Finally I managed to get the time to write this up. Hope you enjoy it, you can find it here.
300B SE Amplifier write-up
All about electronic valves and hi-fi
Finally I managed to get the time to write this up. Hope you enjoy it, you can find it here.
Great write up Ale.
An enjoyable read.
It must be time to think about using a regulated B+ supply….all supplies really.
Morgan Jones recommends their use
Thank you Tim. I’m not fan of regulated +B for the output stage. Only as you can hear them if they have some element of feedback. Hence I’ve always been inclined to use the Cap Multiplier, which sounds very clean in my view. It’s technically not a regulator, but a stabiliser. On the other hand, my mains seems to be very stable and can only notice a couple of mA as much of drift in my system.
Hi again Ale,
If you send me your address (private email) I’ll ship a pair of d3a to you–will be from the same carton but won’t be matched.
Ale,
Is it correct that you use -6dB poles instead of -3dB poles for the frequency plot of the output transformer?
Well spotted Martin! I used the wrong -3dB market reference. It’s on -6dB you’re right. I will correct the plot later. That explains the slightly better response compared to the oscilloscope bode plot which I shared during the initial tests
Uhh Ale, i see that you changed the high frequency -3dB point. Nice, but what about the low frequency point! Still 3,8Hz?
Martin
Hi Martin,
I like your attention to detail! I missed that one which was obviously impacted in the same way by reading incorrectly the cursor. It’s 6.3Hz! Updated thank you for your thorough inspection.
ALe
Hi Ale,
I wonder behind the reasons of the inferior frequency response. It seems both the LF and the HF cut off points are worse. There are two of my primary suspicions:
-Did you load the transformer during measuring with the equivalent resistive load (4R or 8R?
-The tube’s Rp might be a lot higher, but I doubt that.
Best regards,
Alexander,
Hi Alexander
The LF response is end to end including the additional pole added by the 100nF / 510k pair into the Source Follower gate. That is about at 3Hz so will take down your 5Hz corner a bit. The HF response is closer to reality in my opinion. I used a non-inductive 8R load, however left the +2m cables running from my amp into the speakers to connect the test load. That adds inductance as well as parallel capacitance. I think this should explain the lower response from 80kHz to 72kHz.
Yet it sounds fabulous as is and the response is more than outstanding!
Thanks
Ale
Hi Ale,
Yes, your explanation makes sense. Some cable capacitance might affect the result, if you tested the OPTs using a cable.
Thanks!
BR,
Alexander.
Alexandre, do you calculate low frequency point -3dB with Rp parallel with Ra transformer or not? With Ri // Ra the -3db will be lower.
Martin,
I calculate the LF cut-off with a loaded transformer, which of course, is lower.
But the tube resistance is in series to the primary.
Alexander,
I was wondering what would be the best way. For instance the Tango X-10S measurements: Ra=10k Ohm, 80H/70mA and with a 10k resistor you have to make calculations with Ra 10k // 10k Ri = 5k to get the same low frequency response as Tango dos.
So they calculate with Ri in parallel to the transformer.
and it should be in parallel….( see Tango, Hashimoto )