Hiss in Trident 80 Monitor summing amps

[RFSiesta]

Hi,
Anyone really experienced with Trident 80 maintenance?
We spent lots of time with the issue, I'm asking here for new ideas what could be the problem:

Our Trident 80 has too much hiss/noise on the 24 monitor (group) channels.
We reduced the problem to the summing amps that are located in each channel, group bus 1 is summed on the summing section of channel 1, 2 on 2 etc. So the signal gets noisy already at the monitor insert sends on the pb.
Since the hiss is on any channel, yet at different levels, and did go almost away when moving channels/the global pcbs with edge connectors for the channels, but could not at all be reduced to bad contacts/cold solder points or the like - we are running out of ideas now.

Some more information on the desk:
We fitted the mon summing amps (exactly the same design as the remix bus amp) with AD797 instead of 5534 for the first, and OPA134 for second stage.
With this config, the channels were reacting to small movements of the boards, cabling etc with more or less hiss. It actually went away almost completely! So we figured there's a way to solve this obviously mechanical problem, but we could not find a "center": It went away when we rattled on channel 24, and came back when moving channel 5 f.ex. Then we got away by slight pressure on channel 8 or so, and it came back when touching the pcb under channel 21. We did this for days, without any solution. What a f*** joke!
When returning to original5534/071, the hiss was the same amount, but seemed to be not reacting to movement any more. (!?)

We checked all the usual suspects: Ground seems good, edge connectors are clean and tight, we removed the "slate" connections (we have no use for in our current setup), which actually are just a bus that's connected through series resistors directly to each of the buss traces. This has reduced noise a little bit, but not at all satisfied us.

One very important information could be this:
The board is divided in two halfs, 1-18 and 19-36. Inbetween we have 24 additional TSM eq modules. The left and right half of all bus traces are interconnected, gnd and voltages are star connected to several central bridges.
We found that the noise induced in the left half (especially lf content) is PHASE INVERTED to the noise of the right half - which does not make any sense at all to us, since the two blocks are identical...

As I said, we're running out of ideas. People with enough experience with global mixer issues are extremely rare in Austria, we may need to find out ourselfes.
Any help appreciated!

 

[JohnRoberts]

Hi,
Anyone really experienced with Trident 80 maintenance?
We spent lots of time with the issue, I'm asking here for new ideas what could be the problem:

Our Trident 80 has too much hiss/noise on the 24 monitor (group) channels.
We reduced the problem to the summing amps that are located in each channel, group bus 1 is summed on the summing section of channel 1, 2 on 2 etc. So the signal gets noisy already at the monitor insert sends on the pb.
Since the hiss is on any channel, yet at different levels, and did go almost away when moving channels/the global pcbs with edge connectors for the channels, but could not at all be reduced to bad contacts/cold solder points or the like - we are running out of ideas now.

Some more information on the desk:
We fitted the mon summing amps (exactly the same design as the remix bus amp) with AD797 instead of 5534 for the first, and OPA134 for second stage.
With this config, the channels were reacting to small movements of the boards, cabling etc with more or less hiss.
When returning to original5534/071, the hiss was the same amount, but seemed to be not reacting to movement any more. (!?)

That symptom sounds like something may be oscillating, sometimes audio paths will rectify very HF oscillation as hiss. If this is common to all or many channels , look for path through common connection like PS.

A scope might be helpful for troubleshooting HF oscillation.

good luck

JR

We checked all the usual suspects: Ground seems good, we removed the "slate" connections (we have no use for in our current setup), which actually are just a bus that's connected through series resistors directly to each of the buss traces. This has reduced noise a little bit, but not at all satisfied us.

One very important information could be this:
The board is divided in two halfs, 1-18 and 19-36. Inbetween we have 24 additional TSM eq modules. The left and right half of all bus traces are interconnected, gnd and voltages are star connected to several central bridges.
We found that the noise induced in the left half (especially lf content) is PHASE INVERTED to the noise of the right half - which does not make any sense at all to us, since the two blocks are identical...

As I said, we're running out of ideas. People with enough experience with global mixer issues are extremely rare in Austria, we may need to find out ourselfes.
Any help appreciated!

 

[pucho812]

Whenever fault finding on a console start big and work your way down to something manageable.  The console is modular so you can do a channel at a time.   

When was the last time if any the board has had a recap?

 

[radardoug]

When you say hiss, what level is the hiss at? dB below +4 please. As others have said, your revised design could be oscillating.
What test gear do you have?

 

[RFSiesta]

That symptom sounds like something may be oscillating, sometimes audio paths will rectify very HF oscillation as hiss. If this is common to all or many channels , look for path through common connection like PS.

A scope might be helpful for troubleshooting HF oscillation.

We checked for oscillations with a scope and it looks like something at 10mHz is there, though only at a couple of mV. It went away when refitting the original chips, but the noise didn't.

Btw: Noise is at about 30-40dB below zero, but only in the sum of all 24 monitor channels. We really couldn't easily settle to a constant level, because it changed a lot when moving parts.
As said, the stranges thing is that it does not change so much any more, could be that the mechanical (if so) problem has developed to be constant - from our efforts to find the spot by moving things around... sometimes the noise was almost gone by as little a movement as walking across the room or by pushing the desks frame a little from outside.
Is there any chance that the noise is coming from other unstable parts of the circuit? Btw, we seem to have a global grounding issue in the whole studio, which leads to hum problems sometime. We didn't find the reason yet - we have two independent grounders and tried both individually and combined.  There's funny things like 0.4V DC between the grounders, that must come from the earth around our building. We also measure AC on ground, at 150Hz (?) as it seems. Voltage varies from almost zero to 0.3V or so.
We believe it's two different things, but both issues could also be connected.

We already commissioned a technician from radio monitoring and a specialist from electric power company to measure, both didn't find the error yet... Seems like a really hard nut to crack.

cheers Richard

 

[RFSiesta]

We checked for oscillations with a scope and it looks like something at 10mHz is there, though only at a couple of mV. It went away when refitting the original chips, but the noise didn't.

Edit: We do still see HF on the scope at a couple of MHz, something we always measure - and ANYWHERE in the studio.  Even on a straight piece of wire, it seems.
And- with everything switched off except the scope.
The HF at 10mHz is about the same level.
As I said, there could be more than one problem at a time to make things more complicated.

 

[radardoug]

Sounds like you are in a high rf field. That wont help at all. How are you measuring your noise level? With a proper noise meter? Or just guessing it off the scope?

 

[currentstatus]

Does sound like HF interference.. strong RFI in your area maybe?

10MHz could be ham radio or a harmonic from a local LW radio station? Do you have access to a portable RF frequency counter.. even a digital LW radio to monitor anything heavy around 2MHz and/or similar at 10MHz.

I get swamped by my local LW station if gear isn't connected/screened properly.

 

[andow]

Hello everybody and thanks for your help so far! Richard and me are trying to solve this problem and I did some more measurements today.

First, only a loop with some alligator clips to see HF interference. I measured approximately 95MHz, the amplitude changed when moving the wire loop. So in all further measurements this frequency was present.



Next, I measured between the main ground on the desk and the copper bar on the first 18 channels with the power supply switched OFF! There were some transients with 64kHz and a sinewave with 1MHz and other strange artefacts with something like 4,2MHz. And of course 95MHz...






I'll post some more info and measurements later...

 

[radardoug]

Good stuff, so you are in a high rf field. Try recording some tone into your DAW, then noise. Do a gain on a piece of the noise. Some DAWs like Audacity will show you immediately how much gain you can apply. This will give you your audio signal to noise.
Or buy a used HP 334 n&d. Cheap as chips.

 

[RFSiesta]

Good stuff, so you are in a high rf field. Try recording some tone into your DAW, then noise. Do a gain on a piece of the noise. Some DAWs like Audacity will show you immediately how much gain you can apply. This will give you your audio signal to noise.
Or buy a used HP 334 n&d. Cheap as chips.

I've got to add some information on how noise is an issue here: The desk itself, as far as channels and summing are concerned, is in fact clean enough to work with.
BUT:
We use the split monitor section for outboard compressors, especially for parallel compression of groups. It's a super fast and intuitive workflow, I can switch between outboard on the fly and compare ten different compressors within seconds.
Since each of the monitor channels is a summing system of it's own (with the exact same topology as the "remix" master bus), and even get up to 10dB makeup gain through the comps, then adds up by 3dB per channel, the monitor channels should be really as quiet as possible for the setup to be useful.
That said, we could actually live with a bit of white noise on each channel - but the noise we experience is "structured" and includes a lot of hum, crackling and sometimes even tones (around 300Hz?!?) as well...

We will next try to run the whole board on batteries: Should be no issue to use 2*12V batteries on this chip console, logic/routing and LED voltages are not needed for the test.
More information soon!
cheers
Richard

 

[radardoug]

Running it on batteries is just stupid. Think about this for a minute. Trident made lots of these consoles, they worked well. If yours has issues you need to look at the environment you are in. If high rf is a problem, you should be looking to get the rf out, or relocate.

 

[andow]

Running it on batteries is just stupid. Think about this for a minute. Trident made lots of these consoles, they worked well. If yours has issues you need to look at the environment you are in. If high rf is a problem, you should be looking to get the rf out, or relocate.

Well, it's not THAT stupid! (I mean, we sure have done more stupid things... ;D)
Like Richard stated before, we don't have the cleanest earth. So in order to rule out the influence of earth connection and power lines I gave it a quick shot and fired up the desk with +/-13V on the audio rails. And guess what: the desk was quiet! With all group channels up there was there was just white noise and nothing of the hum/crackle/noise/popping/tone-stuff we had before.
I just got a short glimpse on the oscilloscope and saw that there still was some HF-noise on the output, but this doesn't seem to have an influence on the instability we experience when using the power supply.
(Btw. we have the original and a new power supply from GRS and the issue is more or less the same with both, so the power supply itself is fine.)

 

[radardoug]

Well you have solved the problem! Of course, as soon as you hook it up to anything the problem will be back. But you have proved it is NOT the console. Its your crap environment.

 

[RFSiesta]

Well you have solved the problem! Of course, as soon as you hook it up to anything the problem will be back. But you have proved it is NOT the console. Its your crap environment.

No. It's not at all "stupid" to run the console on battery and to rule out the PSU as a source of problems. As I have stated before, we had different high-level experts in the studio - an expert for radio monitoring from Austrian authorities (with 100k€ equipment), a professional for electricity from the largest power supply company in the area, no one found the reason. We have a completely new power installation and we did make new earthing and a new power feed for the building. Our environment is NOT crap, as far as we can tell.
What could still be crappy, is the ground itself we are on. And if it would be, we would have to make even better earthing to cope with that. But it' really very hard to measure without a really clean reference.

What we found out is this:
-It is NOT HF influences through air into the board, otherwise we would have the same issue on battery as on PSU.
-it could still be an issue of the console in combination with power issues, like flawed grounding scheme, rendering the desk too sensitive for ground influences.
-the board could be quiet if we had completely clean ground/voltages
-we will have to find which is easier to achieve: ultraclean ground or a better grounding scheme in the board.

Since we have hum issues in other places of the studio, I guess it could be the ground. Let's hope there's a way to achieve this! Because we have already tried - several times.

 

[RFSiesta]

Still no good news: As of today it seems that we still have strange tones at approx. 300-500 Hz - running on battery.
Amplitude and sound are erratic, not constant. No pattern to be found, sounds random - but I've found the tone to go away if I route remix to master monitor with the remix-to-monitor switch.
Will check the schematics what exactly changes when this button is activated apart from a buffered signal being split and feeding two high-Z opamps inputs  - There should be no difference at all, of course.

And : yes we deactivated the oszillator  :  so it's not crosstalk from there.

... Trident made lots of these consoles, they worked well. If yours has issues you need to look at the environment you are in.

There's still the obvious possibility of electronic devices getting damaged and as a consequence not working well any more.
If so, the idea was to find the damage and repair it.
After some thought, Im still not sure if we have ruled this possibility out!

As often, it could as well be a combination of different reasons: Some of them environmental, some inside the board.
As a matter of fact, much of old equipment tends to be more sensitive to HF issues than new devices, simply because we live in a world with a couple of magnitudes stronger high frequency interferences in the air, than the developers of these devices.

Then again, we hear no or almost no interaction with WIFI, cell phone, switching power supplies and the many sources of interference around us. Switching all off did not make the slightest difference. Not even an extremely trashy Hongkong import LED psu that the radio control tech found to produce high-energy and broadband frequency bursts all over the spectre 8)

 

[radardoug]

Hmm, so you have at least two experts there. What has your rf expert found in terms of stray fields? Are you near any transmitters?
If you have a power expert there, and he has inspected/installed the power system, then if it does not have an adequate ground, he is in violation of your local laws I am sure. Yes, the Trident will not cope with things like cell phones. However as I said,  Trident built lots of these and installed them all over the place. Your mystery tones could be demodulated rf.  The first thing you should do is remove all modules except the master section, and then make sure the master section is working OK.
You say you have hum problems elsewhere. This raises a red flag. I was once involved with a studio next to a factory that had a large power transformer on a common wall. This transformer caused problems in the studio. The answer was move the transformer or move the studio. Guess what, the studio had to move.

 

[RFSiesta]

Hmm, so you have at least two experts there. What has your rf expert found in terms of stray fields? Are you near any transmitters?
If you have a power expert there, and he has inspected/installed the power system, then if it does not have an adequate ground, he is in violation of your local laws I am sure. Yes, the Trident will not cope with things like cell phones. However as I said,  Trident built lots of these and installed them all over the place. Your mystery tones could be demodulated rf.  The first thing you should do is remove all modules except the master section, and then make sure the master section is working OK.
You say you have hum problems elsewhere. This raises a red flag. I was once involved with a studio next to a factory that had a large power transformer on a common wall. This transformer caused problems in the studio. The answer was move the transformer or move the studio. Guess what, the studio had to move.

The RF expert found that the strongest field came from the LED switching supplies. We are not close to transmitters, everything seems to be "normal".
As stated before, we have more than adequate ground from a security viewpoint, close to zero Ω - but we seem to have ripple and a little bit of DC (<1V). The dc between the two available grounding systems could be from oxidation or something. Having some ripple in the earth underneath a building is strange, but (as I found by research) seems to be not too unusual.
What we would maybe need to have is a deeper earth rod that connects to the ground water, as I expect this to be the best available reference in our area.

Yes, the tones could also be coming from rf. But also I'm probing for the mystery tones in context of the lofreq oszillator Trident has built into the 80B - which, in combination with the "slate" switch, has been used to record a subsonic sound to tape that will be recognized as tone when rewinding the tape. Very interesting thing, btw! Maybe something is faulty here, but it's only a very wild guess.
-> The tones only come up if we pull up some of the monitor channels. Usually, they go away when ALL mon channels are on...?!?
So again, it looks like a global problem the desk has, rather than an isolated problem.

Of course we checked all the modules and parts of the desk apart from each other, pulled out all monitor modules, all channels, whatever. We could almost write a book on the combinations we tried - without getting us anywhere so far.

And no, we don't have a transformer close. The next medium voltage transformer is about 50m away from us. But the radio tech didn't really check for 50Hz fields, as this is usually of no importance for the radio control authority. So, still there could be a LF field around.
I plan to build a simple battery powered measuring device for LF magnetic fields, maybe we can see things clearer then...

Thanks so far,
cheers
Richard

 

[radardoug]

Use a guitar with single pole pickups and a portable amp. Instant field finder!

 

[RFSiesta]

Use a guitar with single pole pickups and a portable amp. Instant field finder!

you've got a point here  8)

Hum: We use this simple arrangement that reliably produces hum in amplifiers: A guitar amp, a keyboard and a second device (amp, mixer - doesn't matter) thats attached to the keyboard with a jack. Only the amp needs to be turned on. If we connect the second one to power, the first one hums. So far, so normal: But the hum is more than usual.
This is a classical ground loop, but IF the hum came from magnetic fields, you should expect it to be sensitive to movement. But it's not at all. Again, the only source I can think of is our ground connection.

-> This week, we hope to get the electrician to add earth rods to our existing one to increase depth of grounding. Hope this helps- but at the same time, I doubt it. We've lost so much time. We almost tried everything now. It's extremely frustrating.