Clocking / Noise / Audio Inaudible

[doktor1360]

Let me first apologize to anyone viewing this, I'm doing my best not to clutter up the forum with posts... all related in some sort of referential fashion. Here are the main threads so I can keep this almost in a wiki-type delivery... and moving along accordingly.

http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=2518
http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=2519
http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=2521
http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=2526

So, the latest enigma is circling back around to clocking... to preface this a bit, I have an D8B/HDR audio test project - real basic vocal track so as to quickly check the integrity of an incoming audio signal in the D8B input bloc - just for this kinda thing (ie. bringing up a board for the 1st time, etc). When running this diagnostic project with any piece of gear other than the D8B selected as the master clock generator renders complete garbage for audio output... some inaudible crap with a buzz like elevated noise floor. The console/cpu combo operates as 'normal' other than this latest kick in the shinz. Pretty cool, huh (*rolls eyes*)...

Here's the battery of tests I ran, each designed to isolate any/all edge cases in regards to clocking this gear properly and test the audio from input to speaker(s).

======================================
D8B [set as slave to Rosendahl]
Rosendahl Nanosyncs DDS [set as master clock source]
HDR [set as slave to Rosendahl]
--------------------------------------
* all gear syncs & locks properly
* static-y, garbage audio (inaudible, unrecognizable)

======================================
D8B [set as slave to HDR]
HDR [set as master clock source]
--------------------------------------
* all gear syncs & locks properly
* static-y, garbage audio (inaudible, unrecognizable)

======================================
D8B [set as master clock source]
HDR [set as slave to D8B]
--------------------------------------
all gear syncs & locks properly
completely noise free audio

So, let me go down the triage checklist and review the scorecard...
Rosendahl Clock Generator (6 x word clock out) - Check
HDR - (see D8B below) Check
Monitors - (control room main) Check
Audio Cabling (toslink optical x 6 D8B to HDR) - Check
Clock (RG59/RG6 type 75 ohm) Cabling - Check
D8B (clock out {master} vs clock in {slave}) - Master==Check Slave==Described Audio Problem

{*scratches head...*}

This is the recently purchased console I made from Malcolm Cecil's estate, and it exhibits the exact same symptoms as my original #1 console (soon to be retired), where all this began. Question is, what in the actual audio fuckery is going on here?!? The ADC's in that functional bloc? It does, however, convert the incoming audio, and works perfectly when the clocking is D8B as Master... but I can't completely write that off as of yet. However as indicated above, the D8B audio sounds like shit when clocked as a slave to ANYTHING...

Two different Apogee clock cards, two D8B consoles... no difference, but only one cpu/host involved that is working without issue to all appearances (the other has an apparent power supply issue). Scratching my head at the moment, it's been a series of weird occurances with this gear of late... waiting for Rod Serling to come walking out of the fucking woodwork at any moment (*nervously laughs*). I haven't even introduced the Cranborne R8 into this yet... no sense in adding another variable to the equation at this point - just something else to clock. This is just ugly tho, the clocking according to many posts on the forum should work perfectly fine with the D8B (and HDR) set to slave, and acquiring a clock from the Rosendahl. The HDR performs like a champ in any circumstance, the D8B not so much - it evidently does NOT like seeing an external clock at the Apogee clock card input regardless of what equipment is providing the clock. Syncs fine, just complete garbage for audio being passed. I found these four article links from the forum interesting, but I don't know if they're completely relevant or not to what it is I happen to be experiencing:

http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=126&p=504&hilit=buzz#p504
{peter - adat sync}

http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=277&p=1630&hilit=buzz#p1630
{high c double g - control list / set filter off}

http://www.d8bforum.com/phpBB3/viewtopic.php?f=4&t=2299&p=17187&hilit=clocking#p17187
{y-my-r - clocking}

http://www.d8bforum.com/phpBB3/viewtopic.php?f=2&t=1290&p=10199&hilit=clocking#p10199
{crash - regarding popping & clicking}

Now... IF I'm actually forced, for whatever reason, to use a D8B as the master to sync the Rosendahl... which in turn I think could redistribute the recovered clock to everything else, then so be it. It's just disturbing to me, I'm not one to just glaze over an issue... or better put, whistle thru thru the tech graveyard regarding this at all. Pisses me off, quite frankly...

I'd really appreciate anyone jumping in here, I'm literally looking at all angles at this juncture. I just wanna get this all straightened out again, as It's almost a watershed moment for me... I just can't stomach the thought of shifting away from this gear tho right now. When working correctly, this gear just works... but as we all know, when things start to go sideways {voice trails off}...

Thanx in advance y'all for anything and everything...

I'm just frustrated as a pyromaniac in a petrified forest right now...

Peace

[doktor1360]

Well, for anyone interested I think I have a much better handle on this issue... and ultimately I think it'll work out here that the issue(s) I'm experiencing are due to cabling, and more importantly type and distance. As I stated above, I'm experiencing horrible audio issues whenever a digital word clock is present on the Apogees Clock IN BNC connector... it just will not slave properly to any clock signal. I noticed how the D8B takes a moment or three to sync on fluoro-display '*?' looking for and syncing to master clock siggy when set to be a slave (no issues at all when D8B is the master clock). It makes a lot more sense now, based on further testing and observation. So here's where I figure I'm at, and where I potentially didn't perform the correct due diligence - I essentially used RG-59 cables (75 ohm) that I had from working a gig at Turner Broadcasting Systems two decades ago... the problem is ultimately be in the distance - they're both 25ft in length. Wherein that falls within most distance recommendations for 75 ohm cabling, the type of application certain needs to be accounted for. One, obviously, being digital audio in terms of clocking - NOT coax AES digital audio (different animal entirely). Why? Digital word clocks... jitter... drift... pops, clicks, hash n distortion... yeah, sounds familiar...

OK, so break it down to figure out 'why'... if you're trying to make absolutely sure that the signal gets down the cable in the most efficient manner losing little if any signal strength, well 75 ohm is the correct choice. I read where a good rule of thumb is that if a device is being connected as some sort of receiver (BOOM #1), 75 ohm cable is ideal. OK, kewl... already knew that fact to some degree. Capacitance is something else to consider, the arch enemy of any square wave signal (i.e. digital clock - BOOM #2). Large capacitance values will tend to skew the values of the square wave making it more difficult to reconstruct the clock at the other end. Eye opener...

So, then there is also the consideration of what type of 75 ohm cable to use - either an RG6 or RG59U. The 'U' in the RG-59U indicates 'utility', it's great for general purpose use like TV's, CCTV... generally baseband video apps. They're generally 95% copper shielded (RFI rejection), insulation is OK (compared to RG-6), smaller copper center cable and it can work a bit better at lower frequencies than RG-6. On the other hand, RG-6 cables are constructed with foil and braided aluminum shields, have better insulation properties, the inner copper center cable is thicker, frequency range is better, lower attenuation at longer lengths (significant drawback for RG-59U), but not as efficient when with signals below 50Mhz... fewer signal loss problems (BOOM #3). The thicker conductor provide more extended distance proficiency and a much better bandwidth. It's also got a better frequency range. The important point is the efficiency of RG-6, it's a much more robust cable physically as well (cable jacket difference). RG-6 does, however, cost more than it's RG-59U counterpart... better shielding and insulation, I definitely am in line for purchasing RG-6 cables, cutting back on the length as well. The only BNC cable I need of length would be between the D8B (clock card) and the rack containing the Rosendahl master clock generator... the others shorter due to the HDR and Cranborne R8 being housed in the rack with the clock generator. We'll see, but I think certainly this could be 'problem solved'...

Too much length, wrong type 75 ohm cabling for the application... jitter... drift... pops, clicks, hash n distortion...

Stay tuned...

[Bruce Graham]

Hey Dok;

All spund logical! Standing by!(Good read BTW).

Bruce

[Y-my-R]

It could of course be the cable length (or type, or quality) of the BNC cables you're using... but being as cheap as I am, I ever only bought a single new BNC word clock cable myself (that is over 25 feet long and works fine, btw), and am otherwise only using cables I got free with gear purchases or trades, etc. Some of them don't even LOOK trustworthy, hahaha

For a while, when I didn't have a cable that reached that far away device, I even took a BNC "coupler" of a patch panel and extended two shorter cables with that... still, clean audio (well... jitter in form of a lesser stereo field or worse imaging is possible, but there were no clicks... the thing I'm clocking is just to audition the audio out of an electronic drum kit, that I don't record like that, so, I don't care if I have jitter on that one, that aren't clearly audible as clicks, etc.).

What I'm trying to say is, that the D8B isn't picky about BNC cables for me, so I'm not sure what might be going on here.

One thing I thought about was, that you might have a different digital audio signal that is NOT clocked to the Rosendahl connected to one of the digital inputs, that might mess up your clock. This should only cause occasional clicks, not full on distortion to a point of unrecognizable audio...

...but regardless, my suggestion would be to simplify the test setup to the absolute minimum, to allow for a proper process of elimination.

So, I'd disconnect EVERYTHING from the D8B and HDR, except for the clock cable in a single direction (i.e. Master to Slave), and would also only leave a single digital audio cable connected (e.g. 1 ADAT out from HDR to D8B.... OR 1 ADAT out from D8B to HDR)... no other digital cables should be connected. (Analog cables shouldn't matter, and you of course still need the MIDI and "Console" connections... and power, haha... to even have it run).

Do you get the same result when testing with such a minimal setup?

Is there something that prevents you from moving all these devices close together for a test, so you can try all your BNC cables from the clock master to the D8B's Apogee clock card BNC input (one at a time)?

From the description above, it sounds like you should have at least 3 BNC cables (probably 6, if they were hooked up bi-directionally). You only need a single one to test if the D8B can run as a slave without garbled audio, so I'd just systematically swap them all out one after another, and test again and again.

Oh... and I'm sure you checked, but is the sample rate correct on the Rosendahl? I don't remember clearly, but many years ago, when playing around with a Big Ben and the "super clock" it outputs, I had problems with getting a number of other devices to clock to it. Those would NOT accept the superclock, and intermittently lose clock all the time. If the Rosendahl puts out something like that and allows you to turn it off, I'd try that, too.

FWIW, I always run my D8B via external sync at 44.1 kHz, clocked to a Lucid GENx192. The clock is always on when I turn on the power on the D8B. I can't remember when I last saw the questionmark on the D8B display... it's always locked to external clock before that has a chance to come up.

You already tried different clock cards... otherwise, that would have been my first guess...

Best of luck!

[Phil.c]

The D8B is a strange beast, you say you always have the clock on when you turn on the power, that's totally oposite to me, if I do that, all the faders jump to the top and then back again and I get an era message, I then have to reseat the BFC to get it working.
I have to wait for the flashing ? before turning on the master clock, when I used to use the HDR's clock this was on from the start and I didn't get this problem???

[captainamerica]

not sure if I have anything useful to add but my d8b is the MASTER CLOCK and I have never had any sync issues with my slave MOTU 2408 and MTP gear when I boot up ...clock is always on.