Cool Stuff
Hi-Fi Tweaks
e-mail the grotto
Grotto forum    
Rock Grotto some cool links bed breakfast borders scottish Borders bed and breakfast bed and breakfast st boswells bed and breakfast st boswells bed and breakfast scottish borders best of the borders
WNA MKll Tweaks
Here are a few tweaks you can try, they are very cheap to implement (pennies as opposed to pounds) but bring about major improvements to the sound.. have fun!

___The feedback distributor___

This simply involves fitting a 330K resistor across pins 2 and 6 of the LM6171... improves the musicality of the bass and adds a touch of welcome warmth and smoothness to the sound whilst retaining all the good aspects of the LM6171 sound.

Dr. White Explains: Adding the 330k resistor reduces the gain of the LM6171 from open loop to 330. This increases the bandwidth of the LM6171 gain block so that the global negative feedback loop is effective out to higher frequencies. The high frequency distortion is thereby reduced. This is a standard nested feedback arrangement which I have been using in my mos125 power amplifiers for years. In the latter case there are three rather than two nested feedback loops. I suspect that the reduction in transient intermodulation distortion resulting from this arrangement is more important than the reduction in high frequency harmonic distortion. I was interested to hear that it seems to be generally agreed this improves the bass definition as well as leading to a smoother overall presentation.


___The Zobel Network___

Another cost effective tweak which keeps the amp stable in the absence of an output resistor. Simply connect a 10R resistor and a 100nF capacitor in series and connect one end to the amp side of the output capacitor (not the output side) and connect the other end (of the resistor / capacitor in series) to ground... the zero volt pad is a good place to ground the Zobel network


Dr White Explains: The WNA headamp is very fast ( power bandwidth is several hundred kHz ) and has no compensation capacitors; deliberately so. The resistor, or wire ended lamp, on the output serves two purposes; short circuit protection and isolating the headamp from the load. The amp doesn't mind reasonable duration short circuits like those caused by hot plugging or unplugging the headphones so the resistor/bulb isn't really needed here. The resistor/bulb increases the output impedance of the amp, which is bad, particularly for good bass response where high current is needed, but in the absence of any other measures is needed for stability. This is where the zobel network comes in, it will keep the amp stable in the absence of an output resistor. It works by presenting the amp with a low impedance load at very high frequencies, and this tends to kill any tendency towards oscillation. Interestingly enough you rarely see power amplifiers without zobel networks but the converse seems to be true with headphone amplifiers.


___Increase the quiescent Current___

Increasing the quiescent current also increases the bass slam and gives the amp a lot more grunt.   To increase the quiescent current by 25% you need to replace R5 with 160R and R6 with 100R..... the transistors "must" be fitted with small heatsinks as the increase in quiescent current ups their core temperature from 60C to 80C... this is another very cost effective tweak (a few resistors and heatsinks) which seems to provide more balls to the amps driving ability, well worth a try.


Dr. White Explains: Increasing the quiescent current means that the output stage doesn't switch from class A into class AB operation until later in the day when it is pushed very hard by low frequency, high voltage swings into low impedance headphones. This is why you hear the improvement in bass "grunt".


___Remove the Incandescent output lamps___

Replace the Incandescent lamps (R11) with zero ohm links....... this improves the bass response.... read Dr. White's explanation of the Zobel Network to understand why.


___Remove the Output Capacitors___

I personally like the idea of output caps as DC offset is zero with them in place, anyone who is really opposed to using output caps might like to try the following: Connect the track of a 100k trimmer potentiometer between 0V and -12V then connect a 10M resistor between the slider of the trimpot and pin 3 of the LM6171. Adjust the trimpot for minimum offset, run for 30mins or so then readjust offset to mimimum. The original suggestion of connecting a 10M resistor between -12V and pin3 of the LM6171 usually overcompensated for the offset which would typically go from +200mV with no 10M resistor to -30mV with the 10M resistor. Even with the new mod the offset will drift a little with temperature but will still stay low enough for low impedance headphones to be used safely. Both David White and myself prefer using output capacitors as they offer the ultimate in dc output offset protection, and being non polar caps there will be minimum insertion loss (if any)


___10K input resistors___

Lowering the input impedance to 10k will give ( theoretically ) lower noise and will also reduce the output offset with the LM6171, but not enough so you can dispense with the output caps. The component values that Dr. White uses for his designs are a compromise which will work satisfactorily with the widest range of attached equipment. He generally uses 47k input impedance because that'll pretty much work with anything. For his personal stuff he always uses 10k. That's the beauty of DIY you can tweak your gear to suit yourself.


____Use a fast opamp in the rail Splitter____

Using a fast, high gain opamp in the rail splitter will make it stiffer ( ie: lower output impedance ) and more responsive to musical transients. I would guess this wouldn't affect the sound very much when using "ordinary" audio opamps but will certainly pay dividends with an opamp like the LM6171.



Great Scottish Borders Bed and Breakfast e-mail