So I bought a couple of lundahl transformers, the LL1540. They are 1+1:1+1 high impedance units. They are used as input transformers for a PP amp, performing the phase splitting duties. I connected them as per the schematic: 12k on each secondary winding and the center tap grounded. The primaries are connected in series, and a RCA jack is connected to them, in first instance 'floating'.
Turning the amp on without anything connected to the jacks reveals a 'dead' quiet amp through my test speakers. As soon as I connect a RCA interconnect it starts humming. The interconnect is then plugged into the headphone output of a desktop CD-player. There is still a hum...when I turn on the CD-player the thing starts: heavy audible distortion at low levels. With an alligator clip I shorted the ground bar of the amplifier to the ground of the input jacks: the volume raised a bit and distortion was, apparently, gone. Hitting the pause reveals, again, a dead quiet amplifier.
Wonderful, isn't it? Still, I would actually prefer the input floating, just to create the galvanic insulation between CD player and amplifier. This is one of the main reasons for installing this input transformer anyway. So, what can be going on?
The CD player has a two pin connector (double insulated) and is not in any way grounded. The prototype's main earth is grounded to the 'central heating'.
I can draw it, even take some pictures (which would also reveal details about the new borg collective ) Still I think that the 'cure': non-floating primaries is not an unknown for most of you: I already saw some schematics with a switch to select between 'floating' and 'non-floating'. The switch is the easy cure: I would like a solution which really floats the primary!
Turning the amp on without anything connected to the jacks reveals a 'dead' quiet amp through my test speakers. As soon as I connect a RCA interconnect it starts humming. The interconnect is then plugged into the headphone output of a desktop CD-player. There is still a hum...when I turn on the CD-player the thing starts: heavy audible distortion at low levels. With an alligator clip I shorted the ground bar of the amplifier to the ground of the input jacks: the volume raised a bit and distortion was, apparently, gone. Hitting the pause reveals, again, a dead quiet amplifier.
Wonderful, isn't it? Still, I would actually prefer the input floating, just to create the galvanic insulation between CD player and amplifier. This is one of the main reasons for installing this input transformer anyway. So, what can be going on?
The CD player has a two pin connector (double insulated) and is not in any way grounded. The prototype's main earth is grounded to the 'central heating'.
I can draw it, even take some pictures (which would also reveal details about the new borg collective ) Still I think that the 'cure': non-floating primaries is not an unknown for most of you: I already saw some schematics with a switch to select between 'floating' and 'non-floating'. The switch is the easy cure: I would like a solution which really floats the primary!
Thanks for the contribution! I am now working on a school project, but I will try it out tonight.
The CD player is an original Sony 227ES-D player. I have no idea how the input is done. This is the first time I read that the source can have an influence... I though about trying to feed the power amp from the output of my headphone amp, but this is a parafeed design which is therefore also cap-coupled.
I will try the two resistors and let you know the outcome!
The CD player is an original Sony 227ES-D player. I have no idea how the input is done. This is the first time I read that the source can have an influence... I though about trying to feed the power amp from the output of my headphone amp, but this is a parafeed design which is therefore also cap-coupled.
I will try the two resistors and let you know the outcome!
With a single ended source, you may be doomed to having to ground one end of the input transformer. Your input may be balanced, but do you really think that (the two wire power connection notwithstanding) the impedance to ground of both the output conductors is the same?
FWIW, I float the input to my preamp with no hum issues.
FWIW, I float the input to my preamp with no hum issues.
Hi Stuart
It was exactly that little piece of iron on the schematics of your preamp and your poweramp that inspired me in going for the Lundahl at the input. Also this recent thread made me think that some galvanic insulation would not hurt.
That cap-coupled outputs can cause this problem I didn't know. Also I do not understand why it matters, but Zigzaflux has quite some experience with 'iron', so I accept and will try his recommendation.
Sure my CD-player is single ended output, but if I connect ground and signal to the balanced primary off the transformer there is no unbalanced impedance to ground...well, there is no 'ground' at the transformers input (I would think): all the primary sees is the output impedance of the CD-player.
Maybe different transformers react different to this?
Erik
It was exactly that little piece of iron on the schematics of your preamp and your poweramp that inspired me in going for the Lundahl at the input. Also this recent thread made me think that some galvanic insulation would not hurt.
That cap-coupled outputs can cause this problem I didn't know. Also I do not understand why it matters, but Zigzaflux has quite some experience with 'iron', so I accept and will try his recommendation.
Sure my CD-player is single ended output, but if I connect ground and signal to the balanced primary off the transformer there is no unbalanced impedance to ground...well, there is no 'ground' at the transformers input (I would think): all the primary sees is the output impedance of the CD-player.
Maybe different transformers react different to this?
Erik
Is "E" grounded?
Also, try ungrounding the secondary center tap. It is not necessary, and not necessarilly a good thing. Look at how Cinemag has theirs: http://cinemag.biz/line_input/CMLI-15-15B.pdf
Also, try ungrounding the secondary center tap. It is not necessary, and not necessarilly a good thing. Look at how Cinemag has theirs: http://cinemag.biz/line_input/CMLI-15-15B.pdf
FWIW, I'm not talking about a cap-coupled output in the traditional sense, to remove DC offset from the CDP output. What I'm referring to is if (and I don't know your CDP) your source is fed with a 2 wire power cord, its output is completely isolated, free to float around wherever it wants to. Sure, it may be bonded to the chassis of the CDP, but this chassis would not be earthed if you only have a 2-wire power cord. Further, they may be coupling the negative side of the RCA jack to chassis with just a cheap ceramic cap. Either way, your output is still floating around.
By throwing in some high R values from each lead to ground, you are forcing a reference to earth, eliminating a possible source of noise injection through the capacitance to chassis / neutral / hot, whatever. By definition, a balanced system has equal impedances to ground, regardless of signal balance. Adding some balanced resistance will help swamp out the unequal capacitive reactance I suspect you currently have. I've experienced this before when connecting my waveform generator to my input transformer; it wanted to have a reference.
Anyway, this is just a guess / trial and error thing. Slapping some cheap resistors from your parts bin as recommended doesn't cost anything.
BTW, my experience with iron has been quite mixed, but I still pursue it. When it works, it's beautiful. When it doesn't, it's grounds for depression. Still trying to resolve my current interstage iron issues; making some good progress.
By throwing in some high R values from each lead to ground, you are forcing a reference to earth, eliminating a possible source of noise injection through the capacitance to chassis / neutral / hot, whatever. By definition, a balanced system has equal impedances to ground, regardless of signal balance. Adding some balanced resistance will help swamp out the unequal capacitive reactance I suspect you currently have. I've experienced this before when connecting my waveform generator to my input transformer; it wanted to have a reference.
Anyway, this is just a guess / trial and error thing. Slapping some cheap resistors from your parts bin as recommended doesn't cost anything.
BTW, my experience with iron has been quite mixed, but I still pursue it. When it works, it's beautiful. When it doesn't, it's grounds for depression. Still trying to resolve my current interstage iron issues; making some good progress.
Many thanks to all for the replies!
Hi Dsavitsk
"E" is currently grounded to the earth bus from the amplifier. I also tried the IT's without the center tap grounded, as per your recommendation, but that did not improve matters for the moment: distortion is still terribly high.
Hi Zigzagflux
That is sure worth a thing! I know a lot about the trial and error thing, I spent quite some time trying it different combinations, loads, earthing... I had this gut-feeling that the non-grounded CDP could have an influence on the performance, therefore I cited it in my opening post. Plans are to build an external DAC with balanced output in the near future, so I won't bother with modifying the CD player. The DAC will be grounded, as will the amplifier, so that will probably work!
I just went to the room to test your proposed solution with the resistors, and guess what, turned on the amp and it was playing perfectly with no ground reference at all. I even plugged in the cliplead again to see if it changed the situation, but it kept the same. Also no hum with inputs floating... maybe because its Sunday and she wanted me to listen to her! So I have not tried your proposed mods yet, but I will keep it in mind just in case.
Hi Peter
I measured for DC with my cheap DVM but could not detect any. Still I will keep that in mind as well for future debugging.
Erik
Hi Dsavitsk
"E" is currently grounded to the earth bus from the amplifier. I also tried the IT's without the center tap grounded, as per your recommendation, but that did not improve matters for the moment: distortion is still terribly high.
Hi Zigzagflux
That is sure worth a thing! I know a lot about the trial and error thing, I spent quite some time trying it different combinations, loads, earthing... I had this gut-feeling that the non-grounded CDP could have an influence on the performance, therefore I cited it in my opening post. Plans are to build an external DAC with balanced output in the near future, so I won't bother with modifying the CD player. The DAC will be grounded, as will the amplifier, so that will probably work!
I just went to the room to test your proposed solution with the resistors, and guess what, turned on the amp and it was playing perfectly with no ground reference at all. I even plugged in the cliplead again to see if it changed the situation, but it kept the same. Also no hum with inputs floating... maybe because its Sunday and she wanted me to listen to her! So I have not tried your proposed mods yet, but I will keep it in mind just in case.
Hi Peter
I measured for DC with my cheap DVM but could not detect any. Still I will keep that in mind as well for future debugging.
Erik
My experience with balanced inputs and 2 wire connections (RCA cables) has shown that the chassis of each component must be connected to earth safety ground or you will get distortion and buzz/hum. 2 wire corded sources are very bad this way as they rely on the ground of the interconnect cable to reference the source to earth ground. If you measure the AC voltage (or AC current) between your CD player and earth ground (no interconnects attached, just the line cord plugged in and unit powered up) you may be surprised at the results.
To this end I have added IEC sockets and use 3 wire power cords on all sources and amplifiers in my system.
At this time all of the components except a vintage tuner that is not mine have XLR connectors and use standard XLR cabling. With balanced XLR cabling pin 1 provides chassis bonding while the signal is carried between pin 2 (+) and pin 3 (-). If the components are connected to safety ground the connection on pin 1 can be opened and terminated with ~47 ohms and .01uf for RF grounding.
Jensen recommends that the ground connection on pin 1 be connected to the saftey ground tie point and not to the signal ground. This is to keep ground loop currents between components out of the signal handling stages. This information is available from Jensen Transformers and other pro sound vendors.
See the Tabor schematics on my web page for details on hard ground vs RF ground on pin 1 of the XLR. The pin 1 ground point shown in the schematic should show the saftey ground symbol instead of the signal ground symbol as it does now. More things to update...
Tabor schematic
To this end I have added IEC sockets and use 3 wire power cords on all sources and amplifiers in my system.
At this time all of the components except a vintage tuner that is not mine have XLR connectors and use standard XLR cabling. With balanced XLR cabling pin 1 provides chassis bonding while the signal is carried between pin 2 (+) and pin 3 (-). If the components are connected to safety ground the connection on pin 1 can be opened and terminated with ~47 ohms and .01uf for RF grounding.
Jensen recommends that the ground connection on pin 1 be connected to the saftey ground tie point and not to the signal ground. This is to keep ground loop currents between components out of the signal handling stages. This information is available from Jensen Transformers and other pro sound vendors.
See the Tabor schematics on my web page for details on hard ground vs RF ground on pin 1 of the XLR. The pin 1 ground point shown in the schematic should show the saftey ground symbol instead of the signal ground symbol as it does now. More things to update...
Tabor schematic
Just to expand a little on what Gary said...
Most 2-wire CD players and similar equipment has an EMI filter that contains two caps (the "Y" caps) that connect each side of the line to "ground". In this case "ground" is the case and also probbaly the "ground" side of the RCA jacks.
Since the Y caps form a capacitive voltage divider, then the "ground" of the RCA is at 1/2 the line voltage when referenced to the real ground. It has a fairly high source impedance - a typical Y cap may be ~1nF, which is ~3megs, so it isn't a shock hazzard. But apply this to the primary of a transformer with no reference to real ground, you are asking the transformer to reject ~100V of common mode signal at 50/60Hz! That is a tall order. Interwinding capacitance of just 10's of PF would couple enough noise to be annoying.
You are better off keeping the transformer and connecting the "ground" to a real ground. I haven't found that it matters much if you connect it to safety ground or signal ground... each has some advantages.
Of course the best possible thing to do is to connect the Y caps back to the wall socket ground with a 3-wire cord.
Pete
Most 2-wire CD players and similar equipment has an EMI filter that contains two caps (the "Y" caps) that connect each side of the line to "ground". In this case "ground" is the case and also probbaly the "ground" side of the RCA jacks.
Since the Y caps form a capacitive voltage divider, then the "ground" of the RCA is at 1/2 the line voltage when referenced to the real ground. It has a fairly high source impedance - a typical Y cap may be ~1nF, which is ~3megs, so it isn't a shock hazzard. But apply this to the primary of a transformer with no reference to real ground, you are asking the transformer to reject ~100V of common mode signal at 50/60Hz! That is a tall order. Interwinding capacitance of just 10's of PF would couple enough noise to be annoying.
You are better off keeping the transformer and connecting the "ground" to a real ground. I haven't found that it matters much if you connect it to safety ground or signal ground... each has some advantages.
Of course the best possible thing to do is to connect the Y caps back to the wall socket ground with a 3-wire cord.
Pete
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.