I'm still debating whether to build my EL84 amp as UL or triode or with a switch for both. One consideration, do I need a negative feedback circuit with UL connection? The majority of vintage UL EL84 schematics I have seen do include a negative feedback circuit from the OPT secondary to the input stage, but then I came across this schematic which made me question this:
http://www.electra-print.com/images/PSA_EL84_reva.pdf
Comments? Suggestions?
http://www.electra-print.com/images/PSA_EL84_reva.pdf
Comments? Suggestions?
I recently built my first UL amp (after building a few triode amps) using 6AQ5's. I did not want to use global negative feedback if I could avoid it, but just hooking the screen grids to the UL taps was not enough. UL is a form of local negative feedback, so I tried inserting some series resistance to control it. It worked! I ended up with 300 Ohms in between the taps and screen grids, and it sounds great. I would try with no resistance first, then add resistors in steps until it sounds clean.
My circuit is similar to that one, except that I use a triode for the input stage and use an interstage transformer instead of the cap/choke combination. Either way works. I think you will find that it is easy to get working and sounds very good!
My circuit is similar to that one, except that I use a triode for the input stage and use an interstage transformer instead of the cap/choke combination. Either way works. I think you will find that it is easy to get working and sounds very good!
Thanks! When you say that your design was similar, did you also leave the cathode resistor of the EL84 unbypassed as shown in the above schematic? I guess that is also local feedback employed by that amp.
I have a Leak Stereo 20 which is much modified, but taking off the feedback loop in UL was a breakthrough in sound for me. Much more natural. I agree - try the 300 ohms resistor. Andy
Yes, my understanding is that the bypass caps are not required in the push-pull circuit, so I only use caps on the input tube cathode.
Thanks Andy, for confirming the idea that global neg. feedback is best avoided!
Thanks Andy, for confirming the idea that global neg. feedback is best avoided!
My current amplifier is the circuit you have pointed out from Electraprint. My previous build is a se EL34 but this amp beats it easily for sound quality. The 6au6 in triode mode is a very underated tube. The phase splitting transformer works very well.
The circuit is easy and well thought out. Give it a try and report back your findings.
Phil.
The circuit is easy and well thought out. Give it a try and report back your findings.
Phil.
Hey-Hey!!!,
If the speakers don't do anything screwey, a very high output Z amp can drive them. Think all-pentode. But if you need a bit more NFB and still don't want to do the nasty loop thing...try this one.
It works best with pentodes in the front stage.
cheers,
Douglas
If the speakers don't do anything screwey, a very high output Z amp can drive them. Think all-pentode. But if you need a bit more NFB and still don't want to do the nasty loop thing...try this one.
It works best with pentodes in the front stage.
cheers,
Douglas
Attachments
Yep - a bit is required
Leadbelly,
See my post here for what I do:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=72314&perpage=10&pagenumber=3
I use some shunt feedback (local feedback) and a small amout of global feedback.
Cheers,
Ian
Leadbelly,
See my post here for what I do:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=72314&perpage=10&pagenumber=3
I use some shunt feedback (local feedback) and a small amout of global feedback.
Cheers,
Ian
gingertube - you were threatening to post aschematic of your EL84 amp - would be ijterested to see it + the CCS in the output. Andy
Reply to Andy
Andy,
Posted here:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=72536
The CCS for the output tubes is on page 2. Note I use separate CCS for each tube to allow Class AB operation.
Cheers,
Ian
Andy,
Posted here:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=72536
The CCS for the output tubes is on page 2. Note I use separate CCS for each tube to allow Class AB operation.
Cheers,
Ian
If I may make a few general comments (for regular members some of this would be repitition):
UL is close to triode operation, only with pentode available output. (I often find misunderstanding here.) The main desirable feature is low local internal resistance. As Brandersnatch said, if the speakers are constant impedance ... but where does that exist? Pentodes are distortionwise also quite critical with regard to load impedance. To look from another angle: why would anybody prefer pentode to UL? A lot of hype exists in this regard: much of it unfounded.
I can also not see a basic reason for not using n. feedback (and let us not confuse poor results from injudicious use of often too much feedback, with its correct application). Any amplifier with feedback correctly applied is simply more linear than without it - there are decades of examples supporting this.
Regarding bypassing or not of a common output tube cathode bias resistor: The advantages of fixed bias will be known, and bypassing such a resistor with a high capacity can give resemblance to fixed bias under most music conditions. Only one must be careful here. Mullard shows that optimal EL84 operation for the 2 conditions is different. For (unbypassed) cathode bias, optimum conditions at 16W output are a-a load = 8K.ohm, Ia = 2 x 36 mA for distortion of 2,4%. Using the same values under fixed bias conditions (or heavily bypassed cathode resistor) will give D = 5%. For optimal conditions here Ia needs to be reduced to 2 x 24 mA with an a-a load impedance of 6K.ohm. Then the distortion at 16W will be 1,6%. (The graphs for this come from a brochure; I have unfortunately not yet found them on the internet - perhaps others have.) For the low bias voltage required by EL84s, zener diodes can be used instead of a cathode resistor for full fixed bias.
UL is close to triode operation, only with pentode available output. (I often find misunderstanding here.) The main desirable feature is low local internal resistance. As Brandersnatch said, if the speakers are constant impedance ... but where does that exist? Pentodes are distortionwise also quite critical with regard to load impedance. To look from another angle: why would anybody prefer pentode to UL? A lot of hype exists in this regard: much of it unfounded.
I can also not see a basic reason for not using n. feedback (and let us not confuse poor results from injudicious use of often too much feedback, with its correct application). Any amplifier with feedback correctly applied is simply more linear than without it - there are decades of examples supporting this.
Regarding bypassing or not of a common output tube cathode bias resistor: The advantages of fixed bias will be known, and bypassing such a resistor with a high capacity can give resemblance to fixed bias under most music conditions. Only one must be careful here. Mullard shows that optimal EL84 operation for the 2 conditions is different. For (unbypassed) cathode bias, optimum conditions at 16W output are a-a load = 8K.ohm, Ia = 2 x 36 mA for distortion of 2,4%. Using the same values under fixed bias conditions (or heavily bypassed cathode resistor) will give D = 5%. For optimal conditions here Ia needs to be reduced to 2 x 24 mA with an a-a load impedance of 6K.ohm. Then the distortion at 16W will be 1,6%. (The graphs for this come from a brochure; I have unfortunately not yet found them on the internet - perhaps others have.) For the low bias voltage required by EL84s, zener diodes can be used instead of a cathode resistor for full fixed bias.
- Status
- Not open for further replies.