I shorted one of the speaker outputs out by mistake. This blow one of the fuses (one for each channel) on the HT to the transformer and did not cause any other damage. Glad I did this.
That took a bit of finding. OK the ECC88 are on the driver board and are driving quite a long length of wire to the 6550's. ECC88 are VHF and did not want any chance of building an oscillator from the cathode follower. Only a precaution.
53dB at 1KHz from simulation, drops when the dominate pole comes in at around 8KHz.
Last edited:
From posting #5:
For square wave stability test its important the test waveform is very square, out to many MHz, so that the slew rate is very high and forces the amp in to slew-rate limiting (and thus testing its ability to recover from this situation gracefully)
Thus posting #20 shows the real issues that #5 didn't.
That's not a square wave - that's the output of a sound card band-limited to 20kHz or so.
For square wave stability test its important the test waveform is very square, out to many MHz, so that the slew rate is very high and forces the amp in to slew-rate limiting (and thus testing its ability to recover from this situation gracefully)
Thus posting #20 shows the real issues that #5 didn't.
Yep lesson learnt. Don't use a digital source for this test. The dominate pole is between the gates of the JFETs. Placing it in the usual place betweem the plates often causes slew rate/power limiting in the driver which gives raise to the classic valve fuzz. I hate this. My old K4040 was really bad here. A 10KHz sine wave came out as a triangle.
Last edited:
Hi baudouin0,
I would like to know what are your impressions on this amp after a month listening to it.
Thanks.
I would like to know what are your impressions on this amp after a month listening to it.
Thanks.
![20210318_175237[207].jpg](https://www.diyaudio.com/community/data/attachments/850/850391-a996f487b50cd7aa0a9a1aab4e6fdafb.jpg?hash=qZb0h7UM16 "20210318_175237[207].jpg")
The answer is very neutral, transparent - you would not know you were listening to a valve amp - just works no fuzz or drama.
Thanks, mine is very transparent and not very tubish too. I have to say I prefer the warmth of my BHEL84.
I will try other options for the phase splitter (triode bottom + BJT top), together with better output transformers.
I will try other options for the phase splitter (triode bottom + BJT top), together with better output transformers.
Yes in simulation. The 2sk170 was as the only one to have enough gain for the GNFB and reasonable input sensitivity. I was looking for a 'blameless' amp. With less GNFB many others would work.
Thanks, I'm indeed planning to use other njfets as well, considering I don't use gnfb by now, so I don't need too much gain there. U440 wasn't bad at all, and already matched, but I tried it in another configuration and I need to try it again.
I would like a copy of the .asc files please. You can pm them to me if you don't want to post them. Thanks.
Thanks. Got it.
Not sure what V3 is, but why is C15 so large and no bleed resistor? And what does D7 do?
Not sure what V3 is, but why is C15 so large and no bleed resistor? And what does D7 do?
C15 is large its 4x470uF for both channels the ltspice is illustration only the PSU board does have bleed and timed HT startup (htdelay.asc). D7 is again to test on simulation that there's no switch off thump with this design. A big reservoir is needed to get the 100Hz intermodulation down when 1KHz tone is applied to -50dB below. They are not that expensive (£5 each). You can look at the size of the HT cap if you replace D7 with two diodes and make V3 two phase AC 50HZ sources to mimic the transformer.
Last edited:
Hi baudouin0, I'm looking at your schematic to dimension the PSU as well.
May I ask the reasons of your preference for the 6550 vs the KT88 (sonic, reliability, price?), and which brand you prefer?
Thank you.
May I ask the reasons of your preference for the 6550 vs the KT88 (sonic, reliability, price?), and which brand you prefer?
Thank you.
The supply I don't have an actual schematic for. Its based around a 372LX transformer from Hammond. The HT is half bridge rectified to feed the 4x470uF 450V caps. There is a big 2k2 resistor in series to soft start which is shorted out by a 6V relay after about 10s. There is a circuit for this running off the heater supply but you could use a 555 timer. The 372LX generates the negative supply off a half wave rectifier 1000uF@100V cap through a 220R resistor and to a 470uF@100V cap. All these caps are long life high temperature.