GND Scheme questions

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Balanced connections are easy, it's always the unbalanced ones that cause problems.
Whenever it's possible, ALL inputs should be balanced, including those on RCA phono connectors, which should be complemented with an optional "ground" connection. 1/4" jacks should be TRS. Phono inputs are an exception, but since the cartridge is floating, that should not be an issue.
 
I found this in an old forum. It was written by the designer himself.

The right pin of the Interf connector is a control pin that changes the internal gain of the circuit, allowing the same (unity) gain for both balanced and unbalanced connection.

See: http://jhaible.com/legacy/subtle_chorus/subtle_chorus_main_board_sch_dwg2.pdf

Doesn't have anything to do with plugging an unbalanced output into a balanced input, which *is* ok. Just don't plug the balanced output of the Dimension D into an unbalanced input!--
 
Intuitive experimentation can sometimes be the quickest way to get a good ground scheme.

That seems like a good way to get something which works well in your house but is not necessarily robust in different situations.
I find that understanding electrical principles is the best way to get a reliable scheme.
 
Doesn't have anything to do with plugging an unbalanced output into a balanced input, which *is* ok

Of course it's OK, but the connection shown in the JHaible design is not optimal from a noise/interference rejection standpoint. The design also appears to have pin 3 as hot, which is about 30 or 40 years out of date, so maybe not completely surprising that it is not following best practices for shield connection which have been widely taught for more than 25 years.

The two op-amp instrumentation amp configuration used for the input stage is also a little different than the canonical version, which makes me a bit suspicious that the designer doesn't quite fully understand differential inputs, so I would suggest keep the good parts of the signal processing that you want, but improve the I/O connections to modern standards.
 
Of course it's OK, but the connection shown in the JHaible design is not optimal from a noise/interference rejection standpoint. The design also appears to have pin 3 as hot, which is about 30 or 40 years out of date, so maybe not completely surprising that it is not following best practices for shield connection which have been widely taught for more than 25 years.

The two op-amp instrumentation amp configuration used for the input stage is also a little different than the canonical version, which makes me a bit suspicious that the designer doesn't quite fully understand differential inputs, so I would suggest keep the good parts of the signal processing that you want, but improve the I/O connections to modern standards.
It's a "copy" of the Roland Dimension D. That have 30 years. Haible said, he wouldn't design it that way either. He only respects the original design.
 
What do you suggest me? Don't use shield cables and go straight from pcb in/out to connectors? In that case, what about chassis connection?
 
From the same old forum.

electro-music.com :: View topic - Dim D


The output stage of the "vintage" version is just that: a faithful reproduction of the vintage circuit, for those who want it exactly like the original. Personally, I would not design an output stage like this, powerful output stages with just signal inversion (and no auto-balancing). I'd go for short-circuit protected opamp outputs (my unexpensive version), or a transformer output, or an electronic transformer.
 
I don't know how to wire this thing properly...shield cables, circuit gnd, chassis and safety earth.

I thought everything was clear, but I will summarize as best I can:

  • Read the Bill Whitlock presentation linked previously. It will serve you well now and in the future, and will help you recognize suggestions which are actually dangerous to life and property.
  • Protective/safety earth connects directly to a conductive chassis with a locking washer so it cannot become loose. This design is not high enough power that a protective earth connection would be required in a commercial power supply, so if you use an external power supply, or a commercial encapsulated power supply design, it is possible that you do not have that connection. The rest of the discussion assumes you do have a protective earth connection and conductive chassis, but nothing really changes if you do not have PE. Plastic chassis is a little different story, skipped for now.
  • Cable shield connections connect directly to chassis with a short, wide connection. XLR connectors often have a tab for shell connection so you can connect pin 1 to the shell connection tab with a very short conductor, then the conductive shell makes the connection to chassis.
  • System power supply reference node (aka "ground") connects to chassis near or on top of PE connection to chassis. As "Rock Soderstrom" showed in post #24 some jurisdictions do not allow multiple connections sharing the PE connection, but as far as I know in North America at least you could have a nut holding the PE terminal to chassis, then put the power supply chassis connection on top of that first nut, with a second nut holding that connection, and the requirement would be satisfied, since you could remove the circuit to chassis connection without removing the PE connection at the same time.
That is the basic information you need to avoid interference from other equipment. For single ended connections between different circuit boards within the chassis you can run signal and "gnd" connections beside each other (twisted pair or small coax is good to avoid picking up magnetic interference). The voltage difference between different sub-circuits inside the same chassis should be minuscule, so the concern about leakage current flow is not the same as between different devices.
 
Yes, I undertand this. The confusion is that the pin 1 of the Dim D is not wiring to shell.

Dimension D, and pin 1 of the XLRs are all connected to signal ground, whilst the XLR shells are connected directly to chassis ground. Roland perhaps expected the use of cables where pin 1 is not connected to the shell.

And I don't know if is secure to tie signal ground, and earth ground in this 30 years device.

So, I think that I will wire like the original.

Pin 1 to signal gnd. And XLR shells to chassis.
 

Attachments

  • Screenshot_20221202_094349_Drive.jpg
    Screenshot_20221202_094349_Drive.jpg
    660.5 KB · Views: 0
Roland perhaps expected the use of cables where pin 1 is not connected to the shell.

That is the AES specification for XLR wiring, shield of cable connects to pin 1.

Dimension D, and pin 1 of the XLRs are all connected to signal ground

That is referred to as "pin 1 problem" and has been widely discussed since the popularization of the term in the June 1995 special issue of the Journal of the Audio Engineering Society on grounding and shielding.

The root of the problem is that there will be potential differences (i.e. voltage differences) between different chassis, nothing you can do about that, so you have to deal with the resulting current flow. When you connect pin 1 to the circuit reference node (aka "signal ground") that current flows across the circuit board, and since any current flowing through a resistance generates a proportional voltage (basic Ohm's law), the reference node for the circuitry, i.e. the node against which all your output voltages are defined, is now bouncing up and down with power line related noise. That implies your output is now bouncing up and down with power line noise, and the next device in the chain will see all that noise at the input.
You basically inject power line noise directly into your circuitry when there is no need.

The presentation by Bill Whitlock that I linked and recommended earlier (post #5) discusses pin 1 problem beginning on page 81 and following. You can also search google for "pin 1 problem" and find plenty of links. The old Rane app note on the topic is a pretty good summary.
It has also been covered extensively on this site.

In short, it is ill advised, it has always been a bad idea, it will always be a bad idea in the future, just don't do it.

I think that I will wire like the original.

There is no need to enshrine someone else's ignorance from 30 years ago into a 2022 build.
 
go wireless,

two prong cord safer than three prong in some instances,

(chassis is floating so git player is floating) (not on LSD, i mean electrical)
although i played some of my best stuff on acid, but the fret board started to melt, so i had to switch to shrooms which apparently they are now using to treat depression,

we had a 30 page argument about grounding a while back which alienated a few people , for some reason, grounding is an emotional topic,
I was grounded a few times as a child. Once was for being expelled from school for a week after lighting my volcano science project inside the "high" school. An entire wing of the school filled with smoke which resulted in a school wide fire drill. I mean common man, it was done in the name of science! And the fire drill was a success. I took that as a win-win, but the assistant principal didn't see things my way. I was grounded for that week off of school and my old man made me chop a cord of wood with an axe. Memories of grounding!
 

Latest posts

Back
Top