Unbalanced to balanced input conversion?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi folks,

I'd like to convert this amplifier to balanced inputs (XLR), in the hopes of increasing its CMRR and making it more resistant to ground loop noise.

Would it be possible to bring a phase negative input around to the right side of the JFETs and mix it in without have to add another buffer stage?

I would appeciate any thoughts, and especially specific ideas as to the needed circuit changes.

By the way, the "input from DC servo" is another 10K resistor from a TL-081 based DC servo, so the after-the-op-amp gain of the DC servo is the same as that of the NFB loop.

Thanks,
Paul
 

Attachments

  • psaudio200c.jpg
    psaudio200c.jpg
    98.4 KB · Views: 1,128
SQLGuy said:
Hi folks,

I'd like to convert this amplifier to balanced inputs (XLR), in the hopes of increasing its CMRR and making it more resistant to ground loop noise.


Would it be possible to bring a phase negative input around to the right side of the JFETs and mix it in without have to add another buffer stage?

I would appeciate any thoughts, and especially specific ideas as to the needed circuit changes.

By the way, the "input from DC servo" is another 10K resistor from a TL-081 based DC servo, so the after-the-op-amp gain of the DC servo is the same as that of the NFB loop.

Thanks,
Paul


he he he you make it sound like ground loops are a balanced input issue .....

well its not ....actually is more ground loop techincues and more about grounding properly .... a balanced door will not change much if your installation is poor ....

PA expirience says so ....but this is just my opinion .....
 
What do you suggest for grounding, then? Should I be breaking off ground pins on most of my equipment's power cords?

It's a bi-amped setup with:

Outlaw 990 Preamp
Behringer DCX2496 Crossover
SUMO Andromeda II Amp (Hi/Mid)
PS Audio 200c Amp (Low)
Paradigm Seismic 12 (Sub)
Bryston 2B-LP (Rears)

Every piece of equipment has a ground connection for its power cord.

The Bryston is unbalanced but has a ground defeat switch. With ground defeated I don't get ground loop noise from the Bryston.

The SUMO and the Paradigm have balanced inputs. I don't get ground loop noise from them.

The previous bass amp was an Adcom 555ii. It did not have a grounded power cord. I didn't get ground loop noise from it. The PS Audio ground is not designed to be defeated.

Neccessarily, with this much equipment, and this much power draw, I am not going to plug everyting into one outlet.

The noise is not super loud hum, but it's louder than I'd like it to be. Disconnected from the system, the noise from the PS Audio is about 300uV... connected to the system it's about 5mV. I would have thought that adding the 20dB or so of common mode rejection that balanced inputs would bring would help quite a bit here.
 
SQLGuy said:
What do you suggest for grounding, then? Should I be breaking off ground pins on most of my equipment's power cords?

It's a bi-amped setup with:

Outlaw 990 Preamp
Behringer DCX2496 Crossover
SUMO Andromeda II Amp (Hi/Mid)
PS Audio 200c Amp (Low)
Paradigm Seismic 12 (Sub)
Bryston 2B-LP (Rears)

Every piece of equipment has a ground connection for its power cord.

The Bryston is unbalanced but has a ground defeat switch. With ground defeated I don't get ground loop noise from the Bryston.

The SUMO and the Paradigm have balanced inputs. I don't get ground loop noise from them.

The previous bass amp was an Adcom 555ii. It did not have a grounded power cord. I didn't get ground loop noise from it. The PS Audio ground is not designed to be defeated.

Neccessarily, with this much equipment, and this much power draw, I am not going to plug everyting into one outlet.

The noise is not super loud hum, but it's louder than I'd like it to be. Disconnected from the system, the noise from the PS Audio is about 300uV... connected to the system it's about 5mV. I would have thought that adding the 20dB or so of common mode rejection that balanced inputs would bring would help quite a bit here.

Try a simple check to find how much of your units have earthed their power supply 0V node, with a ohmmeter. Place one of test leads on the earth contact of the mains cable plug. Place the other test lead on the shield of signal input sockets. Make a scratch on a paper: place a dot for each input shield point and for each mains earth point. According your measurement results, draw lines between dots which are shorted. Then you can see any existing gnd loop on the finished scratch. Remember that, avoidance of gnd loops between multiple connected units has the same value like the star point in power supplies gnd node.

Fotios
 
Sorry, but I don't understand the point of this exercise.

I can tell you that the crossover has its chassis ground connected to its signal ground, and that the PS Audio amp also has its chassis ground connected to its signal ground.

As I mentioned before, plugging all of this equipment into a single outlet, which is what it sounds like you're suggesting, is not practical. Currently I have multiple paths to earth ground (through the separate power cords and through the patch cables), and I need a solution that will mitigate that in a pratical manner.

If I had balanced inputs to the PS Audio, I could float the signal shield at the amp and take the signal just from the differential between the XLR + and -. At that point I would only have one path to ground from the amp, instead of two.
 
SQLGuy said:
Sorry, but I don't understand the point of this exercise.

I can tell you that the crossover has its chassis ground connected to its signal ground, and that the PS Audio amp also has its chassis ground connected to its signal ground.

As I mentioned before, plugging all of this equipment into a single outlet, which is what it sounds like you're suggesting, is not practical. Currently I have multiple paths to earth ground (through the separate power cords and through the patch cables), and I need a solution that will mitigate that in a pratical manner.

If I had balanced inputs to the PS Audio, I could float the signal shield at the amp and take the signal just from the differential between the XLR + and -. At that point I would only have one path to ground from the amp, instead of two.

No, no, i am not suggesting a single outlet. I suggest only, the earth of outlets being like a star point. In your case it is not possible because x-over and PS amp. have earthed their 0V supply node. Make a simple external experiment. Place insulating tape (of course temporarily!) over the earth contact of the mains plug of x-over unit in a such way that its chassis is not earthed more. Check it with the buzzer. Be carefull that the x-over chassis is not in contact with any other unit chassis. Try your system again, and check if the noise dissapears. Let me know the result.

Fotios
 
probably a lot easier to add an external isolation transformer, you can buy an XLR adapter with transformer inside. Be sure that your connections do not join pin 1 and chassis.

Some good reading on the subject

RANE

http://www.rane.com/note151.html


Doug Self

http://www.dself.dsl.pipex.com/ampins/ampins.htm

Ted
 
What is the total power draw? I run three power amps, crossover, preamp, and all signal sources from a single power strip. Everything properly earthed and no hum problems. When I've needed a balanced input, it involved adding an opamp as the easiest way to do it. Probably won't solve this problem though. IMHO, something in the collection has poorly designed grounding.
 
To get signal to the amp I was running some custom XLR to phone cables I'd made, with couplers to another pair of regular phono patches. This evening I rebuilt the custom cables with some longer shielded pair wire. I connected shield to XLR chassis, and left the shield unconnected at the phono end. (There is nothing special about the connection... I am just taking ground and signal + from the XLR pins to the inner wires of the shielded pair, and leaving signal - unused.)

The result was quite a bit more hum and noise than with the previous cables. Defeating the ground pin on the PS Audio fixed most of that, and rerouting the patch cables away from everything else helped a bit more... reducing the total noise to about 4mV.

I think I need to look into some better cables, among other things.
 
I read a bit of the content on Doug Self's site (Thanks Ted).

One bit struck me:

"ADVANTAGES OF BALANCING

It discriminates against noise and crosstalk.

A balanced interconnect (with a true balanced output) allows 6 dB more signal level on the line.

Renders innocuous ground-loops, so that people are not tempted to start "lifting grounds" This tactic is only acceptable if the equipment has a dedicated ground-lift switch, that leaves the metalwork firmly connected to mains safety earth. In the absence of this facility, the optimistic will remove the mains earth (not quite so easy now that moulded plugs are standard) and this practice must be roundly condemned as DANGEROUS."

It's that third point I am looking for. Any recommendations on tranformers... especially ones built into panel mount XLR jacks?
 
For domestic use devices, the balanced routing of signal it is some excessive, because the signal cables are short. In large PA systems where the signal cables are of big length, the balanced routing it is necessary to reject unwanted noise. In a domestic system, the main proffit of balanced routing it is the bigger headroom offered. The noise issue it is of small importance.
In your system, you have a misfit unit (compared with the quality of the rest units) which is the Behringer DCX2496. It is really for kicks. When you see a component with a lot of lights, knobs etc which of the cost it is only 300$ it must be suspicious. I have experience with 2 or 3 different units of Behringer (the issue of balanced in/outs is the same although in every unit) and i have observed the following:
a) Noisy outputs due to cheap analog ICs used.
b) Pop-noise due to cold solder points on pcbs in a new unit from the box.
c) Inaccurate produce of balanced signals, either because the cold signal had less amplitude from the hot or a phase shift between hot and cold signals.
d) Curious increase & decrease of signal level during operation; reason? bad assembly of the header on a ribbon cable inside the unit.
And some others....
In general, the final quality inspection of a finished unit of Behringer it is of low level and i am sure that it is also verry hasty.Otherwise the cost would be higher.
RANE, which mentioned by "tedr" offers units of far better quality.
Moreover, your nice Outlaw 990 preamp, it offers the posibility to drive directly your amplifiers, and without the need of using balanced lines as it offers also single outputs. I am wondering why you have added in the chain the x-over unit. Peace is better than a war with signal cables. Try the simplest. Remove the Behringer x-over from the chain, and connect the amplifiers directly to the single ended outputs of your preamplifier. Then check again if any noise issue exists furthermore.

Fotios
 
Fotios, I think you missed the part where I mentioned that this is a bi-amped system. The main speakers are Magnepan Tympani IV's. The PS Audio drives the bass panels to about 350Hz, and the SUMO drives the mid/high panels from 350Hz up.

The Outlaw 990 only has a built-in summed crossover for driving a sub; it does not do bi-amp.

If you are suggesting I switch back to some sort of passive biamp system, then I would say that, from my perspective, that solution is worse than the problem I am trying to solve.

Also, as mentioned above, I did not have ground loop issues with the Adcom 555ii in place of the PS Audio. I've had the DCX2496 for quite some time in my system and I have not seen the problems you have. I do plan to do some mods to it eventually, but I do not believe it is the cause of problems at this point. I have switched back and forth between the Adcom and PS Audio and I am quite sure the Adcom and the Behringer get along just fine.

The previous crossover I was using was a Rowland Research SVX-3. I sold it after running the Behringer for a while. I know some people have had problems with these units, but a lot of people have also had very good experiences with them.

Lastly, a ground loop is a ground loop, and higher CMRR is higher CMRR; I understand that balanced inputs may not be as important in home systems as in pro systems, but I don't agree that that means they don't still have definite advantages over unbalanced connections in home systems.
 
As you wish master...
In a way, I respect the totalitarian war which you have declared against the noise... as well the bi - amping, because is a good thing... moreover, you know better from me what you want exactly.
So, if your Behringer unit its OK, then the simplest and cheaper solution to convert its LO balanced outputs to single ended, are two "D.I. box" units, which offer also a switch for Gnd lift. You can find these in any musical instruments shop. The passive "D.I. box", it contains a transformer which can make the desired convertion.
I wish you success.

Fotios
 
Hi Fotios,

Thanks for the ideas.

It seems to me that it's equally easy, electrically-speaking, to convert to unbalanced at the Behringer end or at the PS Audio end of the connection, but converting at the PS Audio end still gives me the better noise immunity along the way and there is much more room physically to mount a tranformer or diff amp circuit inside the PS Audio.

For now, the noise level is acceptable, though I am not happy with the ground defeated on the amp. I am going to try some other possibilities in terms of how I connect the shield at the crossover end of the connection and see whether that allows me to reground the amp... I may also play around with the grounding inside the amp to see whether there are some other options there to better separate signal ground from chassis ground.

In the longer run I expect that I'll add some XLR's to the amp with, most likely, LME49710's acting as diff amps to convert to unbalanced and feed into the amp's regular unbalanced input. This is similar to what I plan to do inside the Behringer; there I'm planning to use LME49720's to strip the 2.5V DC offset from the DACs and pass the DAC's differential output straight out.

Cheers,
Paul
 
Paul, as i can understand you are very insistent in your target.
If it is allowed, for a little relax, i would like to add an off topic comment.
Your insistence (which is of course a very good idioma for a man) it remembers me a famous phrase of the Romeo senater Cato:

ET PRETEREA CENSEO CARTHAGO DELENDA EST :)

Greetings
Fotios (or Lucio in Roman)
 
nigelwright7557 said:
I have to agree with everyone else, forget the balanced input and sort out the grounds.

My system has one ground at the input to the amp, all other grounds are removed to remove any possible ground loops.

Just as bad as a ground loop is no ground at all !

One ground = no problems.

Which one did you keep, then? I would assume it would be the chassis ground, because of the safety factor.

So, did you just leave the patch shields/grounds unconnected at the amp end and just run the center/signal wires through?

To be clear, with the bass amp I have:

ground between the amp and the crossover twice = left channel patch cable and right channel patch cable

ground between the amp and the household wiring through the grounded power cord

internally the left channel jack shield connects to the left channel PCB at the point where there is a jumper wire that connects to the ground point of the right channel PCB. From there, a jumper goes to the center grounding point on a 1/2" by 1/8" copper ground bar. That center point is also where the ground wire from the power cord, and the center taps from the transformer, connect. The copper ground bar wraps around to the right and left inside the case to provide the ground connection point for the speaker out binding posts. The case is connected to the grounding bar in a couple of places where they're screwed together for support.
 
SQLGuy said:


Which one did you keep, then? I would assume it would be the chassis ground, because of the safety factor.

So, did you just leave the patch shields/grounds unconnected at the amp end and just run the center/signal wires through?


Just connect the the signal grounds to the zero volt line and not the chassis.

I have one earth from the mains which goes to teh chassis and then is connected to the amp input zero volt line in one place.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.