I looked on Mouser & got this. So I quickly lurched to a full stop, gotta be newer & better out there!
I've seen old obsolete solid-state parts soar in price, usually because someone building guitar stomp-boxes thinks that germanium 1N34A diodes, or 2N5457 JFETs, or some similar component once used by Jimi Hendrix or David Gilmour, is absolutely vital.
To be fair, I have not seen these soar into the stratospherically-stupid price range of certain pre-WWII directly-heated triodes.
-Gnobuddy
It really depends on the dynamic impedance of the LED chosen, but in many cases yes. There are both upsides and downsides, they're much better behaved than 60 ohm resistors and 470uF -1000uF capacitors (even shunted) at low audio levels in the front end of phono pre-amplifiers using high transconductance triodes for example, the down side is no DC cathode feedback so the operating point can be quite unpredictable unless you use a CCS load..
It does not act as a bypassed resistor. A cathode resistor stabilizes the operating point, bypassed or not.
The LED instead operates the tube at a fixed grid bias. Very different.
LED biasing is significantly noisier than a bias resistor (4dB in one instance I measured). However, you can bypass the LED with a capacitor to eliminate this.
Last edited:
Thank you. Exactly what I found on my ltspice sim, but since i am not experienced I though i was wrong. When I bypass the led with a cap the noise is the same as the resistor (on ltspice).
At first i was thinking it is because of spice perfect components, so thank you for confirming it.
Rayma pointed out the led operate the grid at a fixed point, does it mean less distortion vs the resistor, I mean less bias shift, does it mean better ac loadline ? (I am talking about led vs resistor both bypassed)
At first i was thinking it is because of spice perfect components, so thank you for confirming it.
Rayma pointed out the led operate the grid at a fixed point, does it mean less distortion vs the resistor, I mean less bias shift, does it mean better ac loadline ? (I am talking about led vs resistor both bypassed)
Last edited:
LED Bias with no other impedance in the cathode is fixed bias. As the tube ages, both operating point
& load line will move. Depending on whether the Diode used is IR, Red, Yellow, Green or Silicon the bias
point arrived at may or may not be in a good place. Also depends on the tube used. And aging will move all that.
So it is a crap shoot whether you get what you want or not.
What are you trying to achieve?
& load line will move. Depending on whether the Diode used is IR, Red, Yellow, Green or Silicon the bias
point arrived at may or may not be in a good place. Also depends on the tube used. And aging will move all that.
So it is a crap shoot whether you get what you want or not.
What are you trying to achieve?
Fixed bias (LED) usually results in more distortion, lower output impedance, and more voltage gain than
with unbypassed cathode resistor bias. Bypass the resistor, and they will be more similar in those respects.
If a bypass capacitor is used on a cathode bias resistor, there will be LF phase shift, and a reduced LF output level
below where the capacitor becomes ineffective. Generally, adding a nonlinear element to a linear circuit
is not the best idea, but bypassing will mitigate that. However, the variability of the quiescent point with fixed bias
is a real concern, both with tube aging and different tubes.
with unbypassed cathode resistor bias. Bypass the resistor, and they will be more similar in those respects.
If a bypass capacitor is used on a cathode bias resistor, there will be LF phase shift, and a reduced LF output level
below where the capacitor becomes ineffective. Generally, adding a nonlinear element to a linear circuit
is not the best idea, but bypassing will mitigate that. However, the variability of the quiescent point with fixed bias
is a real concern, both with tube aging and different tubes.
I just measured a D3a using the "Tavish" protocol. 270R/2200u cathode resistor/bypass cap was 10.1nv/RtHz, Green LED (Vf=1.99V) was 17.2nV.
https://tavishdesign.com/pages/downloads
Scott is on DIYAUDIO from time to time. His phono pre was very favorably commented upon in Stereophile several years back.
Tube is Siemens D3a, current was 7.2mA. It's the quietest tube in my collection.
It might sound horrible also because it's a TL431, with known weird behavior and a tendency to hiss like a snake. Re dynamic impedance for LEDs, I've found that GaAsP red LEDs have a Vf of around 1.6V, and a dynamic impedance of as low as ~1.5 ohms if you bias them up to 20mA or so. GaP yellow-green LEDs have a Vf of around 2V, with a dynamic impedance of 13-15 ohms or so at a bias current of 10mA. IR LEds have a Vf of around 1.2V, and a dynamic impedance of around 10 ohms. You might have to be careful using IR LEDs, as there are several formulations around. I think the ones I measured were GaAs devices.
THX for that. I pulled up data sheets for the D3a, gm can be very high so low noise is a sure thing if hookup is OK.
For a triode equivalent noise is usually given as 2,5 / gm. Many years ago I looked at noise in 6BQ7s used in a preamp
for low level magnetic pickups. I did build one from a circuit in Audio Magazine of that time, it was very quiet.
But many people built successful preamps using only 12AX7s & 6SC7s. Other problems seemed to dominate.
12AY7 & EF86 / Z729 also worked well.
The D3a makes sense for the front end of a 100 MHz receiver tho.
Eons ago the head of the Physics Dept of our high school got a bunch of acorn tubes from the surplus depot in Columbus OH. Would love to have some now just to measure.The D3a makes sense for the front end of a 100 MHz receiver tho.
I am interested in looking at the noise performance of other pentodes in triode config.
To refer back to Scott's work, he references the noise to 390 Ohm (~2.5nV/RtHz), for the Noise Figure in a 30Hz to 8kHz BW. I used the 390 Ohm resistor, but measured the noise at 1kHz and calculated back to noise density, accounting for gain and the filter factor of the AP analyzer and RMS'd out the noise of the resistor.
There's some excellent, but specific, data in https://www.diyaudio.com/community/threads/6sq7-el34-se-amp-design.373976/page-4
Here, jhs posts some actual test data of low-ish DC current distortion contribution of LED bias in the cathode of a high-mu driver valve at roughly 20V RMS, with a specific LED. It points towards a strong need to test the particular LED to be used before giving it the all clear.
Notes on noise: the only usual audio concern of amplifier random noise is in a phono equalizer. Here the RIAA weighted noise is almost always dominated by (undefinable/unpredictable) process dependent 1/f noise. The thermal/Johnson/Nyquist noise contribution in a decent MM cartridge + phono amp system is dominated by the noise of the inductive phono cartridge and the 47K load resistor. This doesn't stop us from compulsing, but we need to keep our perspective. MC cartridges are a different thing: Good Luck and Thanks for all the Fish.
Nobody talks about 1/f noise because it's not amenable to talking. Forums are for talking.
All good fortune,
Chris
Here, jhs posts some actual test data of low-ish DC current distortion contribution of LED bias in the cathode of a high-mu driver valve at roughly 20V RMS, with a specific LED. It points towards a strong need to test the particular LED to be used before giving it the all clear.
Notes on noise: the only usual audio concern of amplifier random noise is in a phono equalizer. Here the RIAA weighted noise is almost always dominated by (undefinable/unpredictable) process dependent 1/f noise. The thermal/Johnson/Nyquist noise contribution in a decent MM cartridge + phono amp system is dominated by the noise of the inductive phono cartridge and the 47K load resistor. This doesn't stop us from compulsing, but we need to keep our perspective. MC cartridges are a different thing: Good Luck and Thanks for all the Fish.
Nobody talks about 1/f noise because it's not amenable to talking. Forums are for talking.
All good fortune,
Chris