[gabriel-dan]

Hi, I recently finished putting together a dynalo amp that's been in the pipeline for quite sometime. Upon powering up all the initial voltages checked Ok as per the construction guide in http://www.djgardner.com/headphone/gilmore/
 
The amp is extremely quite even when using low impedance sensitive headphones. I am using the DC offset servo and the output dc offset never goes over about 5mv for both L and R channels.
 
Everything checks OK except, in both channels the sound signature is on the heavy side of bass compared to the high frequencies. It is acting like a low pass filter sort of speak. Please see attached pics of the dynalo frequency response against
 
other headphone amps that I own. In order to get a decent flat sound signature in foobar and itunes, I have use the equalizer and sharply decrease all the frequencies down from about 1Khz and increase all the frequency up from 1Khz.
 
I borrowed an oscilloscope and function generator and was able to see a sine wave decreasing in size as I increased the frequency, also the problem is happening after the input stage at the "drain" pins of 2SK170BL which connects to the base of
 
the Voltage Amplifier Stage transistors Q3. Please refer to schematic http://www.djgardner.com/headphone/gilmore/dynalo/layout/Rev-C/dynalo_rev_c.pdf. I did replace input stage FET 2SK170BL in one of the channels but it made no difference.
 
I have to admit that I did not match any of the FET's or transistors and probably that is my problem, however both channels exhibits exactly the same behavior.
 
Any suggestion on what might the problem be or things to check is greatly appreciated. Thanks
 

 
 
 

 

[jcx]

always check parts, solder joints, wiring - feedback circuits can "work" even with some major components  not doing what was intended
 
check operating points/bias current with no signal - are the ccs about right, do the drain R currents nearly match, add up to the ccs current? (measure the dV across the Rs involved, calculate the currents, sometimes a kOhm or more in series with a digital multimeter is needed to keep meter input Z and noise from interfering with analog measurements, especially inside a working feedback loop)
 
if you AC couple with a Cap, series limiting R to protect the soundcard input you should be able to see if the fet drains move with nice audio signal even if limited to a soundcard - look for no AC sinewave  signal or clipping
 
global frequency response could be messed up if your feedback lead cap is a few orders of magnitude off in value - possible if hi-k ceramic is misread for value

 

[gabriel-dan]

Hi jcx, I checked the operating points using the schematic found here http://www.djgardner.com/headphone/gilmore/dynalo/layout/Rev-C/Dynalo%20-%20Operating%20Points.pdf as reference, I calculated the ccs currents and got 2.13mA and 2.17 which are a bit high for the recommended 2mA.   
 
As you suggested I did a sanity check of all parts values and found that I have installed the wrong part value for the feedback capacitors C27 and C28. I installed 0.1uF instead of 10pF!.  As you pointed out this might be my problem. Thank you very much for your assistance jcx, much appreciated.