Merlin, if that's the case than an in time shifting operating point (DC) of the cathode would also have a negligible effect on distortion? Than the need for saturation of the LED (to deliver a perfect constant voltage) would probably also be marginal relative to inherent tube distortion? 😎
Check out these diodes (low current) To me it looks like 2 mA operating point would be fine with these??
Data Sheet - HLMP-D150, HLMP-D155, HLMP-K150, HLMP-K155 - T-1 3/4 (5 mm), T-1 (3 mm), Low Current, Double Heterojunction AlGaAs Red LED Lamps (292 KB)
Data Sheet - HLMP-D150, HLMP-D155, HLMP-K150, HLMP-K155 - T-1 3/4 (5 mm), T-1 (3 mm), Low Current, Double Heterojunction AlGaAs Red LED Lamps (292 KB)
The way I understand it is in order to get the benefit of LED bias, with the lowest possible dynamic resistance, you want to select the "cheap red" LED. Going to special types like low current or exotic colors tends to increase the resistance. It's a tradeoff; I'm sure it would work, but would it work as well?
Member
Joined 2009
Paid Member
Nevertheless, people report that changing to LED bias brings a change in sound. Of course doesn't mean it's bad just that we are sensitive to subtle changes.
Zigzag: Looking at the graph of voltage vs. current for these low current diodes it looks to be even more straight up and down than the cheap red ones I can find to look at.
Not that these low current ones are off the charts expensive ~ $.50 each.
Not that these low current ones are off the charts expensive ~ $.50 each.
Output impedance of an amplifying stage depends on cathode impedance which in turn is a function of the value of the stuff that offsets it. i.o.w. changing the cathode impedance changes output impedance, which could react to the input capacitance of the next stage and so on. That's the effect one could hear, not the impedance of the cathode itself.
Last edited:
Like I said, if B+ is not clean, I'd use a CCS to inject the current.
There's a very small, noticeable only on headphones, difference in channel separation. Shared LED system is a bit less 'wide' in soundstage. With speakers it's not perceptible.
The LED’s have other problems which are possible UHF oscillation spikes and a very high increase in IMD due to the audio path is now seeing a two separate motions of audio and DC current. LED’s are little UHF oscillators. If you want these spikes on your audio signal, put them in, Regular diodes will not do this.
Please explain this.
I don't think it's a good idea to allow your audio circuit to pass UHF anyway. I never had any oscillation issues because of LEDs.
I especially like the old silverware chest. It looks exactly (and I mean identical!) to the one my mom's old silverware is in. Perfect!
I'm glad to see you bypassed the heater-coil inductance of the 45's with the orange drop capacitors. Strictly speaking (because I have failed to contract über-capacitor fetish), I'd probably have used Cornell-Dublier CDE 5 μF / 250 V bypasses. But that's just me. I like to give more of the audio spectrum its approaching-zero impedance-to-ground.
(Indeed: since there is a phase tradeoff, it might be wise to at least try larger values)
anyway, here's a toast:
I'm glad to see you bypassed the heater-coil inductance of the 45's with the orange drop capacitors. Strictly speaking (because I have failed to contract über-capacitor fetish), I'd probably have used Cornell-Dublier CDE 5 μF / 250 V bypasses. But that's just me. I like to give more of the audio spectrum its approaching-zero impedance-to-ground.
(Indeed: since there is a phase tradeoff, it might be wise to at least try larger values)
anyway, here's a toast:
I love the 45. My very first amp project was Eric Barbour's 45 driven by 6SN7stereo amp as published in Speaker Builder (or was that Glass Audio?) some years ago. I absolutely loved the sound.
Mr. Curwen, a friend of mine working on those before and that is what he found out about the spike, a regular diode will do much better than led.
Thank you all for the comments. My thinking on the LEDs is that nonlinearity/residual voltage shouldn't be too much of a problem in this design, because the 6SN7s are choke-loaded, meaning that current through the LEDs should be constant. As far as UHF, the output impedance of the 6SN7 forms a lowpass filter with the miller capacitance of the 45, which should prevent any UHF from hitting the power tube (not to mention rolloff from the output tranny and speakers). Maybe in the future I'll switch out the orange drop caps for larger ones.
Many of agree that the balloon 45 or 245 is the Stradivari of vacuum tubes, a musical instrument itself. For stronger bass, I an using a triamp system on 3-way GR Research speakers and the woofers will be powered by a 245 driven 833A with only about 800 volts. The 833A runs at that voltage about 100 ma with zero grid bias so that it can be connected by a 5k inner-stage transformer that will have the primary winding loading the plate of the 245.
All common LEDs exhibit Curvature on the EI plot.
Attached are a couple of HP production from about 35 yrs ago.
As an aside, first LED I ever heard of was around 1960.
Experimental sample, $60 US while I worked on core memory research at U of T Physics. At HP we had Military grade red LEDs for about $10 by 1968.
The price came down very quick. By 1968, 7 or 8 cents in quans of a million.
Attached are a couple of HP production from about 35 yrs ago.
As an aside, first LED I ever heard of was around 1960.
Experimental sample, $60 US while I worked on core memory research at U of T Physics. At HP we had Military grade red LEDs for about $10 by 1968.
The price came down very quick. By 1968, 7 or 8 cents in quans of a million.
Attachments
- Status
- Not open for further replies.