Built a transformer based balancing box- significant differences with channels

[PhilipMarlowe]

Is there an acceptable tolerance between transformers? It’s based on the Lundahl 1527 and balances unbalanced direct outs to balanced direct outs of my mixer.
Unfortunately some channels seem to be way more phase accurate than others and transmit transients better. One channel in particular sounds superb while one has less output and less transient energy.

I already checked cabling and soldering of the PcB the trannies are soldered onto, but it doesnt seem to matter, which just leaves the actual differences inside the transformers.

Is this normal and to be expected from standard non-bespoke transformers?

 

[TwentyTrees]

Assuming everything else is equal and the fault follows the transformer when you change test channels, wiring, etc, then I wouldn't expect that with a higher end manufacturer like Lundahl (though of course, mistakes do happen).

I note the 1527 is 1+1:1+1, is it possible you could have wired one side up with incorrect polarity? How much level difference are we talking here, and is the frequency response the same between channels?

Alternatively, could it be that some channels of your mixer are struggling to drive the transformers for whatever reason? Have you tested the balancing box with different gear?

 

[user 37518]

Is there an acceptable tolerance between transformers? It’s based on the Lundahl 1527 and balances unbalanced direct outs to balanced direct outs of my mixer.
Unfortunately some channels seem to be way more phase accurate than others and transmit transients better. One channel in particular sounds superb while one has less output and less transient energy.

I already checked cabling and soldering of the PcB the trannies are soldered onto, but it doesnt seem to matter, which just leaves the actual differences inside the transformers.

Is this normal and to be expected from standard non-bespoke transformers?

How exactly are you measuring phase? And what are the readings you are getting that made you believe that the transformer is not that phase accurate? Perhaps you could share your measurements so we can get a better idea.

 

[PhilipMarlowe]

How exactly are you measuring phase? And what are the readings you are getting that made you believe that the transformer is not that phase accurate? Perhaps you could share your measurements so we can get a better idea.

I didnt measure phase directly since i lack the equipment but i recorded separate channels into my daw from my mpc line out, as well as the direct outs of my mixer.

 

[gyraf]

..we need much more info to be able to answer this..

/Jakob E.

 

[PhilipMarlowe]

ok, i will post more info later. Thanks

 

[PhilipMarlowe]

This is the PCB used which I ordered directly from Per.
I corresponded with Per and this was his answer pertaining to wiring of unbalanced to balanced connections:

"The assumed configuration is 1:1, where sides are either balanced or unbalanced.

For unbalanced signal to balanced XLR connect as follows:

Source hot pcb 2
Source cold pcb 1

pcb 11 XLR 2
pcb 12 XLR 3
pcb 10 XLR 1 (ground)

If XLR 1 does not have a ground reference you can connect pcb 10 to source cold.
- Per Lundahl"

Is this correct?

 

[user 37518]

I didnt measure phase directly since i lack the equipment but i recorded separate channels into my daw from my mpc line out, as well as the direct outs of my mixer.

It is important to measure phase if you claim that the transformers are not phase accurate. I am not saying you can't hear something different, but it might not be what you think it is. Also, if we are going to discuss objective data (like the tolerance of a transformer), we can't use subjectivity (like your ears) to determine if the transformer is in or out of tolerance.
Regarding your recordings, you can do something with them, you can analyze phase by performing an FFT. You can try recording the signal before and after the transformer, and see how the transformer and the original signal vary in phase, you could also get the group delay from the phase measurements and determine how flat the group delay is; constant group delay is always an indicative of linear phase. It is important to know that transformers, no matter if they are Jensen or Lundahl, will alter the audio signal, it is just part of the nature of transformers. Lundahl transformers are very good transformers, but they are still transformers.

I have a thread here where I have promised I will upload DSP scripts for MATLAB/Octave, I haven't been able to do it since someone in my family died, but I will get to work on them ASAP, and you might be able to use those files for these sort of measurements.

 

[CJ]

Are the transformers brand new? Run some high level audio through them for a couple of hours and recheck.

 

[user 37518]

Are the transformers brand new? Run some high level audio through them for a couple of hours and recheck.

This is good advice, I didn't think of that.

 

[capacitor]

I learn something new every day here.

 

[user 37518]

Are the transformers brand new? Run some high level audio through them for a couple of hours and recheck.

I remember reading a Studer manual which said that a good practice to demagnetize transformers is to apply a very low frequency sine wave and start raising the level from zero, leave it at a high level for a while, and then reduce it gradually to zero again.

 

[HP200CD]

I remember reading a Studer manual which said that a good practice to demagnetize transformers is to apply a very low frequency sine wave and start raising the level from zero, leave it at a high level for a while, and then reduce it gradually to zero again.

I would do that several times, slowly up and down, from zero to a high near the max input signal of the transformer and back to zero, with each subsequent high less than the previous high, until the high is not very high at all. Kind of how we used to demagnetize tape heads.

 

[CJ]

All the Lundahl,s I have seen have had coils that look like a swiss watch. Very precise layers from what has to be computer controlled winding machines.

I remember asking Per about the benefits of a random wound coil while attending an AES show. He thought for a while and then answered that their might be some sonic qualities but that every transformer would be a bit different. So he is really into quality and consistency. This leaves me to believe that the core is the only variable in this equation of sonic differences.

You could measure leakage inductance and capacitance of those transformers. Short the secondary and hook up an inductance meter to the primary. If no L meter simply apply a 1 kc signal to the primary and measure the current draw with the secondary shorted. The currents should all be the same. Use a cap checker to measure primary to secondary capacitance. Be sure to pick the same set of pins.

 

[kbx]

I assume you switched connections to follow the problem, but did you switch the transformers in the actual boards to 100% eliminate all other variables? Apologies if I misread and you already did this.

 

[kidb]

Fascinating, good advice so far so I’ve got nothing to add. Except a tangent! I purchased a lot of marinair 31267 transformers for stereo application, measured them with an analyzer and to my surprise they were all less than a thenth of a dB off in gain, except one which was a half dB off from the rest, my only experience with measuring lundahls is their xlr barrel 1:1 which have about a 1dB tolerance between different barrels.

 

[Walrus]

That Mogami cable has has a conductive carbon sheaf on the centre conductor.
Make sure you strip it back so it doesn't touch the terminal of the jack or pcb, or it will load the signal slightly.

 

[PhilipMarlowe]

That Mogami cable has has a conductive carbon sheaf on the centre conductor.
Make sure you strip it back so it doesn't touch the terminal of the jack or pcb, or it will load the signal slightly.

Sorry, what do you mean by carbon sheaf?

Are the transformers brand new? Run some high level audio through them for a couple of hours and recheck.

They were desoldered by yours truly from a PCB board. I will try to run some low frequency audio through them to see if it helps any. Weirdly enough it's just channel 1 that sounds less phase coherent, while channel 2 sounds significantly more transient heavy than the other two, which sound very similar.

I assume you switched connections to follow the problem, but did you switch the transformers in the actual boards to 100% eliminate all other variables? Apologies if I misread and you already did this.

I cannot desolder these pins from the boards easily without risking breaking both transformer and PCB board so I doubt I will be doing that any time soon. Besides: I won't be able to get a refund since these were bought used. I will just have to deal with it. Regardless, this seems curious to me as it must be the transformers that are different somehow...

 

[Walrus]

If you look at this site, you will see how the cable is constructed High impedance guitar cable (2524) – HHB

From Mogami site.....

Most musical instrument sound pick-ups, for example those in electric guitars, are comprised of high impedance circuits driven mainly by voltage, with very little electrical current flow. That is why handling noise (microphonics) can be a problem for guitar cables. Microphonic noise is caused by the minute voltage generated when a cable is flexed, stepped on, etc. Guitar cables must be designed to prevent this, so a conductive PVC layer is placed under the shield conductor to drain away this voltage. Note: This conductive layer must be stripped back when wiring, or a partial short will result.

 

[PhilipMarlowe]

If you look at this site, you will see how the cable is constructed High impedance guitar cable (2524) – HHB

From Mogami site.....

Most musical instrument sound pick-ups, for example those in electric guitars, are comprised of high impedance circuits driven mainly by voltage, with very little electrical current flow. That is why handling noise (microphonics) can be a problem for guitar cables. Microphonic noise is caused by the minute voltage generated when a cable is flexed, stepped on, etc. Guitar cables must be designed to prevent this, so a conductive PVC layer is placed under the shield conductor to drain away this voltage. Note: This conductive layer must be stripped back when wiring, or a partial short will result.

Oh I see. Yea of course i stripped that back. Thanks