JLM VU Buffer Troubleshooting Help! (Finished w/ Pics!)

you need to measure the voltage of your DA output and compare to 0 VU. I believe 1.23V RMS = 0 VU.

I measured the output of my audio interface. A 1kHz sine @ 0dBFS gives 3.89V, while @ -20dBFS gives 0.36V. In order to get it to 1.23V (+4dBu) I have to set it to -10dBFS. Weird, huh? I thought -20dBFS was 0VU. Oh well...  ;D

How are you measuring it? across the hot and cold connection? Also 0Vu can just about any level in dBFS, it all depends on what you calibrate to. For example digidesign/Avid does -18dBFS = 0Vu = 1.228V/1.23V, etc, etc. The old Tascam da-88's did -16dBfs = 0Vu,etc,etc. Yamaha on their 02r's was doing -21dBFS = 0Vu,etc,etc and in film often they use -20dBFS = 0Vu, etc,etc.

Also What is your audio interface? does it have balanced output?

My interface is an M-Audio ProFire 2626 (balanced outs). I also found out that nominal level varies from interface to interface. I measured across hot and ground. Is this wrong?

EDIT: Measuring across hot and cold or across hot and ground? Which one is the correct?  :-[ (embarrassing ignorance)

Firstly hot-to-cold.

Secondly, ANYTHING dBFS can correspond to 0VU... VU has no place in digital as a level reference for recording... Think of 'VU' as standing for "Virtually Useless". -This is why there's no digital audio recording gear made with VU meters on it. Never has been, never will be.

We LOVE standards... thats why there are so MANY of them.

-14dBFS
-16dBFS
-18dBFS
-20dBFS

But it doesn't MATTER... you shouldn't be using the VU meter to set levels in digital... -it's just a wiggly thing to make you feel good, it can't tell you when you're overloading.

Digital doesn't CARE about average energy, it cares about PEAK content. VU can't tell you that, nor can it tell you that a sine and a square have the same peak level.. it'll sow very different levels for the same signal... When the Vu says they're the same (i.e. when the square is reduced to have the same 'average' energy as the sine) then the sine has a much higher peak level.

If you want to know what the LEVEL is in digital terms, a VU is a complete waste of time... so therefore you can set it to whatever offset makes it wiggle in a pleasing area of the scale for most of the time... which won't be all the time, no matter what you do.

ytsestef said:

I measured the output of my audio interface. A 1kHz sine @ 0dBFS gives 3.89V, while @ -20dBFS gives 0.36V. In order to get it to 1.23V (+4dBu) I have to set it to -10dBFS. Weird, huh? I thought -20dBFS was 0VU. Oh well...  ;D

Click to expand...

Multimeters are not always correctly reading AC with freq > 50/60 Hz. Check again with lower/different frequencys...

Depending on the output design, a lot can change of course... however...

-10dBFS signal giving +4dBu when measured over ONLY ONE LEG of a 'simple' electronically balanced output would -for example- mean that the ACTUAL level is 6dB louder.

Thus -16dBFS would produce +4dBu (Notice that I'm using dBu instead of dBm since this is NOT a 600 ohm load and we're not talking milliwatts here, but the unterminated equivalent, hence dBu).

As I mentioned earlier, -16dBFS equaling zero VU is one of many "standards" which people choose, so nothing would be 'wrong' in that case.

Keith

I never expected it to do anything more than "bounce back and forth"  ;D so we're good on it being purely cosmetic.
BTW, my multimeter reads the same using 1kHz and 50Hz tones, but still very good advice, thanks pan!

On topic, I measured hot-to-cold as Keith said and this time in order to read 1.228V I had to adjust the output to -16dBFS (as opposed to -10dBFS giving same results when measuring hot-to-ground). The problem is, when calibrating the VU's at -16dBFS, a lot of commercial mixes make the needle only move in the +4 to +5 area, which is unpleasing aesthetically (damn loudness wars!)...

Thanks everyone!!! FINISHED! Pics:  ;D ;D ;D

Nice looking build!

i like this  nice box

Thanks mitso! I try to get better with each build!
This box is awesome considering it costs EUR12.50!!

Where did ya get the box?

EDIT:typical American spelling error.

I really like this build, and the veroboard software looks really cool too.

great stuff!

also, how did you get your cutouts so good? thats some crazy dremmel skills!

I bought the case from a local shop here in Athens, Greece. I don't know what the brand is but I could ask them. The rectangular holes were done with a standard dremel. I just cut the holes just a little bit smaller than required, then trimmed them to perfection using flat and 90° corner files.

Sorry to revive this thread, but I would like to build this with a couple of Nissei VUs that have internal rectifiers and had a question: What modifications would I need to do on this circuit to do so? I figured others who have found this (awesome) post and who are novices might like to know as well. Thank you!

Nice build - something to consider is the effect of having the 'thru' connectors so close to a psu transformer.

It may not be a problem for you, but I would recommend checking the 'thru' signal on your soundcard with a 'real time analyser' application and look for excessive amounts of mains-induced noise.

Generally speaking, it is preferrable not to run unshielded, untwisted audio wiring close to a psu transformer, either toroidal or not.

I would think that this would add quite a bit of mains harmonic noise to the 'thru', which presumably is your main mix.
I have certainly had this issue with signal wires to my builds when using jlm's vu buffer.
Repositioning of the wires mad a BIG difference.

Anyway - something to think about.

PS - the internal diodes in true VU meters works fine with the adustable gain of the jlm vu buffer circuit 'in practice'