Rockbox mail archiveSubject: input pad tolerances (was: line in clipping when recording)
input pad tolerances (was: line in clipping when recording)
From: spidermagnet <rokbokser_at_imtarget.cotse.net>
Date: Thu, 15 May 2003 18:36:56 -0400 (EDT)
Jerry explained pads far more technically than I did, and he has a good
design for a roll-your-own pad.
I would recommend against Radio Shack resistors, though. If you are
recording stuff you consider high-value, you can get very significant (=
very audible) improvement by spending (literally) just a few dollars more
on quality resistors. Also, the quality of cable you use--and especially
how good your soldering is--can make a difference, particularly if thie
thing is going to see field use.
Why resistor choice is important:
- Resistors do not impede all frequencies evenly; they are not perfectly
"flat". The result is that some frequencies are more attenuated than
others, changing the color of the signal. The carbon and metal-oxide
resisitors sold by Radio Shack are particularly bad. If you use them,
expect to lose some high end. You can get excellent metal-film resistors
for a buck or two each.
- Resistors vary widely in quality. Standard Radio Shack carbon or
metal-oxide resistors have a 5% tolerance; for a 1K resistor, that means
that you could end up with 950 ohms on one channel, and 1050 on the
other. For audio applications, consider looking for 1% or better yet 0.1%
tolerances. It might cost you .50 USD more per resistor.
- I got _my_ initial info on them from this article on resistor fidelity:
- Holsworthy precision metal-film are pretty good. I know mouser.com has
them; or try a local electronics supply place. 1-2 USD if you buy just a
- Vishay Sfernice Bulk Metal Foil resistors are reputedly among the best
you can get for audio applications (I haven't used 'em myself). They're
probably more than 5 USD each. www.vishay.com.
REALLY IMPORTANT WHEN YOU PUT IT TOGETHER:
- Resistors are almost all quite sensitive to temperature. Overheating a
resistor while soldering is often the cause of odd circuit behavior (i
know). Solder carefully, so that the resistor does not get too hot.
all good things,
Jerry Van Baren said:
> Disclaimer: I have not done this myself, YMMV.
> You should be able to make a "resistive pad" to reduce the line level
> externally using four resistors in-line with the line input. I would
> suggest buying a stereo interconnect cable (or sacrifice the one you
> already are using), a small project box, a terminal strip (solder or
> screw type, depending on your capabilities), and four resistors (two
> per channel). Everything is available off the shelf at Radio Shack or
> For a 11:1 pad (-21dB) use 1K and 100 ohm (or 10K and 1K) resistors For
> a 110:1 pad (-41dB) use 10K and 100 ohm resistors
> Line Out RockBox line in
> 1K or 10K
> | | 1K or 10K | | |
> | | | |
> | 100 ohm > < 100 ohm |
> | < > |
> | | | |
> The ends are 1/8" stereo connectors (buy a cord and cut it in half).
> The "tip" of the plug is one channel ("<-" above), the first "ring" is
> the second channel ("o" above), and the big ring is ground ("O" above).
> This forms a 11-to-1 or 110-to-1 voltage divider. Note that your
> source must be on the left side and your Rockbox on the right side
> (label your box!).
> You may want to try 10K/1K for higher impedance but it should not be
> needed (line out drives into a 600 ohm load and we are above 1K with
> the 1K/100 configuration). The RockBox input impedance should be very
> high. If my assumptions on impedance are wrong, you will get more
> attenuation than the specified -21dB but, other than that, it should
> work fine. Theoretically, 1K/100 ohms will give less noise (our golden
> ear friend might be able to tell the difference, but I'm pretty sure I
Received on 2003-05-16