So glad it fixed itself!
I have mixed up the input/output a few times myself.
A new review of this amp just came out today on headfonics.com:
http://headfonics.com/2017/06/the-class-a-nhb-pocket-amp-by-xrk-audio/
What Michael at Headfonics said.
If you're looking to build a system of 'smalls', take a serious look at this DAC:
PupDAC Overview. I just finished mine,
and am using it with this USB decrapifi'er:
DoodleBug Overview
They are both available in kit form here: Beezar Audio
I am very happy with mine, and definitely recommend them as a great match with
the 'X' amp.
If you're looking to build a system of 'smalls', take a serious look at this DAC:
PupDAC Overview. I just finished mine,
and am using it with this USB decrapifi'er:
DoodleBug Overview
They are both available in kit form here: Beezar Audio
I am very happy with mine, and definitely recommend them as a great match with
the 'X' amp.
Great review X! After reading this I will have to upgrade to the Uber version.
I agree with Michael : "I hope XRK971 decides to build his own legitimate business and takes his knowledge of circuitry into a much bigger platform for sale in the community"
No. The input to the JFET gate requires a DC bias provided by R4. It's about 1.5v to 1.8v and allows the JFET to control the bias setpoint of the MOSFET. The AC coupling allows the DC bias to work.
You have to AC couple the output anyhow. If you notice the review by Michael, there was no shortage of bass, treble, or sound stage/imaging despite being cap coupled.
Congrats on the excellent Headfonics review, X! He compared the amp favorably to portable and desktop rigs costing many times the price! And he definitely nailed it when he highlighted the impact and tonality factors. You might consider also getting one of your amps into the hands of Tyll Hertsens or Zeos eventually...if you're ready for a deluge of orders, that is.
Diyrookie, I'm so glad your problem turned out to be an easy fix and now you get to enjoy your amp! I'll admit I've made the same mistake before too.
Diyrookie, I'm so glad your problem turned out to be an easy fix and now you get to enjoy your amp! I'll admit I've made the same mistake before too.
Thank you, but I did not really fix anything. Would still like to know what the cause was - it might occur in future again.
Could it be heat? I have read about thermal shutdown somewhere, but how would a mosfet behave if it gets too hot? I was running on 19,3V at the time (when the distortion in one channel started) and as I said - the mosfets could only be touched for 2/3 seconds max. I have since tuned it down to 18V. Considering to glue small heatsinks to the mosfets for my uber version.
That might be heat. Those MOSFETs are rated at 3w max dissipation but I would guess the little on-board copper plane is good for maybe 1w max. Running at 19v and say 75mA is 1.42w ea (mosfet and resistors) so I would not be surprised it it was too hot. Those resistors are rated 1/8th watt x 4 is 0.5w. Assuming half the dissipation is from the resistors, they would have been running too hot. The resistance changes when it is too hot - not sure if it goes up or down for metal thin films - I am guess up. So it may have causes a very low bias situation because R goes real high?
Here is my latest unit designed to drive a 13.8ohm load with 115dB sensitivity (Andromeda's). With two 1000uF Oscons and a 10uF Silmic II on output and 10uF Silmic and 2.2uF Wima on input. With LTSpice sim I get 0.05% THD at 106dB SPL with 96% being H2 and balance just H3 and. Ofhing else. Frequency response is flat to 15Hz at -0.5dB. This unit is going out to Poland tomorrow after a 4hr burn in. The sound is perhaps the deepest bass impact Inhave heard on my 55ohm OB-1's. Actual test at 270ohm load and 700mV (about same drive level as 106dB on Andromeda's) measured only 0.02% THD and all H2, H3 almost non existent (-90dB).
Attachments
Last edited:
Awesome X. My current cans are also 55ohm, so I assume this cap configuration is pretty much ideal for low impedance headphones? I have been reading through the thread just yesterday with the view to order caps for some rolling. Thanks, now I know what to get, unless Stellar has a different opinion.
Here are the cap part numbers from Mouser. You need two per channel or 4 total.
16SEPF1000M:
http://www.mouser.com/ProductDetail/Panasonic/16SEPF1000M/
The 10uF 35v Silmics (RFS-35V100ME35) are here:
http://www.mouser.com/ProductDetail/Elna/RFS-35V100ME35/
16SEPF1000M:
http://www.mouser.com/ProductDetail/Panasonic/16SEPF1000M/
The 10uF 35v Silmics (RFS-35V100ME35) are here:
http://www.mouser.com/ProductDetail/Elna/RFS-35V100ME35/
I've heard great things about the F4, if you have the means to drive them adequately.
Now I think I might have to build an F4, and another headamp to use as a pre-amp.
Might want to simulate those conditions before giving an official nod. Quick sims I did suggest distortion raises dramatically. Not to mention that the F4 is a 25W amp in its stock configuration, but the headphone amp would only allow ~7W of output because of voltage limitations in the headphone amp. Of course, one could scale up the power supply voltage, if one is careful with MOSFET dissipation...
You are right - It would take a total redesign to do properly and require at least a CCS and another stage of transistor. Maybe a JFET input followed by a CFP slave stage and a bias controlled CCS on the CFP to reduce distortion. Rail voltage would need to go up to allow 20v peak to peak output into, say a 4k7ohm load. So BF862 may be out due to voltage limitations unless cascoded.
Might be a fun project as a specific high gain (selectable 12dB and 20dB) preamp.