiver, then use the rec out to go into the mic input, then calibrate.

Yeah but then you've added the response of the receiver lein-in to your cal file, which you don't want to do......

brucek

[brucek]

I have no idea what that means, at least as it relates to a level attenuator. Can you help me out?

When you are trying to control a very high gain stage with a pot, the amount of change for even the slightest movement of the pot would be huge. You would likely be working with the pot almost in the fully CCW position, and the distance from too little to too much would be very small and would be considered twitchy.............. it would need a pad.

rbucek

[brucek]

I just have a hard time believing and trusting that we can expect linear response from a budget product. Am I behind the times?

No, I just think that in this case we would be compensating for the line-out stage, and the line-in stage, and would only miss the mic preamp (which feeds the line-in stage). Would it really make that much difference that we have to fabricate resistor pads for a dB or two ? When I look at the difference between my standalone soundcard cal file, and then the soundcard plus Xenyx802 mixer soundcard cal file, it isn't enough to even bother with. Do you think it would matter much if I added the mic stage into that file? I don't, and I don't bother..... I can move my microphone an inch or two when measuring and change the level by quite a few dB, so I don't think a dB at 10Hz being lost in my Mic preamp matters - the line stages are enough.

brucek

[etc6849]

brucek wrote:

Well, the 3.9K is fine (between the RCA output and mic input), but the 1K to ground is too high. This becomes your output impedance and would not work.
I suggest about a -32dB attenuation, which would be 3.9K (between the line-out and mic-in) and then a 100 ohm to ground (on the mic side of course). This would be about a -32dB pad (fairly typical for a mic pad), and would offer 100 ohm output impedance.


Yeah but then you've added the response of the receiver lein-in to your cal file, which you don't want to do......

brucek

You are correct. This test jig's output impedance is very important to the mic's input. What is the typical impedance value for a omnidirectional condenser microphone? Specifically, I'm using the ECM8000. Mine hasn't arrived yet, but will be here tomorrow. I found this from google for the ECM8000:

Type: omni electret condenser
Impedance: 600 ohms
Sensitivity: -60dB
Frequency response: 15Hz-20kHz
Connector: gold-plated XLR
Phantom power: 15-48V
Weight: 4 oz.

I suppose I should use a calculation like this to find the output impedance (of resistors and output of US-144) as seen by the mic's input:

100ohms (output impedance from US-144) + 3.9ohms = 4kohms. 4kohms in parallel with 100ohms gives 4kohms||100ohms ~ 97.5ohms. So really the mic input just sees 97.5ohms.

Now, if I want to closely match the impedance of the ECM8000 I might try this (edited as I should have 4k = 3.9k + 100 (output imp from US-144):
If my mic is 600 ohms, maybe a 750ohm resistor is ok to use as this would give (4*.68k)/(4+.68k)= .581kohms = 581ohms as seen looking from the mic's input into the test jig+output of the US-144.

Maybe I'm mis-understanding the mic's specifications... Am I missing something? Is the impedance of a condenser mic constant over it's bandwidth; ie should I be designing to this 600 ohms output impedance (test jig including output impedance of the US-144)?

[brucek]

100ohms (output impedance from US-144) + 3.9ohms = 4kohms. 4kohms in parallel with 100ohms gives 4kohms||100ohms ~ 97.5ohms. So really the mic input just sees 97.5ohms.

Yes, that's correct. We're using this interface as a voltage bridge. It's a typical low output impedance feeding a high input impedance. The pad as I specified would honor that.

Yeah, the mic is 600 ohms, but the impedance of the mic is not important enough to modify the pad to change the values as seen by the line-out in the soundcard cal routine. We are maintaining the 'rule of thumb' of >= 10:1 for a voltage bridge for both situations, and so it should work fine with the resistors specified.

brucek

[etc6849]

brucek wrote:

Yes, that's correct. We're using this interface as a voltage bridge. It's a typical low output impedance feeding a high input impedance. The pad as I specified would honor that.

Yeah, the mic is 600 ohms, but the impedance of the mic is not important enough to modify the pad to change the values as seen by the line-out in the soundcard cal routine. We are maintaining the 'rule of thumb' of >= 10:1 for a voltage bridge for both situations, and so it should work fine with the resistors specified.

brucek

Bruce, what you are saying makes a lot of sense the more I think about it. Greatly appreciate your help and time. I'm going to use 3.9k and 100ohm resistors like you state. Thanks everyone for teaching me something. I must say hometheatershack is a pretty friendly forum.

Cheers,

Ellery

[Wayne A. Pflughaupt]

brucek wrote:

When you are trying to control a very high gain stage with a pot, the amount of change for even the slightest movement of the pot would be huge. You would likely be working with the pot almost in the fully CCW position, and the distance from too little to too much would be very small and would be considered twitchy.............. it would need a pad.

Makes sense. However, a mixer like the one I pictured is fully capable of accommodating a situation like that. In addition to the main output attenuator, you have post-preamp attenuators on the individual channel strips. Both are fully capable of reducing the incoming signal to zero. To address the "twitchy" situation, you can easily reduce the pre amp signal to a level that's extremely low via the channel strip fader (i.e., near full CCW). That leaves the main output fader with a wide sweep to fine tune the signal that's left.

Regards,
Wayne

[etc6849]

I'm kind of wondering the same thing. Guess I'll find out tomorrow when the US-144 and mic come in the mail. I'll post the cal results on here both ways (through input and also through resistor network and the mic) and hopefully one of you guys will take a look :-)

Cheers,

Ellery

Wayne A. Pflughaupt wrote:

I dunno, I just have a hard time believing and trusting that we can expect linear response from a budget product. Am I behind the times? I mean, I know that technology "trickles down," but has it trickled down all the way to the bottom end?

Regards,
Wayne

[Wayne A. Pflughaupt]

brucek wrote:

No, I just think that in this case we would be compensating for the line-out stage, and the line-in stage, and would only miss the mic preamp (which feeds the line-in stage). Would it really make that much difference that we have to fabricate resistor pads for a dB or two ? When I look at the difference between my standalone soundcard cal file, and then the soundcard plus Xenyx802 mixer soundcard cal file, it isn't enough to even bother with. Do you think it would matter much if I added the mic stage into that file? I don't, and I don't bother..... I can move my microphone an inch or two when measuring and change the level by quite a few dB, so I don't think a dB at 10Hz being lost in my Mic preamp matters - the line stages are enough.

No doubt, that's laudable performance on the part of the Xenyx 802. Maybe I'm just overly skittish, but I have my doubts that every other budget mixer or interface on the market is as good as Behringer's. There are an awful lot of them out there...

On the other hand, nothing in the signal chain here is laboratory grade to begin with, so maybe I'm being overly concerned...

Regards,
Wayne