electro-music.com   Dedicated to experimental electro-acoustic
and electronic music
 
    Front Page  |  Radio
 |  Media  |  Forum  |  Wiki  |  Links
Forum with support of Syndicator RSS
 FAQFAQ   CalendarCalendar   SearchSearch   MemberlistMemberlist   UsergroupsUsergroups   LinksLinks
 RegisterRegister   ProfileProfile   Log in to check your private messagesLog in to check your private messages   Log inLog in  Chat RoomChat Room 
 Forum index » DIY Hardware and Software » Lunettas - circuits inspired by Stanley Lunetta
Simple Single Supply CV Summing
Post new topic   Reply to topic Moderators: mosc
Page 1 of 1 [23 Posts]
View unread posts
View new posts in the last week
Mark the topic unread :: View previous topic :: View next topic
Author Message
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Mon Aug 26, 2019 11:29 am    Post subject: Simple Single Supply CV Summing Reply with quote  Mark this post and the followings unread

Hi all,

I'm not sure if this falls entirely within the scope of lunettas, so forgive me if I'm stretching it too far with this one:

I had in mind to build a simple summing mixer for CV's, so it could be used to sum LFO signals, or hooked up to the output of a shift register for some pseudo-sequencing, or... whatever else comes to mind and seems like it might be fun, without being fixed into using a R2R. While searching for hours (or maybe days at this point), I've found more than one post suggesting that making a summing mixer for CV's on single supply, even if only for positive voltages, is a no-no, but little explaining why. What's the deal?

What I've found (but have yet to try) are:

- the datasheet for the lm324 shows a DC summing amp running on what appears to be a single 5V supply: http://www.ti.com/lit/ds/symlink/lm324-n.pdf (page 15)

- PHOBoS has CV summing for the VCO of their Moon Base Explorer, but this uses a virtual ground with an inverting amp: http://electro-music.com/forum/phpbb-files/moon_base_xplorer_vco_877.gif

Is the circuit from that datasheet viable, and if so, is there a reason why you would choose to use PHOBoS' setup instead?

Thanks
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Mon Aug 26, 2019 2:15 pm    Post subject: Reply with quote  Mark this post and the followings unread

I think it depends on what voltages you use for the input signals, and what you want to connect it to.
With a single supply everything has to be referenced to a different voltage than GND, usually it's 1/2 the supply voltage as
this gives the most headroom but you can choose another voltage. For the Moon Base Xplorer VCO it's actually around 5.5V

Because of this reference voltage your input signals have to be biased around this voltage as well, if you are already working
with a single supply that's usually not a problem. More importantly the output will have a DC offset voltage added to it. So if
you turn the pots of the mixer for the MBX VCO down its output is not 0V but 5.5V.

With audiorate signals you can use capacitors to remove a DC offset voltage and bias the signal around GND, although it's not
as ideal as using a bipolar supply. For low frequency signals this doesn't really work, but depending on what you want to use it
for a single supply CV summing mixer can work.


As for the circuit in the datasheet, it shows a non inverting and anverting summing amplifier in one circuit and the important
part is "(V 1 + V 2 ) ≥ (V 3 + V 4 ) to keep V O > 0 V DC". If you'd use V3, V4 than V1, V2 have to be connected to a voltage
higher than 0V, which is the reference voltage mentioned before. If you'd use V1, V2 instead you have a non-inverting summing
amplifier which can work but if I remember correctly it has a bunch of drawbacks which is why you won't see it as often.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Wed Aug 28, 2019 5:04 am    Post subject: Reply with quote  Mark this post and the followings unread

Thanks a bunch for your reply!

Quote:
With a single supply everything has to be referenced to a different voltage than GND, usually it's 1/2 the supply voltage as
this gives the most headroom but you can choose another voltage.


I understand this for AC signals, but does this still hold true for DC? It seems like there would be a lot of lost range in this... ie, if you had a supply of 5V and Vref is 2.5, assuming I'm understanding this correctly, you'd only be able to adjust between 2.5V and 5V?

If halfway isn't a necessity for DC only, what establishes the lowest Vref can be? The op-amp's minimum output voltage?

Quote:
As for the circuit in the datasheet, it shows a non inverting and anverting summing amplifier in one circuit and the important
part is "(V 1 + V 2 ) ≥ (V 3 + V 4 ) to keep V O > 0 V DC". If you'd use V3, V4 than V1, V2 have to be connected to a voltage
higher than 0V, which is the reference voltage mentioned before.


My plan was to only use V1 + V2 (and add two more inputs in the same configuration). I figure I'll lose the ability to have an inverted input, but at least I won't have to worry about said voltage relationship.

Quote:
If you'd use V1, V2 instead you have a non-inverting summing
amplifier which can work but if I remember correctly it has a bunch of drawbacks which is why you won't see it as often.


I was under the impression that a virtual ground summing amp self-compensates in terms of gain (to a degree) as you add or take away input channels. Non-inverting requires the circuit to be designed for a set number of inputs, and I believe it's poor (noisy?) with summing a large number of channels. I'm sure there are additional differences, but I'm not sure what the ramifications of using one or the other are for a 4 channel DC mixer, hence all the questions! Very Happy
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Wed Aug 28, 2019 10:41 am    Post subject: Reply with quote  Mark this post and the followings unread

dk wrote:
I understand this for AC signals, but does this still hold true for DC? It seems like there would be a lot of lost range in this... ie, if you had a supply of 5V and Vref is 2.5, assuming I'm understanding this correctly, you'd only be able to adjust between 2.5V and 5V?

For an ideal opamp in a inverting summing configuration it would be between 0 and 5V.
If you would connect the non inverting input to GND and the input voltage is +5V the output would be -5V (assuming a gain of 1).
Of course if it's only powered by +5V the output can never go below 0V. If instead of GND you use +2.5V and input +5V then for the opamp
the voltage is 2.5V above the reference voltage so the output will be 2.5V below the reference voltage, which in this case is 0V. Similarly
if you input 0V it is 2.5V below the reference voltage so the output is 2.5V above it which is +5V. However, keep in mind that an opamp isn't
ideal so with a supply voltage of +5V, a Vfref of 2.5V and an input voltage of 0V the output isn't actually going to get to +5V. So for maximum
headroom you actually have to put Vref a bit lower than 1/2 the supply voltage. (single supply opamps do tend to output pretty close to 0V)

Quote:
If halfway isn't a necessity for DC only, what establishes the lowest Vref can be? The op-amp's minimum output voltage?

Minimum output voltage and the input voltage range.

Quote:
I was under the impression that a virtual ground summing amp self-compensates in terms of gain (to a degree) as you add or take away input channels. Non-inverting requires the circuit to be designed for a set number of inputs, and I believe it's poor (noisy?) with summing a large number of channels.

Yes, you are correct. I think you could buffer the inputs so their impedance stays the same but I am not sure if that makes it more practical than
just reinverting the signal after summing. Also for non-inverting the gain is always higher than 1 but you could lower the input voltage first with
a voltage divider.

Of course you also have to keep in mind that if you have multiple inputs the voltages add up but it can't go higher than the supply voltage. 2x +5V
would of course not give you +10V on a +5V supply. To make this a bit easier for myself I added a switch to the mixer of one of my synths so
I can set it to the number of inputs and it will adjust the gain accordingly.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Wed Aug 28, 2019 3:07 pm    Post subject: Reply with quote  Mark this post and the followings unread

Ah, yes, you're totally right about Vref needing to be centered for an inverting configuration. I don't why that didn't register, but it totally makes sense!

So here it starts to get fun. I just breadboarded the circuit from the lm324 datasheet and tested it. With only 2 100k input resistors and 2 100k pots in front of them, all pots turned off gives me an output of 0.8mV. On a 5V unipolar supply and an input of just over 3.7V, each pot turned all the way up adds ~1.24V to the output. That got me thinking... is the circuit on the datasheet not just a passive averager, with the op-amp serving as a voltage follower? There should be a resistor from the inverting input to ground for it to have gain, right? If so, then the output should be the average of the 2 pot inputs plus the 100k to ground on the non-inverting input, which would explain why the output looked like it did, or am I missing something?

I guess the first drawback that's occurring to me is that if I add another two inputs, each pot won't be capable of generating an output more than 1-1.5V on it's own, assuming an input voltage of 5V and that I can remove the single 100k strapped from the non-inverting input to ground. I guess without adding gain it would limit it's usefulness, depending on how it gets used.
Back to top
View user's profile Send private message
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Sat Aug 31, 2019 12:07 pm    Post subject: Reply with quote  Mark this post and the followings unread

Juts a quick update. This is a schematic of what I've come up with so far. It has a switch for selecting whether the input CV's are mixed at 1/4, 1/2, or full strength. The only thing that's a bit uncomfortable is that at anything but 1/4 strength, there's the potential for the first amp stage to attempt amplifying the mixed CV signal over it's own supply voltage. With audio, that would just result in clipping, but I'm not sure what happens when it's DC? I'll assume for now a lot of dissipated heat and a fried chip...

Do you have any recommendations PHOBuS (or anyone else)?


EDIT: There are a few faults in this schematic, as discussed below. See below for an updated version.


Simple Single Supply CV Summer schematic.png
 Description:
 Filesize:  140.37 KB
 Viewed:  555 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

Simple Single Supply CV Summer schematic.png



Last edited by dk on Mon Sep 02, 2019 1:34 pm; edited 1 time in total
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Sat Aug 31, 2019 3:07 pm    Post subject: Reply with quote  Mark this post and the followings unread

dk wrote:
is the circuit on the datasheet not just a passive averager, with the op-amp serving as a voltage follower? There should be a resistor from the inverting input to ground for it to have gain, right?

I think that would be the resistors connected to V3,V4. which you could connect to 0V or another voltage which will create an offset.
But yes, you're correct, if you would leave those resistors out/unconnected it would be a voltage follower.

Now about the Simple Single Supply CV Summer schematic.
First thing is that you should add a resistor between the output of the first opamp and the diodes. Right now if the voltage gets higher than 5V
the diodes wil start clamping it down while the opamp will fight to get it higher. If you add a resistor the voltage difference between them can be
dropped across the resistor and they're both happy.

I am not sure if you'll get any problems with having the pots connected to the non-inverting input since that would change the input impedance.
Also I am curious as how you got those resistor values. Currently doing some calculations myself.

As for AC or DC, clipping is clipping. Unless your input voltages are higher than the supply voltage of the opamp it shouldn't really do any harm.
If you use an opamp as a comparator you're doing the same thing. Although sometimes opamps do behave a bit weird if you get close to the
supply rails. The only difference between AC and DC is frequency related. AC could have a very low frequency and practically be DC afterall.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Sat Aug 31, 2019 3:58 pm    Post subject: Reply with quote  Mark this post and the followings unread

If I am not mistaken then instead of a 75K resistor you'd need another 100K resistor. That would get you a gain of 2x.
_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Mon Sep 02, 2019 1:17 am    Post subject: Reply with quote  Mark this post and the followings unread

Quote:
If I am not mistaken then instead of a 75K resistor you'd need another 100K resistor. That would get you a gain of 2x.


I think you're right. My bad!

Quote:
First thing is that you should add a resistor between the output of the first op-amp and the diodes.


This is exactly why it's great to have someone to run ideas by! Very Happy Does it need to be a specific value? Or will a "standard" 1k do the trick? While we're on resistors, I guess I can leave out the 100k in the feedback loop of the output voltage follower, right? I think I only carried it over there when I copied the first op-amp stage...

So far, I've tried it with just 2 pots attached and 2 100k's just going to ground. I haven't noticed any problems connecting it to both unbuffered CMOS outputs as well as transistor buffers, but I guess it'll be good to keep in mind the change in impedance.
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Mon Sep 02, 2019 4:14 am    Post subject: Reply with quote  Mark this post and the followings unread

Quote:
Does it need to be a specific value? Or will a "standard" 1k do the trick? While we're on resistors, I guess I can leave out the 100k in the feedback loop of the output voltage follower, right? I think I only carried it over there when I copied the first op-amp stage...

1K should be ok, it's not very crictical but if it's too large the zener can't do its thing.
For the voltage follower you can leave out the 100K and connect the inverting input to the other side of the 1K resistor, so your output.
This will add current limiting while keeping the voltage stable when you connect a load.

Quote:
So far, I've tried it with just 2 pots attached and 2 100k's just going to ground. I haven't noticed any problems connecting it to both unbuffered CMOS outputs as well as transistor buffers, but I guess it'll be good to keep in mind the change in impedance.

sounds promising Very Happy
Is the response linear ? I know that with an inverting summing mixer you need to keep the pot values low and the summing
resistors high or it will change the response. So setting a pot in the center position will not result in half the voltage. I also wonder
if you can get cross-coupling/bleed-through between channels. With an inverting summing mixer this is prevented because the
summing point is held at virtual GND, with a non-inverting summing mixer it's pretty much floating.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Mon Sep 02, 2019 6:39 am    Post subject: Reply with quote  Mark this post and the followings unread

Perfect... I'll update the schematic in repost. I'll try to measure crosstalk and see what the response looks like coming out of the pots, but based on ear alone (feeding a expo vco), my guess is that it's linear.

Just out of curiosity, I'll have the other half of the lm324 and just enough panel space for a banana socket and a knob, so I was thinking of trying to add a second output with a glide control, more or less taken right off the output section of Ray Wilson's 10 step sequencer (http://musicfromouterspace.com/analogsynth_new/TENSTEPSEQUENCER/pdf/tenstepschempage3_schem.pdf). If I were to take everything between U6-D and U6-C, excluding CV2 out which I wouldn't use, and try to stick it on to mine, what would be the past way to connect it? My first guess was to take the output off my circuit and connect to U6-D's non-inverting output on Ray's. If that's the right way to go about it, would the portamento/glide circuit have an impact on my 5V max level (ie would it be able to somehow go higher)? I realize that this makes everything a tad more complicated, but it seems like a pity to waste the space...
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Mon Sep 02, 2019 6:50 am    Post subject: Reply with quote  Mark this post and the followings unread

That should work, you could connect it at the same point as your second opamp so after the diodes.
Voltage won't go higher, if anything it would be slightly lower as a capacitor has somewhat of a resistance but it's negligible.

For measuring crosstalk you'd have to connect at least 2 signals/voltages to the mixer and measure if it changes at the source.
Think of it as using a passive multiple with one output going to the mixer and measuring the other ouput.
Since the resistor values are fairly large and your signals likely buffered it's probably not a problem, but it could be with the
standard 1K resistor used on outputs.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Mon Sep 02, 2019 1:36 pm    Post subject: Reply with quote  Mark this post and the followings unread

Cool. I'm not sure if I'll manage to check it tonight, but if not, for sure tomorrow.

In the mean time, here's an updated schematic. Is this what you meant with keeping the output resistors inside the feedback loop? Also, how strict do you think the value of that 20R resistor in the glide circuit is? I'm not sure a value that low on hand, but obviously when I can get back to the breadboard I'll try whatever I do have. Is its function similar to that of the resistor between the first op-amp stage and the zener?


Simple Single Supply CV Summer schematic.png
 Description:
 Filesize:  170.64 KB
 Viewed:  519 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

Simple Single Supply CV Summer schematic.png


Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Mon Sep 02, 2019 2:13 pm    Post subject: Reply with quote  Mark this post and the followings unread

thumleft

yes, with the pot at 0 ohms the 20R limits the current that is the result of the voltage difference between the capacitor and the opamp, since the capacitor
can't follow the voltage directly (which is how the circuit works of course). Its value isn't very critical but you want it to be small otherwise you get too
much glide at the minimum setting. 100nF isn't very large though and it probably won't damage the opamp (at least not an LM324) if you'd leave it out,
but it might cause a bit of noise on the power rails so I'd leave it in.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Tue Sep 03, 2019 1:50 pm    Post subject: Reply with quote  Mark this post and the followings unread

Ok, I guess crosstalk is inherent in this design. Depending on what position the switch is in and what I was feeding it, I got anywhere from 1-10mV of crosstalk. For mixing a handful of LFO's together or general Lunetta-ing, I guess this will pass, but if someone were out to make something more precise (ie as part of a sequencer), this probably would be too imprecise.

The pots do in fact work linearly.

Quote:
yes, with the pot at 0 ohms the 20R limits the current that is the result of the voltage difference between the capacitor and the opamp, since the capacitor
can't follow the voltage directly (which is how the circuit works of course). Its value isn't very critical but you want it to be small otherwise you get too
much glide at the minimum setting. 100nF isn't very large though and it probably won't damage the opamp (at least not an LM324) if you'd leave it out,
but it might cause a bit of noise on the power rails so I'd leave it in.


Cool. I think I might have a few 10R's floating around, so I guess I'll just wire them up in series.
Back to top
View user's profile Send private message
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Thu Nov 21, 2019 3:02 pm    Post subject: Reply with quote  Mark this post and the followings unread

An update and a question:

This does work fine for non-critical applications, but it's pretty dodgy when used in "sequencer" applications, where the knobs are independently stepped through. With the switch at it's highest gain, moving non-active knobs can shift everything by over half a volt, which is certainly quirky. Use for mixing LFO's, audio (my audio and CV's are both 0-5V), or other stuff where it doesn't matter works fine. The only adjustment from the last schematic I posted is that I have the glide pot wired backwards.

I do have a question about doing this with a bipolar power supply and the op-amps in inverting configuration, though. If you were to put multiple summing amps on the same chip, and/or multiple chips on the same board (PCB, vero, etc), do the ground lugs of the pots for each mixer need to be connected as close as possible to where the non-inverting input of their summing amp is, or can all pots of all the mixers be ganged together and connected to one ground point on the board?
Back to top
View user's profile Send private message
JovianPyx



Joined: Nov 20, 2007
Posts: 1988
Location: West Red Spot, Jupiter
Audio files: 224

PostPosted: Fri Nov 22, 2019 7:14 am    Post subject: Reply with quote  Mark this post and the followings unread

Grounds (and really power rails too) are always best done as a "star".

This means that one point is picked as the ground node.

This is usually the place where ground is "created" at the PSU.

From that point, high current conductors* radiate out in a star to connect things that need to be grounded.

Daisy chaining grounds can cause a build up of noise along the daisy chain's length.

This may not be the most physically convenient method, but electrically it is the best.

* "high current conductor" means a conductor heavy enough to conduct well more current than will flow. The reduced resistance of the heavy conductor reduces noise and other problems.

_________________
FPGA, dsPIC and Fatman Synth Stuff

Time flies like a banana.
Fruit flies when you're having fun.
BTW, Do these genes make my ass look fat?
corruptio optimi pessima
Back to top
View user's profile Send private message Visit poster's website
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Fri Nov 22, 2019 7:40 am    Post subject: Reply with quote  Mark this post and the followings unread

Hi Jovian, thanks for your answer.

Just to confirm, the ground lugs of the pots should directly connect to the star ground point of the module, and not to where ground is connected to the positive terminal of the summing amp, right? If this were just one summer per board, it would be fairly obvious, but with 4 (2 chips and 2 summing amps per chip) I wasn't sure if I wouldn't get some sort of interference this way...

Thanks again!
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Fri Nov 22, 2019 9:42 am    Post subject: Reply with quote  Mark this post and the followings unread

dk wrote:
Hi Jovian, thanks for your answer.

Just to confirm, the ground lugs of the pots should directly connect to the star ground point of the module, and not to where ground is connected to the positive terminal of the summing amp, right? If this were just one summer per board, it would be fairly obvious, but with 4 (2 chips and 2 summing amps per chip) I wasn't sure if I wouldn't get some sort of interference this way...

Thanks again!



A proper starground would actually be directly on the powersupply, usually where the capacitors are connected, you'll often see it used in audio
amplifiers. In a modular system where you use some sort of busbar for power distribution there is no such thing. As a matter of fact a busbar
is partially daisy chaining the supply which isn't a great method to begin with and also why it should have a very low resistance. What is important
is to know why something like a starground (not just limited to ground btw also for general power distibution) is used to begin with.

In an ideal world it would absolutely not matter how you would connect things together as long as there is an electrical connection. However that
is of course not the reality and the main problem here is resistance. There is also capacitance and inductance but as long as you are not messing
around with very high frequency circuits and/or very long cables that can mostly be ignored. Basically every trace, cable, connection has a resistance
and if there is a current passing through it this creates a voltage. Think of what would happen if you have a supply of 12V and you connect a
module to it through a resistor. Depending on the resistance and the current there is a voltage drop so the voltage that the module 'sees' will be
lower than 12V. If the current is constant and not very high this isn't that big of a problem. But if you have something like an oscillator this current
varies causing the voltage to fluctuate. If you then connect another module to the same connection (after the resistor) it will get a supply that
fluctuates which can cause problems. An example of what often happens is that you can hear an LFO through a VCO even if they aren't connected
to eachother and only share power. Another thing that can happen is oscillators locking to eachother when they are close in frequency causing
them to oscillate in sync. This is what happens with daisy-chaining and why it is better to use a seperate connection to power for each module/circuit.
Even for a single circuit it is sometimes necessary to have seperate paths and this is why you sometimes see circuits that have a seperate analog and
digital ground.

How you connect things together in practise and what you can get away with depends on the current and sensitivity to fluctuations of the circuit
itself. For your circuit the current through the pots is very low and since they are probably all mounted pretty close together there would hardly be
any resistance between them if you just connect them all together, at least as long as you don't use a very thin cable. (see Scott's note on "high
current conductor"). This is of course also much more practical than using a seperate wire for each. However, if you also have some flashing LEDs
that have a connection to GND it is better to keep that connection seperated (this would be an exampe of analog and digital ground).

There is something else though which is noise interference from outside the circuit. Wires act as antenna and if you have a sensitive low current
input this can become very noticable. Opamps usually are sensitive and low current and on top of that often used as amplifiers so can pick up and
amplify this noise. To keep most of it out you can use shielded wires and if you use metal pots also connect the housing to ground to act as shielding
and it is better to use non-conductive knobs. If you have a metal frontpanel it usually takes care of grounding the pots. If you use shielded wires you
already have a seperate ground connection available for each pot and it is tempting to use this and I have connected pots this way without any
issues. It is not entirely the correct method though. Shielding should not carry any current and therefor should only be connected on one side and
not be used as an electrical connection. This is actually a flaw in modular synths (unless you use bananas but then you don't have any shielding).
Ideally you should only connect the ground to input connectors and not to output connectors and if you want to connect things together that don't
share the same power supply there should be a seperate (very low resistance) ground connection directly between the supplies.

to sum it up you have some options:
1. use one ground connection from the circuit to your pots and wire all of them together. It is the most practical solution and will likely work.
2. use seperate ground connections for each pot. Less practical and will probably not make a noticable difference with option 1.
3. use shielded wire for each pot and use the shielding as your ground connection. Not ideal but is an improvement over non-shielded wire.
4. use shielded wire for each pot but only connect the shielding on one side and wire all the pots together with a seperate ground connection.
5. use shielded wire that has an extra core that you can use for ground (this is how proffesional balanced systems are connected together)
note that using shielded wire can become unpractical for short distances and unless it is close to an electromagnetic noise soure won't really add
much of an improvement.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube

Last edited by PHOBoS on Fri Nov 22, 2019 11:37 am; edited 2 times in total
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
JovianPyx



Joined: Nov 20, 2007
Posts: 1988
Location: West Red Spot, Jupiter
Audio files: 224

PostPosted: Fri Nov 22, 2019 9:44 am    Post subject: Reply with quote  Mark this post and the followings unread

Correct. Every ground should theoretically be it's own conductor back to the star's center (which should be the PSU ground). This is considered best practice, but may not always be convenient.

I should also add that while daisy chains are not the best, when they are very short or done with fairly heavy wire, they are fine. When daisy chaining numerous pots, it can help to daisy chain a few together and wired back to the star center. Example - for 5 pots, I would wire their grounds together and wire that to the star center from the middle of the 5 pots. But practicalities being what they are, do the best you can.

_________________
FPGA, dsPIC and Fatman Synth Stuff

Time flies like a banana.
Fruit flies when you're having fun.
BTW, Do these genes make my ass look fat?
corruptio optimi pessima
Back to top
View user's profile Send private message Visit poster's website
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Fri Nov 22, 2019 10:40 am    Post subject: Reply with quote  Mark this post and the followings unread

Thanks for the descriptive reply. I think I'll try grouping the pots by row (4 each) and connecting each row to the ground point of the board for now, and see there that ends up.

Quote:
unless you use bananas but then you don't have any shielding


I am! Very Happy Obviously, no shielding has some potential for trouble, but sometimes it's fun to live dangerously.... until you have to ask on here how to ground pots for a cv mixer/sequencer Very Happy
Back to top
View user's profile Send private message
PHOBoS



Joined: Jan 14, 2010
Posts: 5591
Location: Moon Base
Audio files: 705

PostPosted: Fri Nov 22, 2019 11:49 am    Post subject: Reply with quote  Mark this post and the followings unread

banana banana banana banana banana banana are great for lunettas since you generally aren't using very low voltage levels and the signals
are digital so much less prone to noise to begin with. Also inputs usually have a resistor to GND which makes it less sensitive and
this is not only the case with lunettas. For long cables or cables close to an electromagnetic source (like a transformer) shielding can
make a lot of difference. I remember breadboarding something on top of my laptop and it took me a minute before I realized where
all the horrible buzzing was coming from.

_________________
"My perf, it's full of holes!"
http://phobos.000space.com/
SoundCloud BandCamp MixCloud Stickney Synthyards Captain Collider Twitch YouTube
Back to top
View user's profile Send private message Visit poster's website AIM Address Yahoo Messenger MSN Messenger
dk



Joined: Feb 12, 2019
Posts: 115
Location: Europe

PostPosted: Fri Nov 22, 2019 1:02 pm    Post subject: Reply with quote  Mark this post and the followings unread

Quote:
I remember breadboarding something on top of my laptop and it took me a minute before I realized where
all the horrible buzzing was coming from.


I totally feel you here. I think the issues I had posted early this year in the "Why Isn't My Lunetta Working" thread was caused by my patch cables touching my laptop while resting on my synth. It seemed they were accumulating a static charge, so every time I plugged one in everything went haywire. Moving the cables to the other side of the synth away from the computer was enough to fix the issue. Rolling Eyes
Back to top
View user's profile Send private message
Display posts from previous:   
Post new topic   Reply to topic Moderators: mosc
Page 1 of 1 [23 Posts]
View unread posts
View new posts in the last week
Mark the topic unread :: View previous topic :: View next topic
 Forum index » DIY Hardware and Software » Lunettas - circuits inspired by Stanley Lunetta
Jump to:  

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum


Forum with support of Syndicator RSS
Powered by phpBB © 2001, 2005 phpBB Group
Copyright © 2003 through 2009 by electro-music.com - Conditions Of Use