Hey
This question has been asked on diyhifi.org too, but I'm trying here cause the activity is much higher. I'm totally new to DIY hifi or tube amps. I've picked up Morgan Jones' famous book and thought I'd give the Bevois Valley amp a try. I haven't got through most of the book yet, but I have taken a sneak peek at this amp. I notice it's only 2x10 Watts which will be a little measly. I'd like to have much more. In the book the author suggests what he calls a 'cracking solution' to the power problem: Use more el84 output tubes in parallel. After some searching on the net it seems this is called parallel push-pull? Are there any disadvatages with this approach? The author mentions only advantages. I have searched and found that some component values must be reconfigured, but haven't found which ones? Are there any books/internet resources detailing how to reconfigure a push-pull to a parallel push-pull? I understand you have to get a different output transformer. The loadline of the output tubes must be reconfigured maybe?
Also, does anyone have any experience with Bevois Valley, how does it sound?
I will get through the rest of the book before I start building it, and will probably figure out the answers sooner or later, I'd just like to keep my brain pondering on the 'ultimage goal' while I learn...
This question has been asked on diyhifi.org too, but I'm trying here cause the activity is much higher. I'm totally new to DIY hifi or tube amps. I've picked up Morgan Jones' famous book and thought I'd give the Bevois Valley amp a try. I haven't got through most of the book yet, but I have taken a sneak peek at this amp. I notice it's only 2x10 Watts which will be a little measly. I'd like to have much more. In the book the author suggests what he calls a 'cracking solution' to the power problem: Use more el84 output tubes in parallel. After some searching on the net it seems this is called parallel push-pull? Are there any disadvatages with this approach? The author mentions only advantages. I have searched and found that some component values must be reconfigured, but haven't found which ones? Are there any books/internet resources detailing how to reconfigure a push-pull to a parallel push-pull? I understand you have to get a different output transformer. The loadline of the output tubes must be reconfigured maybe?
Also, does anyone have any experience with Bevois Valley, how does it sound?
I will get through the rest of the book before I start building it, and will probably figure out the answers sooner or later, I'd just like to keep my brain pondering on the 'ultimage goal' while I learn...
Hello.
I,m kind of new to this too but let's see what we can do.
First of all, if you havn't listend to low powered tube amps
don't assume that ten watts per channel is measly. My old
transistor amp put out 40 wpc channel but I wanted to
check out tube sound. I built a single ended triode amp
that puts out less than 2 watts and hooked it to original speakers. They are rated at 90db/watt so are not particularly
effiecient. It drives theese speakers plenty loud. If you do much
reading about tube amps, you will see it writen time and again
that a lower powered tube will kick *** over a high powered transistor amp.
If you really want to change the design to parralell tubes:
1) catode bias resistors need to be half the original value.
so they carry twice the current. They also will dissapate
more power so will run hotter.
2) Bypass capacitors need to be much bigger so that their
reactance at the lowest bass frequency you care about
is low compared to the new value of the bias resistors.
3) Power supply is now having to put out twice the current
(almost) so may have to be upgraded as well.
4) You will need to make sure the driver stage has enough
current to overcome the input capacitance of two tubes
or high frequncies may suffer.
I am not completly sure but I think you would use half the primary impeadeance output trafo. The project will definatly get much mmore complex and expensive. You might be quite surprised what a "measly" 10 watts can do! why not try it as
is.
Good luck.
I,m kind of new to this too but let's see what we can do.
First of all, if you havn't listend to low powered tube amps
don't assume that ten watts per channel is measly. My old
transistor amp put out 40 wpc channel but I wanted to
check out tube sound. I built a single ended triode amp
that puts out less than 2 watts and hooked it to original speakers. They are rated at 90db/watt so are not particularly
effiecient. It drives theese speakers plenty loud. If you do much
reading about tube amps, you will see it writen time and again
that a lower powered tube will kick *** over a high powered transistor amp.
If you really want to change the design to parralell tubes:
1) catode bias resistors need to be half the original value.
so they carry twice the current. They also will dissapate
more power so will run hotter.
2) Bypass capacitors need to be much bigger so that their
reactance at the lowest bass frequency you care about
is low compared to the new value of the bias resistors.
3) Power supply is now having to put out twice the current
(almost) so may have to be upgraded as well.
4) You will need to make sure the driver stage has enough
current to overcome the input capacitance of two tubes
or high frequncies may suffer.
I am not completly sure but I think you would use half the primary impeadeance output trafo. The project will definatly get much mmore complex and expensive. You might be quite surprised what a "measly" 10 watts can do! why not try it as
is.
Good luck.
I agree with rman's reply above but I would add that doubling the available power does provide more headroom for transients, which is always nice to have. Also, paralleling OP tubes, as Morgan Jones says, requires half the load impedance, meaning fewer turns on the OPT primary, which makes the OPT easier to design and make for high quality.
The only downside of paralleling that I can think of is the risk of current 'hogging' (one tube of a parallel group taking more than its fair share of the plate current). You can guard against this, to some extent, by adding a small resistor (100 ohms or so) in the plate circuit of each output tube, but matching of tubes is also important.
The only downside of paralleling that I can think of is the risk of current 'hogging' (one tube of a parallel group taking more than its fair share of the plate current). You can guard against this, to some extent, by adding a small resistor (100 ohms or so) in the plate circuit of each output tube, but matching of tubes is also important.
Thanks rman. Yes, I have read that low power rated tube amps can give plenty of power. I have read that you should multiply with approx 3 times to get tranny equivalent.
I am currently using a NAD rated at 2x50 watts. NADs are too known for feeling much more powerful than what they are rated as (by tranny standards). I LOVE the punch and dynamics I get from this amp (forceful playing, though I don't always play loud), so I reckoned I'd need a little more than 2x10 watts. Then again, I realize that this may not be a beginners project...
Yeah, I thought about what you said about the driver stage, that it may not be sufficient for parallel output tubes. Luckily, I see that Morgan Jones include a schematic for a phase splitter/driver stage especially for higher powered output stages.
I guess I'm not going to play so *loud* all the time, I just want a little headroom for demanding music with lots of dynamics. You say it will get more complicated, and I just hope with lots of computer simulations and very careful planning using Jones' two books and forums, I will succeed...
There is just one important precaution I have: I want to keep HT voltage fairly low, so it is at least safer than 500-1000 volts...
I am currently using a NAD rated at 2x50 watts. NADs are too known for feeling much more powerful than what they are rated as (by tranny standards). I LOVE the punch and dynamics I get from this amp (forceful playing, though I don't always play loud), so I reckoned I'd need a little more than 2x10 watts. Then again, I realize that this may not be a beginners project...
Yeah, I thought about what you said about the driver stage, that it may not be sufficient for parallel output tubes. Luckily, I see that Morgan Jones include a schematic for a phase splitter/driver stage especially for higher powered output stages.
I guess I'm not going to play so *loud* all the time, I just want a little headroom for demanding music with lots of dynamics. You say it will get more complicated, and I just hope with lots of computer simulations and very careful planning using Jones' two books and forums, I will succeed...
There is just one important precaution I have: I want to keep HT voltage fairly low, so it is at least safer than 500-1000 volts...
Hi there...
I've been a lurker here for a while now, this is my first post.
I've built a modified version of that amp (regulated everything, bigger coupling caps, no GNF, line stage added w/12AX7's, bass and treble control). I think the thing is the cat's pajamas. I'd also say it's loud enough for me. I regularly get complaints from my wife that she can hear it at night when she's trying to sleep 2 floors above me, and I think my neighbors can hear me, too.
Pete
I've been a lurker here for a while now, this is my first post.
I've built a modified version of that amp (regulated everything, bigger coupling caps, no GNF, line stage added w/12AX7's, bass and treble control). I think the thing is the cat's pajamas. I'd also say it's loud enough for me. I regularly get complaints from my wife that she can hear it at night when she's trying to sleep 2 floors above me, and I think my neighbors can hear me, too.
Pete
ray_moth said:
I always design with parallel UL p-p esp 100W + without any trouble. As RayM mentions, the obvious benefit is the o/p tranny, with lower leakage inductance implies a high frequency response and sure it does. It's much easier for the tranny designer.
An E & I laminated type should expect a response -3dB down at 50Khz+....compared to 35Khz with standard p-p output stage.
I've had no evidence of tube currrent hogging and there is no run-away. There will always be 5-10mA anode currrent difference caused by the gm unless one selects matched pairs or quads.
Tip: when ordering matched o/p tube pairs /quads, make sure you specify the working voltage. If you don't, the vendors will use 250V B+ which the gm ( which ends up as current match) can be miles out for 450V+.
The only real change is the mains transformer now has to supply double the current compared to a standard o/p pair.
Once you get into the amp building business one will soon appreciate how annoying it is to series-chain the psu caps to achieve a B+ headroom. By using a single 500V electrolytic for a 450V B+ in parallel push-pull will easily give 100W poke and piles of headroom.
rj
That's an interesting PPP design with current sinks under the EL84s.
I'm asking myself whether you could use one shared CCS under each pair of EL84s rather than a CCS for each. Presumably this would force class A operation, which might not be a bad thing since we are in PPP and have more power available.
Any thoughts?
I'm asking myself whether you could use one shared CCS under each pair of EL84s rather than a CCS for each. Presumably this would force class A operation, which might not be a bad thing since we are in PPP and have more power available.
Any thoughts?
CCS in output stage cathodes is fashionable, but recovery from overload is often questionable. If it were me, I'd use a more conventional bias scheme, such as the original one shown in the Bevois design. Paralleling output tubes will work very well, but they need to be matched if a single cathode RC bias network is used. Additionally, the feedback compensation may need to be slightly tweaked because each extra set of output tubes increases the output stage grid capacitance. MJ runs through the calculation for a single pair, so it's pretty easy to follow along and multiply the output stage capacitance by the number of paralleled tubes.
As for the sound... well, there's a reason that this sort of design is considered classic.
As for the sound... well, there's a reason that this sort of design is considered classic.
SY said:
I, for one, am glad we have resources such as SY and DIY to keep us well balanced between 'fashion' and performance. It can be tempting to use oodles of the latest find (CCS-10M45S-DN2540) in an attempt to maximize our designs. The tradeoff I would have never seen, but the warning was enough to make me really consider what to do.
Keeping with tradition is not the answer; knowing when to break from it IS.
$0.02
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.