Hi SY,
Is there an easy method to measure the dynamic impedance of a mystery LED? I have trays full. But I have very little in the way of specialized measurement instruments or signal gens, etc. Just a few DMMs.
I'm not sure what the definition of dynamic impedance is. I thought all impedance was pretty dynamic.
Note to self: Ask manager about the unused oscilloscope in the surplus parts room at the office.
..Todd
Is there an easy method to measure the dynamic impedance of a mystery LED? I have trays full. But I have very little in the way of specialized measurement instruments or signal gens, etc. Just a few DMMs.
I'm not sure what the definition of dynamic impedance is. I thought all impedance was pretty dynamic.
Note to self: Ask manager about the unused oscilloscope in the surplus parts room at the office.
..Todd
Hi Todd,
simple PC-based curve tracer: http://www.syscompdesign.com/AppNotes/curve-tracer.pdf
"In-circuit" LED impedance testing: SYclotron Audio Red Light District, p2
Best regards,
Fred
simple PC-based curve tracer: http://www.syscompdesign.com/AppNotes/curve-tracer.pdf
"In-circuit" LED impedance testing: SYclotron Audio Red Light District, p2
Best regards,
Fred
What fhs said. In a pinch, you can substitute a 12V (or so) transformer for the signal source in my test jig and use the DMMs to measure AC voltages, but... if you want to do electronics, you need a scope. It's not optional. May as well try to work on cars without wrenches.
FHS,
The pictures in that curve tracer doc were worth a thousand words. Thanks for that. It all makes sense now.
..Todd
The pictures in that curve tracer doc were worth a thousand words. Thanks for that. It all makes sense now.
..Todd
Hi Todd,
if you happen to get a scope (with X/Y capabilities), you can take the route SY proposes or you can use a very simple circuit like the one on the bottom of the following page: Curve Tracers
Fred
if you happen to get a scope (with X/Y capabilities), you can take the route SY proposes or you can use a very simple circuit like the one on the bottom of the following page: Curve Tracers
Fred
I remember reading somewhere that there wasn't any point in using LED biasing for single ended stages. Think it might of been in a thread discussing the RLD.
Is that true? Could someone enlighten me a little? Or did I not read that at all?
Thanks
Is that true? Could someone enlighten me a little? Or did I not read that at all?
Thanks
I used a bunch of old red LEDs (26x4 per output triode) in biasing SE 6C4C (eq 6B4G) - it sounds way much better than biased with RC. Totally new amp.
I often use a single red led in the cathode and there can be slight variances between makes varying from 1.6-1.9V. I use the slightly higher volt ones. Tubes ideal for this are cheap ex TV line video, EF184, 12BY7A,12GN7 and other types, so these have become my favourite workhorse pentodes configured as triodes which pushes up the curves nice and linear.
As SY quotes, the working curves are closer to the bottom end but there fantastic benefits of a little closeness and the red led can operate right in the curve ballpark. It's worth it.
There is little benefit putting a shunt cap on red types, remember the intrinsic AC impedance of the current source shunt is already very low. Going to a yellow led (25ohms) depending on circuit one might need a cap. Better with two red's in series.
Once you learn't to use these magic current sources....one never turns back.
Morgan Jones, valve amps 4th ed. page 170. DC bias problems.... Read it !
richy
As SY quotes, the working curves are closer to the bottom end but there fantastic benefits of a little closeness and the red led can operate right in the curve ballpark. It's worth it.
There is little benefit putting a shunt cap on red types, remember the intrinsic AC impedance of the current source shunt is already very low. Going to a yellow led (25ohms) depending on circuit one might need a cap. Better with two red's in series.
Once you learn't to use these magic current sources....one never turns back.
Morgan Jones, valve amps 4th ed. page 170. DC bias problems.... Read it !
richy
Can somebody summarise the main advantages that led biasing gives in different cases compared with (resistive) cathode bias ?
If you mean "resistor bypassed with big electrolytic," then better overload recovery, lower cost, flatter gain/phase at low frequencies, potentially lower distortion (if one believes that capacitors impart certain sonic characteristics). And of course, a visual indication that a stage is operating. If you mean "just resistor," add lower source impedance, higher gain, lower order distortion spectrum.
At first, I have not used LED's in my circuits, so I just give my opinions as an assumption.
Compared with auto bias with bypassed resistor:
- lower cost : definitely much in percentages
- flatter gain/phase at low frequencies: I think similar result can be achieved with both methods, but I addmit this requires very big bypass capacitor and is easier to do with the LED
- better overload recovery: I can not see this happening in typical voltage amplifier circuit, ( or if voltage amplifying stages are already overloaded, the even more serious overload exist at the following stages, especially at output stage )
- lower distortion: difficult to see how this would be achieved, well I know there are guys who hear the difference but any test gear do not indicate it
And without bypass cap.
- lower source impedance, higher gain: Sure
- lower order distortion spectrum: I don't know about the "order" of the distortion, but all my circuits, intentionally done without bypass capacitor exhibit lower total distortion, obviously due to local NFB
Do we lose something ?
- Led bias is fixed, then the compensating effect of the cathode resistor to DC current variations is lost.
- I have often observed that the linearity of the voltage amplifying stage can be improved by fine tuning the cathode resistor from the "calculated" value. This is almost impossible to carry out with LED's
Compared with auto bias with bypassed resistor:
- lower cost : definitely much in percentages
- flatter gain/phase at low frequencies: I think similar result can be achieved with both methods, but I addmit this requires very big bypass capacitor and is easier to do with the LED
- better overload recovery: I can not see this happening in typical voltage amplifier circuit, ( or if voltage amplifying stages are already overloaded, the even more serious overload exist at the following stages, especially at output stage )
- lower distortion: difficult to see how this would be achieved, well I know there are guys who hear the difference but any test gear do not indicate it
And without bypass cap.
- lower source impedance, higher gain: Sure
- lower order distortion spectrum: I don't know about the "order" of the distortion, but all my circuits, intentionally done without bypass capacitor exhibit lower total distortion, obviously due to local NFB
Do we lose something ?
- Led bias is fixed, then the compensating effect of the cathode resistor to DC current variations is lost.
- I have often observed that the linearity of the voltage amplifying stage can be improved by fine tuning the cathode resistor from the "calculated" value. This is almost impossible to carry out with LED's
Blocking and discharge time, a clearly audible problem in any circuit which could clip. This is a particular issue in phono stages, feedback circuits, and output stages. By contrast, LED recovery is instantaneous (or very nearly so, order of nanoseconds). Clipping is often not audible, but the hangover from blocking clearly is.
Exactly so, and the nonideality of a large cap there aside (from both electrical and physical considerations), there's a tradeoff- the larger the cap, the lower the LF phase shift and response error but the longer the overload recovery time.
Yes, the small amount of local feedback reduces the overall magnitude but of necessity increases the order of distortion. There's a model calculation in the famous Baxandall Wireless World paper based on a square law (i.e., pure second harmonic) device which is good reading. I'd also spend a bit of time with a spectrum analyzer to see these effects.
I've also found that there's a synergy to the combination of LED bias and CCS plate loading. The distortion magnitude and spectra of my designs bear this out.
I couldn't agree more and I found this out in my power amps.
The standard anode-grid Baxandall feedback circuit makes a good examination candidate, esp the association of the treble control boost response and 2nd harmonic.
The tube audio world nearly left it too late in waking up to exactly to the benefits which you mention. Fortunately, the tubes that respond best to both CCS & LED are still around in surplus. The magic is a little bit more of running current.
richy
I'd never tried this until this afternoon. I have a Push-pull 6550 driven by 6922. Just switched the resistor and cap out of the cathode of the 6922 for 2x red "cooking" LEDs , and the difference is wonderful... even to my tin ears the bass seems better controlled.
My home theatre system (still awaiting the OPTs and case!) has now got LED biasing in it's 6N6P driver stages.
My home theatre system (still awaiting the OPTs and case!) has now got LED biasing in it's 6N6P driver stages.
Remember, tube power amps which use highly strung amounts of global feedback (25-35dB) from o/p tranny to the 1st stage cathode;-> removing the cap from (I assume from) the first stage cathode, one is really modifying the 6dB droop pole from the sub LF response, which modifies the phase shift stability bode plot at the o/p transformer lower freq end. In essence, without the cap reactance the open loop gain will try to correct and force the amp gain to a near DC response; (near impossible due to other coupling caps) but the first stage cathode cap does modify the LF phaseshift degrees imposed by the output transformer inductance to avoid the transfer function gain from completing the circle; ie tendancy to instability and subsonic cone warbling.
If one don't notice the nasty symptoms of horrendous phaseshift complications, fine, leave it and put the covers back on.. Simple signs of LF instability are observing the LF LS cone after a hefty single beat note transient and see if the cone settles instantaneously.
In many design cases most won't notice the problems as there is enough phaseshift margin engineered in; but theoretically with a lower quality o/p tranny all may not be quite perfect. A simple solution to correct is to reduce the quantity of global feedback.
It may sound even better, but in reality THD will rise at the lower end, at the expense of increasing the residual noise level of the amp. This mustn't be done indiscriminately, in 3db increments and with it the global nfb step cap will have to be trimmed too, another nasty surprise. This requires complete user oscilloscope familiarisation & skill; older users will be familiar at the art of XY Lissajous patterns...but one has to know what one is looking at. It's all revealed in the picture !
1st stage cathode decoupling cap :MJ Valve amps 4th ed page 77. Read it!
richy
- Status
- Not open for further replies.