Latest Schematic – updated May 2020
|47k Input||–||Vishay S102|
|13k8 made up of 12k and 1k8||–||Takman Rex 50 carbon film|
|3k18 made up of 3k and 180R||–||Takman Rex 50 carbon film|
|510k||–||Kiwame 2W carbon film|
|10k D3a anode and 3k9 5687 tail||–||Mills MRA12 wirewound|
|0.1u||–||Russian FT-3 teflon|
|100n EQ||–||Russian silver mica 0.4%|
|34n2 EQ made up of 33n and 1n2||–||Silver mica 500V 1% and 5%|
|22u and 150u||–||Black Gate VK|
|220u+220u||–||Black Gate Wkz|
|Grid Stoppers||–||Allen Bradley CC|
|Screen and PSU decouplers 390R, 1k & 47k||–||Kiwame 2W carbon film|
I got enthused again to revisit this after seeing a few build threads on the HiFiHaven Forum. The end to end layout was going to become an issue so decided to rebuild with connectors at the back in the same slide in/out case.
A prejudice of mine was parallel components and so had chosen to live with 33nF in the EQ instead of the calculated value of 34.2 nF. I had altered the values in the EQ to get it within 0.3dB 20 to 20kHz which I thought was good. Reworking the whole EQ was a worthwhile improvement with a much flatter response in the model and slayed a prejudice as well.
I wanted to secure valves for the foreseeable future and sadly, the MV rectifiers had become unobtanium. I had heard good reports about direct heated rectifiers so secured some DW4 and 5R4GY. The DW4 was easy to slot in as it has the same 4V filament. I didn’t think there was much if any difference to the MVs which was a relief.
I increased the current capacity in the HT to 200mA; revised transformers and chokes which would allow me to run things a bit hotter
After all these years, I felt the battery bias should be tested again as I felt it might not be performing as well as it could so tried an LED which was OK, a more relaxed sound to the battery so preferable but then tried the CREE Schottky diode favoured by Ale Moglia. This produced a much hotter bias point which upped the gain a few dB and I liked this very much. Gain is now modelled at about 48dB excluding the stepup transformer.
Also tried the D3a as triode again but preferred the pentode. It has a lot more dynamic jump from the increased gain but is leaner sounding than as a triode. I find that as a triode the sound is more shut in but could see in a hot sounding system, it might be fine; just not for me. Another reason for the pentode is reduced input capacitance. This article by Dave Slagle of Intact Audio is a good read.
I had gone back for a bit to cap coupled as the D3a anode wasn’t giving the 5687 enough voltage to work properly; it being a compromise I had lived with for the direct coupling option. With better PSU capacity I could run the 5687 hotter. The diode bias D3a anode voltage was a bit lower opening up direct coupling again and I liked the result and this time, it wasn’t a compromise.
As the design is fairly settled these days, it was time to revisit the resistor types to improve things from the ordinary metal films and try some of the “made for audio” carbon types. The results of that are on this Audio Talk Forum thread.
Inside bottom and top of slide in chassis