SY said:The suggestion won't bother me a bit as long as they don't go into my preamp.
The Siliconix CCS diodes are very mediocre current sources. Convenient, yes, but poor performance and no cheaper than a discrete bipolar cascode.
SY,
Being a novice in this area, do you object to the noise or regulation characteristics?
I tried the CR series "diodes" a few years ago. I used some old Motorola brand diodes and came to a similar conclusion. I didn't need to make any fancy measurements. They just sounded bad.
After testing a lot of CCS's, I settled on the IXYS IXCP10M45S chip. These work well at currents from 5 to 100 mA. The Supertex DN2540 mosfet makes a decent CCS also.
After testing a lot of CCS's, I settled on the IXYS IXCP10M45S chip. These work well at currents from 5 to 100 mA. The Supertex DN2540 mosfet makes a decent CCS also.
What am I possibly doing wrong if I use an array of red LEDs (1.7V x 8 in parallel) on the cathode of an SV83 and shortly after power up, the array flashes on and off continuously. I'm not applying any signal .
My op. point:
My op. point:
An externally hosted image should be here but it was not working when we last tested it.
SY said:
No, that was my first thought. They stay lit connected through a battery / resistor . I've temporarily disconnected the LED array and replaced it with a cathode resistor, resulting in an approx 2.5 V bias.
I notice that I hear a very high freq. noise with or without speakers connected, so it's probably oscillation (and yes, I do have grid stoppers) !!! With speakers connected to the OPT I can hear music at a low level.
Debugging time, . I plan to first disconnect the feedback loop (this is an RH84 using 12AT7/SV83 ), and then the individual driver and output stages.
Matter for another thread, I guess. Thanks for the help.
I'm prepared to try my own now, the present cathode voltage of the voltage amplifier tube is 3 volts thru a 820 ohm resistor with 80mf cap and the plate is 105 volts passing about 3.6ma. So, 2 red led give 3.2volts bias, can I just remove the resistor and solder in the 2 leds in its place? Or, do I need to adjust the B+ before solder them in? Thank you
Hi Arnold, I'm using that config as the output of a Scott tuber tuner. 5687, a shared pair of green LEDs in series in the cathode (not too worried about separation in this application), Hammond 156C for plate loads. Vp ~ 125 Vdc, Ip ~ 8ma. Into a 10 kohm load good for about 4 volts p-p at just under 0.1% pure second. Should drop dramatically into 50k. Hard to say using FM, especially given the questionable headphone amp in monitoring use, but upstream changes are clearly audible. Gain too high for a pre, still confirms the topology. (Edit: IMO.)
Now I report what I found. The original cathode voltage of 3 volts after replaced with red led it was 1.87volts, but, the sound is much cleaner and dynamic seems bigger, I like it. Now, for the cathode of 2A3, I dream of making similar change, but, with led its a bulky leds lot..... can its place be replaced with tude diode, like the damper diode that can pass the 60 ma current?
Tubesin said:Now I report what I found. The original cathode voltage of 3 volts after replaced with red led it was 1.87volts, but, the sound is much cleaner and dynamic seems bigger, I like it. Now, for the cathode of 2A3, I dream of making similar change, but, with led its a bulky leds lot..... can its place be replaced with tude diode, like the damper diode that can pass the 60 ma current?
Try fixed bias with 5 x 9V transistor radio batteries in series to bias the output tube(s), lift grid resistor(s) and connect -45V here, ground +45V, remove 2A3 cathode resistor(s) and bypass cap(s) and ground that node through a 10 ohm 1/4W resistor so you can measure the quiescent current which should be between 50 - 60mA per tube. (It will also act as a fuse should the bias voltage fail for some reason.) Tweak the power supply voltage as necessary to get down to 250V on the 2A3 plate.
To bias a single 2A3 with leds would require about 60 - 70 leds per tube in series parallel, and it is not as good as pure fixed bias in this case due to the significant dynamic resistance of all the leds in series.
Tube diodes do work, but suffer from more sonic issues than the led string. The one time I tried this it did not sound good and it took a bunch of tube rectifiers in series to get the right voltage.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.