Trident Series 65 grounding: theory and practice

[leigh]

I put your file into Audacity, and apply gain to it. I get 70 dB of gain to take it from where it is to digital full scale. But without knowing what the reference zero was, I can't tell how noisy it really is. I can tell you it is 70 dB based on your file.

No.

Noise measurements are done as RMS. When you normalize my file in Audacity, you are gaining it up until the highest peak hits 0dBFS. That is not the same thing.

Attached is a screenshot of the noise file analysis as done by Audiofile's Wave Editor app. As you can see, the peak level is -70 dBFS, but the RMS level is -87.1 dBFS.

 

[leigh]

You need to record some tone at your VU meter zero into the front of the file, and then the noise. Then a simple gain exercise gives us the signal to noise.

Sounds solid. Will do that.

And with all the confusion over the math needed to translate levels for the -10 dBV converter input trim setting, I will repeat the test but with the converter at the +4 dBu setting.

 

[leigh]

While measuring with the DAW has some merit, go and get (lend) a Neutrik Minilyzer XL2.

Like one of these? I'll see about borrowing one.

can you measure your individual channel thru a direct out? that would be interesting in relation to the bus only and 16 channel routed measurements. to check if the numbers add up correctly.

Hadn't thought to do that, that'd be an interesting comparison.

The Trident S65 specs published in the original manual were:

Overall System Noise
-75dBu one channel routed to one group, faders set for unity

Line Input
-84dBu (Eq in, direct output)

 

[radardoug]

No.

Noise measurements are done as RMS. When you normalize my file in Audacity, you are gaining it up until the highest peak hits 0dBFS. That is not the same thing.

Attached is a screenshot of the noise file analysis as done by Audiofile's Wave Editor app. As you can see, the peak level is -70 dBFS, but the RMS level is -87.1 dBFS.

No, you are not correct. If you had a proper noise meter you would see. Yes, when I gain up in Audacity, I end up with the peak noise signal. However it is pink or white noise with a low crest factor. To prove that I allowed Audacity to amplify with clipping, and I can add another 6dB of gain and filter out the random peaks. You can then allow another 3dB for the rms conversion, and get -79 dB.  But it still depends on your original reference level which we don't know.

 

[leigh]

No, you are not correct. If you had a proper noise meter you would see. Yes, when I gain up in Audacity, I end up with the peak noise signal. However it is pink or white noise with a low crest factor. To prove that I allowed Audacity to amplify with clipping, and I can add another 6dB of gain and filter out the random peaks. You can then allow another 3dB for the rms conversion, and get -79 dB.  But it still depends on your original reference level which we don't know.

Okay... so you want me to believe that multiple RMS meters in Pro Tools, as well as file-based analyses by Wave Editor and Triumph (attached), are all wrong about the RMS level, and that the numbers you are pulling out of thin air to convert peak level to RMS level are right.

Add another 6dB to clip the normalized file, because that "filters out the random peaks"... what kind of method is that? Are you just adjusting gain based on the waveform view on the screen? And the simple 3dB adjustment you toss out to convert between peak and RMS only works with sine waves.

Pink or white noise can have all sorts of crest factors... not that this is either, since it contains a low-level 60Hz hum and its harmonics.

I am all for healthy debate and learning, Doug. You seem to be determined to shoot me down here, however, and defending my methods in the face of this bias is not a good use of my time.

 

[radardoug]

Sorry, I can only go by the data file you gave me. If you look at that file, near the front there is one solitary noise spike which sets the peak level.  But it is nowhere near the 17 dB difference between the peak level and your rms level.
It was my understanding that you were interested in the noise level of your console. I am concerned that the tools you are using are not giving you an accurate result. I do not take Protools as the definitive authority on anything, and certainly not measurement.
However I take it that you wish to shoot the messenger, so I will consider myself shot.

 

[leigh]

Sorry, I can only go by the data file you gave me. If you look at that file, near the front there is one solitary noise spike which sets the peak level.  But it is nowhere near the 17 dB difference between the peak level and your rms level.

Yes, there happens to be one abberant spike in that file, around 200ms in. Which is completely inconsequential to anything I have said, since I have been talking only about RMS level, not peak level. (Indeed, if I chop the first 500ms off the file and re-analyze it, the RMS value remains at -87.1 dB while the peak level drops slightly to -72.8 dB.)

I do not take Protools as the definitive authority on anything, and certainly not measurement.

I am not taking Pro Tools as "an authority" on anything here. The analog front end of Pro Tools is as good as the A-D converters you pair with it, and the Lynx Aurora I am using is pretty darn good. After the signal is in the box, I am simply saving it as a file to disk, a task during which I trust Pro Tools will not change the level of the digital signal unless I ask it to. And then various software tools, from Avid as well as from other vendors, are being used to measure the signal.

However I take it that you wish to shoot the messenger, so I will consider myself shot.

Far from it. I welcome the messenger; I will however shoot the obfuscator. Rather than perform a real RMS analysis of the file I sent (either with software, or by sending it back out a -10dBV output and reading a hardware noise meter), you came up with some contrived method of converting the peak value to an "RMS" value, and tried to pass it off as factual. That is not helpful in any way.

 

[blake1234]

Okay I just discovered this entire thread a few days ago.  I see nobody has posted since October, I'm hoping someone will see this.  I'm fairly new with all the electrical stuff, so please forgive me if I miss something simple but important.

I have a Trident Series 65 which I am having some noise issues with.  It has 24 input channels and 8 group sends.  Other than the number of channels, I'm assuming that all the schematics/mods discussed here are perfectly relevant to my board as well (please correct me if I'm wrong).  My problem is that I'm getting noise in my monitors - from the main stereo outputs - when everything is muted, unassigned, faders down and everything is unplugged from the rear.  With the master fader down in this situation, there is a small amount of noise in the monitors: what I would call white noise, although I'm not sure if that's technically correct.  It is a hissing/rushing-water kind of sound: I suppose the words buzz and hum could both be used to describe it.  As I pull the master fader up, this noise increases.  When the master fader is at unity, this noise is unacceptably loud.  The noise persists even if I mute the master channel.

So first off I should ask if it would be plausible that the modifications discussed here could help my situation?  Regardless of the answer to that question though, I have a few questions about this thread for Leigh.


- When you say you converted insert send jacks to TRS (for ground separation) I believe you're referring to the GROUP inserts, but would this also help the channel insert sends/returns?

- Did the last schematic you posted for impedance balancing on October 6th work?  I got lost about the status of that situation as the conversation turned to other things.

- Did you leave the “momma to momma” jumper wires for power and ground in place, or remove them, and substitute with ONLY the 18ga run from the barrier terminal?

- Did you find the ground mod you did with spare pin 7 on the groups helped?  Also, as for the configuration of this I'm a little lost.  One end of the wires connects to motherboard pin #7 and the other end (running toward the top of your picture) connects where?

- How exactly did you mount the copper bus bar?  This needs to be electrically isolated from the chassis right?


Thanks SOO much, this forum has been very helpful! 

 

[radardoug]

You need to do MEASUREMENTS and post them here. Every amplifier circuit has noise, its a question of how much.
If your monitoring is badly set up you may have too much gain in the monitor path, and you will hear noise.

 

[leigh]

Okay I just discovered this entire thread a few days ago.  I see nobody has posted since October, I'm hoping someone will see this.  I'm fairly new with all the electrical stuff, so please forgive me if I miss something simple but important.

hi Blake,

Welcome and I'm glad you find this thread helpful. I'm still subscribed to the thread, so yes I caught your questions. I don't have time today to get to answering them, but I will soon!

In the meantime, also see here on my tech blog for related posts about my Trident overhaul project:

http://fishboytech.tumblr.com/tagged/trident

cheers,
Leigh

 

[leigh]

I have a Trident Series 65 which I am having some noise issues with.  It has 24 input channels and 8 group sends.  Other than the number of channels, I'm assuming that all the schematics/mods discussed here are perfectly relevant to my board as well (please correct me if I'm wrong).

Yes, definitely the same schematics!

My problem is that I'm getting noise in my monitors - from the main stereo outputs - when everything is muted, unassigned, faders down and everything is unplugged from the rear.  With the master fader down in this situation, there is a small amount of noise in the monitors: what I would call white noise, although I'm not sure if that's technically correct.  It is a hissing/rushing-water kind of sound: I suppose the words buzz and hum could both be used to describe it.  As I pull the master fader up, this noise increases.  When the master fader is at unity, this noise is unacceptably loud.  The noise persists even if I mute the master channel.

First question - are your monitors really plugged directly into the main stereo outputs? Or are they running through a DAW or monitor controller, which you are then sending to your monitors?

The noise persists even if I mute the master channel.

The master "mute" button affects only the monitor outs (on the schematic, it's labeled as "control room mute"). The main stereo outs ("remix o/p left" / "remix o/p right" on the schematics) are not affected by this button.

So first off I should ask if it would be plausible that the modifications discussed here could help my situation?

Maybe, too early to say...

I will get to some of the specific questions here, but for others I'd need to look back through notes and don't have quite enough time to find those right now.

When you say you converted insert send jacks to TRS (for ground separation) I believe you're referring to the GROUP inserts, but would this also help the channel insert sends/returns?

Yes, the same mod would help the channel insert sends & returns. However, I don't have any channel inserts wired up in my setup, so I didn't bother with them.

Did you leave the “momma to momma” jumper wires for power and ground in place, or remove them, and substitute with ONLY the 18ga run from the barrier terminal?

I did cut the momma-to-momma jumper wires for power and ground. In noise measurements before and after, I don't recall there being a significant difference (although I could double-check that when I have time), so if I was doing it again I would probably leave them in for redundancy.

How exactly did you mount the copper bus bar?  This needs to be electrically isolated from the chassis right?

If you use 1/2" bus bar, you can squeeze it in along the top of the motherboard's metal mounting plane. Yes, the bar does need to be electrically isolated from the chassis. See the attached photo for a detail shot. (In this photo, the motherboard has been removed from the console.) I took out some of the existing small nuts along the top of the metal mounting plane, that attach it to the chassis bottom, and substituted them with the white plastic bolts and nuts that you see here (with plastic spacers to keep the bus bar lifted to about the same height as the metal mounting plane). With mounting holes drilled roughly centered in the bar, that leaves a consistent gap of about 1/8" between the bus bar and the motherboard mount.

It was kind of fiddly getting everything lined up right for that. If I'm remembering right, in order to get those mounting holes drilled in the right spots along the bar, I took all the metal nuts off the top of the motherboard mount, laid the copper bus bar flat along the top of the mount, temporarily taped it in place, and then used a Sharpie to mark the spots for the holes on bus bar from the bottom (poking up through the existing holes in the bottom of the chassis). Especially if you have several motherboards in your console (which, for a 24 channel Trident, you would), you cannot count on the mounting holes for one motherboard lining up exactly with the holes for another motherboard, and with only an 1/8" gap to play with, you need to be precise.

cheers,
Leigh

 

[leigh]

...you can learn a lot from simple relative measurements. Better is always better,  as long as you know  what you are looking at, thus my advice to bench mark the bench noise floor etc.

I just realized that in the ensuing debate, I didn't respond to this. Yes, I am in full agreement, which is why I took lots of DAW recordings before starting work, so I'd be able to benchmark improvements.

And, for the record, when it comes to reducing the noise in the mix bus, the improvements made by adding the copper bus bar are shown as before/after shots here:

http://groupdiy.com/index.php?topic=57113.msg734246#msg734246

 

[leigh]

- Did you find the ground mod you did with spare pin 7 on the groups helped?  Also, as for the configuration of this I'm a little lost.  One end of the wires connects to motherboard pin #7 and the other end (running toward the top of your picture) connects where

Those pins are still unused. During testing, the far end of the wires connected to the Ring of the new TRS output jacks for the group inserts. And the near end of the wires (to pin 7, on the group's PCB), was jumpered to various ground points on the PCB, trying for the lowest noise on the insert send.

But in the end, if I recall right, I didn't find a ground point on the PCBs that was any better than tying the group insert send jacks' Ring connection to the central "star ground" point of the console (through 100 ohm resistors, to impedance match the signal being sent out the Tip of the TRS jacks).

So pin 7 on the groups is still open for mods. For example, I was thinking of adding silent mutes on the groups, using vactrols, so that spare pin could be used to run the control voltage to the vactrols. (The Trident's switches are all direct/mechanical pushbuttons, and they pop audibly enough that I wouldn't switch them in and out during a mix.)

 

[blake1234]

I thought I replied to this a while back, but I guess I did not.  Thanks Leigh for the help!  To respond to your questions for me - I am not using a DAW controller, I am going straight from board to monitors.  I am having the same noise problem coming out of both the stereo output and the monitor output.

I have a couple more questions for you. 

-My master module's PCB looks a little different than yours,  I will attach a picture of it.  I'm wondering what exactly are all the points that you've grounded to the 12 AWG local bus bar?  I know you say the master fader's ground reference is tied directly to this, I've put a red circle around what I believe is my master fader's ground (the brown wire coming off the bottom of the fader).  Is that right?  Yours shows no jumper on that point.

-What are the orange (yellow) capacitors doing on your master module picture and what value are they?  The bottom left corner of my master module does not have what appears to be an IC chip, to which you've connected a capacitor and a purple jumper.

-I'm reading through this forum again and towards the beginning you mentioned on Aug 28th: "And, when it comes to sensitive points in the circuit, like the summing amps' 0V reference (+ input), I would also give those points direct connections to the bus bar. No more sharing PCB traces with power bypass caps."  Did you do this in the end?  If so, what other "sensitive points" did you connect to the bus bar?

-With chassis grounding pin 1 on Line and Mic inputs, I'm just wondering exactly how you did this? (mainly where signal ground goes).  The signal ground is the internal cable shield right?  You mentioned something about a 12K resistor going to the signal ground.  On my board, there is no 12k on pin 1 of the mic inputs, just two 12K resistors from pin 2&3, both go to the same tab on the phantom power switch.

-In the discussion about impedance balancing the the unbalanced 1/4" outputs, I got lost.  I'm wondering if you could clear up exactly what you did for these as there seemed to be a lot of resistors being thrown in.  A picture would be fantastic!  If the impedance balancing works, I would like to do the same, I just don't want to screw it up.

Thank you so much for your help and patience!
-Blake

 

[blake1234]

Okay I just discovered this entire thread a few days ago.  I see nobody has posted since October, I'm hoping someone will see this.  I'm fairly new with all the electrical stuff, so please forgive me if I miss something simple but important.

hi Blake,

Welcome and I'm glad you find this thread helpful. I'm still subscribed to the thread, so yes I caught your questions. I don't have time today to get to answering them, but I will soon!

In the meantime, also see here on my tech blog for related posts about my Trident overhaul project:

http://fishboytech.tumblr.com/tagged/trident

cheers,
Leigh

Hello Leigh.  I am trying to send you a personal message, but your inbox is full, so I hope that you will see this reply and maybe I can PM you soon.

 

[leigh]

Uh oh, I hit 50 messages in my Inbox, and couldn't receive any others. Welcome to the mid-90's!

I'll PM you my email address, or you could just ask your questions here on the forum so everyone can benefit from the conversation.

 

[blake1234]

A couple questions about impedance balancing/ground separation.  As soon as I plug my power supply in, all of my audio grounds are connected to the chassis (as per multimeter continuity test).  Without the PSU plugged in, audio and chassis ground appear to be separate.

- I'm wondering if your Acopian power supply has somehow separated the signal ground and chassis ground, or if there's something I'm missing here? 

Basically as long as my power supply is plugged in, I don't see any way you could separate signal and chassis ground.  This has me confused because you stated that chassis grounding pin 1 was one of your most notable improvements. 

- I'm still trying to figure out if there is any advantage to impedance balancing the insert RETURN jacks.  Would converting to TRS and having a 100Ω resistor on the ring help anything when it's receiving a balanced signal?  Or should I just chassis ground the unbalanced TS of the return (leaving audio ground hanging)?  Or do nothing? 

You say in the forum that you had to plug dummy patch cords into the group insert returns on the patch bay in order to see the terminating resistor behave as it should.  I have installed insert switches onto my board and I'm wondering if that will change this scenario.  I cut the return pcb trace right at the molex pin.  When the switch is in "bypass," the insert is wired directly from the send molex pin to the return's pcb trace (bypassing the return molex completely).  With the patch bay hooked up, the send is simultaneously going out the send jack, 1/2 normalled to the return at the patch bay, and all the way back to the return jack/mol  ex (which is disconnected when the switch is in "bypass"). 

- Will the whole patch bay loop (especially the return path) induce noise when my switches are in bypass simply because the send is electronically connected all the way through?  Or was your problem in seeing the 100Ω resistor's effectiveness because the patch bay normal connected to the rest of the channel circuitry at the return end?

Thanks

Blake

 

[leigh]

Blake,

Please have a look at the links in the "bibliography" I posted earlier in this thread for some useful background on grounding and interconnections.

A couple questions about impedance balancing/ground separation.  As soon as I plug my power supply in, all of my audio grounds are connected to the chassis (as per multimeter continuity test).  Without the PSU plugged in, audio and chassis ground appear to be separate.

- I'm wondering if your Acopian power supply has somehow separated the signal ground and chassis ground, or if there's something I'm missing here?

That sounds fine. All "grounds" (chassis grounds, audio grounds, power supply grounds) have to eventually reference back to some ground point, let's call it "ultimate ground". So I believe my Acopian power supply does connect the two grounds you are asking about, although they are kept separate within the console itself.

If I recall the theory correctly, the basic goal with using separate grounds is that they connect in one and only one spot. And that one spot is far "downstream" from points sensitive for audio signals picking up noise. For example, if you have stray currents on a console's chassis ground (say, from "catching" incoming RF interference), then those currents flow all the way back to the "ultimate ground" point on their own conductor path (chassis metal, wires, etc). They therefore do not flow across the same wires used for audio grounds, and they therefore (in theory) don't show up as noise on the console's internal, unbalanced (zero-volt-referenced) audio signals.

I'm still trying to figure out if there is any advantage to impedance balancing the insert RETURN jacks.  Would converting to TRS and having a 100Ω resistor on the ring help anything when it's receiving a balanced signal?  Or should I just chassis ground the unbalanced TS of the return (leaving audio ground hanging)?  Or do nothing?

You can't "impedance balance" an insert RETURN jack, because that is an INPUT. You could convert it to a balanced input if you care to (by adding a balanced input circuit, or by slapping a transformer on the input before the existing, unbalanced input circuit). And this relates to the next question...

You say in the forum that you had to plug dummy patch cords into the group insert returns on the patch bay in order to see the terminating resistor behave as it should.  I have installed insert switches onto my board and I'm wondering if that will change this scenario.  I cut the return pcb trace right at the molex pin.  When the switch is in "bypass," the insert is wired directly from the send molex pin to the return's pcb trace (bypassing the return molex completely).  With the patch bay hooked up, the send is simultaneously going out the send jack, 1/2 normalled to the return at the patch bay, and all the way back to the return jack/mol  ex (which is disconnected when the switch is in "bypass").

The reason I had to plug dummy patch cords into the group inserts returns on the patchbay, is that my insert SENDS were normalled to the UNBALANCED insert RETURNS. Impedance balancing an output only makes a whit of difference if it is connected to a BALANCED input. So, I had to break that patchbay normal in order to measure the noise difference (for my measurements, the insert sends then connected ONLY to the balanced inputs on my A/D converter).

I do not fully understand the patchbay wiring setup you are describing with your board, or what your question is about the insert bypass switches you installed. (If you have the switch in bypass, then you're shunting the audio signal internally, so you don't have to worry about balanced or unbalanced interconnects.)

Leigh

 

[blake1234]

Okay, thank you Leigh. 

My question about the insert switches in bypass is that even though it's being shunted internally, the insert send/return jacks on the back of the board are still plugged out to the patch bay (which has a normal).  The way I've wired the switches, the send signal is always connected to the send jack, I just tapped off of the send's molex pin for the input of the switch.  There is electrical continuity from the send all the way to the return via the patch bay.  I am concerned about noise from the return line as per this electrical continuity.  Are you saying that the internal shunt overrides the snake to the patch bay because it is the shortest path and therefore I needn't worry about the dummy patch cords?
 
I have one more simple question right now.  At the beginning of the forum you say you swapped the summing amps for LME49710's.  Are these just IC20 and IC22 on the master section?

 

[blake1234]

Also, just to clarify, it is advantageous to chassis ground all the jacks on the rear of the console whether balanced or unbalance, input or output, XLR or 1/4"?  Or are there some that this is not a good idea?

Blake