As I promised, I took the HP amp back to the workshop for a bit of abuse on the bench. Here are some interesting measurements which correlate with some of my listening impressions so far.
I added a set of 300Ω dummy loads at the output to simulate the HPs. Firstly I tested the optimal circuit which has the Sowter 8665 output transformer in Parafeed mode (4:1) to drive the load. You can get a very clean and flat response across the audio band from 3Hz up to +90kHz. There is a minor hump to be tuned due to the RLC circuit formed by the parafeed circuit. I used a 4.7μF Mundorf EVO oil cap. It’s pretty flat to me, so will leave it there.
The performance of the circuit is amazing. Distortion is very low, up to 0.07% at 100mW (which is VL=5.47Vrms):
Harmonics from H3 onwards are very low level and distortion is primarily driven by H2. This measures as good as it sounds. Great stage.
Well, let’s look at the stage with the 100μF Mundorf EVO oil cap instead. This is a completely different beast. We are asking the hybrid mu-follower to drive a 300Ω which is pretty tough. The stage has low output impedance and current drive. The maximum current drive for 100mW will be around 18mA. Anyhow, the load will be steep and a tough one for this stage without the output transformer. Although the sound is very nice at low level, I could easily prefer the Sowter circuit during the listening tests. Lets see why.
As you can expect from the hybrid mu-follower topology, the frequency response is very good:
However, distortion is significantly higher: THD is 1.7% for 100mW in 300Ω (5.47Vrms). You should expect 0.5% or less at normal listening levels though.
Hi Ale,
Thanks for the great work., it’s a pleasure to read. A question regarding output impedance and matching with 300\600 ohms headphones. These headphones show a impedance peak around 100hz that is around twice the nominal Impedance. So, i think, if amp’s output impedance isn’t really low this results in augmented current output in the impedance peak. When you listen bass is probably too heavy. In conclusion and if i’m right: could you measure the frequency response on the real headphone ( not 300ohm resistor).
Thank you in advance
Marco
Of course i’m referring to the 100uF output configuration; nevertheless it’d be of interest to compare the frequency sesponse on real load for the two configurations.