LED phase splitter, in need of knowledge

[Theashe]

I have a pile of 3mm LEDs left over from another project, and I'm still tweaking my first tube amp almost every day, so I'll also give this a shot and add my two cents to the pile.

Although, it's two cents Canadian, so it's worth half that south of the border.

And we completely phased out pennies, so it rounds down here too...

[Theashe]

Gentlemen, checking in. I replaced the V1a and V1b cathode resistors and bypass caps with blue LEDs in my amp, and took some video for you to decide the differences yourselves. You'll probably come to the same conclusion I did, which is that it makes little to no difference. Maybe a slight mid boost, but you'd really be stretching to hear it.

It's not professionally recorded by any means, but there's two comparison videos: clean tone and distortion tone. Sorry about the sloppy chops - I'm a fingerpicker, and I play acoustic!

Clean: https://www.youtube.com/watch?v=BgiC2Glrj80
Distortion: https://www.youtube.com/watch?v=BDBe8nyzvr8

My amp doesn't normally hum as bad as it did in those videos. It's probably from being wide open with alligator clips dangling out, and my unshielded cable.

EDIT: For the record, I kept the LEDs in the amp to have fewer parts. I'm going to play around with putting more LEDs in the amp on another day.

[Theashe]

Update 2:

The blue LEDs were dropping about 2.6V, so I switched to 1.7V red LEDs to bias the tubes warmer. Not too much of a difference noticed.

I also did an unbypassed gain stage and the phase inverter. Now, I had those stages biased really warm, so putting LEDs in them moved them closer to center-bias. And since those tubes were unbypassed, I expected an increase in gain. Definitely a big difference. More clean volume, definitely more mids. I liked the change to the distortion tone, so I'm keeping it.

[Theashe]

Guys, after experimenting with different LEDs for a few days now, I gotta say I really like them. I tried unbypassed, partially bypassed, and fully bypassed cathodes with Rk/Ck on several of my tubes, and 1.7V red LEDs (center bias), 2.7V blue LEDs (cold bias), one 1N4007 at 0.6V (very hot bias), and two 1N4007s at 1.2V (warm bias).

No matter what change, I always kept coming back to the 1.7V red LEDs. The distortion sounds nicer, the cleans are warmer, and the amp breaks up more smoothly to my ear. I scoped my tubes with the LEDs in them, and despite being roughly center biased, the LEDs were clipping very asymmetrically for large signals (probably from being pushed into cutoff for large positive swings).

It's a keeper.

[JazzGuitarGimp]

Hi Theashe,

Thanks for your research. I am curious how the red LED compares to 1.5K || 5uF in terms of frequency response - does the LED produce a bigger bottom end?

Thanks,
Lou

[rp]

I tried this to on the PI of my 5C8, I didn't hear any difference but I didn't really try that hard. I used a common green tha gave me 2.7V rather than the 2.1V it had before. After about 10 minutes the bass started distorting - wtf, there's always got to be something! I decided to leave well enough alone and put the r/c back. If you like bling the LED light will shine up from around the tube sockets (mine had rubber grommets) and any other little light leaks. I think next I'll need to add water cooling

Thinking if you make your tube knock outs one size up, use bottom mounts with a nut as a spacer you'd have some nice running lights bling - but where to put the spinners?

Here's a little aid if thinking of trying, the cheap common leds are the one's you want I think, they seem to vary best buy a bunch of different colors.

[Theashe]

Hi Lou,

I did two tubes with 1.5k || 4.7u vs red 1.7V LEDs (red because I invariably liked the sound of them better than LEDs that broke down at higher voltages in some of my earlier tests). These tubes were the second and third gain stages in the amp.

The differences between fully bypassed cathodes and LEDs were very subtle. At full distortion, I couldn't really tell the difference. Clean and slightly overdriven, the sound was a little warmer, which may be because of the extremely low resistance of the LED allowing more bass frequencies to pass. Even at 100Hz, a 4.7uF cap has an impedance of 338 Ohms.

An interesting behaviour I noticed with my multimeter. With resistors in the cathode, cranking the volume and hitting a big chord made the cathode voltage rise (as expected), while with LEDs in the cathode, hitting a big chord made the cathode voltage fall. This is just an observation of mine, and definitely take this with a grain of salt, but it makes me think that for larger signals, the LED moves toward warmer biasing. More clean headroom, but asymmetric clipping when overdriven? Might be a win-win, but I might be totally wrong.

[gingertube]

Green LEDs typically 1.8 to 2.0 Volts.
Winner is RED LED with voltage of about 1.4 Volts and lower dynamic resistance than green, use if you want the stage to run a little hotter.
Infra Red LEDS work really well too and the HIFI guys tend to use them.
Cheers,
Ian

[rp]

Some follow up on my attempt to try a red LED my 5C8 clone PI and getting distortion on the bass notes. I figured I did something wrong, cursed, and yanked it right out. After further web reading I went and checked the current on that stage and it's .88mA per triode. Would that have accounted for the PI with the LED borking? From what I garnered on the DIYA forum LEDs work best if the tube is drawing above 5mA minimum, and higher the better as long as you don't fry the LED. Note that most 12AX7s in the Fenders and Marshalls we play with are <2mA.

I also tried it on the input 6SL7, 1.7V on the cathode and the Vp looked nice but I had a slight buzz (not hum) and sounded a little worse than the bypassed resistor there before. I checked the current through that plate resistor and had 2mA - maybe that was it - enough for the LED to work but not well?

Have a green LED on the 6J5 which the schematic called for 1500/250mf 2.7V so a green LED looked perfect there. Drawing 5.5mA working and sounding great, so it's staying, or 'till I get bored again