The relationship between output voltage and sine wave frequency

[A27Hull]

Hi all,

Upon setting up a signal generator in Pro Tools LE for AD/DA calibration, I noticed a peculiar thing.

I had my multimeter across Pin 2 & 3 of a main analog output of my 002R, trying to manipulate the signal generator plugin to read 1.228v on the MM.

I noticed that (keeping the internal PT level constant, at +0.1) when I swept the sine wave frequnecy, the output voltage varied relative to the selected frequency.

I am confused about it, and I am really interested in finding out why this occurs.

It could be something as simple as a physics relationship I have forgotten or never learned, but guidance is greatly appreciated.

Can anyone tell me where I can find reading material on this "phenomenon?" All my electonics and sound books come up short.

[CoteRotie]

Many inexpensive multimeters are calibrated to read AC Sine RMS voltage only for low frequencies (around 60 Hz.) A few can read accurate (and even true) AC RMS voltages into the KHz region or beyond, but unless you know the bandwidth of your meter you won't know how high a frequency you can measure accurately. Check your meter's specs, and if AC measurement bandwidth isn't specified you can't trust the readings for higher frequencies. Use a scope instead.

Regards,

John

[A27Hull]

Thanks John,

I have a Fluke 177, but no scope...i think its fairly accurate, but there may be other things at play beside the accuracy of the meter.

My main question was why there is that relationship, as the sine wave gets lower in frequency, all other things constant, the voltage on the output increases. Likewise, it decreases as frequency rises.

I want to say that it is because low frequency has more energy than higher frequencies, but I feel that is incorrect, and on the verge of confusing the ears response to sound with the voltage relationship of sinewaves and harmonic content.

Could it be related to wave length, in that a large wave (lower frequency) has more inherent energy than a shorter wave?

[mikeg]

The Fluke 177 has a freq response that only goes to 1K. In reality anything beyond 500-600 Hz will not give a real accurate reading.

[A27Hull]

Thank you for your reply, but again..



Why does the output voltage increase as the sine wave frequency goes downward, and vice-versa. What is the scientific principle or theorem called?




I can measure a 980 HZ tone @ -13.9 Generator level, Unity internal fader level, and get a 1.228v sustainable reading on my MM across the left analog output pins 2&3...(and still be +/-(2%+3 counts) in varying degrees of accuracy from 500 to 1KHz, according to the manual.)

I've been thinking about it, and though I'm probably only slightly closer to the truth, here's what I think:

A lower frequency has a longer wavelength, a more sustained elevation above/below the 0 axis, giving it an inherent energy potential higher than those of shorter, higher frequency waves.

There's got to be a right explanation to this, I just don't know where to look.

Thanks to all for your replies and views.

[David Kulka]

You're getting yourself all sidetracked with incorrect assumptions. Believe in the good answers that you have already been given. The output voltage is probably not increasing. Your meter is not designed to test audio, and you'll get confusing results if you use it to check frequency response.

If you use a meter that is designed for measuring audio levels, I think you'll find that your Protools outputs are passing signal correctly.

[A27Hull]

Alright, thank you David.

I guess I'll have to test it again with an accurate meter to reproduce what I experienced, if it exists at all.

I'm still open to others opinions.

[A27Hull]

Originally Posted by David Kulka

Your meter is not designed to test audio, and you'll get confusing results if you use it to check frequency response.

David, you are the third person to tell me that my meter is wrong, so for all I know, it is wrong. I'll admit that my meter is messed up/inaccurate, and that I don't know how to use it in audio applications.


If the audio signal alternates between pin 2 and 3, is that not the + and - of the sine wave?

If that is not true, I am wrong globally in my thought of audio as an electronic voltage.

If it is true, would not the potential difference between pin 2 and 3 be measure as the AC voltage?

[loujudson]

They are telling you that the reading you get at higher frequesncies is wrong! Borrow a VTVM, or send it to a knonw audio level meter, and see the difference.

Your meter drops off at higher freqs...

Lou

[nosebleedaudio]

Years ago I bought a true RMS meter calibrated in db with a response that is flat above 1 megahertz, A Leader RMS meter, around $600.00 if I recall.
Just for this and beyond purpose, bias current on tape heads is a good example, over 180Khz for some machines. A scope is not ideal when looking for a dip/peak in the .2db range.
Even a Fluke if not designed for high freq. will not respond...
Like stated above they are for 50 to 60 hz.

[A27Hull]

I apologize to all for my rudeness and stubbornness.

I thought my meter was actually accurate, and given what I paid for it, I am somewhat dismayed to be enlightened to its limitations.

I seriously thought that my meter was not lying to me, despite what people who obviously know about this stuff have said.

So, to answer my own question, as you all suggest,

There is no relationship between output voltage and frequency, all other things constant. Meaning, output voltage should not change, regardless of the frequency the tone generator is producing.

Is that what you all are saying?

[dkatz42]

Originally Posted by A27Hull

There is no relationship between output voltage and frequency, all other things constant. Meaning, output voltage should not change, regardless of the frequency the tone generator is producing.

Is that what you all are saying?

Yep. The RMS voltage of a pure sine wave is sqrt(2)/2 * the peak voltage, independent of frequency. The peak voltage in your case is set by the output of your DAC, which should be very flat across most of the range, and thus the RMS voltage should be constant until you start getting close to the Nyquist frequency.

[Tim Farrant]

If the system you are measuring has a flat frequency response, you are correct in your assumption. If you eq the system, boost or cut, then the frequency to output voltage will vary, this is what eq is.

[Rick Sutton]

I'm suprised that you don't have other meters around to double check the output of your signal generator. I know I'm "old school" but it seems as if any recording environment has multiple meters, either hardware or in the computer, mine must have a couple of hundred possible alternate metering points. no other meters in the house? Just curious.

[CoteRotie]

Originally Posted by A27Hull

I apologize to all for my rudeness and stubbornness.

I thought my meter was actually accurate, and given what I paid for it, I am somewhat dismayed to be enlightened to its limitations.

I seriously thought that my meter was not lying to me, despite what people who obviously know about this stuff have said.

So, to answer my own question, as you all suggest,

There is no relationship between output voltage and frequency, all other things constant. Meaning, output voltage should not change, regardless of the frequency the tone generator is producing.

Is that what you all are saying?

Well, the voltage and frequency of a sinewave are independent variables. You can have a constant voltage at any frequency, and a constant frequency with any voltage. What any particular system does is dependent on the design of the system. Your multi-meter has a low-pass filter built in, so that as you go higher in frequency the apparent reading drops, even though the input stays relatively constant.

Regards,

John

[A27Hull]

Thank you so much guys, and again I apologize for my rudeness and stubbornness.

Rick,

The Fluke is the most accurate DMM I own, I have a small radio shack pocket DMM that I never use, and I also have an old analog Micronta which I can barely use for anything other than checking continuity. The reason is that it is a ranged meter that has difficult to understand meter labels, and I don't have a manual. It was a hand-me-down.

The tone generator is a plug-in in PT, I just wanted a 1.228v tone to calibrate my ADDA unit, so I tweaked the internal levels and bussed it one of the main analog outputs. The PT hardware output was the first and only metering point, besides that, and as I have accepted, it is likely that my meter is not accurate enough to judge the signal leaving the output.

[loujudson]

Just a comment that Protools is not a metering thing at all, you have to interpret the levels by the colors, and the peak lights have nothing to do with actual overs at all in my book. You can read headroom directly in -dB, but need meters on the outputs to calibrate the voltages.

I check it against my Sound Devices 744, a variety of DAT machines that have great metering even though they won't play tapes, a Masterlink, and so on.

One can hardly ever find a VTVM any more, but they are the best reference for actual voltage from any device... I feel extemely lucky to have learned in the good old analog days how to calibrate audio devices, and to check their frequency repsonses with RTAs, Oscilliscopes, and VTVMs. Anyone remember Heathkit? ;-)

Lou

[Rick Sutton]

I can see your situation. I'm just used to having plenty of VU meters around to send signals to for quick checks. Any outboard gear (say a good compressor) with a decent VU meter (with a needle movement) that can read input could be a handy addition to your setup.
Best regards,
Rick

[David Kulka]

The Fluke 177 is not "inaccurate". It's just not designed for measuring audio frequencies.

You cannot measure the weight of a truck on a postage scale, but that doesn't mean the postage scale is inaccurate. Any piece of measurement equipment is designed with certain functions and ranges in mind, and cannot be expected to be useful outside of those ranges.

[Wavebourn]

Mine is younger than me, but older than some of you guys... It still works fine on audio frequencies!




I paid $5 on e-Pay +shipment.

[nosebleedaudio]

I have an old Triplett meter a D’Arsonval meter, ran a freq. gen. thru it from 15 hz to 30khz and it was pretty flat. So maybe a Radio Shack meter will work as well.

[Brian M. Boykin]

I like this thread. I recently ran into this problem while calibrating my otari mtr-10. When I set up repro with my mrl the vu's zeroed out just fine. When I set up the record eq I was checking the vu response angainst my mm on the outputs. Despite the vu reading being right and zeroing out, the mm was wrong down low and up high. I knew the heads were not the problem because JRF had seen them and said they were practically brand new. I thought my vu's or the repro electronics had crapped out. I discussed this with John French and he was baffled too. It sounded fine when I played music back on it so I wrote it off for about half a day. Then I played back my mrl which I know is referensed correctly. The vu's would zero out at 100, 1k, 10k and 16k, + or - 1/2db, but the mm was still off down low and up high. So I ran the outputs into my tascam and its vu's confirmed that the Otari's were accurrate. I came to the same conclusion as A27hull, different frequencies must require different amounts of power (dbm) to accomplish the same db reading. Now I know my mm sucks!!!! I used my mm to set the bias because it's accurate to the decimal point. Time for new testing tools. Great thread!!

Thanks,

Brian

[nosebleedaudio]

Originally Posted by Brian M. Boykin

I like this thread. I recently ran into this problem while calibrating my otari mtr-10. When I set up repro with my mrl the vu's zeroed out just fine. When I set up the record eq I was checking the vu response angainst my mm on the outputs. Despite the vu reading being right and zeroing out, the mm was wrong down low and up high. I knew the heads were not the problem because JRF had seen them and said they were practically brand new. I thought my vu's or the repro electronics had crapped out. I discussed this with John French and he was baffled too. It sounded fine when I played music back on it so I wrote it off for about half a day. Then I played back my mrl which I know is referensed correctly. The vu's would zero out at 100, 1k, 10k and 16k, + or - 1/2db, but the mm was still off down low and up high. So I ran the outputs into my tascam and its vu's confirmed that the Otari's were accurrate. I came to the same conclusion as A27hull, different frequencies must require different amounts of power (dbm) to accomplish the same db reading. Now I know my mm sucks!!!! I used my mm to set the bias because it's accurate to the decimal point. Time for new testing tools. Great thread!!

Thanks,

Brian

When you say bias do you mean the 150+Khz bias for the record/erase heads?
If so your DMM will be useless...

[Brian M. Boykin]

No, I set the overbias for each channel using the dmm. When the overbias is supposed to be 1.5 db, it is easy to just watch the dmm on the output and not rely on the vu. You can set the overbias to the decimal place. What I do is find the peak on the dmm, then I use the output pre amp to bring the vu meter to 0 vu. My dmm will read +4 dbm, which is my output reference level at 0 vu. I double check my peak, make any corrections, and then overbias until it reaches + 2.5 dbm. The difference is 1.5 dbm, which, according to the thinking on this thread, should equate to 1.5 db, or a 1.5 db drop on the vu. Let me know if you see a problem with this approach. I'm totally self taught and dreamed that technique up.

Thanks,

Brian

P.S. I eventually want to get a distortion analyser and set overbias by watching the analyzer for the least amount of distortion.

[Rick Sutton]

Originally Posted by Brian M. Boykin

P.S. I eventually want to get a distortion analyser and set overbias by watching the analyzer for the least amount of distortion.

It's been a while since I've done a lot of tape machine set up but IIRC using a low frequency sine wave and your ears and you can do it just fine.

[Brian M. Boykin]

I have done that; 40hz sine wave. According to Eddie Ciletti it's the magic frequency for setting overbias. Just play with the overbias until you get a clean signal, ie: no fuzz.

[A27Hull]

Originally Posted by Brian M. Boykin

I like this thread. Thanks, Brian

Thank you too Brian!

As you can see I was rather resistive to believe others who knew more than I at first. I thinl I'll take their word for it now.

Still, I'm glad that someone else out there experienced a similar phenomenon.

From now on, rule #1 for me is listen to your elders.

[Jim Williams]

Originally Posted by Brian M. Boykin

No, I set the overbias for each channel using the dmm.Thanks,
Brian
P.S. I eventually want to get a distortion analyser and set overbias by watching the analyzer for the least amount of distortion.

I use the Audio Precision to do bias checks by using a 50~10k sweep on the THD vs frequency plots. It's the most accurate way of doing this. I found that all the recommended bias settings by either the machine maker or tape maker were wrong. I sweep the frequencies while twisting the bias pot for lowest THD. You can optimize it for the mids, lows or highs by how much you use. I usually set 1k hz at the lowest THD, the tops and lows will rise on either side.

With quality record/play electronics I have reduced THD from the standard .55% at +3 (250 nw) with 226/456 tape to .15% at +9 on GP-9. That allowed the Dolby SR's to be turned off.

Jim Williams
Audio Upgrades

[nosebleedaudio]

Bias is also the art of compromise, the higher bias setting will lower the high freq. output.

[Brian M. Boykin]

With quality record/play electronics I have reduced THD from the standard .55% at +3 (250 nw) with 226/456 tape to .15% at +9 on GP-9. That allowed the Dolby SR's to be turned off.

That's what I'm talking about!!!