500 series design questions

[Bobby Baird]

Going back to my question in the first post, which pin on the card connector (1, 5, or 13) would be the ideal place to connect an internal shield or ground connection of a transformer? I would think that an external shield like a mu metal case or similar would go to chassis, but would the same apply to something like pin 6 in the attached image, or would that be better off going to audio ground? I know there's probably a certain amount of "just try each and see if one gives better results," but since I'm laying out a board, I'd like to try to get it right as early as I can.

I would attach transformer shield to chassis pin 1.

 

[dmp]

Going back to my question in the first post, which pin on the card connector (1, 5, or 13) would be the ideal place to connect an internal shield or ground connection of a transformer?

Pin 5 & 13 are connected together and are audio ground. Pin 1 is chassis ground.
Shields are typically best connected to chassis ground - I would think including an external transformer shield.
A core ground wire (if that is that pin 6 is in your image) I think should go to audio ground, since noise in the core can pollute the audio. Maybe Jensen has an application note for this? This datasheet implies BOTH the core ground wire and external case should go to audio ground

 

[ccaudle]

which pin on the card connector (1, 5, or 13) would be the ideal place to connect an internal shield or ground connection of a transformer?

I am going to disagree with the other posters who recommended connecting internal transformer shield connections to the chassis connection.
The purpose of chassis is to keep EMI out of the inside of the chassis, and by connecting an internal component shield to the chassis, you have now made a connection between the outside and inside of the chassis, which bypasses the point of having a shielding chassis to begin with.

To determine where the best place to connect an internal shield would be, you need to determine where the circulating currents would be sourced, and make the connection which returns those currents most directly to their source.
That brings up the exception to the rule I mentioned above, which is EMI filtering components. Those might technically be inside the chassis, but the source of the current is coming in on the external wires, so the filtering components should be as close to the input connectors as possible (ideal would be on the input connectors) and return to the chassis connection.

So the first question to ask is what is the shield shielding (i.e. what is the sensitive node which is being protected), and what is the shield shielding the sensitive side from (i.e. where is the source of the interference)?
In the case of an input transformer, or possibly an output transformer, you may decide that the shield is actually part of the EMI protection, in which case it may be appropriate to connect to the chassis connection. The best solution in those cases is dual shield windings, one around primary and one around secondary. You see that in some Jensen input transformers (note also that the Jensen diagrams indicate whether the shield winding is closer to the secondary winding or closer to the primary winding by the way the symbol is drawn).

In other cases you are trying to prevent e.g. power supply noise from getting across from a higher voltage winding to a lower voltage secondary winding, so the primary winding is the driven side, and the transformer shield winding should be connected back to the ground for the primary side to return parasitic currents back to the source.

In the example screenshot you provided, it appears to be a signal transformer with 1:1:1:1 windings, so it is not clear how the transformer is used, what is driving and what is receiving. If the two sides are truly galvanically isolated it can be better to tie the shield to the secondary side ground reference, because if you tie it to the primary side then any common mode noise (i.e. primary side ground reference moving in relation to the secondary side ground) will be imposed on that shield winding, and couple through parasitic capacitance into the secondary side windings.

Sorry the question is not quite as straight forward as you were hoping, but for best performance the details matter.

 

[JMan]

Sorry the question is not quite as straight forward as you were hoping, but for best performance the details matter.

It never is!

Well, I should have given a little more context for my question - the transformer in question is set up 1:1 (10k:10k) on the input of the module, and is used to unbalance the signal going into a passive inductor-based EQ circuit. The picture I included was of the UTM2571 (similar to Carnhill VTB 9071), which has no external shielding but does have this earth pin which I took to be an internal shield of some sort.

I’m thinking audio ground might be the best choice here, but I also might just add a three pin header with a shunt to the board so that I can easily try out both audio and chassis ground separately and see what changes. Maybe this is weird, I don’t know, but it’s the idea I came up with.

 

[richiyobs]

is the attached the right card edge connector 15/30 pin footprint for 500 Goldfinger?
it s Edac 307-xxx
found it on Mouser and the folder contains different Ecad platform symbols and models...
is this a good and cheap alternative? https://it.aliexpress.com/item/32395403249.html?gatewayAdapt=glo2ita

thanks for your attention
best

 

[abbey road d enfer]

Well, I should have given a little more context for my question - the transformer in question is set up 1:1 (10k:10k) on the input of the module, and is used to unbalance the signal going into

With just this part of your answer, I would have answered: shield connected to the input stage's "ground"...

a passive inductor-based EQ circuit.

...but with this complement of information I must take time and sit. Since it's a passive EQ, EMI/RFI is not likely to interfere with its operation, so the question now is, what about RFI/EMI that goes through and where to stop it. Indeed, it must be stopped at the output, which can be done with two caps connected to chassis, maybe with the help of inductors. If there is an output xfmr with a shield, it should be grounded to chassis.
I would say in this particular case, the input xfmr shield should be grounded to chassis.
Or is there an active stage in this passive EQ?

 

[JMan]

Since it's a passive EQ, EMI/RFI is not likely to interfere with its operation, so the question now is, what about RFI/EMI that goes through and where to stop it. Indeed, it must be stopped at the output, which can be done with two caps connected to chassis, maybe with the help of inductors. If there is an output xfmr with a shield, it should be grounded to chassis.
I would say in this particular case, the input xfmr shield should be grounded to chassis.
Or is there an active stage in this passive EQ?

So what I'm putting together is a somewhat simplified version of one of Ian's EQs. The EQ itself is fully passive, but I am indeed feeding it into a make-up amp that uses an IC prior to the output transformer. The OPT does not have a connection for an internal shield, nor does it have a mu metal or other shielded housing.

 

[abbey road d enfer]

So what I'm putting together is a somewhat simplified version of one of Ian's EQs. The EQ itself is fully passive, but I am indeed feeding it into a make-up amp that uses an IC prior to the output transformer. The OPT does not have a connection for an internal shield, nor does it have a mu metal or other shielded housing.

In that case, I would start with the shield connected to the audio ground of the active stage.

 

[estocastico7]

I want to make 500 series projects,but i cant find 500 series veroboard,where can i find?thanks

 

[JMan]

There’s someone in the Black Market selling 500 series veroboards. AML also has them.

 

[firerunner]

I have a similar question about the multiple returns on the 500 series.
In some designs I have seen post regulator filtering/reservoir caps tied to the signal ground.

As some diagrams in this thread have mentioned, is it more appropriate to tie the larger filtering caps to Power GND (pin 13)? Is the longer return path through the power supply negligible?

And with that I have smaller(10uf+0.1uf) local decoupling for DOA's tied to signal ground (Pin 5), that sound ok? Or should every power cap tie to pin 13?

Thanks!

 

[abbey road d enfer]

I have a similar question about the multiple returns on the 500 series.
In some designs I have seen post regulator filtering/reservoir caps tied to the signal ground.

As some diagrams in this thread have mentioned, is it more appropriate to tie the larger filtering caps to Power GND (pin 13)? Is the longer return path through the power supply negligible?

Not a simple question.
If you connect reservoir caps to the audio ground, whatever ripple there is on the rails will transfer to the audio ground. How much depends on many factors, value of the caps, line resistance...
If you connect reservoir caps to power gnd, the voltage may not be as clean as you wish, but most audio stage have a relatively good PSRR at least those that use NFB) so it shouldn't be a problem. Even less if there are regulators.

And with that I have smaller(10uf+0.1uf) local decoupling for DOA's tied to signal ground (Pin 5), that sound ok? Or should every power cap tie to pin 13?

The optimum point for decoupling caps is always the point where the load current returns. But it's not always feasible, so one has to make a choice.
I've seen decoupling caps returning to audio ground, and I've seen them returning to a dedicated "dead dog" bus. I have never seen one being objectively better than the other. I'm glad other people have tried that for me, so I could learn from their mistakes.

 

[firerunner]

Not a simple question.
If you connect reservoir caps to the audio ground, whatever ripple there is on the rails will transfer to the audio ground. How much depends on many factors, value of the caps, line resistance...
If you connect reservoir caps to power gnd, the voltage may not be as clean as you wish, but most audio stage have a relatively good PSRR at least those that use NFB) so it shouldn't be a problem. Even less if there are regulators.

The optimum point for decoupling caps is always the point where the load current returns. But it's not always feasible, so one has to make a choice.
I've seen decoupling caps returning to audio ground, and I've seen them returning to a dedicated "dead dog" bus. I have never seen one being objectively better than the other. I'm glad other people have tried that for me, so I could learn from their mistakes.

Thanks for the response Abbey. I may try doubling up on the reservoir caps as a testbed, with one pair connected to AGND and the other to PGND with some resistance between them. I'm currently using 820uf tied to PGND with 10ohm resistors.

Sounds good on the decoupling, I'll keep them tied to AGND for now. Would you mind elaborating on the dead dog bus? My searches yield sad results about actual dogs, poor puppers!

 

[abbey road d enfer]

Would you mind elaborating on the dead dog bus? My searches yield sad results about actual dogs, poor puppers!

Dead dog bus was a colloquial term used in restricted circles.
In essence it's a "ground" bus running in parallels with the audio ground, but to which only the decoupling capacitors were connected.
The idea was that noise induced on the power rails would transfer to this dirty bus, and thus not pollute the "noble" audio ground.
Of course it just doesn't make sense when you analyze how currents travel.

 

[firerunner]

Dead dog bus was a colloquial term used in restricted circles.
In essence it's a "ground" bus running in parallels with the audio ground, but to which only the decoupling capacitors were connected.
The idea was that noise induced on the power rails would transfer to this dirty bus, and thus not pollute the "noble" audio ground.
Of course it just doesn't make sense when you analyze how currents travel.

Very interesting. Thanks for clarifying!