Why do something simple if you can make it complicated for twice the money? Err wait, that didn't come out right...
The whole purpose of this thread is going as simple as possible while still retaining a healthy power output and great sound.
One of my monoblocks went silent a week or so ago and life without music just isn't ok. So here begins my journey to restoring music...
- Time is short
- I have virtually no tools
- I've never built anything with tubes
- Money is scarce
- I'm a chronic contrarian
Old amp: Hypex nCore monoblock
DSP: 10db Linkwitz transform to extend the low end
Speakers: 2-way, 95db sensitivity, 5-6 ohm,passive x-over (15" midrange on woofer duty).
Source: Buffalo II DAC with transformer output.
Mission statement: As simple as possible but not any simpler than that!
Objective: Single stage, push/pull, 10-15W.
There is no such thing as a free lunch and I'm very much aware that I can't get extra everything without paying a premium, so let's see how far we can get when we reduce a power amp to it's bare essentials. I'll be using SY's Red Light District amp as primary source of inspiration, unless something more suitable shows up along the way.
Since I'm a complete to0b no0b, I will gratefully welcome any help, tips and tricks you are willing to share. I expect the design to evolve along the way, so nothing is set in stone.
From what I can tell EL84 seems like a good candidate for this project. I'd like to try and keep it a single stage design with a transformer taking care of splitting and voltage amplification. Using a balanced source should help in reducing the work load for the tranny.
Ok, fire away.
The whole purpose of this thread is going as simple as possible while still retaining a healthy power output and great sound.
One of my monoblocks went silent a week or so ago and life without music just isn't ok. So here begins my journey to restoring music...
- Time is short
- I have virtually no tools
- I've never built anything with tubes
- Money is scarce
- I'm a chronic contrarian
Old amp: Hypex nCore monoblock
DSP: 10db Linkwitz transform to extend the low end
Speakers: 2-way, 95db sensitivity, 5-6 ohm,passive x-over (15" midrange on woofer duty).
Source: Buffalo II DAC with transformer output.
Mission statement: As simple as possible but not any simpler than that!
Objective: Single stage, push/pull, 10-15W.
There is no such thing as a free lunch and I'm very much aware that I can't get extra everything without paying a premium, so let's see how far we can get when we reduce a power amp to it's bare essentials. I'll be using SY's Red Light District amp as primary source of inspiration, unless something more suitable shows up along the way.
Since I'm a complete to0b no0b, I will gratefully welcome any help, tips and tricks you are willing to share. I expect the design to evolve along the way, so nothing is set in stone.
From what I can tell EL84 seems like a good candidate for this project. I'd like to try and keep it a single stage design with a transformer taking care of splitting and voltage amplification. Using a balanced source should help in reducing the work load for the tranny.
An externally hosted image should be here but it was not working when we last tested it.
Ok, fire away.
Doesn't look like you'll have enough gain from the OP tfmr. EL84's from memory are biased at about 7-8v, so that's 14-16v at least needed g1-g1, which with a tfmr of x2 gain means you'll need a music source capable of providing 8v RMS. Most line level music is 500mV -1v roughly.
An LED won't provide a bias of 7v ish, at around 1.5v bias the EL84's are going to be cooking, hope you have a few hundred spare, they won't last long.
There's also no NFB, haven't done the math's but I think the design as is will have high THD. Dare say other forum members will have more to contribute.
Andy
An LED won't provide a bias of 7v ish, at around 1.5v bias the EL84's are going to be cooking, hope you have a few hundred spare, they won't last long.
There's also no NFB, haven't done the math's but I think the design as is will have high THD. Dare say other forum members will have more to contribute.
Andy
Thanks for joining in. The LED is actually meant to be a connection to the LED array as outlined in the Red Light District amp, and it'll bias the tubes at 11'ish V. I plan on driving the amp with a balanced source, that gives us 4V input signal, and 8 Vrms at the grid (that should be enough to max out the tube). The reason I opted for a 1:2 transformer on the input was because I'd like to try and avoid issues with capacitance and impedance matching. I might end up on the low side if the source isn't up for the task, you're right on that one for sure. B+ is planned at 320V, so it'll be running hot. I honestly don't know how to implement feedback, and especially with a circuit this tiny.
Bumping the input transformer up to 1:4 could be an option, but I don't know if that'll put me in the danger zone? It would definitely help with bringing the voltage up. If I buy a 1+1:2+2 transformer, it would give me the option to try both variants.
Bumping the input transformer up to 1:4 could be an option, but I don't know if that'll put me in the danger zone? It would definitely help with bringing the voltage up. If I buy a 1+1:2+2 transformer, it would give me the option to try both variants.
Last edited:
I'm going to be a "party pooper", sorry. Without loop NFB of some kind, UL mode O/P tubes will frequently exhibit an inadequate damping factor.
By definition, a "balanced" source already splits the phases. Also, magnetics are expensive. I suggest a differential gain block at the I/P. Perhaps the so called plate to plate topology can be employed for the NFB that improves both distortion performance and damping factor.
By definition, a "balanced" source already splits the phases. Also, magnetics are expensive. I suggest a differential gain block at the I/P. Perhaps the so called plate to plate topology can be employed for the NFB that improves both distortion performance and damping factor.
No worries Eli, that's why I posted the thread to begin with. I'm looking to learn, and I suspect there are plenty of pitfalls ahead of me.
So, let's take this one step at a time. Since I'm new to this, it'll take some hand holding.
Pete Millett has a spud design that is very similar to what I suggested above, but biased differently. Could it be that those caps connected between the center taps of the transformers are giving you that feedback that you requested?
I would prefer to keep the one-stage design if possible.
If a second stage becomes necessary, I could probably go with SY's RLD or tubelab's SPP without any mods at all. The only real reason I'm trying to do this my self is to swap the tube input stage vs a transformer. Iron is expensive but it is also idiot proof and there is zero upkeep.
So, if you take a peak at Pete's design, would that satisfy the requirements? Could I work with that? The drawback is a different bias and I was hoping to give the diode bias a try. Nothing is decided though, and I'll adapt to whatever topology is needed if I have to.
So, let's take this one step at a time. Since I'm new to this, it'll take some hand holding.
Pete Millett has a spud design that is very similar to what I suggested above, but biased differently. Could it be that those caps connected between the center taps of the transformers are giving you that feedback that you requested?
I would prefer to keep the one-stage design if possible.
If a second stage becomes necessary, I could probably go with SY's RLD or tubelab's SPP without any mods at all. The only real reason I'm trying to do this my self is to swap the tube input stage vs a transformer. Iron is expensive but it is also idiot proof and there is zero upkeep.
So, if you take a peak at Pete's design, would that satisfy the requirements? Could I work with that? The drawback is a different bias and I was hoping to give the diode bias a try. Nothing is decided though, and I'll adapt to whatever topology is needed if I have to.
Pete employed triode wired "finals". There is no loop NFB in that setup. Also, the EL802 is rated for a 6 W. plate dissipation. The 6BQ5/EL84 is rated for a 12 W. plate dissipation. PP UL mode EL84s will comfortably yield 15 W.
AFAIK, both the RLD and the SPP are configured for an "unbalanced" I/P. To use a "balanced" I/P, some modification is required.
AFAIK, both the RLD and the SPP are configured for an "unbalanced" I/P. To use a "balanced" I/P, some modification is required.
Aha, so there is such a marked difference between triode wiring and UL operation? On paper they look pretty similar, so I didn't realize it would make such a big difference.
In itself there is no real value in balanced connections for home equipment like this. The reason I've stressed the use of balanced connections is because it allows you to get higher input voltages "for free". For a commercial/public unit, it makes much more sense to have both single ended and balanced connections interchangeable.
So, I guess in the end it looks like I have a choice to make. Build a single-stage amp with lower output, or accept a two-stage amp with higher output.
And if I remember correctly, the RLD is wired for pentode operation? I don't know what would be needed to be done in order to change over to UL? The SPP is already set up for UL, and there are prefabbed PCB available. The SPP looks like a faster and easier build since you get most of the planning served on a platter.
If it wasn't for me fearing 6W being too low, I would have gone with a smaller amp. Now... I dunno?
It's not that I need gobs of power, I just need it to be enough. If anything, the closer to "just enough" the better. Decisions decision... ;D
In itself there is no real value in balanced connections for home equipment like this. The reason I've stressed the use of balanced connections is because it allows you to get higher input voltages "for free". For a commercial/public unit, it makes much more sense to have both single ended and balanced connections interchangeable.
So, I guess in the end it looks like I have a choice to make. Build a single-stage amp with lower output, or accept a two-stage amp with higher output.
And if I remember correctly, the RLD is wired for pentode operation? I don't know what would be needed to be done in order to change over to UL? The SPP is already set up for UL, and there are prefabbed PCB available. The SPP looks like a faster and easier build since you get most of the planning served on a platter.
If it wasn't for me fearing 6W being too low, I would have gone with a smaller amp. Now... I dunno?
It's not that I need gobs of power, I just need it to be enough. If anything, the closer to "just enough" the better. Decisions decision... ;D
"The LED is actually meant to be a connection to the LED array as outlined in the Red Light District amp" That makes more sense, sorry, not familiar with that amp.
"Aha, so there is such a marked difference between triode wiring and UL operation?2 I n a nutshell you get more power out with UL connection, it's supposed to be the best of both worlds, a lot of folk don't like it for some reason.
"I don't know what would be needed to be done in order to change over to UL?" Just remove any dropping resistors and caps and connect to the tfmr taps, it's pretty straight forward although you might need to put a resistor as near as poss to the valve base pin to stop oscillation and to make sure G2 is within specs.
"If anything, the closer to "just enough" the better" It's always good to have a bit more power or headroom, the nearer an amp gets to clipping the more distortion usually.
With any design you'll need to adjust, chop and change as you test on the bench, a tfmr with various ratio's would be a good idea.
Andy.
"Aha, so there is such a marked difference between triode wiring and UL operation?2 I n a nutshell you get more power out with UL connection, it's supposed to be the best of both worlds, a lot of folk don't like it for some reason.
"I don't know what would be needed to be done in order to change over to UL?" Just remove any dropping resistors and caps and connect to the tfmr taps, it's pretty straight forward although you might need to put a resistor as near as poss to the valve base pin to stop oscillation and to make sure G2 is within specs.
"If anything, the closer to "just enough" the better" It's always good to have a bit more power or headroom, the nearer an amp gets to clipping the more distortion usually.
With any design you'll need to adjust, chop and change as you test on the bench, a tfmr with various ratio's would be a good idea.
Andy.
UL has been called partial triode. Triode mode "finals", without loop NFB, can exhibit an adequate damping factor in combination with a fair number of speakers. However, plenty of speakers, particularly of the sealed variety, need damping factor galore. Triode mode "finals" will need help from NFB, to work well in combination with the "trouble makers". Proper matching of amps and speakers is a huge part of getting good sound.
You need to oversize the outputs to have a comfortable power triode wired . EL34 as Marantz 8b is possible . The bellow circuit is from @Salas Kofi Annan in: "Tube Amp for Multi-Way Speakers"
Attachments
After careful consideration, I think going with Tubelab's SPP is the best option. So, I'll be sending George an email...
I couldn't find where/how to edit my previous post? Anyway, where do you go for parts these days? Is there any Asian one-stop-shop where you can get most of what you need in one go and at a great price?
Living in Europe makes buying anything from the US horribly expensive. I'm not joking, the last items I looked up would cost me twice as much as the list price due to shipping, VAT, customs duties, etc.
I'm thinking James transformers might be a good option if ordered directly from the source. But the rest of the parts???
I couldn't find where/how to edit my previous post? Anyway, where do you go for parts these days? Is there any Asian one-stop-shop where you can get most of what you need in one go and at a great price?
Living in Europe makes buying anything from the US horribly expensive. I'm not joking, the last items I looked up would cost me twice as much as the list price due to shipping, VAT, customs duties, etc.
I'm thinking James transformers might be a good option if ordered directly from the source. But the rest of the parts???
Pretty sure James transformers have been OOP for some time. What about Lundahl? Lundahl Transformers • Audio Transformers • Home
jeff
Have they? I just checked their web page out the other day... :/ Ah well, life goes on. Lundahl are awesome, and they are located almost next door. The only negative part is that they don't offer "pretty" transformers that you mount on top. Everything is industrial look for built-in purposes. If I could get Lundahls in a nice looking packaging, that would be my go to choice without a doubt. I was hoping to mount all the Iron on the outside of the chassis...
Edit. My bad, it appear they now sell "covers" for their tube amp transformers, I'll have to look into that.
Edit. My bad, it appear they now sell "covers" for their tube amp transformers, I'll have to look into that.
You have a limited time (15 minutes??) to edit your words here.
Apparently in the past some folks mis-used the Edit function and made trouble.
A pair of low cost 6LU8 tubes can make for a low cost but high quality amplifier.
Vertical deflection tubes for audio output?
tony's 6LU8 push-pull power amp build
Vertical deflection tubes for audio output?
tony's 6LU8 push-pull power amp build
And with a pentode connection it becomes a current amp.
While Eli’s use of the word “inadequate” applies to 99.9% of loudspeakers but not universally. Some speakers work better with highish output impedance amplifiers (like a UL connection), and a few are happy with current amps.
Flat frequency response requires a quite flat impedance curve on your loudspeakers (or a rise at the extremes where normally the response would start to slump).
What does the impedance of your speakers look like?
I wouldn’t hesitate to connect the EL84s in triode, should be enuff for your speakers. And that would make the amp simplier… is the OPT from your stash?
dave
How to describe my speakers... ehm... There was a fairly well known speaker project some years back, but I can't for the life of me remember any names.
It's basically a 15" mid/bass + PA tweeter in a wave guide with a passive x-over in the 1200-1300 Hz area (AE TD15M, SEOS15). The impedance is surprisingly flat (stays pretty well between 5-6 ohm), nothing is perfect but really nice for such a simple concept.
What sets them apart from other siblings and the original project, is the box and using a DSP to extend the low frequency response (Linkwitz 10db). The boxes are sealed and as small as you can make them and still fit the 15" driver. Think slightly larger than a 15" cube...
Heavily braced internally, but with thin plywood walls in a sandwich construct. Works like a charm, no problems what so ever with resonances or anything. And I can easily move them around if needed.
Googling the TD15M, I see now the 1WSPL is 97.8dB.
The only reason I've been so focused on getting the wattage up, is because of the frequency compensation. With compensation, they stay flat down to 35'ish Hz, without I think they start rolling off at 70-100 something Hz. Maybe one could look at it like them needing close to the same power as a pair of speakers with 88dB 1WSPL, due to the compensation (since most on the power is used by the low end).
So, how much power do I really need? I dunno? I've been using Hypex nCore monoblocks and for this particular application, that can be equated to unlimited power more or less. I quit long before speakers or amps quit...
If I could get away with a smaller amp, that would suit me perfectly. ;D
It's basically a 15" mid/bass + PA tweeter in a wave guide with a passive x-over in the 1200-1300 Hz area (AE TD15M, SEOS15). The impedance is surprisingly flat (stays pretty well between 5-6 ohm), nothing is perfect but really nice for such a simple concept.
What sets them apart from other siblings and the original project, is the box and using a DSP to extend the low frequency response (Linkwitz 10db). The boxes are sealed and as small as you can make them and still fit the 15" driver. Think slightly larger than a 15" cube...
Heavily braced internally, but with thin plywood walls in a sandwich construct. Works like a charm, no problems what so ever with resonances or anything. And I can easily move them around if needed.
Googling the TD15M, I see now the 1WSPL is 97.8dB.
The only reason I've been so focused on getting the wattage up, is because of the frequency compensation. With compensation, they stay flat down to 35'ish Hz, without I think they start rolling off at 70-100 something Hz. Maybe one could look at it like them needing close to the same power as a pair of speakers with 88dB 1WSPL, due to the compensation (since most on the power is used by the low end).
So, how much power do I really need? I dunno? I've been using Hypex nCore monoblocks and for this particular application, that can be equated to unlimited power more or less. I quit long before speakers or amps quit...
If I could get away with a smaller amp, that would suit me perfectly. ;D
Nice find, that does look interesting. I would have to take some time to study the schematics, but it definitely looks interesting. It would make the tube total 2 tubes per channel, manageable heater current, and super simple PSU. If anyone has heard this build and can vouch for the sound quality, it would certainly be a contender. It must be the simplest schematics I've seen while still packing a respectable punch.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.