Hi all,
All schematics I could find about LED bias on a typical SET amplifier 6N1P + KT88 (a.k.a. Abdellah) show 2 red LEDs in series, drop in volatge being ~1.66V / LED.
My question is could I use a blue LED from Vishay part number TLHB4400 instead of 2 red LEDs? The spec shows 3.9V@20mA (from the graph ~3.5V@4mA). I know blue LEDs are different internally than red LEDs but for this application will this be ok?
Thanks,
Alex
All schematics I could find about LED bias on a typical SET amplifier 6N1P + KT88 (a.k.a. Abdellah) show 2 red LEDs in series, drop in volatge being ~1.66V / LED.
My question is could I use a blue LED from Vishay part number TLHB4400 instead of 2 red LEDs? The spec shows 3.9V@20mA (from the graph ~3.5V@4mA). I know blue LEDs are different internally than red LEDs but for this application will this be ok?
Thanks,
Alex
The trick is to find a suitable voltage drop with the smallest resistance possible.
Cheap red leds seems to have the lowest resistans/voltage ratio.
Ultra bright leds and blue ones is less good because of higher internal resistance.
Voltagedrop depends on the current so check manufacturers spec sheet to find a suitable led.
Cheap red leds seems to have the lowest resistans/voltage ratio.
Ultra bright leds and blue ones is less good because of higher internal resistance.
Voltagedrop depends on the current so check manufacturers spec sheet to find a suitable led.
What he said!
You can find the dynamic resistance of the LED by looking in the datasheet. The dynamic resistance is the slope of tangent to the forward voltage vs forward current curve at the chosen operating current.
~Tom
You can find the dynamic resistance of the LED by looking in the datasheet. The dynamic resistance is the slope of tangent to the forward voltage vs forward current curve at the chosen operating current.
~Tom
Loading triodes with a LED in the cathode is interesting.
Are there any drawbacks versus using a resitor and a cap?
Are there any drawbacks versus using a resitor and a cap?
It's essentially the same as fixed bias so it comes with all of the pros and cons that entails.
LEDs need a few mA to reach the constant Vf region. LED bias doesn't work with 12ax7, 6sl7 and other similar low cathode current applications.
Cathode resistors adjust better to scatter in tube parameters (the classic "auto bias"), whereas LEDs nail down the bias to a fixed voltage. This can be a problem with certain high transconductance types or sloppy tube tolerances.
If you shunt the LED, it can work for 12ax7 and 6sl7. You just need a really quiet DC source.
I drew some mA's off my regulated DC heater source to power up the cathode bias LED in one amp. It worked quite nicely.
I had measured and selected the LED's in advance, as well as the "high transconductance" valves/tubes that were used.
I drew some mA's off my regulated DC heater source to power up the cathode bias LED in one amp. It worked quite nicely.
I had measured and selected the LED's in advance, as well as the "high transconductance" valves/tubes that were used.
Last edited:
There are low current LED.
I found this LED 5mm RED with a nominal foward current of only 1mA.
Which means it has its voltage already well below 1 mA.
Even more common are the LED with 2mA forward current.
Noramally a 5mm RED has 5mA for turning on well.
Bias LEDs should not be fed with extra current; it increases distortion in the valve. LED colour also does not matter; the dynamic resistance is always negligible compared with the valve's internal cathode resistance. Basically, just pick an LED and use it, simple as that. The 'constant Vf region' is a red herring.
An LED will add noise, however, so if this is a noise-critical situation then you'll want to bypass the LED with a large capacitor (>100uF).
An LED will add noise, however, so if this is a noise-critical situation then you'll want to bypass the LED with a large capacitor (>100uF).
Different color LEDs generally have different voltages for the same current in the LED.
Many Red LEDs are less voltage than some of the other colors at the same current, i.e.
voltage at 3 mA or 10 mA, etc.
If you are afraid of the noise when using LED bias, then you might not want to use an LED plus bypass cap. That is because one of the major reasons for using LED bias was to eliminate the bypass cap of a 'resistor and bypass cap' bias circuit.
Another major reason for using LED bias was to have a specific bias voltage.
You could use a resistor instead. Calculate the necessary resistor:
(R = V/I, where V is the specific bias voltage you want, and I is the current you want).
Then use a current source in the plate load that is set to the current you want.
The tube plate voltage will adjust according to the bias voltage and current that you want.
I have used this circuit with very good success. Gain approaching u, low distortion, and large voltage swing.
You may want to use a bypass cap across the bias resistor (I do).
The load on the plate is the current source impedance in parallel with the Rg of the next stage.
Many Red LEDs are less voltage than some of the other colors at the same current, i.e.
voltage at 3 mA or 10 mA, etc.
If you are afraid of the noise when using LED bias, then you might not want to use an LED plus bypass cap. That is because one of the major reasons for using LED bias was to eliminate the bypass cap of a 'resistor and bypass cap' bias circuit.
Another major reason for using LED bias was to have a specific bias voltage.
You could use a resistor instead. Calculate the necessary resistor:
(R = V/I, where V is the specific bias voltage you want, and I is the current you want).
Then use a current source in the plate load that is set to the current you want.
The tube plate voltage will adjust according to the bias voltage and current that you want.
I have used this circuit with very good success. Gain approaching u, low distortion, and large voltage swing.
You may want to use a bypass cap across the bias resistor (I do).
The load on the plate is the current source impedance in parallel with the Rg of the next stage.
And it can keep one from turning off the LED on peak excursions of the input signal.
Bypassing the tube and anode resistor with resistor to maintain bias voltage across the LED is a good thing.
Bypassing the tube and anode resistor with resistor to maintain bias voltage across the LED is a good thing.
Design a tube stage with a specific bias voltage, specific quiescent current, and specific plate voltage (from the set of tube curves) so that it will not go too low of a current unless it is at or near clipping.
If you need more dynamic range from that tube stage to drive the next stage, start the first stage design over, or use a different tube and start from scratch.
Clipping (and near clipping) usually does not sound good in Hi Fi, it usually only sounds good on a guitar amp when that is intended.
One advantage of the LED bias is simplicity, extra resistors to keep a minimum current in the LED reduces simplicity.
The advantage of resistor bias plus bypass cap is that you are not dealing with dynamic resistance as in an LED, just the resistance, and the ESR and capacitive reactance of the bypass capacitor.
Just decide which is better for your circuit design.
As Wavebourn says: "Nothing in the universe is perfect"
And as I say: "All generalizations have exceptions"
If you need more dynamic range from that tube stage to drive the next stage, start the first stage design over, or use a different tube and start from scratch.
Clipping (and near clipping) usually does not sound good in Hi Fi, it usually only sounds good on a guitar amp when that is intended.
One advantage of the LED bias is simplicity, extra resistors to keep a minimum current in the LED reduces simplicity.
The advantage of resistor bias plus bypass cap is that you are not dealing with dynamic resistance as in an LED, just the resistance, and the ESR and capacitive reactance of the bypass capacitor.
Just decide which is better for your circuit design.
As Wavebourn says: "Nothing in the universe is perfect"
And as I say: "All generalizations have exceptions"
I am currently using LED Bias on the 6N1P on this DIY Single-Ended (SE) KT88 / 6L6 / EL34 / 6CA7 Tube Amplifier
without a capacitor. I may have to try a capacitor on it, at the time I couldn't find any information about doing it with a capacitor.
invaderzim,
If you want to use a capacitor across a couple of LEDs that are in series, keep in mind what the circuit was:
The 6N1P had a 1k Ohm resistor, plus a bypass capacitor.
Now suppose that you use 2 LEDs in series, and suppose each has a dynamic resistance of 100 Ohms (200 Ohms for the two in series).
1k/200 = 5 You will have to bypass with a capacitor that has 5 times the capacitance.
Please look at my posts #11 and #15 above. I discuss the ins and outs of resistor bias versus LED bias, as well as the issues of a bypass capacitor.
If you want to use a capacitor across a couple of LEDs that are in series, keep in mind what the circuit was:
The 6N1P had a 1k Ohm resistor, plus a bypass capacitor.
Now suppose that you use 2 LEDs in series, and suppose each has a dynamic resistance of 100 Ohms (200 Ohms for the two in series).
1k/200 = 5 You will have to bypass with a capacitor that has 5 times the capacitance.
Please look at my posts #11 and #15 above. I discuss the ins and outs of resistor bias versus LED bias, as well as the issues of a bypass capacitor.
Wavebourn,
I should have said the dynamic resistance might be 10 Ohms or even less.
Depends on the LED and the current.
Ivaderzim,
The 6N1P transconductance is about 4,400 uMhos (cathode impedance is no lower than about 225 Ohms).
If the plate load is 47k Ohms, the additional cathode impedance is 47k/33 = 1,425 Ohms. Therefore using a bypass cap is a moot point, unless you want to get rid of some LED noise, because you can only bypass the LED, the LED is already 'bypassing' the cathode.
But 1000uF capacitive reactance is about 10 Ohms at 20 Hz, and will not get rid of any extremely low frequency noise, since the LED impedance is about 10 Ohms there.
I should have said the dynamic resistance might be 10 Ohms or even less.
Depends on the LED and the current.
Ivaderzim,
The 6N1P transconductance is about 4,400 uMhos (cathode impedance is no lower than about 225 Ohms).
If the plate load is 47k Ohms, the additional cathode impedance is 47k/33 = 1,425 Ohms. Therefore using a bypass cap is a moot point, unless you want to get rid of some LED noise, because you can only bypass the LED, the LED is already 'bypassing' the cathode.
But 1000uF capacitive reactance is about 10 Ohms at 20 Hz, and will not get rid of any extremely low frequency noise, since the LED impedance is about 10 Ohms there.
Last edited:
Because increasing the LED current increases the forward voltage, meaning cooler bias for the valve, which means higher valve distortion, even though the LED impedance is lower. Yes, even with 100 ohm LEDs! See my book for measurements.
The decrease in dynamic resistance is paid for with increasing valve bias, meaning larger and more non-linear internal cathode resistance, which works out worse. The internal resistance of the best valve is more nonlinear than the worst LED, so current boosting ultimately works out as a false economy (though it may only be a very tiny degradation, depending on application).
Last edited:
- Status
- Not open for further replies.