I put about 3.3K after the transformer so that there is a combined load of around 1.5K seen on the primary. It doesn't seem too critical. A Zobel could also work, but I find the 3.3K flattens the impedance just fine.
How much DC imbalance do you think there is? How much current will flow thru the primary?
Easy enough to measure; put a DCmV meter across the transformer primary. There isn't any. You can connect a transformer directly across the output pins and you're fine.
Uh... yeah. I was hoping you might know something that I don't. From your answer it just looks like wild speculation on your part. If I get a chance, I'll measure it.
Thanks anyway.
Pano, DON'T measure the resistance of a transformer with a multimeter. Indeed, sending DC through the coil (as the multimeter will) may magnetize it and that will increase distortion and decrease induction.
It depends on the type of transformer whether the effect is significant or not, but it's worst for low signal low level transformers like input transformers.
SE output transformers necessarily have DC going through them and they have an air gap to avoid the effects of DC.
jan
Are you only taking one phase from the DAC to say pin 2, and leave the other phase alone?
That should be no problem.
jan
Thanks Jan. It may be too late for some of my transfos. Do you really think the tiny current (microamps) of the DVM will hurt it?
As for DC imbalance, I was going to measure DC flow thru the primary with the DC powered on, but no signal. By measuring a few different chips, I hoped to find an average value.
When I've had Bud Purvine make transformers for me he has said that they can stand a little DC current without trouble. How much, I don't remember - because I wasn't planning to run DC thru them. I'm interested to know how much is too much. And if there is anywhere near that much imbalance on typical DAC chips.
Right, I understand that and agree. I just find it hard to believe that there is enough current DC offset on the pins of the DAC to cause significant current flow. That's the speculation I was talking about. I thought maybe you had some actual numbers or experience with it. Seems not.
It's a good point, but I wonder how much it matter in reality. Some measurements might tell me.
I think this was more of an issue in the days of 20,000 ohm/volt VOMs. Remember, Jan is even older than I.
My experience with the DCX2496 has never shown any offset differences on the DAC outputs, so I agree that would be a non-issue. Although if you look look close enough you'll find some uV probably.
As to measuring the DC resistance, yeah it may not be an issue with modern very low current multimeters. I can't find the data right now but I think that even 1mA DC might be harmfull for low-level input xformers.
jan
that's why the manufacturers specify the distortion of high level low frequency as being much worse than low level higher frequencies.
The short term effect of sustained one way current due to big bass signals distorts the output.
I think it has more to do with the B-H curve of the core material. At low levels or high frequencies you are on the linear portion of the curve, so distortion is at a minimum. When the frequency is decreased to the sub-20Hz region and signal level is sufficiently high, you start to flat top on the curve, leading to distortion.
There is no one-way current of bass signals in music; they have the same area above and below the curve, though they may look much different above and below the zero axis.
You can have pretty good results demagnetizing your transformers by running them at 10 Hz for a period of time with a strong signal (even into saturation), then slowly reduce level and/or increase frequency. Resets the net flux to near the origin of the B-H curve. On a practical level, even minor offset on the B-H curve is still in a linear region on a well-designed xfmr that is not run near its limits, so you end up with good performance.
Agreed.
Agreed
Why? I think because the "flat top" as you put it, is the direct equivalent to adding a DC current to a smaller AC current in that "half" of the waveform. i.e. the flat top whether due to DC saturation or due to extended period of long wavelength music signal sounds the same. The core has become saturated.
If I consider the "half waveform" and make it of long duration, then it is a one way current for the period that the waveform is all to one side of the zero current axis.
Almost agreed. There can be a small area offset from second to second. Most of the time these small offsets cancel out to leave a net near zero offset in the long term.
Last edited:
Improvements to the 'Pilgham Audio' Upgrades?
Hi There,
I have, for some time now, the full compliment of upgrades offered by Pilgham Audio installed in my DCX..
I'm just wondering if I could improve its performance further with the following change:
-Replacing the LM4562's on the 'Jan Didden' i/o board with Burson Discrete op-amp modules? Would they work as a direct drop in replacement?
I know the actual Burson modules are a direct replacment for LM-4562's in general, but is there anything special or different about the Jan Didden active i/o that would prevent them from being a drop in replacment?? (not including physical space for the modules)
Thanks!
CM
Hi There,
I have, for some time now, the full compliment of upgrades offered by Pilgham Audio installed in my DCX..
I'm just wondering if I could improve its performance further with the following change:
-Replacing the LM4562's on the 'Jan Didden' i/o board with Burson Discrete op-amp modules? Would they work as a direct drop in replacement?
I know the actual Burson modules are a direct replacment for LM-4562's in general, but is there anything special or different about the Jan Didden active i/o that would prevent them from being a drop in replacment?? (not including physical space for the modules)
Thanks!
CM