I would suggest regular schottky junction diode such as the 1N5822 rather than a regular pn junction device for lower noise but more devices are required due to lower forward voltage drop.
Last edited:
Hi Rod,
That was exactly my thought initially, but I decided to test it and listen to the stage which was revealing. I went back to my SiC tracing and zoomed the low current section. Between 5-6mA area where I'm currently operating at, seems to be quite linear:
However, I agree fully that in filament bias will may be even better due to the current level will push down the dynamic resistance below 1ohm for sure.
Cheers
Ale
Not a problem and good to see you around here. Is a personal decision. I prefer the gyrator load, however chokes (if properly selected) are very good as well. On the flip side if you are looking at low current and high Ra DHTs, the choke isn't the best option and gyrator will work better overall. It sounds best in my opinion on those specific DHTs .Hey Ale, love your site, some really nice things on there that I want to try outRod's here too, have a bunch of his filament/bias boards here ready to go into amps.
Still on the fence about CCS/Gyrators in the anode circuit, have always preffered inductive loads and no solid state in the signal path. Maybe you'll convince me one day
sorry for the threadjack
Ale
Looks good, Ale - and filament bias looks the way ahead. I am sure we will be hearing from those that try it with filament bias soon.
As for other didoes to try, yes the 1N5822 is certainly worth a look. It is low-cost, and given the quantity of diodes Andy was discussing, it is worth a try.
Normal PN diodes are not noisy - in forward conduction mode, which is 100% of the time, in this circuit. Any sample should measure <1µV rms over the audio band, with 250mA flowing. They are worth investigating - even the standard-recovery parts - to reduce the number of parts.
Example: the ultra-cheap 1N4001 gives 0.8V forward voltage at 250mA (the 01A filament current) and <0.5Ω at that current, looking at the Fairchild curves. Much smaller than the others, too.
As for other didoes to try, yes the 1N5822 is certainly worth a look. It is low-cost, and given the quantity of diodes Andy was discussing, it is worth a try.
Normal PN diodes are not noisy - in forward conduction mode, which is 100% of the time, in this circuit. Any sample should measure <1µV rms over the audio band, with 250mA flowing. They are worth investigating - even the standard-recovery parts - to reduce the number of parts.
Example: the ultra-cheap 1N4001 gives 0.8V forward voltage at 250mA (the 01A filament current) and <0.5Ω at that current, looking at the Fairchild curves. Much smaller than the others, too.
Forward voltage drop will also increase with more current, the number of diodes may need to be reduced (lower dynamic resistance bonus) to maintain bias.
There is also an option to use a choke as Rmu on the gyrator.
'm glad you posted this, the 10Y was going to be one of my tube variants to try.
I think I'll stick with the 01A
Here's a snapshot of current/voltage using Ale's model on 6 diodes
Attachments
Last edited:
We only heard SIC diode bias in the 01A. Doesn't mean that there won't be equivalent benefits for all the other DHTs. I'm rebuilding a 2P29L preamp at the moment to see what that gives.
There is something that my ignorant person doesn't understand:
1-Filament of 01A current 250mA voltage 5V
2-If used filament bias resistor as per Moglia schematic 4V across 20 ohms = 200mA (starved) that's OK
3-If substitute the filament bias resistor for diode bias we can get the voltage needed but how can reach the 200mA of current for filament current?
4-Or the 200mA are supplied by Rod Coleman reg. and the string of diodes it's only to reach the voltage grid bias + current plate?
TIA
Felipe
For a 26 preamp you set up the Rod Coleman regs normally, not in filament bias. This may mean changing one of the resistors (was R8 on my older versions of Rod's regs). So you need the usual filament current and voltage from the reg - 1.5v at 1A or a bit less if you're starving it. You then set up the string of diodes to give you the required bias voltage across it, e.g. 10v.
We're also talking about using the diodes in filament bias, which is another way of doing it and may require heatsinks. But that's another story at the moment.
We're also talking about using the diodes in filament bias, which is another way of doing it and may require heatsinks. But that's another story at the moment.
As I said, if your 26 is in filament bias right now you need to change the supply voltage and one of the resistors in Rod's regs. If your 26 is on normal bias then as you say, you just change the resistor for the diode string.
any ideas why the diodes are improving listening pleasure?
Experimenting is the word! I've been trying for hours to get my 2P29L version with SIC bias to work. Since these are plastic bodied devices, I started off with the "good idea" of putting a long M3 screw through the holes in the 12 diodes and keeping them apart with M4 nuts. Please explain this to me - can these devices inductively couple or what? Anyway, voltages were all over the place.
I then took a length of 28 way solder tag strip and soldered the 12 diodes onto that. This immediately gave me the 10v bias I was after.
WTF.... Ideas?
Last edited:
Last edited:
I was connecting the diodes correctly when they were all held together by the long M3 screw. I can only guess that there was something inductive going on.
Anyway, all is working now and I have my 2P29L back sounding nice. The SIC diode bias gives a little more clarity. I think it's a worthwhile mod.
Anyway, all is working now and I have my 2P29L back sounding nice. The SIC diode bias gives a little more clarity. I think it's a worthwhile mod.
These TO220 devices have a plastic case so no dead shorts. I also used some CSD01060A which does have a metal case, and for that I used short nylon spacers which worked fine. Something about this long M3 screw going through all the devices. I just realised I didn't ground the screw - it was floating. Maybe that would fix it. But something "screwy" going on here. When the devices are mounted on tag strips they work fine.
Interesting. I wonder what range of voltage fluctuations you got with spaced stacking on the M3 screw? Good the issue is fixed now. I imagine ideally you would have wanted all the diodes to be the same temperature and running cool for best voltage stability. So you probably wouldnt have kept them stacked anyway as the inner diodes would theoretical have warmed up more with infrared radiation from the adjacent devices. Practically do the diodes warm up at all with such tiny loads?
More importantly with the SICs on both the 01A and 2P29L, which valve do you like more? Does the venerable thoriated 01A still have the magic. Thanks again.
More importantly with the SICs on both the 01A and 2P29L, which valve do you like more? Does the venerable thoriated 01A still have the magic. Thanks again.