I have been struggling with some serious xtalk in a preamp based on Stevens and Billington tx102 mk2 transformer volume control for the past 2 weeks. Any help is greatly appreciated.
The preamp is connected to the balanced output of a dcx2496 digital xover .
Let's consider just one amp for simplicity. With one channel (e.g. right) muted or disconnected, I nonetheless hear sound coming from the corresponding (e.g right) speaker. The sound is lower in volume relative to the connected (e.g left) channel but does increase as I raise the volume on the muted (right) channel.
The problem is there only if the preamp is connected to a single ended amp but disappears if connected to a balanced amp. Also the problem is not there if the preamp input for the non-connected channel is shorted between hot and cold (pin 2+3).
The preamp layout is the following. The transformer is wired for unit gain and the primary wires to the inputs are twisted. The transformer screen wires are brought to a single common ground point (a screw on the chassis). Input and output connectors grounds (pin 1) are also connected to such point but can be floated (input and output separately) by means of switches. Floating or grounding makes no difference as far as xtalk is concerned.
As for the connection between the preamp and the single ended amp (a class D Charlize): preamp xlr pin 2 to amp rca hot, pin 3 to rca ring, pin 1 to amp chassis.
I am stumped
and cannot think of any possible solution. Most of all I cannot understand why the xtalk goes away if a balanced amp is used. Given that it increases with volume I would expect it to be happening on the input (primary). So why does the amp matter?
TIA
Giulio
The preamp is connected to the balanced output of a dcx2496 digital xover .
Let's consider just one amp for simplicity. With one channel (e.g. right) muted or disconnected, I nonetheless hear sound coming from the corresponding (e.g right) speaker. The sound is lower in volume relative to the connected (e.g left) channel but does increase as I raise the volume on the muted (right) channel.
The problem is there only if the preamp is connected to a single ended amp but disappears if connected to a balanced amp. Also the problem is not there if the preamp input for the non-connected channel is shorted between hot and cold (pin 2+3).
The preamp layout is the following. The transformer is wired for unit gain and the primary wires to the inputs are twisted. The transformer screen wires are brought to a single common ground point (a screw on the chassis). Input and output connectors grounds (pin 1) are also connected to such point but can be floated (input and output separately) by means of switches. Floating or grounding makes no difference as far as xtalk is concerned.
As for the connection between the preamp and the single ended amp (a class D Charlize): preamp xlr pin 2 to amp rca hot, pin 3 to rca ring, pin 1 to amp chassis.
I am stumped
and cannot think of any possible solution. Most of all I cannot understand why the xtalk goes away if a balanced amp is used. Given that it increases with volume I would expect it to be happening on the input (primary). So why does the amp matter?TIA
Giulio
Hi Giulio!
I'm not sure how exactly you measured crosstalk but it might not be a bad idea to repeat the procedure. The input of the channel which is measured should be terminated by a resistor whose value is equal to the output of the preamplifier/source. If it is a solid state source, the input may be shorted because the output impedance of the source is very low. If the output of the source is balanced, you should shorten pins1, 2 and 3 at the input XLR connector. This should give you much more accurate crosstalk measurements.
The difference you have noticed between unbalanced and balanced connections clearly demonstrates the advantages of balanced connections. It is obvious that the crosstalk between the channels is induced in both signal wires equally and can be sucessfully eliminated by balanced connections, but not by unbalanced connections.
Finally, I wouldn't worry too much about crosstalk. Regardless of the connection, proper crosstalk measurement procedure should give good results.
Regards,
Milan
I'm not sure how exactly you measured crosstalk but it might not be a bad idea to repeat the procedure. The input of the channel which is measured should be terminated by a resistor whose value is equal to the output of the preamplifier/source. If it is a solid state source, the input may be shorted because the output impedance of the source is very low. If the output of the source is balanced, you should shorten pins1, 2 and 3 at the input XLR connector. This should give you much more accurate crosstalk measurements.
The difference you have noticed between unbalanced and balanced connections clearly demonstrates the advantages of balanced connections. It is obvious that the crosstalk between the channels is induced in both signal wires equally and can be sucessfully eliminated by balanced connections, but not by unbalanced connections.
Finally, I wouldn't worry too much about crosstalk. Regardless of the connection, proper crosstalk measurement procedure should give good results.
Regards,
Milan
Milan,
thanks a lot for your reply. I do not have any way of measuring crosstalk but it was clearly audible. The disconnected (or muted) channel was definitely not 60db below the one playing (closer to 20db or less, the total attenuation allowed by the TX102 being -46db).
Connecting pin 1 (ground) and 2 (-ve) on the output xlr sorted out the problem completely. The muted channel is now dead silent. My understanding is that leaving the output of the preamp floating implied the single ended amp (battery powered) lacked a path to ground. Connecting the preamp output pins 1 and 2 grounds the amp through the DCX. In fact if the preamp output ground is floating relative to input ground the problem persists.
This was not true for the balanced amp which has its own earthing.
I hope this post hoc ergo propter hoc explanation makes sense...
Best
Giulio
P.S. I have heard very good things about you from my good friend Brian C. (the Spuddo master). I can no longer find your xover page though.
thanks a lot for your reply. I do not have any way of measuring crosstalk but it was clearly audible. The disconnected (or muted) channel was definitely not 60db below the one playing (closer to 20db or less, the total attenuation allowed by the TX102 being -46db).
Connecting pin 1 (ground) and 2 (-ve) on the output xlr sorted out the problem completely. The muted channel is now dead silent. My understanding is that leaving the output of the preamp floating implied the single ended amp (battery powered) lacked a path to ground. Connecting the preamp output pins 1 and 2 grounds the amp through the DCX. In fact if the preamp output ground is floating relative to input ground the problem persists.
This was not true for the balanced amp which has its own earthing.
I hope this post hoc ergo propter hoc explanation makes sense...
Best
Giulio
P.S. I have heard very good things about you from my good friend Brian C. (the Spuddo master). I can no longer find your xover page though.
giulio said:
You've been very kind Giulio. Thank you. Say hello to Brian for me.
The original MOX thread is a bit difficult to trace (a ton of stuff has been written on the MOX in the meantime) but here it is:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=27651&highlight=
Mind you, I myself am not using the old MOX any more. I've developed a new and a more compact version, with ICs instead of discrete op-amps. You may check it out here:
http://www.moxtone.com/mox.htm
The text is in Croatian but the pictures and schematics should be self-explanatory. Anyway, I'm here if you need more info.
Regards,
Milan
Hi Milan,
thanks for the link. The picture are indeed self-explanatory. I am using something similar (w/ dip sockets to plug in the appropriate resistors and capacitors) for prototyping.
At the moment it does just baffle step given that I am using the DCX2496 for the rest. Have you seen people raving about the new LM4562 opamp?
Best
Giulio
thanks for the link. The picture are indeed self-explanatory. I am using something similar (w/ dip sockets to plug in the appropriate resistors and capacitors) for prototyping.
At the moment it does just baffle step given that I am using the DCX2496 for the rest. Have you seen people raving about the new LM4562 opamp?
Best
Giulio
giulio said:
Yes, it looks very promising. I tried to get it from the National through their sample program but the shipping cost was prohibitive (around USD30) so I dropped the whole thing (and it's all because of those SOBs who sell their samples
). Anyway, I certainly plan on giving the LM4562 a try as soon as I get it. Have you tried it yet?
Regards,
Milan
There is a group buy here at DIYAUDIO. I think it is still open. Same story for me as far as samples go. In all fairness National has never been too kind to me, as opposed to AD...
I have ordered 12 to try in my baffle step eq + balanced buffer circuit and a pair of pimeta headphone amps.
Best
Giulio
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.