Can anybody in the know, clear this up for me.
- What are advantages and disadvantages of both implementations in power amplifier output stages?
- Why would one choose one over the other?
I am only asking, since both are as complex and costly as each other.
Also "fixed bias" means a negative supply with trimmers to adjust bias.
- What are advantages and disadvantages of both implementations in power amplifier output stages?
- Why would one choose one over the other?
I am only asking, since both are as complex and costly as each other.
Also "fixed bias" means a negative supply with trimmers to adjust bias.
Last edited:
The two are completely different..
And dependant on power..
The LED idea does not use electrolytic caps so there is no aging and cap failure..
Fixed biasing requires a DC bias supply..how are they as expensive as each other?
A DC supply requires supply caps and trimmers and fail safe design. (Fixed bias)
LED in low power can cause problems not enough current to stabilise or light the LED..
Regards
M. Gregg
And dependant on power..
The LED idea does not use electrolytic caps so there is no aging and cap failure..
Fixed biasing requires a DC bias supply..how are they as expensive as each other?
A DC supply requires supply caps and trimmers and fail safe design. (Fixed bias)
LED in low power can cause problems not enough current to stabilise or light the LED..
Regards
M. Gregg
Last edited:
I take "small signal class A" as a valid answer for cheaper and simpler.
I suppose I should have added:
" In Power amp final stage" as it is mostly there that one may consider either.
Say 10W to 40W amps.
Also as "fixed Bias" I meant negative supply with trimmers to set the bias.
With LED bias, some possible voltage is 'wasted', so output swing (i.e. output power) is reduced compared with fixed bias.
You need a lot of LEDs to handle the current and acheive a useful bias voltage with most power valves. Mounting all those LEDs can be inconvenient. I'd say this qualifies as 'more costly' than simple fixed bias.
good points, but we are counting pennies.
What I am after is , are there real benefits in LED biasing that I can not see?
Compared to overall cost of an amplifier, and the effort to put one together, the cost/effort difference between two implementations is real small.
Why did anyone think of LED biasing?
What issue(s) were they trying to fix?
What new issues did they create?
LED biasing provides very low dynamic resistance in the cathode circuit of a tube output stage.
It is considered an alternative to cap bypassed resistor in fixed bias.
They were addressing the capacitor and it's deficiencies.
for a power output stage, it takes an array of LEDs. Several strings of LEDs in series have one small resistor for balance when the strings are tied in parallel. The LEDs need to be select3ed for (1) low dynamic impedance and (2) matching voltage drop.
It is considered an alternative to cap bypassed resistor in fixed bias.
They were addressing the capacitor and it's deficiencies.
for a power output stage, it takes an array of LEDs. Several strings of LEDs in series have one small resistor for balance when the strings are tied in parallel. The LEDs need to be select3ed for (1) low dynamic impedance and (2) matching voltage drop.
I am confused, aren't you describing cathode bias?
Fixed-bias as in no cathode resistor, and a separate supply and pots to adjust bias (in power-amp output stages).
Call it manual bias if you like.
What you are actually referring to is "Grid bias". Grid bias is when you adjust a negative supply with a trimmer to reference the signal input. "Fixed bias" requires the cathode to be elevated with a (hefty) resistor, while the grid is referenced to ground.
Grid bias is a nice solution when you don't have enough B+ to do "Fixed bias". It has the disadvantage that you need an additional negative supply secondary from your mains transformer. It also has the disadvantage that you immediately introduce new AC noise into your audio signal.
Fixed bias : Has the advantage that you won't be introducing new AC noise into your signal, such as with "Grid bias". You also don't need an additional negative supply winding off your transformer. It has the disadvantage that it will need higher B+ than "Grid bias". So perhaps not something for the novice builder... Also has the disadvantage that designs often need to bypass the cathode resistor with an electrolytic capacitor. I won't go into all the things you might not want this electrolytic capacitor though. You also need a substantially rated cathode resistior to do Fixed bias.
In any case, I would chose Fixed Bias over Grid Bias any day for anything you aspire to become Hi-Fi. Just my opinion though...
Ian
Last edited:
I stand corrected - thank you.
I do mean grid bias.
If I understand it correctly (help me here) grid bias current requirements are so low, that providing a relatively noise free supply is not so difficult.
If it happens to be a push-pull design, even less of an issue.
I understand LED biasing is mainly used on tubes such as EL84, which needs a low bias voltage. Providing tens of bias volts is impractical with LED's.
A separate low power transformer, few resistors and pots etc. will not cost much more than tens of LED's wired together, not mush harder to put together either.
So my question remains:
Why should I use one over the other?
Are there any performance issues?
Any reliability issues?
anything else?
Last edited:
Grid bias current requirements are low. If you want to use Grid bias then of course it needs to be nicely filtered, and it must be from a nice stable source, ideally shunted to ground. Most of the current flows to ground and you reference your grid to some tap off that flow. True, it can be made quiet. Yes it can be made dead silent if need be. You still need an additional secondary tap for it though. And if you do it for push-pull, you might discover that turning one pot can influence all the other grid volts too (depending on your design). So perhaps my true dislike for it is from having to fiddle with it on cheapish amplifiers.
If you want to do grid bias well, then you need to do it separately for each channel too. So even more fiddly parts..
LED's as they are often used in pre-amps, are used for Fixed bias. Same goes for the Red Light Distric EL84 amp that I think you might be referring to. I would suggest that LED bias is rare for end-stages. Most Fixed bias amplifiers can't be reasonably done with LED's.
Fixed bias and Grid bias can both work very nicely when implemented in the right design. I still prefer Fixed bias. It depends on what you want to build though and what you have on hand. Not sure if that makes sense...
I have seen many KT88 push-pull designs with grid bias. They sounded very pleasant once tuned. I have never seen a 2a3 push-pull with grid bias though. I would never consider building one either...
If you want to do grid bias well, then you need to do it separately for each channel too. So even more fiddly parts..
LED's as they are often used in pre-amps, are used for Fixed bias. Same goes for the Red Light Distric EL84 amp that I think you might be referring to. I would suggest that LED bias is rare for end-stages. Most Fixed bias amplifiers can't be reasonably done with LED's.
Fixed bias and Grid bias can both work very nicely when implemented in the right design. I still prefer Fixed bias. It depends on what you want to build though and what you have on hand. Not sure if that makes sense...
I have seen many KT88 push-pull designs with grid bias. They sounded very pleasant once tuned. I have never seen a 2a3 push-pull with grid bias though. I would never consider building one either...
Last edited:
Rarely is it remembered that the overwhelming advantage of capacitor-bypassed-cathode resistor bias is that it automatically adapts to a relatively large range of tube characteristics all by itself, something which both (or "neither") LED bias or fixed B- bias manage to do worth beans.
How so?
Simple enough: the combination of the cathode resistor and its bypass capacitor act as a variable-voltage power supply, the equivalent of the B- supply or nearly equivalent LED bias voltage … except that through [ E = I·R ] Ohms law, the E part is directly dependent on the I and the R. As a tube ages and degrades, its operating point(s) will change. The R+C solution adjusts to the degradation in mu, in a compromised fashion. LED bias (and fixed bias) will result in a far larger change in quiescent current flow as a tube ages.
If you read about, you will find that the B- bias supply issue is significant: there is plenty of advisories that indicate it is sensitive, and ends up needing periodic adjustment. You will also note that no one ever seems to advise that R+C cathode bypass needs much adjustment … except possible for the final output stages (if it is done that way).
The most important Idea (I think) as to why output stages are often the only stage which has a separate B- bias, is because it really widens the range of the current that the finals can control, or if you prefer, the voltage swing they'll produce. Save those precious watts.
Or, if you want to distill this thought somewhat, cap bypassed cathode resistance bias works great for the front end, and fixed bias works well for finals, esp. push-pull. Intermediate solutions such as LED bias also work well especially for new tubes at the front end, but the strings required for finals … are unwieldy. Strictly speaking, you could go for the MOSFET equivalent of a LED, of course.
I admit fully that the above is my opinion and most likely not shared by the majority of the Brainiacs of this DIYaudio community.
GoatGuy
How so?
Simple enough: the combination of the cathode resistor and its bypass capacitor act as a variable-voltage power supply, the equivalent of the B- supply or nearly equivalent LED bias voltage … except that through [ E = I·R ] Ohms law, the E part is directly dependent on the I and the R. As a tube ages and degrades, its operating point(s) will change. The R+C solution adjusts to the degradation in mu, in a compromised fashion. LED bias (and fixed bias) will result in a far larger change in quiescent current flow as a tube ages.
If you read about, you will find that the B- bias supply issue is significant: there is plenty of advisories that indicate it is sensitive, and ends up needing periodic adjustment. You will also note that no one ever seems to advise that R+C cathode bypass needs much adjustment … except possible for the final output stages (if it is done that way).
The most important Idea (I think) as to why output stages are often the only stage which has a separate B- bias, is because it really widens the range of the current that the finals can control, or if you prefer, the voltage swing they'll produce. Save those precious watts.
Or, if you want to distill this thought somewhat, cap bypassed cathode resistance bias works great for the front end, and fixed bias works well for finals, esp. push-pull. Intermediate solutions such as LED bias also work well especially for new tubes at the front end, but the strings required for finals … are unwieldy. Strictly speaking, you could go for the MOSFET equivalent of a LED, of course.
I admit fully that the above is my opinion and most likely not shared by the majority of the Brainiacs of this DIYaudio community.
GoatGuy
Besides simplicity (even the array I made for my power amp was simple and cheap), low noise, and stable voltage, the big advantage is instantaneous recovery from overload. LED-biased stages don't have blocking distortion.
See this:
The Red Light District: A 15W Push-Pull Amplifier
See this:
The Red Light District: A 15W Push-Pull Amplifier
Hmm, some confusipn here...
Fixed bias
Is when you force a fixed voltage difference between the cathode an grid. A LED string in the cathode circuit and the grid tied to ground OR cathode at ground potential and a negative voltage on the grid are both examples of fixed bias, they do behave a bit different at clipping and I cannot say that one would be better than the other.
Self bias, auto bias or cathode bias
Grid at ground potential and a resistance in the cathode circuit, bypassed or not. Simple and safe.
Grid leak bias
Grounded cathode and usually a very large resitance from grid to ground, not to many tubes are made for this and it most commonly used in condenser microphones and similar circuits.
BR,
Anders
Fixed bias
Is when you force a fixed voltage difference between the cathode an grid. A LED string in the cathode circuit and the grid tied to ground OR cathode at ground potential and a negative voltage on the grid are both examples of fixed bias, they do behave a bit different at clipping and I cannot say that one would be better than the other.
Self bias, auto bias or cathode bias
Grid at ground potential and a resistance in the cathode circuit, bypassed or not. Simple and safe.
Grid leak bias
Grounded cathode and usually a very large resitance from grid to ground, not to many tubes are made for this and it most commonly used in condenser microphones and similar circuits.
BR,
Anders
Very interesting reply, thank you.
Just to mention, that I am one of those Capacitor Haters you described elsewhere!
As to your point regarding grid-bias (fixed-bias) maintenance issue that requires periodic readjustments - remember where we are " DIY audio".
At DIYaudio, we periodically poking into our equipment for fun, upgrades etc. - so checking the bias is not much of an burden. Correct that if it was a commercial amp, as manufacturers we had a duty to design a maintenance free product, but not here!
I reckon you don't believe in the HiFi-devil within capacitors, but in input stage biasing, are there any sonic benefits with LED biasing compared to cathode biasing with a large cap?
KenTajalli - You might not need a cathode bypass capacitor if your design is well considered. I don't like capacitors, but they are a necessary evil. I avoid electrolytics and use film caps.
Look at post #12 here: http://www.diyaudio.com/forums/tubes-valves/259333-6b4g-ecc83-2.html#post4256787
No cathode by-pass, not even on finals. No coupling capacitors. No transformer coupling except to speakers, only film caps... but it is not push-pull.
Ian
Look at post #12 here: http://www.diyaudio.com/forums/tubes-valves/259333-6b4g-ecc83-2.html#post4256787
No cathode by-pass, not even on finals. No coupling capacitors. No transformer coupling except to speakers, only film caps... but it is not push-pull.
Ian
Last edited:
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.