Connecting grounds in guitar FX Line mixer

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
Any reliability concerns wrt GFCI / RCD ?
I don't want to be the "I found this on the web" guy ;D but I ...errr... found this on the web...


http://www.ecmweb.com/basics/how-gfcis-work

In addition, GFCI protection devices fail at times, leaving the switching contacts closed and allowing the device to continue to provide power without protection. According to a 1999 study by the American Society of Home Inspectors, 21% of GFCI circuit breakers and 19% of GFCI receptacles inspected didn't provide protection, leaving the energized circuit unprotected. In most cases, damage to the internal transient voltage surge protectors (metal-oxide varistors) that protect the GFCI sensing circuit were responsible for the failures of the protection devices. In areas of high lightning activity, such as southwest Florida, the failure rate for GFCI circuit breakers and receptacles was over 50%!
 
There is a standard about RCD in EU
IEC/EN 61008

More info can be find here (use google to translate it to english).
https://www.schrack.hr/know-how/energija/fid-isklopna-karakteristika/
 
wall socket GFI's have a space limitation so they try to do everything with a chip or two,

hysteresis pins are found in some comparator chips so no RC needed.



http://www.onsemi.com/pub/Collateral/FAN4149-D.pdf
 

Attachments

  • g.png
    g.png
    67.6 KB · Views: 8
ok that GFI spec with 150 ms back a page is an industrial app, (machinery) here is a spec for a wall socket (human) app.

from>

https://www.nema.org/Products/Documents/NEMA-GFCI-2012-Field-Representative-Presentation.pdf

as far as the original question on grounding in the mixer? try this, connect all grounds  to individual wires with ring terminals, pwr supply, in/out jacks, signal gnds, then just start plugging away at different combinations grounding points at various places on the chassis, one experiment worth 1000 theories is the mechanism here.

it will not take that long, and you will learn a lot for future projects.
 

Attachments

  • t.png
    t.png
    77.2 KB · Views: 6
> I wonder what causes this time dependancy.

The tolerance of people and practical issues.

You can take a big shock for a short time or a small shock for a long time.

Narrow 5mA spikes are not uncommon, and should not cause false trips.

The early National Semiconductor GFI chip (they are all chips, not dumb relays) had an R-C network to set time*current response. This isn't an exact match for the accepted human tolerance data, but can be set a bit quicker over most of the range so it never trips too slow.

This set of curves with AC-1 to AC-4 is widely published in several lands. I believe it is some committee's approximation to the thin data available. Clearly cleaned-up "straight" (on log-log). I think "all" human-safety GFI/RCB design is done to these curves, but the nominal "name" 5mA or 30mA is taken rather arbitrarily or from some specific type-test.
https://groupdiy.com/index.php?action=dlattach;topic=67448.0;attach=55060;image

I have seen the US excerpt that CJ posts. The 500 Ohms test is a field test, and not a test of any 5mA standard. (Note that a US GFI "can" carry 5mA leakage forever; but 500r on 120V is 240mA, and should be an "instant" snap.)
 
Hey guys!
It's been a long time since my last massage here case of my full time job..Jeez!
Now I've got some spare time to modify my mixer design a bit.

I started to think about balanced IOs and even found a great example - Rane SM26S Line mixer and there's is its schematic! Awesome!
Full schematic of the device - https://app.box.com/s/f9zo3iuvn22nd1hrkgw7c0857gfh4qv1
Description and block diagrams - http://www.rane.com/sm26s.html

So I've erased unwanted features and channel strip now looks like this:
https://app.box.com/s/eidqrwfydgzmvjsjxjhgu816nkalxjky

Mix out:
https://app.box.com/s/hh938w718byg1uwwg2290vwy8mnfdcnr

I've got several questions.
1) What are those components right beside IO jacks with designators Z1, Z4 & Z33 thru Z36 with value of 2200pF? The block diagram reads they are some RFI filters, but what are they componentwise?

2) Ok, I've got my TC electronic G-Force FX processor with has balanced outputs,  but the dry signal comes from unbalanced SEND jack of an amp. Also MIX outs are connected with unbalanced inputs of Mesa Boogie 395 stereo power amp. Can it be done? Or there are some "traps for young players"? Or, perhaps, better stick to the original design with unbalanced connection? xD


user 37518 said:
May I ask which software did you use to draw your schematic?

Yeah, sure, the package I use is Altium Designer, the schematic font is "Perfect DOS VGA 437"


Best Regards,
Phil


PS. Here's my last revision of the original design with added active volume and increased output cap value
https://app.box.com/s/xeg4l6m1nyslwrmdd3xe58ld45kdb922
 
Phil smith said:
So I've erased unwanted features and channel strip now looks like this:
https://app.box.com/s/eidqrwfydgzmvjsjxjhgu816nkalxjky

I've got several questions.
1) What are those components right beside IO jacks with designators Z1, Z4 & Z33 thru Z36 with value of 2200pF? The block diagram reads they are some RFI filters, but what are they componentwise?
I don't know but it might be related to the TVS diodes right next to those symbols. It is interesting that they use those because they are pretty high capacitance. Datasheets for TVS diodes sometimes show capacitance in the almost 10n range. That could be difficult for something to drive or, with any significant source resistance, it will cause high frequency loss. And the capacitance changes depending on the voltage across them which means you could get distortion. I have also seen these diodes used on a demo board for a fancy audio part.

I have explored using TVS diodes but the super low-capacitance ones used for DSL modems. Specifically CDSOD323-T12C-DSL. I purchased a few but other than measuring their capacitance with a multi-meter I have not actually tried them (I was going to carefully test the build without and then put them in and do my measurements again to see if there was a difference and observe how they clip).

Phil smith said:
2) Ok, I've got my TC electronic G-Force FX processor with has balanced outputs,  but the dry signal comes from unbalanced SEND jack of an amp. Also MIX outs are connected with unbalanced inputs of Mesa Boogie 395 stereo power amp. Can it be done? Or there are some "traps for young players"? Or, perhaps, better stick to the original design with unbalanced connection? xD
It can be done but it is very common to get hum from ground noise in these situations. It might be very worthwhile to make special cables that are TRS on one end for the FX processor and then, at the unbalanced end for the amp, connect + wire to tip but then the - wire to the sleeve of the TR plug and leave the shield unconnected (it's connected to the FX processor). If it helps reduce noise, make two. But label them because without the sleeve connected they cannot be used with anything that doesn't already have it's own ground (like a guitar or guitar pedal running on battery).
 
> What are those components right beside IO jacks with designators Z1, Z4 &

Feedthrough caps.
 
squarewave said:
I don't know but it might be related to the TVS diodes right next to those symbols. It is interesting that they use those because they are pretty high capacitance. Datasheets for TVS diodes sometimes show capacitance in the almost 10n range. That could be difficult for something to drive or, with any significant source resistance, it will cause high frequency loss. And the capacitance changes depending on the voltage across them which means you could get distortion. I have also seen these diodes used on a demo board for a fancy audio part.

I have explored using TVS diodes but the super low-capacitance ones used for DSL modems. Specifically CDSOD323-T12C-DSL. I purchased a few but other than measuring their capacitance with a multi-meter I have not actually tried them (I was going to carefully test the build without and then put them in and do my measurements again to see if there was a difference and observe how they clip).

Yeah, quite a strange arrangement for a line mixer input, imho.  :eek: I guess better stick to the plain input design.


squarewave said:
It can be done but it is very common to get hum from ground noise in these situations. It might be very worthwhile to make special cables that are TRS on one end for the FX processor and then, at the unbalanced end for the amp, connect + wire to tip but then the - wire to the sleeve of the TR plug and leave the shield unconnected (it's connected to the FX processor). If it helps reduce noise, make two. But label them because without the sleeve connected they cannot be used with anything that doesn't already have it's own ground (like a guitar or guitar pedal running on battery).

Sounds like a plan :) So I basically shorting tip to ring and leave the shield open at the unbalanced device side to eliminate ground loop. Thanks! 8)

PRR said:
> What are those components right beside IO jacks with designators Z1, Z4 &

Feedthrough caps.

Thank you!
 
Phil smith said:
Hey guys!
It's been a long time since my last massage here case of my full time job..Jeez!
Now I've got some spare time to modify my mixer design a bit.

I started to think about balanced IOs and even found a great example - Rane SM26S Line mixer and there's is its schematic! Awesome!
Full schematic of the device - https://app.box.com/s/f9zo3iuvn22nd1hrkgw7c0857gfh4qv1
Description and block diagrams - http://www.rane.com/sm26s.html

So I've erased unwanted features and channel strip now looks like this:
https://app.box.com/s/eidqrwfydgzmvjsjxjhgu816nkalxjky

Mix out:
https://app.box.com/s/hh938w718byg1uwwg2290vwy8mnfdcnr

I've got several questions.
1) What are those components right beside IO jacks with designators Z1, Z4 & Z33 thru Z36 with value of 2200pF? The block diagram reads they are some RFI filters, but what are they componentwise?

2) Ok, I've got my TC electronic G-Force FX processor with has balanced outputs,  but the dry signal comes from unbalanced SEND jack of an amp. Also MIX outs are connected with unbalanced inputs of Mesa Boogie 395 stereo power amp. Can it be done? Or there are some "traps for young players"? Or, perhaps, better stick to the original design with unbalanced connection? xD


Yeah, sure, the package I use is Altium Designer, the schematic font is "Perfect DOS VGA 437"


Best Regards,
Phil


PS. Here's my last revision of the original design with added active volume and increased output cap value
https://app.box.com/s/xeg4l6m1nyslwrmdd3xe58ld45kdb922

A few quick observations :

You don't need the unity gain inverting stage around U2A since you're not using it to get a balanced output as in the Rane schematic.
If you did want to keep the phase the same as from your schematic you can just swap the inputs on the front end diff amp.

Also in your last rev of the original design you do have a phase inversion from input to output.
You can avoid this by making the 'middle' inverting amp op amp stage non-inverting instead.
 
Newmarket said:
A few quick observations :

You don't need the unity gain inverting stage around U2A since you're not using it to get a balanced output as in the Rane schematic.
If you did want to keep the phase the same as from your schematic you can just swap the inputs on the front end diff amp.

Also in your last rev of the original design you do have a phase inversion from input to output.
You can avoid this by making the 'middle' inverting amp op amp stage non-inverting instead.

Oh, right! Corrected!
https://app.box.com/s/xeg4l6m1nyslwrmdd3xe58ld45kdb922
Thank you!

PS. I think the best option now is to build my simple version cause I've got all parts already purchased.

 
Phil smith said:
Sounds like a plan :) So I basically shorting tip to ring and leave the shield open at the unbalanced device side to eliminate ground loop. Thanks! 8)
Not quite. There's no "shorting tip". Tip goes to tip, ring goes to sleeve and sleeve goes to nothing (left unconnected at unbalanced end).

Like #11 here:

Rane Cable Interconnect Example Diagrams

But it may not be necessary. If both pieces of gear are properly wired, it shouldn't be necessary. If you don't have hum, you don't need special cables. But if you're going to have a problem, it's probably going to be when trying to connect a digital FX processor to a high gain guitar amp.
 
squarewave said:
Not quite. There's no "shorting tip". Tip goes to tip, ring goes to sleeve and sleeve goes to nothing (left unconnected at unbalanced end).

Like #11 here:

Rane Cable Interconnect Example Diagrams

But it may not be necessary. If both pieces of gear are properly wired, it shouldn't be necessary. If you don't have hum, you don't need special cables. But if you're going to have a problem, it's probably going to be when trying to connect a digital FX processor to a high gain guitar amp.

Now I see, thank you *thumbs up*
I'll make one stereo channel and mix bus amps on a bread board to try this thing out.
 
Phil smith said:
Oh, right! Corrected!
https://app.box.com/s/xeg4l6m1nyslwrmdd3xe58ld45kdb922
Thank you!

PS. I think the best option now is to build my simple version cause I've got all parts already purchased.

Stop  :)
That's not how to 'do' a non-inverting op amp stage !
 
Phil smith said:
Oh, right! Corrected!
https://app.box.com/s/xeg4l6m1nyslwrmdd3xe58ld45kdb922
Thank you!

PS. I think the best option now is to build my simple version cause I've got all parts already purchased.

Also, you need a resistor in series between the op amp (U100B / U101B) and the following potentiometer to limit the ("attempted") gain. As shown with the pot wiper at the top (ie the op amp output) then the input impedance of the inverting stage would be Zero (disregarding the track resistance  :) ). It won't end well.
Look at the example active volume stage you posted previously and you'll see 1K0 in series.
 
Newmarket said:
Stop  :)
That's not how to 'do' a non-inverting op amp stage !
Newmarket said:
Also, you need a resistor in series between the op amp (U100B / U101B) and the following potentiometer to limit the ("attempted") gain. As shown with the pot wiper at the top (ie the op amp output) then the input impedance of the inverting stage would be Zero (disregarding the track resistance  :) ). It won't end well.
Look at the example active volume stage you posted previously and you'll see 1K0 in series.

Thank you, Newmarket , for your input!
Yes, my bad. So I've started studying opamp math little bit and this is what I've come up with (see attachment)

Non-inverting stage is what I have in basic diagram, and after is my design of non-inverting stage with same gain value.

I try to use the same component values as much as possible.
Initial design has roll-off frequency of 102 kHz. Top achieve the same value of f, the cap C shout be 65p, say 68p (E12 value).
With 47p cap roll off frequency would be 141kHz - is it ok, or C value should be changed to 68p?

And question about R1 in non-inverting circuit - should I have 10k resistor in series between previous opamp stage output and this stage + input, or, as illustrated, direct connection from output to the +input and 10k to Ground?
 

Attachments

  • 20180520_141705.jpg
    20180520_141705.jpg
    1.5 MB · Views: 14
Phil smith said:
Thank you, Newmarket , for your input!
Yes, my bad. So I've started studying opamp math little bit and this is what I've come up with (see attachment)

Non-inverting stage is what I have in basic diagram, and after is my design of non-inverting stage with same gain value.

I try to use the same component values as much as possible.
Initial design has roll-off frequency of 102 kHz. Top achieve the same value of f, the cap C shout be 65p, say 68p (E12 value).
With 47p cap roll off frequency would be 141kHz - is it ok, or C value should be changed to 68p?

And question about R1 in non-inverting circuit - should I have 10k resistor in series between previous opamp stage output and this stage + input, or, as illustrated, direct connection from output to the +input and 10k to Ground?

Looks okay to me. I wouldn't worry about  changing the cap value.
No need for series resistor between op amp stages.
I'm assuming the two stages are on the same pcb / chip and nothing on datasheet app notes indicates you need anything on the input ( I can't recall what op amps you are using ).

What I would say is that the resistor values look rather high to my thinking ( = more hiss) so I'd look to scale them lower whilst staying within the op amp capabilities and power supply current available.
Also if you want 'exactly' (component tolerances notwithstanding) a gain of 2.3 then I'd suggest simply using two resistors  to give the value required.
 
Newmarket said:
Looks okay to me. I wouldn't worry about  changing the cap value.
No need for series resistor between op amp stages.
I'm assuming the two stages are on the same pcb / chip and nothing on datasheet app notes indicates you need anything on the input ( I can't recall what op amps you are using ).

What I would say is that the resistor values look rather high to my thinking ( = more hiss) so I'd look to scale them lower whilst staying within the op amp capabilities and power supply current available.
Also if you want 'exactly' (component tolerances notwithstanding) a gain of 2.3 then I'd suggest simply using two resistors  to give the value required.

Thank you!
Yeah, all stages are on the same board.

You mean, for example, replace 24k and 10k resistors in opamp feedback with 2.4k and 1k? That way I should change and cap as well to be in the same ball  park of roll-off frequency.

To achieve the gain value closer to  3.3  I can use R2=10k, R3 = 20k+2.7k = 22.7k, then gain = 22.7k/10k + 1 = 3.27. But does making of chain of resistors in the feedback increase a hiss?
 
Back
Top