Bal/unbal fx box connection advice

shop deals on amazon
Advertise with us
Shop deals on ebay
GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.
dirty1_1garry

dirty1_1garry

Well-known member
Joined
Oct 3, 2011
Messages
511
Location
Moscow
HI folks!

Want to build a guitar pedal type compressor/fx box so it can get both balanced line level signal and unbalanced guitar level signal on same TRS jack.
One more aspect is 9Vdc wall adapter power supply typical for guitar pedals.

For supply ICs I’m planning to use Traco DC converter which create +15Vdc and -15Vdc from 9Vdc with separated ground.
For bal line in/out - That 1246/1646 drivers
For unbal in/out - simple opamp buffers
For switching between balanced and unbalanced inputs/outputs planning to use group of relays.

schematic sketch attached.

question is:
how it’s better to organise Sleeve pin connection in an aspect of that balanced line Sleeve pin is  typically connected with a chassis, but in unbal connection its typically connected with signal ground (especially that signal ground is separated by DC conversed from 9Vdc supply ground)
 

Attachments

  • FX box.jpg
    FX box.jpg
    150.8 KB · Views: 27
Not quite the question you ask but from a quick glance you don't need a separate opamp to provide the unbalanced output.
The THAT driver can be used single ended simply by grounding one output. A mono TS jack (ie a 'guitar jack plug' will do this automatically.
 
Newmarket said:
Not quite the question you ask but from a quick glance you don't need a separate opamp to provide the unbalanced output.
The THAT driver can be used single ended simply by grounding one output. A mono TS jack (ie a 'guitar jack plug' will do this automatically.
Click to expand...

And any decent balanced in device will do the same at the imput.

Cheers

Ian
 
ruffrecords said:
And any decent balanced in device will do the same at the imput.

Cheers

Ian
Click to expand...

Yes indeed.
But I guess that the OP wants a high Impedance - into the MegOhms - available for the guitar input.
On that point the input buffer shown won't work as it is missing a bias resistor from the non-inv terminal to 0V for DC bias.
Move the 2M2 resistor to the other side of the cap or if you are using the 2M2 for input capacitor discharge to minimise 'pops' then fit another resistor after the capacitor.

But I think the OPs original question was about what to do with the Screen / 0V connections on the TRS / TS jack.
Back to the old 'Pin1' issues / isolated (or not) jack sockets / ground loop breaking circuits etc.
Complicated by the fact that using a TS jack into a TRS shorts out the RS connection.
 
As Newmarket says: do not omit the grid-leak.

Buffer BOTH inputs to MegOhm. Balanced sources will not mind. The extra 11 cents is less than the cost of the switching.

Let the output of the '1646 do what it was designed to do. It will adjust for one side grounded.

No switching needed.
 

Attachments

  • Garash-hack.gif
    Garash-hack.gif
    13.6 KB · Views: 27
Newmarket said:
Yes indeed.
On that point the input buffer shown won't work as it is missing a bias resistor from the non-inv terminal to 0V for DC bias.
Click to expand...
as I understand, DC bias should be if opamp os supplied by single rail supply, but not in situation when opamp is supplied via bipolar supply. Am I wrong?)

Newmarket said:
Move the 2M2 resistor to the other side of the cap or if you are using the 2M2 for input capacitor discharge to minimise 'pops' then fit another resistor after the capacitor.
Click to expand...
useful recommendation

Newmarket said:
original question was about what to do with the Screen / 0V connections on the TRS / TS jack.
Back to the old 'Pin1' issues / isolated (or not) jack sockets / ground loop breaking circuits etc.
Complicated by the fact that using a TS jack into a TRS shorts out the RS connection.
Click to expand...
that's the main question, yes) attempts to get rid of ground loops
 
PRR said:
Buffer BOTH inputs to MegOhm. Balanced sources will not mind. The extra 11 cents is less than the cost of the switching.
Let the output of the '1646 do what it was designed to do. It will adjust for one side grounded.
No switching needed.
Click to expand...
smart and simple ideas as usual , PRR! 8)

What if we will add MegOhms for 1246 inputs directly? Or it's better to buffer 1246 additionally?
 

Attachments

  • Garash-hack2.gif
    Garash-hack2.gif
    15 KB · Views: 9
peterc said:
Given that most guitar pedals are for low level signals, you might want to consider a pot after the input stage.
Or even a pad before the input stage.

Peter
Click to expand...

yes, of course! but that's the simplest question in comparison with the questions that I encountered and try to solve here  :)
 
> What if we will add MegOhms for 1246 inputs directly?

I had assumed you did your homework.

Look at the THAT1246 datasheet and figure what the input impedance is.
 
PRR said:
> What if we will add MegOhms for 1246 inputs directly?

Look at the THAT1246 datasheet and figure what the input impedance is.
Click to expand...

simple reason, yes!))
input impedance of 1246 is 24K, thats the reason of additional buffer with megaohm resistors. Go it, thanks for a tips!
 
dirty1_1garry said:
as I understand, DC bias should be if opamp os supplied by single rail supply, but not in situation when opamp is supplied via bipolar supply. Am I wrong?)
Click to expand...

Yes.
Inputs must have a source of dc current for the bias current requirement.
In a single supply circuit the voltage used is usually half the supply voltage to allow for ac signal around that. In a dual supply +/- system it's simply biased to 0V - which is again half the supply voltage.
 
dirty1_1garry said:
that's the main question, yes) attempts to get rid of ground loops
Click to expand...

Difficult to be definitive here as your box is trying to allow for different uses.
So...in an ideal world balanced connections should be able to have all screens intact allowing bonding of S connection to (metal) chassis and all is good.
BUT it doesn't always work out like that - pin 1 / Screen gets mixed up with signal ground and we get hum and buzz and the rest.

With unbalanced input  you are inevitably mixing the Screen function with the Signal 0V.

There's loads of stuff on this on the web - google 'Pin 1 problem' 'Neil Muncy'

But you want to know what to do now  :D

Okay - google 'Bo Hansen DI' and look at the scheme there.
The output connector screen connection (XLR pin 1 = S on TRS) is bonded directly to chassis (and nowhere else).
Input connector S connection is isolated (eg plastic body TRS jack) from chassis but connects to chassis through a parallel resistor / capacitor network. Some largeish diodes there too but may not be needed in your application.

You could put a switch in to bypass this if you wanted to although it's probably not necessary.
Input screen connection is compromised but it does work well.

There is an argument to be had about where the ground loop break should be in the chain but practically this seems to 'do the job'.

Others may have different approaches...
 
Newmarket said:
Difficult to be definitive here as your box is trying to allow for different uses.
So...in an ideal world balanced connections should be able to have all screens intact allowing bonding of S connection to (metal) chassis and all is good.
BUT it doesn't always work out like that - pin 1 / Screen gets mixed up with signal ground and we get hum and buzz and the rest.

With unbalanced input  you are inevitably mixing the Screen function with the Signal 0V.

There's loads of stuff on this on the web - google 'Pin 1 problem' 'Neil Muncy'

But you want to know what to do now  :D

Okay - google 'Bo Hansen DI' and look at the scheme there.
The output connector screen connection (XLR pin 1 = S on TRS) is bonded directly to chassis (and nowhere else).
Input connector S connection is isolated (eg plastic body TRS jack) from chassis but connects to chassis through a parallel resistor / capacitor network. Some largeish diodes there too but may not be needed in your application.

You could put a switch in to bypass this if you wanted to although it's probably not necessary.
Input screen connection is compromised but it does work well.

There is an argument to be had about where the ground loop break should be in the chain but practically this seems to 'do the job'.

Others may have different approaches...
Click to expand...

one more idea that I had in my mind before posting these post was - in unbal connection connect R and S pins to signal ground and in bal connection connect S pin to chassic and (may be) with signal ground through resistor/capacitor as Bo Hansen sch shown.
May be I just need some sort of ground lift switch that will connect Spin differently and allow to minimise any hum troubles.
Any way I should experiment a little bit to find the best way in exact situation.
Thanks for nice tips!
 

Similar threads

L
Fender Solid State Stage 100 Amp huge buzz Problem... help needed ...
2
Replies
26
Views
2K
scondon
S
L
double Bo Hansen active DI Box with modifications
Replies
6
Views
3K
low.end.b
L
Whoops
Help Design an unbalanced consumer Line level to Balanced interface
5 6 7
Replies
131
Views
13K
Whoops
Whoops
C
Passive DI / Reamp Box design queries
Replies
5
Views
3K
crduff
C
P
Studio Cue System Project
3 4 5
Replies
89
Views
7K
abbey road d enfer
A

Latest posts

Back
Top