[luvdunhill]

What's an acceptable amount of DC offset to have during turn-on / turn-off of a headphone amp?

Here's what I have:

This referenced to ground.

Which brings up a few questions. In a bridged amp with four identical channels (i.e. "balanced"), is there any way to predict the polarity of DC offset at turn on? I mean is it possible I would see double the above offset, or perhaps zero offset (i.e. cancellation)?

 

[johnwmclean]

http://www.head-fi.org/forums/f6/saf...offset-308295/

This may help, there’s a discussion at the back end of the thread regarding turn-on dc off-set.

 

[tomb]

Great link, but I'm not sure constant DC offset is applicable for this question. For most of the relay-delay circuits around here, it's not unusual to see offset in whole numbers of volts - like 4.0VDC for a split second. It depends on whether the output is loaded or not. If unloaded, you'll see something like that 4VDC. If loaded with a low-impedance phone, the spike is much smaller and quicker. For a high-impedance phone, the spike is slower, but easily absorbed by the higher impedance load.

 

[scompton]

This makes me ask the question, can DC offset be measured with a DMM?

 

[johnwmclean]

Quote:

Originally Posted by tomb /img/forum/go_quote.gif Great link, but I'm not sure constant DC offset is applicable for this question.

There is discussion at the back end of the thread regarding turn on dc offset.

 

[johnwmclean]

Quote:

Originally Posted by scompton /img/forum/go_quote.gif This makes me ask the question, can DC offset be measured with a DMM?

Yes,

Measuring DC offset

 

[scompton]

Thanks

 

[tomb]

Quote:

Originally Posted by johnwmclean /img/forum/go_quote.gif There is discussion at the back end of the thread regarding turn on dc offset.

Sorry - missed that.

EDIT: Turn on offset is often addressed with delay-relays, but turn off offset is not:
http://headwize.com/ubb/showpost.php...63212&fpage=12 (There's a small discussion before and after this post.)

As for measuring it, it depends on how fast the DMM is - a good Fluke may be able to measure this, but a cheaper meter may not. We also found that raising the relay voltage from 12 to 24V (means using a different relay and circuit) increased the speed of turn off so that it can't be measured - with my Fluke at least. However, as you can see from the link above, it's often dependent on the connected load and any load at all can radically reduce the voltage spike that occurs.

Ultimately, in a DC-coupled amp, I think the turn off offset is dependent on the capacitors and how much charge exists in the amp - but this is not the case in an AC-coupled amp. Conceivably, the output capacitors should discharge slowly enough to continue to block the offset, but this is not really what happens - they won't stop the fast transient DC that occurs - only if the relay closes faster (assuming you use one).

 

[wink]

scompton asked:-
This makes me ask the question, can DC offset be measured with a DMM? .

If you are enquiring about the DC offset at turn-on the answer is yes, but you need a digital multimetet with a min/max mode such as found on a Fluke 187 or an Agilent 34401A.
The min/max mode allows the meter to store the highest reading and display it rather than just keeping updating the readings as it normally does.
Hope this helps.
An oscilloscope can do it too, but you need to trigger it and capture the display - can be difficult. It depends on the oscilloscope.

 

[wink]

luvdunhill,
What's an acceptable amount of DC offset to have during turn-on / turn-off of a headphone amp?

Here's what I have:



This referenced to ground..


Compared to what a headphone gets in normal usage, that's no different to some ckicks or pops you get from a scratchy record.

The problem only needs to be addressed when there is a substantial amount of DC voltage present continuously.

The DC voltage can overheat the voice coil and burn it out.
An AC voltage is subject to back EMF (the reverse voltage generated by the change in current through the coil) to limit the current flowing through the voice coil.
Hence the voice coil can tolerate much higher AC voltage than DC voltage.
Low frequency AC voltages can be a problem as they don't generate as much back EMF as high frequency voltages.
Usually there's nothing below 20Hz in normal content you're likely to be listening to, nor will it be there for any appreciable time - so no real problem.

 

[tomb]

Quote:

Originally Posted by wink /img/forum/go_quote.gif <snip>Compared to what a headphone gets in normal usage, that's no different to some ckicks or pops you get from a scratchy record. The problem only needs to be addressed when there is a substantial amount of DC voltage present continuously.</snip>

I can't see the pics (I think there are supposed to be some there.), but agree with the above - good statements.

 

[luvdunhill]

In this case, there is very little capacitance. About four 100uF polymer caps, or 200uF per rail after the PSU. There is 1000uF per rail after the rectifier.

Also, the above graph was unloaded.

What I'd like to do is figure out if I could use one-shot mode on my DSO to capture this... or figure out a way to to determine the frequency and number of cycles. Anyone know how to do this? I've never ventured into the one-shot mode, and not sure how to handle the triggering.

Also, any opinion on what happens in the balanced case? I suppose I could measure it