Pre newb.. advice please. Confused by inputs and cables.

champagne_charl

New member
Hi,

Ok so my first post.. I’m hoping for some advice if poss.

I am trying to monitor noise with sound level meters and want to capture the audio to file.

What I'm using is an mbox2. I did have some brief success capturing quality audio then something went pear shaped. I tried another meter I got the same issue.

So the outputs on each sl meter differ.
1) rca
2and 3) 2.5
4)3.5

With device 2 the manual is explicit saying tip outputs ac of 0.4v Vrms or 0.94 Vrms dependant on h/l settings

Ring 2 on the same cable is dc output 0.1v/db to a max of 1.4v

Third ring is ground..

With the other devices I have no further info.

I started using a selection of purchased cables and adapters and terminated at the Mbox using an XLR adapter.

I read shielded was needed and also wanted longer (up to 10m) so picked up some dmz cable silver solder and some decent connectors..

I made the first cable today and used only tip and ground. I soldered shield to ground. And connected to the mbox. Mostly all I got was noise on the line.

Note this had 3.5mm either side and was connected to an rca or 2.5 adapter one side and xlr adapter the other..

I checked wiring of xlr adapter and it bridges pins 1 and 3. Normal I think?

As I had no success getting anything other than interference.. I connected the 3.5 output to a little boss Bluetooth speaker. The sound was crystal clear.

I then used my phone as source and had no issues.
I feel I'm doing something fundamentally wrong. But can't quite grasp what.
The mics/Noise meter do have preamps and I can only assume these are not bypassed..
I'm considering getting some 1/4 inch plugs but thought I'd seek advice first..
OH I didn't switch on the 48v phantom.. didn't want to fry anything.

I'd be grateful if somebody could point out where I've tripped up.
 
Last edited:
Not sure what you are trying to accomplish nor the procedure you are using to do what??? As for your assumption about bridging pins on xlr this is completely wrong. You need to do some research on the difference between "balanced" and "unbalanced" audio cables. An RCA should not be wired to an xlr . They are very different animals.
 
Thanks for your efforts although not terribly helpful.

Issue found. The xlr adapter I purchased maps tip to pin 1 instead of pin 2 then shorts it to pin 3. Xlr connectors arriving tomorrow should fix this.
 
"Thanks for your efforts although not terribly helpful."

And that comment will not garner you any friends either. People come here to help as best they can and for FREE! Your post is frankly a a confusing mess but I shall reserve judgement since I suspect English is not your first language?

I do not know what the noise measurments are for but if any kind of accuracy is needed you have to understand that "noise" is far from a simple matter!

First of all, to get any meaningful result you must define the measurement bandwidth and the frequency response of the whole system must be scrupulously flat. That of course rules out any kind of signal processing such as MP3 or Bluetooth.

There are various ways to measure noise. RMS*, Peak, Quasi-peak, and various weightings, A,B.C. The method chosen depends upon the use the data is to be put (gear makers usually chose the one that gives the best result for the AdPuff dept!)

*This is a time dependent/ statistcal measurement and different meters give different readings on the same noise signal!

Dave.
 
Thank you Dave for your comments.

Thank you for pointing out the lack of clarity in the post. I shall correct the ambiguities and confusing typos shortly. Spelling has always been a major weak point.

With regards gto comments. I do apreciate his efforts. Please excuse my bluntness with regards the content or usefulness of the same. His comments.

1) xlr>rca is wrong. .. How does rca differ to any other unbalanced mono source? Should we just bin all unbalanced kit?
Further more rca although not ideal is the manufacturer's choice not mine. Even todays 7k$ kits still come with no balancing.

2) jumping pin 1 to pin 3. In xlr. This seems common place.. with reference to the same being rife when using balanced to unbalanced devices.
Mention of ground loop problems would be more helpful and sugestions in regards to sudo balancing and avoiding groundloups may have been more usefull Than simply stating it is completely wrong.

Again thanks for your post. Yes noise/ spl is highly complex. I'm only scratching at the surface. The purpose of my audio setup is simply to record audio. the analytical side will for now be procesed by the meter. At a later stage results from the type 1 meter will be used to analyze the audio files and other type 2 meters. I'm hoping offsite calibration will help in terms of frequency response specs of the various mics and . Preamps.

Anyway.. primary objective here is to get a decent quality audio files. Hopefully Not using adapters and as such knowing exactly what goes to what will help. The use of potentiometers and resistors at unbalanced end will come later if deemed necessary.
 
Well, bluntness might be the way you communicate in general but in forums, especially THIS one, a certain degree of forebearance and politeness usually gets better results.

It would help me at least to have the make and model of the sound level meters. I could then look for user manuals and hopefully find the exact nature of the output regimes? You could of course save us all even more time by posting a link to manual .pdfs! But to cut to the chase!

An unbalanced source such as RCA going to a balanced input normally works best if the third "cold" pin (3 for XLR almost always and ring for TRS jack) is tied to the screen (1 for XLR) You can wire as "impedance" balanced when the cold pin is connected to screen via a resistor ideally equal to the output resistance of the source. For most purposes 100 Ohms is near enough.

Earth/ground/hum loops can only arise if a piece of gear is mains earthed and connected to another piece of earthed kit via an audio screened cable. Since SPL meters are all battery powered, earth loops cannot happen.

Have you considered calibration? The signal from the meters will show up in the DAW as "dBfs" levels and you will need, it seems to me, some form of audio source to generate a noise signal to set a level. Such as "meter 1 shows 50dBA SPL and that is -18dBfs on the DAW meters". Ideally you want a meter calibrator but them's expensive!

Dave.
 
Well, bluntness might be the way you communicate in general but in forums, especially THIS one, a certain degree of forebearance and politeness usually gets better results.

Actually no its not my normal manner.. not really. not unless i feel i'm addressed in that way or similar. That said. You are right. Point taken and understood.


Thanks again for your comments..

It would help me at least to have the make and model of the sound level meters. I could then look for user manuals and hopefully find the exact nature of the output regimes? You could of course save us all even more time by posting a link to manual .pdfs! But to cut to the chase!
cel 440 : manualsdir.com/manuals/get?doc=750022&h=0240c1547c0165d6793b13a52f6319f17cc4703d

cel 630 casellasolutions.com/content/dam/casella/site/support/Sound-Level-Meter-CEL-63X-Handbook-English.pdf
cel 240 casellasolutions.com/content/dam/casella/site/support/Sound-Level-Meter-CEL-24X-Handbook-English.pdf
cel 1405 Too old to find digital manual..


An unbalanced source such as RCA going to a balanced input normally works best if the third "cold" pin (3 for XLR almost always and ring for TRS jack) is tied to the screen (1 for XLR) You can wire as "impedance" balanced when the cold pin is connected to screen via a resistor ideally equal to the output resistance of the source. For most purposes 100 Ohms is near enough.

Thank you. I read this and its what prompted me to get DMX cable

Would the resistor go on either or just source side. Im guessing just the source.

Earth/ground/hum loops can only arise if a piece of gear is mains earthed and connected to another piece of earthed kit via an audio screened cable. Since SPL meters are all battery powered, earth loops cannot happen.
And that there solved the issue.. or at least diagnosed the issue and also explains why i did at one point get good sound quality. I've been using these with a DC charger. When my laptop is plugged in and the SLM is plugged in is when im getting the noise. If either is disconnected then I have no noise. Thanks.. but now how do i get around that. the battery's don't have enough run time. and comp side that's even worse.. I could maybe make a few packs of Liion batterys and use a diode to drop the V down slightly. Any thought on a simpler method would be great.

ohh just thought the 240 'needs' to be plugged in to store results as it has no mem. Software is rather limited but i found that buy monitoring the info over the usb i don't just get a and c weightings but also the i result and get around 10 results/ second instead of one every second.. Not sure why they didnt put the effort in with the software. The only thing i can think of is splitting the power and data lines and having the power feed run off liion pack. .

Have you considered calibration? The signal from the meters will show up in the DAW as "dBfs" levels and you will need, it seems to me, some form of audio source to generate a noise signal to set a level. Such as "meter 1 shows 50dBA SPL and that is -18dBfs on the DAW meters". Ideally you want a meter calibrator but them's expensive!

I had.. or have. I have 2 calibrators a 110/2 and an older sadly both type 2. 114dB@ 1Khz

I Just tested with one meter.. the 440. The problem i have is the range on the device needs to be set much higher than i would like.. So when amplifying what has been captured lots of noise appears.. Ideally id have something around the 50DBa to reference against.. i had found an old one on fleabay which was in ozz.. I may have to try and find it again unless i have more success with the other devices.

I just tried the 630 and it's full range. so no issues with that.

The 240 will be limited also im sure.
 
Checked cable impedance which is 110ohms.. So purchase some resistors to match..

Where did you get that figure from? Unlikely to be the output resistance of one of the meters since the only specification even remotely akin to that for one meter states "load resistance 22k or higher"(see more on "specs" further down)

I suspect that "110R" is in fact the characteristic impedance of a cable since CATx for instance is specified at 110 Ohms, SPDIF cable 75R and the Rubber Ducks brigade's antenna cables 50R. These are not "real" resistances (except when terminating a line as e.g. video) We audio people have no truck with such things.

Specifications: Poor really innit? One of those links wants me to run an exe file? No chance! The information I have managed to grasp shows me that at least one of the meters records the events as .wav files. That would mean you don't need any external recording gear and can just dump the files onto a PC AFAICT, via USB or from an SD card?

The high, 114dB level of the calibrator is a problem but you could put a precision attenuator between the output of the meter and the interface. Say you attenuated by 20dB then you would know that whatever level was shown in the DAW it was reading 20dB below the actual SPL. All gets very messy I know and attenuating "precisely" is hard to do when source and sink resistances* are largely unknown. The way to do it is a unity gain buffer op amp, NE5534, and a low resistance attenuator after that.

But I am sorry friend! Not having the equipment before me and never having used SPL meters of that class I am rather lost. Another snag I see looming is the fact that although the AI and DAW should easily have a dynamic range of 100dB and so be able to capture a wide range of levels, the meters have at least two ranges and Sod's Law predicts that the signal YOU want spends most of its time between ranges!

*I say out and in "resistances" because we assume and WANT the signal to be independent of frequency? Not therefore an "impedance".

Dave
 
For a better indication of what i wish to achieve.
I live in a leasehold property where the floorboards of the flat above along with joists are in disrepair. They creak like hell. and also bang dependent upon where. EH have no backbone and believe it or not don't understand SPL or frequency weightings despite it being supposedly their field.
Furthermore, they are only concerned with noise created by antisocial behaviour type noise. Therefor i have to gather evidence. The flat is small so a noise emitted from one room can be heard in the next. By having several meters working in parallel i can identify the source of the noise and the spl of the incident LZpk LAi LZi . I can also zoom into the noise itself and take a closer listen to it. The levels are generally 30 to 58dBa (i) and rarely get above 94LZpk but occasionally a bang from the neighbours dropping or banging something will hit 108 Lzpk. Generally, i want sensitivity and ability to analyse the noise pattern as opposed to just the statistical analysis. The view is to prove the disrepair through noise as opposed to just look at noise element

Where did you get that figure from? Unlikely to be the output resistance of one of the meters since the only specification even remotely akin to that for one-meter states "load resistance 22k or higher"(see more on "specs" further down)

I suspect that "110R" is in fact the characteristic impedance of a cable since CATx for instance is specified at 110 Ohms, SPDIF cable 75R and the Rubber Ducks brigade's antenna cables 50R. These are not "real" resistances (except when terminating a line as e.g. video) We audio people have no truck with such things.

cable accu - cable 5pdmx 110ohm 22awg 2 pair low impedance shielded


So, for the440, should i be looking to put a 22kohm resistor across shield and ground and 2.2kohm or the 630? Can these resistances be verified with a multi meter? I think I tried and was in the mega scale.
Whichever resistor is used? Could you confirm at which end it is required/ and if the ground should be bridged to shield at both ends? I think this may be what you are trying to clarify in your final point.
If I was to take an attempt at logical but clearly ignorant view, i would connect shield to ground at one point only and put the resistor at the XLR for the sake of ease and space (compared to 3.5 jack) . ... Equally if im not mistaken reading what you said the resistor should match the output and in doing so balance the resistance between shield and ground and signal and ground. therefor neither path would be easier for the noise signal to follow and it should remain balanced across both lines.

Some specs...

CEL 440:
Specs: AC output signal for tape recording or headphone.
0.5 V RMS Maximum with minimum load impedance 22 kΩ.


AC input signal for analysis inserted via a line input terminal in the bottom panel.
0.5 V RMS Maximum with maximum source impedance 22 kΩ, or AC input signal inserted via a CEL-516 Input Adaptor, 10 V
RMS Maximum with maximum source impedance 100 kΩ.



CEL 630:
Due to the instruments wide dynamic
measuring range, the AC outputs full scale
signal is adjustable to suit High or Low noise
levels. The AC output range may be selected
within the system tools screen:-
• On the low range, a full-scale AC output of
approximately 0.4 Vrms corresponds to a
maximum sound pressure level of 96 dB. On
the high AC output range, a full-scale output of approximately 0.94Vrms
corresponds to a level of 140dB. The AC output signal corresponds to
the Z-weighted response of the instrument and has an output
impedance of approximately 2.2 kΩ.
If you use the AC output, you should make certain the load impedance is
as high as possible, and you should use a screened or co-axial cable of
length 0.5 m to 10 m.

240

AUX Socket (2.5mm stereo) AC output provided for DAT tape / PC wav file recording or headphone applications. Approx 0.5V RMS Full Scale Deflection (FSD) ‘A’ weighted output on selected range. Minimum load impedance 22kΩ. (Optional DC Output via internal configuration, 0 to 3.3V DC for FSD on selected range. Output corresponds to selected weighting, 2kΩ Output impedance). AC input used for electrical calibration, switched on via configuration menu AUX Socket (2.5mm stereo) AC output provided for DAT tape / PC wav file recording or headphone applications. Approx 0.5V RMS Full Scale Deflection (FSD) ‘A’ weighted output on selected range. Minimum load impedance 22kΩ. (Optional DC Output via internal configuration, 0 to 3.3V DC for FSD on selected range. Output corresponds to selected weighting, 2kΩ Output impedance). AC input used for electrical calibration, switched on via configuration menu

Specifications: Poor really innit? One of those links wants me to run an exe file. No chance! The information I have managed to grasp shows me that at least one of the meters records the events as .wav files. That would mean you don't need any external recording gear and can just dump the files onto a PC AFAICT, via USB or from an SD card?

Peculiar re the exe. I certainly don't blame you! However, it should not be the case. all terminate with pdf file extension except the first and i just checked that. however, it points to the following : Casella CEL CEL-440 User Manual | 70 pages
I’d got the link from my download history…
Re the recording of the wav on the device.
The wav file is highly compressed in comparison. I prefer audio captured through output... well as long as i get rid of this noise. . Sadly, it is the only one of the 3 that does save to wav. I will probably save it for comparative purposes.

The high, 114dB level of the calibrator is a problem but you could put a precision attenuator between the output of the meter and the interface. Say you attenuated by 20dB then you would know that whatever level was shown in the DAW it was reading 20dB below the actual SPL. All gets very messy I know, and attenuating "precisely" is hard to do when source and sink resistances* are largely unknown. The way to do it is a unity gain buffer op amp, NE5534, and a low resistance attenuator after that.


Not sure if this would really work with the calibrator unless I was to crack it open. the calibrator has a socket into which the mic fits. no space for attenuators. ALTHOUGH... The 440 could be calibrated by using an attenuator between the output of the 630 and its input. Maybe the comparative method with alternate sound source would be better.

But I am sorry friend! Not having the equipment before me and never having used SPL meters of that class I am rather lost. Another snag I see looming is the fact that although the AI and DAW should easily have a dynamic range of 100dB and so be able to capture a wide range of levels, the meters have at least two ranges and Sod's Law predicts that the signal YOU want spends most of its time between ranges!

Fortunately with this one. murphy hasnt stuck his nose in. 20-90DBa is fine. OTher than the calibration aspect. I need a type 1 94DB callibrator.
*I say out and in "resistances" because we assume and WANT the signal to be independent of frequency? Not therefore an "impedance".

sorry... but woosh. despite a little more reconnaissance. over my head. which frequency are you referring to? The frequencies of the various (intentionally) captured noise? Im clearly missing a lot of basic knowledge.. Sorry to be dome here but could you please explain slightly differently?

How should i handle the ground loop issue/ noise from PSU? If i can reduce the source of that it would help also.
Thanks again.. Highly appreciated.
 
To be brutally honest I do not think those sound levels are high enough for you to have a hope in H of getting anything done about them but that is your business! I shall stick to the technicals.

Impedance: The word is bandied about freely and no doubt you have seen the impedance curve of a loudspeaker? (properly called the "modulus" of impedance) The term is connected with how a component affects the flow of alternating current. A capacitor allows current to flow more and more easily as frequency rises. An inductor, the opposite. Now we almost always want our signals to be transferred with NO CHANGE in level ref frequency! So, the output of a device, be it SPL meter or mic pre (say) SHOULD deliver the same voltage (into a load) over the whole range of audible frequencies. Most do and so it cannot have an output "impedance" but an output "resistance".

Some things, the passive electric guitar is the main culprit, have a gaggle of interconnected inductors, capacitors and resistors inside them and the best you can do is present them with a very high "impedance". The magic meg' (which is in fact a physical 1mOhm resistor in most cases with a tiny bit of shunt capacitance that does FA at audio frequencies)

Sadly, just about everyone talks in terms of input and output impedances. (aka "Z")

You do not want to make a "Z" balanced cable with 22k in the cold/shield circuit. The 22k refers to the minimum load the device can handle before inaccuracies creep in. If they don't tell you there is no way to determine the output Z (unless you have a sine wave generator, an AC voltmeter and a selection of load Rs. Bit of math, Ohms law as well)

The calibrator attenuator: I meant you attenuate the output of the SPL meter a precise amount so as to bring the reading in the DAW's meters into a better place. Since the OPZ of the meter is unknown you cannot MAKE a precision attenuator unless you buffer the output first.

Hum loops: These are almost impossible to diagnose "off site"! I shall therefore draw you up the ideas I have for your intended rig and hope you can adapt things to your particular situation.

Dave.
 
Thank you for such a comprehensive reply.. again..

I'll reply fully soon..

For now..
The mic in the 630:
http://site.jjstech.com/pdf/Casella-pdf/CEL-251-Measurement-Microphone-Capsule_Datasheet.pdf

Could one problem with the audio capture of the spl during calibration and using attenuation to reduce be that the problem actually lies with the SLM reducing the voltage depending on the range setting.

Re the sine wave generator.. in essence I think i have one. The older calibrator... Circuit board is simple (in design). although i need to study it and learn as to learn how it actually works..

PHOTO_20190602_232613.jpgPHOTO_20190602_232637.jpgPHOTO_20190602_232733.jpgPHOTO_20190602_232757.jpg

But surely if i reduce the amplitude and frequency somehow (research needed, then i can verify both using the 630.


Testing the ac output i got the following:

Range | mV @114db @1Khz | db(a)
-----------------------------------------------------
70-140 | 24.7 | 114
60-130 | 77.8 | 114
50-120 | 245.0 | 114
40-110 | 495.0 | 110

I didnt take it any lower although 2 steps remain with a min of 20 or 30Db.

With the audio that was recorded, to amplify sufficiently in the DAW as to hear other sounds the result was major distortion and lots of noise.
 
Please excuse the crudity of attached. Eyes not what they were, hands not as steady.

Fig 1 shows a notional classic setup that is likely to cause a ground loop. There are various solutions.

1) Find a PSU that does not carry the mains earth thru to the audio device. N.B. You only want ONE PSU earth free. Not good to have the whole system NOT earthed although that might be electrically safe with the correct PSUs. What is NOT safe is the lifting of and EXISTING mains earths.

2)Run meter or laptop on battery.

3)Use a 1:1 line transformer between the devices. Not really an option here as I suspect the output Z of the meter is too high to give a flat frequency response driving a transformer (could use an op amp buffer as per the attenuator)

4) Make up the cable as shown. See the break in the screen ONE end? It is shown at the AI end but might give lower noise tother way. The capacitor "C" is optional but serves to keep the screen continuous at RF. A value of 220pf to 1000pf (1nF) will be fine and the working voltage very low, 63V up. A tiny ceramic cap is favourite and it should be possible to stuff it inside a jack plug.

5) I have include details of what I have dubbed a "shielded return" cable. This requires twin core (mic) cable and I hope you can see the idea? At one end the cold core is connected to the screen but at the other only the cold core is connected.
This cable regime gives a high degree of protection in high RF fields.

I am sorry but your last post just confused me further! In essence a signal even of 24mV is decent enough, around -30dBu and typical of electric guitar signals and we have had little trouble amplifying and recording those for best part of 3/4 of a century!

Also not sure why you have posted an "exploded view"? I think it best to choose a meter and stick to that (assembled!) and solve the noise/level problems.

BTW I have since had a though as to a solution of "range changes"? Set one meter low and a second higher. Say #1 captures 40dB to 80dB and #2 60dB to 120dB?
Feed each meter to a separate track. Careful note needs to be made re gain settings and relative levels.

I wish I had an SPL meter with a signal out but I don't, just cheap wee thing I calibrate my monitors with.

Dave.
 

Attachments

  • SPL.pdf
    32.8 KB · Views: 4
Thank you very much. Far better than i can sketch,

Sorry for the late reply.

Finally received some xlr plugs and a digi 002

I have wired as follow

tip > Hot > XLR 2
Sleeve >Cold >XLR 1
Not conneted > Shield > XLR 3

No hum. WOuld a capacitor/resistor improve the above? Should it go between 1 and 3 on XLR?

I did find this: Sound System Interconnection which was also useful but again. I haven't sat down to digest it properly.

I suspect the problem (or the major one) was the XLR adapter.

Im still looking at improving and will fiddle.


The exploded view was in case you could see anything that may assist in lowering the amplitude. I should have checked a little more. Today i spotted a couple of pots. That combined with a few 110ohm resistors.. I was able to reduce DB to 96. I have some 500ohm resistors now on order. Ideally id like to get the signal to < 90Dba. The idea behind this is to adjust DAW so that when i calibrate I know what calibration Db output corresponds to. Its not that i want high or low ranges. Just to have a common datum/ without need for resetting ranges.

Re the 24mV issue. The issue i was trying to explain was not the 24mV at 114Db (which sounded crystal clear) but the mV output at 30-40dba and how it was in the range of the faint line hum/interference being picked up (tested when on batteries). a mV figure at lower dB would have made it clearer.


Question(s):
The resistor previously mentioned, that has been replaced now by the capacitor i presume? For same or similar purpose?
Fig 2... should signal ground be connected to anything there?
The XLR plugs have a chassis ground.. can i incorporate that? or is it already internally once plugged in? .. I suppose i could check continuity and find out that way.


Again sorry not have sent a thanks earlier.. i wanted to get around to sending some positive results first.. and time passed..
 
Back
Top