Tony said:
cathode bias is still the the safest way to go imho.
isn't ss the safest way to go?
seriously I agree with you on cathode bias being safe, but I have been there so many times that I have lost the feeling. figuring out fixed bias satisfies my need for excitement. i'm sure you've been there before.
I've finished my tube rectified, fixed biased Williamson and so it is now time to make an ss rectified fixed bias amp.
trying to get something useful out of the irons and parts I have laying around for so long.have a nice one Tony,
JojoD
Unlikely, IMO. They made other amps with fixed bias. The most populat tube power amp ever made in the US, the Dyna Stereo 70, used fixed bias. And the marketing penalty for a lower rated power is severe- most audiophiles, and all nonaudiophile consumers, don't have an appreciation of how small a level change 3 dB is. All they know is that THIS amp is 25 watts and THAT amp is 50 watts!
SY said:
thanks SY. We await the full and final version.
BTW SY I got an idea....
what if in include the crossover inside the power amp as follows....
ECC88 based Steve Bench Crossover. 1 tube per channel.
mated to a ckt simiar to the HIF 87 but eliminating the ECC83 as one does not need gain with today's sources. it is just that my design skills do not allow me to figure out where to put the NFB loop.
6SN7 Phase splitter. 2 tubes per channel one for Low pass and one for High pass.
EL34 Power. 4 tubes per channel. 2 for high pass (Triode PP) and 2 for low pass (UL PP).
now this is 7 tubes per channel.
the preamp would be a passive input selector and volume control.
Navin, that balancing setup is not what I use, it was just to illustrate the right way to do what Jojo was trying. My own preference is for 4 separate pots, one for each output tube. No advantage either way that I can see, it's just a matter of how you like to do your idle current checks.
The Eico's extra winding with a solid state bridge rectifier and a moderate cap (100uF/150V) give a nice 100V. The 7027A will need roughly -40 to -45V. So the string could easily be set with a potentiometer sandwiched between two resistors of the same value as the pot. For example, I happen to have some very cool double trimmers in a dip package, 20K each. So two of these packages gives me 4 trimpots. If they are all in parallel (other than the wipers!), they form a 5K resistor. So a string of a 4.7K, the four pots in parallel, and the second 4.7K gives me plenty of adjustment range. If it's too twitchy to adjust, I can diddle the series resistor values to narrow the adjustment range.
If the output stage will be run as UL or triode without regulation, the bias supply should likewise be unregulated.
The Eico's extra winding with a solid state bridge rectifier and a moderate cap (100uF/150V) give a nice 100V. The 7027A will need roughly -40 to -45V. So the string could easily be set with a potentiometer sandwiched between two resistors of the same value as the pot. For example, I happen to have some very cool double trimmers in a dip package, 20K each. So two of these packages gives me 4 trimpots. If they are all in parallel (other than the wipers!), they form a 5K resistor. So a string of a 4.7K, the four pots in parallel, and the second 4.7K gives me plenty of adjustment range. If it's too twitchy to adjust, I can diddle the series resistor values to narrow the adjustment range.
If the output stage will be run as UL or triode without regulation, the bias supply should likewise be unregulated.
SY said:
SY et al
(Sigh) aint that too true! But going back slightly (I have been away for awhile): This is also true in most cases for fixed bias. Meaning that one does not readily get more than +3 dB going from cathode to fixed bias, the rest being equal. Fixed bias does have its charm (I am using it too), but in reality, what is the actual gain we get from it?
It must be admitted that cathode bias has its conveniences - all listed here before, except for perhaps also the advantage that a 5x higher G1 resistor is allowable - that can count in some circuits! Also, when I examine signals at the crucial points while using large capacitors (including cathode bypass of 10 000 uF), and for the mark-space ratio of most music, I find very little change in voltages. That means something remarkably close to fixed bias; much less than 3 dB advantage overall.
This may not be so relevant to high-power jobs, say many 100W (I keep to 100W max. with tube amps), but every time I used fixed bias I had to convince myself that it is not just "man's vanity" with me.
And going forward, as 2005 draws to a close (at least here in the RSA): My sincere wishes to all for a good 2006 - even though we are told that the year is now shorter because of the fault that caused the severe tsunami in the east - by a full 3 mS! After all the technology expertly reported here, may there be sufficient time to relax and listen to our efforts, as well as to live performances.
Regards from RSA,
Johan
SY said:
Sy, I am not using the 7027.
Yes I need feedback but an uncertain hoe to implement it if I eliminate the AF tube (ECC83 in the HIF ckt).
Do you know of any EL34 ckt that eliminates the AF tube?
Given today's suroces (CD Players, Ipod, etc..) Do we need so much gain anymore?
I hope to make my recitification all Solid State and use individual pots for each tube's bias. It is easier to set bias this way esp for a newbie.
BTW do I need bias pots for the phase splitter and other tubes as well or only for the output tubes?
navin said:
OK, sorry, can't keep track of all the amp builders without a scorecard. Anyway, the argument doesn't change much for EL34s- you still need 30V or so of drive, signal sources are 2-3V, and most phase splitters have lowish gain. And you need excess gain to implement feedback. Figure 15dB (x6) as a minumum. So if you were to do it as a single stage, say a CC diff amp, you'd need to have a stage gain of 90 or so. That's going to be tough to do!
Conventional circuits got that way by being the best engineering solutions.
Depends on the topology you're using. For cascaded diff amps (e.g., Ray Moth's design or Morgan Jones's Crystal Palace), my personal preference runs to 6SL7->6SN7. But there are a lot of other fine possibilities. For a single-ended stage feeding a long-tail phase splitter (e.g., the Eico circuit), I might still think of 6SL7->6SN7. But a 12AT7 will also perform great in the first hole while showing a lower input capacitance.
SY,
Here's the revised schematic for the bias section....
http://img.photobucket.com/albums/v70/jojod818/Dump/EICOHF87FixedBias2.gif
Will this be ok now?
BTW, my brother asked me for the 7027 so I may use 6L6 if not EL34.
Thanks!
Here's the revised schematic for the bias section....
http://img.photobucket.com/albums/v70/jojod818/Dump/EICOHF87FixedBias2.gif
Will this be ok now?
BTW, my brother asked me for the 7027 so I may use 6L6 if not EL34.
Thanks!
I'd swap the 47K and 100K resistors. The way you have it drawn, the effective grid resistance is under 60K. Swapped, you get about 110K as the resistance, a little gentler on the drive stage, and this will also extend the low frequency f3 from 12Hz to 6Hz. The latter will help stability, important given the relatively high feedback.
ooppsss... that is what I used in my Williamson ala fixed bias tube amp and i think that sometimes it's a little bit bass shy. 
i'll have to work on it too then.
so have I forgotten anything in my schematic? I will swap the resistors in the schematic. a 12au7 can also be used in place of the 6sn7 in the phase inverter?
i'll have to work on it too then.so have I forgotten anything in my schematic? I will swap the resistors in the schematic. a 12au7 can also be used in place of the 6sn7 in the phase inverter?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.