Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Is there something wrong with this picture?
New Posts  All Forums:Forum Nav:

Is there something wrong with this picture?

post #1 of 37
2/20/09 at 7:19pm
Thread Starter 
If you think the case should not be grounded this way (ie to the transformer bolt) please be specific as to why this is a problem in this build.

Reply
post #2 of 37
2/20/09 at 7:51pm
I don't recall the theory behind it but this has been advised against ever since I started on this hobby a decade or 2 ago. I have it in my mind that it has to do with ground noise and the inductive possibility at the center of the core.

It has looked like an obvious place to me in the past but I have resisted the temptation. I would like to get to the bottom of this too..dB

ps - nice build btw
Reply
post #3 of 37
2/20/09 at 8:09pm
If the other end of the wire touches the case it effectively creates a 'shorted turn' around the transformer, and you can guess the rest...

Do not do it!
Reply
post #4 of 37
2/20/09 at 8:18pm
Since it appears to be ground on the IEC connector I say it doesn't matter one bit. It won't touch anything in the power circuit until it gets back to the breaker box anyway, where it will be tied to neutral.
Have you listened to it yet?
Reply
post #5 of 37
2/20/09 at 8:22pm
Thread Starter 
Quote:
Originally Posted by Leny View Post
If the other end of the wire touches the case it effectively creates a 'shorted turn' around the transformer, and you can guess the rest...

Do not do it!
I think we can assume everyone knows what a shorted turn is. If not here's a description:
WikiAnswers - What is a shorted turn

It's true that if the black wire on the IEC touches the case the turn is then shorted. However, it is not touching the case, and if it did break and contact the case it would simply blow the 250mA fuse before any damage was done (actually it is a safety feature as it you can't operate the amp with the case ungrounded ).

Also, no offense personally, but please no statements of "DON'T DO IT" etc. I'd like this to be about learning and not dogma.

Quote:
Originally Posted by dBel84 View Post
ps - nice build btw
Thanks

Quote:
Originally Posted by digger945 View Post
Have you listened to it yet?
Of course!
Reply
post #6 of 37
2/20/09 at 8:27pm
If it was a shorted turn, which I don't think it is, then it's obvious why this would be bad. Of course you would already have a fire so...

As to your response on the other thread about how is it different from the mount being grounded. I'm not sure, but perhaps the mount and bolt is better at NOT picking up noise then a wire. By attaching the wire to the mount, you are placing it very close the transformer. Knowing how a transformer works, it's easy to see why it would be a major cause of EMI and introduce noise to the signal. I'm not saying that in your pic, there is a problem. Again, like I said in the other thread, if nothing is too hot and it sounds good to you, then enjoy it and don't worry about it If it were me, I just wouldn't attach it there. If you're worried that the standoff isn't as good a connection, make a connection for it. I think the main goal would be to move as many wires away from the transformer as possible.
Reply
post #7 of 37
2/20/09 at 8:34pm
Thread Starter 
Quote:
Originally Posted by Hayduke View Post
By attaching the wire to the mount, you are placing it very close the transformer. Knowing how a transformer works, it's easy to see why it would be a major cause of EMI and introduce noise to the signal. I'm not saying that in your pic, there is a problem. Again, like I said in the other thread, if nothing is too hot and it sounds good to you, then enjoy it and don't worry about it If it were me, I just wouldn't attach it there. If you're worried that the standoff isn't as good a connection, make a connection for it. I think the main goal would be to move as many wires away from the transformer as possible.
Care to explain precisely how it is a 'major cause of EMI'?

I will move the ground if convinced that there is good reason to. I don't believe in doing something just because so and so says I should. I want reasons (please)
Reply
post #8 of 37
2/20/09 at 8:37pm
Agreed - your build is nicely laid out and attractive. Well done!

I wanted to reply to a few comments that you made in 'that other thread'

1. 'It doesn't matter where you ground the IEC to the chassis'
-I assume you are not referring to a circuit with a star ground. In most cases it certainly does matter where you ground the IEC. Just a few weeks ago I cured a tiny line-frequency buzz in a ß22 by relocating the IEC ground wire from a random bolt in the middle of the case to the chassis ground pad on the ß22 backplane.

2. 'I chose the transformer bolt for the lowest impedance path to the chassis'
-If you're after the lowest impedance path, why not just crimp a ring terminal on the end of the wire and bolt it right to the chassis? Also, I'm surprised that you measured the transformer bolt as having lower impedance than the standoffs. Regardless, bolting directly to the chassis will have lower impedance than going through a bolt or standoff.

Even when there's only a minute chance of something bad happening, if there exists an alternate solution that reduces the chances to virtually zero at no additional cost, I'm always inclined to take that safer solution.
Reply
post #9 of 37
2/20/09 at 8:50pm
Thread Starter 
Quote:
Originally Posted by MoodySteve View Post
Agreed - your build is nicely laid out and attractive. Well done!
Thanks


Quote:
Originally Posted by MoodySteve View Post
I wanted to reply to a few comments that you made in 'that other thread'

1. 'It doesn't matter where you ground the IEC to the chassis'
-I assume you are not referring to a circuit with a star ground...
No, I'm just referring to my build. Star grounding is not used here.

Quote:
Originally Posted by MoodySteve View Post

2. 'I chose the transformer bolt for the lowest impedance path to the chassis'
-If you're after the lowest impedance path, why not just crimp a ring terminal on the end of the wire and bolt it right to the chassis? Also, I'm surprised that you measured the transformer bolt as having lower impedance than the standoffs. Regardless, bolting directly to the chassis will have lower impedance than going through a bolt or standoff.
It is just a better mechanical connection in my opinion. Also, see my comment about 'safety feature' above. Besides that it allows for a neater layout.


I want to try to clarify what's going on here with a drawing. I have drawn the partial loop in blue.

Reply
post #10 of 37
2/20/09 at 8:51pm
Quote:
Originally Posted by rds View Post

Also, no offense personally, but please no statements of "DON'T DO IT" etc. I'd like this to be about learning and not dogma.

Also, no offense personally, but as a professional electrical engineer I feel obliged to discourage bad design practice. My 'Do not do it' stands firm. Naturally it is the builder's choice if he / she chooses to ignore reasonable and experienced advice. However I would advise consideration of potential future events; for example you might short the turn while fault-finding, or while modifying the unit, or you might sell it to someone who does not understand, or you might make a replica for someone who pokes about inside it... all of which suggest that the initial design should be wired in a relatively safe way. It is so easy to change to a conventional grounding point that there is no logical reason at all to use the transformer bolt now that it has been estabished that it is potentially dangerous. I'd like this to be about learning and not potentially dangerous mains related wiring.

L.
Reply
post #11 of 37
2/20/09 at 9:07pm
Quite a bit more than a supermodded Ear.
It looks great, I wish someone was bringing one to the upcoming meet(or a Buffalo).
Reply
post #12 of 37
2/20/09 at 9:16pm
Quote:
Originally Posted by rds View Post
It is just a better mechanical connection in my opinion. Also, see my comment about 'safety feature' above. Besides that it allows for a neater layout.
There is almost no better mechanical connection than a good crimp. Bolting a crimped ring terminal is a far more robust connection than squeezing the wire between the bolt head and fastening plate.

I hope you're not referring to your fuse as the 'safety feature'; relying on a fuse for fault protection is never a good idea.

I agree with Leny - there's just no real reason to ground the IEC there - it achieves what can be easily achieved elsewhere but poses an unnecessary (albeit extremely unlikely) risk.

I think your signal wiring and transformer location is optimal, although there will be an orientation of the transformer that has minimum induced noise (in theory - if it's all below audibility you could argue it's just numbers).

Please don't get me wrong - this is a great build - we're just trying to make it greater-er

I've been diddling around with my ß22 for a year because of that very tendency FWIW.
Reply
post #13 of 37
2/20/09 at 9:31pm
Thread Starter 
Quote:
Originally Posted by MoodySteve View Post

I hope you're not referring to your fuse as the 'safety feature'; relying on a fuse for fault protection is never a good idea.
No, after all it only works if the broken ground touches the case. That would be a terrible thing to rely on.

Quote:
Originally Posted by MoodySteve View Post

I agree with Leny - there's just no real reason to ground the IEC there - it achieves what can be easily achieved elsewhere but poses an unnecessary (albeit extremely unlikely) risk.
I resist what Leny is saying on the grounds that he is arguing as an EE rather than explaining his argument. When someone tries convince me by their authority they will surely fail (call it stubbornness).

I have taken a course specifically on (electrical) medical equipment design (where standards are particularly stringent), and never once did I hear mention of considering what might happen if a layperson opens up and starts messing with a power supply. After all there is exposed hot and ground wires. I think a shorted turn is the least of their worries.

Quote:
Originally Posted by MoodySteve View Post
I think your signal wiring and transformer location is optimal, although there will be an orientation of the transformer that has minimum induced noise (in theory - if it's all below audibility you could argue it's just numbers).
Please don't get me wrong - this is a great built - we're just trying to make it greater-er
I really do appreciate the comments and I am not offended at all by criticism, but I want us to walk away with a deeper understanding of what precisely is wrong. If all we gain is to say don't do it cause so and so said don't, then imo this has been totally pointless.
Reply
post #14 of 37
2/20/09 at 10:03pm
Quote:
Originally Posted by rds View Post
No, after all it only works if the broken ground touches the case. That would be a terrible thing to rely on.
The broken ground wire touching the case is one of the very faults that requires either prevention (which an alternate grounding point would accomplish) or protection, though. In this case, you are relying on the fuse for protection against this fault, since the chance does exist (again, however remotely).

Engineers are trained to absolutely minimize risk - it goes without saying that eliminating it when possible is ideal. In this case, there's an easy way to completely eliminate the chance of a shorted turn. Why not take that approach?
Reply
post #15 of 37
2/20/09 at 10:17pm
I have basically no authority, and Im pretty sure everyone thinks Im full of **** at least half of the time. In short Im an idiot, so I'l give it a go too.

The argument of why NOT to ground BOTH ends of the bolt has been explained. As it sits now, it does not appear that you have a shorted turn BUT anything connecting ground to the case in a second location will create a loop (this includes stuff OUTSIDE of the chassis touching it... The "case end" of the bolt touching the top of a metal computer case for very possible real world example, or paper clips left on a metal desk. I could go on) and a shorted turn.

Its a weak and unsafe practice at best where better methods in every respect are accepted.

You have made a design/construction decision which puts the overall durability of your gear in question to save the expense of a nut, bolt, a couple washers and a crimp connector.
Quote:
Originally Posted by rds View Post
I have taken a course specifically on (electrical) medical equipment design (where standards are particularly stringent), and never once did I hear mention of considering what might happen if a layperson opens up and starts messing with a power supply.
you are correct. Hopefully warning labels will protect you from the lawsuits when unskilled people open your gear and get zapped.

This piece of gear on the other hand can be made to fail catastrophically without opening the case. it is an accident waiting to happen.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › DIY (Do-It-Yourself) Discussions › Is there something wrong with this picture?