MUFF WIGGLER Forum Index
 FAQ & Terms of UseFAQ & Terms Of Use   Wiggler RadioMW Radio   Muff Wiggler TwitterTwitter   Support the site @ PatreonPatreon 
 SearchSearch   RegisterSign up   Log inLog in 
WIGGLING 'LITE' IN GUEST MODE

Improving common eurorack GND while staying compatible?
MUFF WIGGLER Forum Index -> Music Tech DIY  
Author Improving common eurorack GND while staying compatible?
clorax hurd
I read the threads with Graham Hinton explaining the proper way and how it's not possible in eurorack.

Now I wonder, what are the best (least wrong) things one can do about this when designing a module, without becoming incompatible?

I got few ideas and would be very grateful for feedback.

1) Having 2 separate ground planes on module. One for 0V reference and second for either jack shields or metal panel + jack shields together.
Connecting them together at the power connector.
Could that help a bit?
It wouldn't be 0V to chasis GND single point connection per case, but at least per module. Until we connect jacks to other "bad" modules, but if everybody started doing this and half of the modules in system would have it like this. Wouldn't it help?

- if it makes sense, then the question is: only the jack shields or even the metal panel?

- we could push the point of 0V to chasis GND connection from power connector to busboard, if we connected 0V to only 5 of the 6 GND pins on power connector and use the sixth pin to connect chasis ground. If this makes some sense, the what would be reasonable ratio? 5:1 or rather 4:2 or even 3:3?
Yes, the IDC connectors and cables have some resistance and by using less pins, we make it bit worse, so I got another idea:

- IDC power cables with 3 connectors. One connector, exactly in the middle of the cable, would be connected into module and the other two ends would connect to 2 bus board ports next to each other. That way, even using just 4-5 pins for 0V, we would have better connection than normally.
I'm just not sure, if this isn't a ground loop and I don't know why are those bad. But the resistance between two ports on a busboard is usually very tiny... especially compared to jack shields connecting ground points so far from each other => huge loops.

I'm very curious to hear if any of these ideas are bad or good and why and if they could cause trouble if used only for few modules in system where other modules are regular eurorack with jack shields connected to 0V plane.
Monkizzle
Very involving question.
Graham Hinton
clorax hurd wrote:

1) Having 2 separate ground planes on module. One for 0V reference and second for either jack shields or metal panel + jack shields together.
Connecting them together at the power connector.
Could that help a bit?


It is better not to join 0V and Chassis at the module power connector because then they share the same wire(s) back to the distribution.

I have only ever made one module with a Eurorack power connector before I decided that was a bad idea. I have been using a modified MOTM power connector with pin 2 = 0V and pin 3 = Chassis, but have now decided that it is better not to try to bodge poor design, but do what was was really needed in the first place. All my modules are now going over to this system Hinton Power Specification and may be provided with a discrete wire cable to any distribution system. Obviously Euro busboards have to have thinner wire than I would like, but very few people use my modules without my power distribution. Please feel free to use it, but not to alter it.

Quote:

It wouldn't be 0V to chasis GND single point connection per case, but at least per module.


Don't think in terms of a physical single point, think in terms of resistance and the currents flowing. The aim should be to connect each to a very low resistance which are referenced together. A busbar across a case with short wires is a very low resistance and better by orders of magnitude than long thick wires all going to one place and is topologically equivalent to a single point. Think of the resistance between any two modules and think down to the microohm level.

Quote:
Until we connect jacks to other "bad" modules, but if everybody started doing this and half of the modules in system would have it like this. Wouldn't it help?

- if it makes sense, then the question is: only the jack shields or even the metal panel?


Most module metal panels are not electrically bonded to anything or form a screen. Just because a cable is called screened does not mean that it acts as one. If it is carrying the signal return current it won't be (what screens the "screen"?).

When you connect two modules the cable screen is in parallel with the 0V power wiring between those modules. Screens are supposed to be an equipotential surface carrying no current. In practise they may act as an aerial or have currents induced by magnetic fields, but those are usually tiny currents and not a problem unless they are amplified.


Quote:

- we could push the point of 0V to chasis GND connection from power connector to busboard, if we connected 0V to only 5 of the 6 GND pins on power connector and use the sixth pin to connect chasis ground.


Eurorack only gets away with ribbon because there are six 0V wires in parallel. Forget IDC connectors. Forget ribbon cable. Forget pcb distribution. None of them scale up. Forget all 0.1" pitch data connectors that were never intended for power use.

Quote:

I'm just not sure, if this isn't a ground loop and I don't know why are those bad. But the resistance between two ports on a busboard is usually very tiny... especially compared to jack shields connecting ground points so far from each other => huge loops.


A true ground loop picks up the magnetic fields enclosed in its area. Unless you are right next to a power sub-station these tend to be small currents. Most so-called ground problems are really common impedance coupling in the 0V. The solution to a ground loop is to flatten it and the solution to CIC is to reduce the critical resistances.

Quote:

I'm very curious to hear if any of these ideas are bad or good and why and if they could cause trouble if used only for few modules in system where other modules are regular eurorack with jack shields connected to 0V plane.


Good audio grounding practise compatible with EMC requirements is recommended in AES48 for both screened and non-screened equipment. All the examples given assume that the equipment is self powered and has balanced interconnections because the authors assumed that no-one would be daft enough to build a large unbalanced system with low grade distributed power. That doesn't mean that it is not applicable to synthesziers, which are subject to the same laws of physics and man, just that some interpretation is necessary.

Ideally modular synthesizers would come in screened cases with integral mains power supplies and balanced I/O like all pro audio equipment. So it is a non-ideal world and we have to live in it, but we can improve it. The first most important thing is to understand that if you have an unbalanced system everything in it has to have the same common 0V reference and the larger the system the harder that is to achieve. The system is not just the synthesizer, but everything connected to it too. That means that every piece of equipment should have a chassis bonding terminal and what I'm calling a "ZeroVolt" connector. Most commercial equipment has neither which means that common connections default to mains cable Earth wires and cable screens, both are too high resistance.

Since I first introduced busbar power distribution I have installed hundreds of medium to large systems. What I can tell you is that improving the power system will improve the system performance of most existing modules, but not improving it will negate the benefits of better designed modules. That doesn't mean don't improve modules, you should do both.
clorax hurd
Thanks for the answers!!

Graham Hinton wrote:

but have now decided that it is better not to try to bodge poor design, but do what was was really needed in the first place.


I agree and I like your standard and I would love if it became mainstream.
But I'm afraid that with busbar being 50e/piece and needing 3, it won't be affordable for many people... (i hoped it would be chaper. piece of metal with holes. isn't it?).
(and by the way, regarding the separate return paths for + and - ... what about capacitor of passive lowpass stage? It should be connected to ground. But both - and + currents will go there, so which one to choose? (probably not both, as they would not be separated anymore).Or which ground to should be connected to the + input of inverting opamp stage?)

But lets say that we don't have a choice. Lets say we have control only over a module design and the module is to be mounted into regular eurorack case with other eurorack modules and connected to busboard, and we know it's a nightmare, but we still want to make our best designing the module.

Graham Hinton wrote:

It is better not to join 0V and Chassis at the module power connector because then they share the same wire(s) back to the distribution.


When you say "chasis" you mean both panel+jack shields. right?
(cause that's the proper way if I understand it correctly.... but mentioning "chasis" in eurorack is confusing, as metal panels are usually disconnected while jack shields go directly into 0V.

So is it better to have the jack shields not connected to anything? (so they don't share the "power" cable wires)
Or is it better to sacrifice them one of those 6 pins and have them share only the busboard? (which is order of magnitude better)

And what about panels? If jack shields are already polluting the 0V anyway, couldn't we as well connect all the panels to 0V and have better ground connections thanks to the panels touching each other and all screwed to metal rails too.

Quote:

Just because a cable is called screened does not mean that it acts as one. If it is carrying the signal return current it won't be (what screens the "screen"?).


I understand. Some current returning through screens is surely bad.
How big / significant are those currents? It's very difficult for me to imagine, with so many chaotic interconnections.
Could it be, that with many patch cables, modules are returning less current through their own power connector than through sum of all those parallel paths (through screens and power cables of all other modules)?
Isn't the eurorack busboard perhaps just a decoration and screens are the real busboard? :) Will have to measure.

But doesn't all those interconnections also reinforce the poor 0V distribution among modules? Doesn't so many 0V interconnection through screen also help to lower the differences between 0V in different modules?

Quote:

Eurorack only gets away with ribbon because there are six 0V wires in parallel. Forget IDC connectors. Forget ribbon cable. Forget pcb distribution. None of them scale up. Forget all 0.1" pitch data connectors that were never intended for power use.


Sure, but let's pretend we are bound to eurorack and signal connector/cables used for power. Wouldn't it help a bit, if we used 2 ports on the busboard and two ribbon cables to connect one module? Surely not order of magnitude better, but is almost 2x better pointless? (almost, unless we used two parallel connectors also on the module)

Quote:

everything in it has to have the same common 0V reference and the larger the system the harder that is to achieve
...
every piece of equipment should have a chassis bonding terminal and what I'm calling a "ZeroVolt" connector.


So if somebody's bound to inferior equipment, isn't it perhaps better to forget about chasis entirely (if it can't be done properly anyway), and instead connect any and all grounds together using anything that's available, including the screens?

Quote:

What I can tell you is that improving the power system will improve the system performance of most existing modules, but not improving it will negate the benefits of better designed modules. That doesn't mean don't improve modules, you should do both.


Sad to hear that it's not possible to do much, if power distribution is wrong.
Anyway, I'm curious about what makes a "better designed module".
I can only think of basic things like proper power filtering with big caps, having thick power paths and planes, making separate power paths for noisy/hungry parts of circuit (maybe even filtering them separately).
What are some less obvious things not to forget about?
What are the things that people most often forget to do properly?
ricko
This is just the difference between professional and consumer grade technologies, isn't it? The professional gear has proper points for earthing chassis, connecting shields and uses balanced connectors. The consumer gear just keeps anything with signal clear from the chassis.

If you have an US-style two pin 110V power supply, you don't have any ability to earth the chassis anyway, so your only choice is the consumer approach unless you actually want to make your own earth, but running a wire to some copper pipe that goes into the ground. (I know of professional studios that do this, even in countries with three-pin AC power, because the earth of the 240V power supplies are so full of hash noise.) If you are in a country with three pin AC power, you can use the earth pin.

But I am not sure that the earth circuit is something you really want to distribute on the bus. It seems to violate that inside/outside way of thinking. What a lot of professional equipment does it provide a lug on the chassis which you can attach an earth wire: if you are also using shielded cable you can connect the shield up to that lug as well. I found it really useful to hookup any fixed audio cable, such as long runs between mixers and whatever, and for hi-fi hookups.

(But I do suspect eurorack would have been better to make the sleeve into an earth connection for shielding rather than the 0V connection. Topo late now. I don't think adding an extra band on the jack is feasible, and adding a second prong for earth would just take up valuable real-estate on the module.)
cretaceousear
Could someone explain what Graham's "return voltages" are. They seem to be at 0v for each..
What would these connect to on a module/ pcb?
Grumble
A return voltage is a return to the power supply that measures the voltage at a certain point to regulate the voltage at that point.
Say you want an exact 12 volt at a certain point, you take the return voltage from that point to the power supply, and the power supply regulates the voltage to exactly 12 volt
ricko
The best electrical circuits are often star arrangements, where the 0V used in each section has a separate return to a common point as close to the power supply as possible.

The reason is to isolate each individual circuit as much as possible, for example so that a high speed digital clock does not add noise or signal onto the 0V line that is used by an audio amp. This problem can show up as noise or crosstalk, but also by making VCOs sync up with each other too readily (people implementing dual oscillator designs, such as ARP Oddysey boards have reported this.)

With Eurorack you have six of the 16 pins of the connectors used for ground, but they are always tied together, giving a larger effective 0v wire, so the star point for each module is either at the bus connector or perhaps even just at your module.

I guess the six pins on the connector means you could use each of the six returns for a different 0V, but that wouldn't get you all the way to the power supply. But it might be better than nothing if the bus board treated each of the six 0V lines separately and gave them separate decoupling caps or brought them together at the last moment. I guess.

So is that the best practice for Eurorack modules? Have modules that don't have a common star and treat each of the six 0V lines as separate returns, and have a bus board that does the same, so that at least as far as the bus board's power connections you have a better star. ???
Graham Hinton
cretaceousear wrote:
Could someone explain what Graham's "return voltages" are. They seem to be at 0v for each..
What would these connect to on a module/ pcb?


They are not "return voltages" they are rail voltage return paths. They all go back to the distribution 0V, but through separate wires. By defining them the way I have it makes it possible, if required, to separate two circuits within a module so that they don't interact. The 0V or 0Vs on a module are separated from the system 0V by the resistance in the power cable wires which means there is a voltage drop due to the instantaneous current used. This results in a dc shift plus an ac component to the module 0V and may cause, e.g., an LFO or digital circuit with flashing lights to effect a sensitive analogue circuit in the same module.

Grumble wrote:
A return voltage is a return to the power supply that measures the voltage at a certain point to regulate the voltage at that point.


You are confusing this with regulator sense wires.
Graham Hinton
clorax hurd wrote:

But I'm afraid that with busbar being 50e/piece and needing 3, it won't be affordable for many people... (i hoped it would be chaper. piece of metal with holes. isn't it?).


Well a tyre is just some tree sap with sulphur added, but have you tried making one?
Let me know if you find anyone making them cheaper.

Quote:

(and by the way, regarding the separate return paths for + and - ... what about capacitor of passive lowpass stage? It should be connected to ground.


There are not separate return paths for +ve and -ve, there are two primary wires and two secondary wires. If you don't have need a separate secondary circuit that is four 17awg 0V wires in parallel getting the power cable resistance down to near 1milliohm.

Quote:
Or which ground to should be connected to the + input of inverting opamp stage?)


To the 0V of the circuit it belongs to. Don't call it ground, that only leads to confusion.

Quote:

But lets say that we don't have a choice. Lets say we have control only over a module design and the module is to be mounted into regular eurorack case with other eurorack modules and connected to busboard, and we know it's a nightmare, but we still want to make our best designing the module.


You have a choice between using as is and making improvements, but it is a waste of time and money if you do that with such restrictions.

Quote:

When you say "chasis" you mean both panel+jack shields. right?


Yes and any enclosure.

Quote:

(cause that's the proper way if I understand it correctly.... but mentioning "chasis" in eurorack is confusing, as metal panels are usually disconnected while jack shields go directly into 0V.


Connecting cable shields to pcb 0V has been considered bad practise since 1995. A lot of audio products aim to conform to AES48 now and you will be connecting your synthesizer to them. That's when the crunch comes and your £10,000+ heap of poorly executed design degrades a properly designed <£1000 mixer.

Quote:

So is it better to have the jack shields not connected to anything? (so they don't share the "power" cable wires)
Or is it better to sacrifice them one of those 6 pins and have them share only the busboard? (which is order of magnitude better)


Neither and most busboards are not an order of magnitude better.

Quote:

And what about panels? If jack shields are already polluting the 0V anyway, couldn't we as well connect all the panels to 0V and have better ground connections thanks to the panels touching each other and all screwed to metal rails too.


No.

Quote:
How big / significant are those currents? It's very difficult for me to imagine, with so many chaotic interconnections.


A 5V signal going to 100k input will draw a 50microamp current which will return to source through whatever paths it sees in parallel, but some of those paths will be carrying other currents that may combine and become part of the signal.

Quote:

But doesn't all those interconnections also reinforce the poor 0V distribution among modules? Doesn't so many 0V interconnection through screen also help to lower the differences between 0V in different modules?


It will change it if it is a similar value which means that module 0Vs will change as you patch. If the distribution 0V is a lower resistance by a factor of 100 then 99% of the current will take that path.


Quote:

Sure, but let's pretend we are bound to eurorack and signal connector/cables used for power. Wouldn't it help a bit, if we used 2 ports on the busboard and two ribbon cables to connect one module? Surely not order of magnitude better, but is almost 2x better pointless?


2x 28awg is pretty pointless when you could simply use a discrete wire connector with 20awg instead.



Quote:
So if somebody's bound to inferior equipment, isn't it perhaps better to forget about chasis entirely (if it can't be done properly anyway), and instead connect any and all grounds together using anything that's available, including the screens?


That is what makes it inferior equipment.

Quote:

What are some less obvious things not to forget about?
What are the things that people most often forget to do properly?


1k output resistors should be deprecated for a start.
The signal noise ratio of most audio paths could be improved by paying attention to low noise design techniques.
clorax hurd
Graham Hinton wrote:

2x 28awg is pretty pointless when you could simply use a discrete wire connector with 20awg instead.




That connector looks good.

By the way, there are better ribbon cables than 28awg. 26 & 24 commonly available. I even found 22awg one by amphenol on mouser, i'm just not sure if even that one can be used for crimping IDC connectors onto it. but 1,27mm pitch seems ok: www.mouser.com/ProductDetail/Amphenol-Spectra-Strip/111-2214-010

But 20AWG would be surely even better. Who makes these connectors? How do I find them?
Even though, not so cool contact resistance (looking at specs on your web, says 30mΩ which is same as IDC).

So if these are cable resistances (per one wire)
10cm AWG28 20 mΩ
10cm AWG26 13 mΩ
10cm AWG24 8 mΩ
10cm AWG22 5 mΩ
10cm AWG20 3.3 mΩ
10cm AWG17 1.6 mΩ

And connectors are usually 30mΩ and there is two / one on each side of the cable. then connectors resistance are 2x worse bottleneck than even the thinnest ribbon cables.

So I feel like it wouldn't be pointless to put more power connectors on a module. After the design is finished and some unused space allows it then why not.
Would work with one cable, but if unused slots were available on busboard near the module, then multiple connections could be made to make it better.

If upgrading 30mΩ connectors to 10mΩ mini-fit Jnr makes sense, then using double connections to get 15mΩ also shouldn't be totally pointless.

Graham Hinton wrote:

You have a choice between using as is and making improvements, but it is a waste of time and money if you do that with such restrictions.


Maybe time, but most of the things I suggested doesn't cost much.
For example keeping jack screen signal separate is a design decision and most panel PCBs are not very dense anyway.
Adding 20cent connectors to make double connections is also cheap.

Quote:

Neither and most busboards are not an order of magnitude better.


You are right. It seems to be in the same ballpark. Just measured around 25mΩ between ground pins at ends of a commercial busboard... or around 15mΩ on my DIY made of stripboard tinned with a thin layer of solder. (could improve by making the layer thick))

Quote:

If the distribution 0V is a lower resistance by a factor of 100 then 99% of the current will take that path.


I see. Makes me think, because I just measured around 80mΩ on between sleeve pins of two jacks with 10cm doepfer cable between them.
Path through IDC + ribbon + IDC is also 80mΩ per pin.
And there are six pins...
So if I make 6 connections between two modules with jacks, then it's 13mΩ through the correct path or 13mΩ (sleeves) + 13mΩ (power con of the other module) = 26mΩ through the incorrect path ...
as much as 33% of the current returning to ground through the sleeves then, polluting 0V reference of the other module, right? Seems like a lot. (maybe I made a mistake somewhere?).

Why is it then really worth using sheilded cables at all / worth connecting the sleeve pin of sockets to anything, if it can make so much mess?
Especially with relatively hot signals...and if banana based systems seem to work OK without it too.

Or once again, why not make modules with separate plane for sleeve pins and giving them half of the power connector's 0V pins, while also putting two power connectors on the module to allow dual connection. It wouldn't make the things worse in any way, cause 6 pins still available for 0V... and at the same time it would allow to connect jack screens to 0V at least 2x closer (by 2x closer I mean 1/2 resistance) to PSU than when we are connecting it to the module's 0V plane directly from jack socket's sleeve pin.
Why wouldn't this be worth it? Would cost little and significantly help avoiding mess from one module polluting 0V of other module... or not?

Quote:

1k output resistors should be deprecated for a start.
The signal noise ratio of most audio paths could be improved by paying attention to low noise design techniques.


Thanks. So perhaps also lowering the 100K input impedance a bit, to get less noise?
Could it cause any trouble to just ignore that part of the standard?
(using lower resistors after the input stage comes without saying...)
cretaceousear
For bus bars I had the cheap-ass idea of using these connector blocks -


Screw in to your wooden case a set of 3 connector terminals for each module :
0 - - - - - 0 - - - - - 0 - - - - - 0 - - - - - 0
0 - - - - - 0 - - - - - 0 - - - - - 0 - - - - - 0
0 - - - - - 0 - - - - - 0 - - - - - 0 - - - - - 0

m/ule . . m/ule . . m/ule . . m/ule . . m/ule

and connect the terminals one to another with.. wait for it.. coat hanger wire.

A good size 10 amp connector block would allow you to poke in the wire end that leads to the module along with the coat hanger bus bar, and you have the screw fixing to clamp it all up.
Comments?
Nasty action between copper wire, steel connector terminal and coat hanger wire ?
Are coat hangers coated with any lacquer?
nurbivore
cretaceousear wrote:
For bus bars I had the cheap-ass idea of using these connector blocks -


It's saying something about the state of Eurorack power distribution that this probably would be better...

I don't think that the cheap wire hangers are coated, but if they are, you should be able to sand or strip it off fairly easily.

That said, I don't think you'd get a very reliable connection that way. A small, stranded wire isn't going to sit well between the "bus" wire and the screw. You could get really janky and find/make some kind of really small spade connector that could curve around the wire and slide into the terminal block... I haven't seen anything that would fit the bill, though.
Graham Hinton
clorax hurd wrote:
But 20AWG would be surely even better. Who makes these connectors? How do I find them?


Lots of manufacturers. Search for "0.1" board to wire" connectors. You will need a ~£300 tool to crimp the contacts though.

The cable housing pictured is what is supposed to mate with the unshrouded headers used for Eurorack power.

Quote:

Even though, not so cool contact resistance (looking at specs on your web, says 30mΩ which is same as IDC).


30mΩ is worse case, but that is why a power connector should be used for power rather than a data connector.

Quote:

And connectors are usually 30mΩ and there is two / one on each side of the cable.


That is a consequence of using ribbon cable: you have to have a ribbon connector both ends. With discrete wire you don't.

Quote:

So I feel like it wouldn't be pointless to put more power connectors on a module.


Why put another connector when you can just have a better connector?
The fact that synthesizer modules don't plug in directly to a fixed distribution (like most industrial electronics), but use a cable link means that you can change either end of that link as long as you have a suitable cable. I have been making modules that don't have Eurorack power connectors that may be used in Eurorack cases or other cases and it isn't a problem.

Quote:

For example keeping jack screen signal separate is a design decision and most panel PCBs are not very dense anyway.


That should be done.

Quote:

Just measured around 25mΩ between ground pins at ends of a commercial busboard... or around 15mΩ on my DIY made of stripboard tinned with a thin layer of solder. (could improve by making the layer thick))


Solder has quite a high resistivity, you should be looking at copper or aluminium. Making heavy copper plane pcbs is expensive and you can never get enough copper or even approach the resistance that can be achieved with a simple busbar. Soldering connectors becomes difficult and industrial backplanes use push in connectors.

Quote:

as much as 33% of the current returning to ground through the sleeves then, polluting 0V reference of the other module, right?


The 0V of each module is more "polluted" by the module itself. Any current due to the power consumption will shift the 0V relative to the distribution. Some modules now, typically digital ones regulating the +12V, can put as much as 250mA into the 0V.

Quote:

Why is it then really worth using sheilded cables at all / worth connecting the sleeve pin of sockets to anything, if it can make so much mess?


The mess is caused by the incorrect connections.

Quote:

Especially with relatively hot signals...and if banana based systems seem to work OK without it too.


Screening works in both directions, it protects other signals too. Banana systems don't work OK, you can modulate any signal by gripping a patchcord.

Quote:

Would cost little and significantly help avoiding mess from one module polluting 0V of other module... or not?


Not. Get away from this idea of modifying the use of the existing connector. You can't polish a turd.

Quote:

So perhaps also lowering the 100K input impedance a bit, to get less noise?
Could it cause any trouble to just ignore that part of the standard?


It is just a convention, not a standard.
I would write a standard in terms of what should be achieved, not by giving a specific circuit. E.g. if you say that all outputs should deliver the correct voltage into loads down to 2k then that is achievable with common circuits and allows for enough loads in parallel even if they are 10k. The thing that prevents using lower input resistances is the 1k outputs forming a potential divider.
Actually there is an IEC standard which says that all audio outputs should be 50ohms or less.
JohnLRice
Graham Hinton wrote:
You can't polish a turd.
This is a common misconception, you just have to freeze them first, but I digress. hihi
Grumble
cygmu
JohnLRice wrote:
Graham Hinton wrote:
You can't polish a turd.
This is a common misconception, you just have to freeze them first, but I digress. hihi


But what you end up with for your trouble is a shiny turd.
KSS



seems you can polish a turd. It just has to be really old. Gives a new answer to what is that shite you're wearing. Is it TRex, or Allosour? ;-)
ricko
By the way, you can follow the earlier thread discussing Hinton's power arrangement that he is proposing as a standard for professional quality eurorack equipment, at https://www.muffwiggler.com/forum/viewtopic.php?t=202200 which includes a link to the actual doco on his site.




(Here are some more on power, which is kinda a Hinton's greatest hits!)

Some photos at: https://www.muffwiggler.com/forum/viewtopic.php?t=191493

Some discussion on the "worst possible" arrangement, supposedly, at https://www.muffwiggler.com/forum/viewtopic.php?t=135536

A discussion of signal ground (I prefer 0V, or signal common), chassis ground and earth ground (e.g. mains earth) is at https://www.muffwiggler.com/forum/viewtopic.php?t=78451 and https://www.muffwiggler.com/forum/viewtopic.php?t=102801. You don't want your local signal ground to connect to the chassis ground (if possible) so you don't become part of the circuit when you touch the panel or jacks; you don't want your signal ground to be connected to the mains ground because it may have lots of noise on it; but youDO want your chassis ground to be connected to an earth ground so you don't get zapped.

And https://www.muffwiggler.com/forum/viewtopic.php?t=135351 has a nice diagram showing that if you have multiple systems each with power supplies, you are best bringing each of their 0V out to a common point (i.e. a star), using heavy cable and a busbar as the centre of the star.

What kind of wire and crimps? https://www.muffwiggler.com/forum/viewtopic.php?t=151443 https://www.muffwiggler.com/forum/viewtopic.php?t=149590
Graham Hinton
We seem to have more interest in scatology around here than engineering.

ricko wrote:
By the way, you can follow the earlier thread discussing Hinton's power arrangement that he is proposing as a standard for professional quality eurorack equipment,


I am not "proposing" anything for Eurorack. There is no way of improving existing modules, they are a lost cause. The power distribution may be improved considerably, but a system will always be compromised by the module connectors.

Quote:

You don't want your local signal ground to connect to the chassis ground (if possible) so you don't become part of the circuit when you touch the panel or jacks; you don't want your signal ground to be connected to the mains ground because it may have lots of noise on it; but youDO want your chassis ground to be connected to an earth ground so you don't get zapped.


However carefully I try to explain it there is always someone who doesn't get it. It is all connected, but it is the order that it is connected in and the resistances involved that makes the difference. The 0V signal ground should be connected to the mains Earth for a potential reference, but not through cables from the PSU carrying large varying currents. When you connect two separately powered pieces of equipment there may be a common path through part of the mains Earth wiring, but only as far as the mains distribution. The mains Earth is connected, but it is not part of a signal circuit and noise does not magically disappear in to it.
MUFF WIGGLER Forum Index -> Music Tech DIY  
Page 1 of 1
Powered by phpBB © phpBB Group