SE/Parallel Lonestar-based build questions

[Gainzilla]

Hi All,

10 or so years ago I built an AX84 High Octane. It was a fun little build and I learned a lot in the process, but it's truly a rats nest in there. I've been wanting to re-wire it properly for a while now. I've also been wanting a clean(ish) amp that I can switch over to so I can keep my other single channel amp optimized for high gain (hehe).

I've cobbled together a re-wire of the amp as a proof of concept but found the volume was significantly lower than before. I assumed it was due to loading down the first stage with the tone stack. Since I had a spare triode, I wired it up as a cathode follower after the first stage and fed the tone stack from there. It was late last night when I checked but the volume issue seems to be much better now. Tone seems to be relatively close too.

Overall I'm pretty happy with the tone, but plan on adding a loop, and possibly playing with the power section for more flexibility (Thinking about parallel EL84s and I would switch out the ground to one for half power, or possibly going octal and swapping out different tubes for flexibility/tone…).

Anyway, my questions are mainly:

• Does my layout look ok? This is my first "real" layout. Any glaring issues so far? I'll add the power supply and power amp next.
• I was hoping to run 2 el84s cathode biased, but individually cathode biased (does that make them parallel?), then lift the ground on one of the cathodes for half power SE mode. Is that crazy?
• The power supply is from an OLD P1 High Octane document that I can't even find in the archives over at AX84. It uses a tube rectifier. I want to reclaim the spot in the chassis for the other EL84. Can I just replace the Tube with a couple diodes on the board? I've seen one on each side, serial pairs on each side... Can I just pop in a couple of diodes or should I pattern it after something specific?

Thanks in advance, and Cheers!

[eleventeen]

Some, probably not all your Q's:

• Does my layout look ok? This is my first "real" layout. Any glaring issues so far? I'll add the power supply and power amp next.
• Seems OK.
• I was hoping to run 2 el84s cathode biased, but individually cathode biased (does that make them parallel?),
• No. If you mean two separate cathode resistors...I don't imagine that accomplishes a whole lot of anything other than letting you use slightly lower power resistors. The tubes would still be in push-pull config, the only thing changed would be that they each get their own private cathode resistor. I don't see that doing anything, assuming the current through each tube stayed about the same.
• No then lift the ground on one of the cathodes for half power SE mode. Is that crazy?
• Without additional switching, you are going to be completely missing one phase of your input signal as produced by the phase splitter. Offhand I don't know of a down and dirty way to do this (eg; change a push-pull output stage to SE) but I'll bet some other forum members might have some ideas. My basic point is that if you do this as you propose, after the phase splitter, you'd be missing one phase. So you would have to also take a signal before the signal is split and use that to drive your single output tube. Sure seems possible, but clearly, two poles of switching would be needed.
• The power supply is from an OLD P1 High Octane document that I can't even find in the archives over at AX84. It uses a tube rectifier. I want to reclaim the spot in the chassis for the other EL84. Can I just replace the Tube with a couple diodes on the board? I've seen one on each side, serial pairs on each side... Can I just pop in a couple of diodes or should I pattern it after something specific?
• In general you can change a tube rectifier to full-wave dual diode. Sometimes you have to take precautions as your B+ might rise as much as 50 volts using (SS) diodes. 
  
  Concept-wise, from your "plan" for adding a cathode-follower to make up for tone-stack loss-of-gain, you will not be getting any additional gain from a CF as the gain is always less than 1 from a CF stage. CF's are generally used for either impedance matching or boosting output current from the prior stage. You need more volts and won't get any more volts from a CF.
  
  

[Gainzilla]

Thanks for the thoughtful reply, eleventeen. Ok, in retrospect I might be trying to do too much with this little amp. The thing about the power section is that, while I like the sound of the amp as it is, I don't think it's going to get very loud and stay very clean. The dilemma seems to be:

• If I convert from tube rectifier to solid state, I can either reclaim the socket and build a 15 watt push pull...
• OR, wire them in parallel for something closer to 10 or 12 watts. Something tells me that the parallel won't stay clean any better than a single SE tube.
• OR, swap the socket for an octal socket for a SE EL34, 6L6, or KT88 build, which would help keep some of the clean headroom.

Any thoughts about this? I'm not really expecting to compete with a loud drummer (I have other amps for that). I would like to be able to use it as the clean amp in a compact recording/rehearsal rig. I'll admit, the idea of swapping octal tubes seems a little appealing.

Good to know about the voltages when switching to diode rectification.


Cheers!

[HotBluePlates]

... while I like the sound of the amp as it is, I don't think it's going to get very loud and stay very clean. ...

It's a Champ, but using an EL84. How loud & clean do you really need it to be? If much, much louder you probably need a push-pull pair of something.

... The dilemma seems to be:

• If I convert from tube rectifier to solid state, I can either reclaim the socket and build a 15 watt push pull...

Have you played a Princeton Reverb? It won't be any louder/cleaner than that amp. Call it maybe-12w (to give an honest power output rating). If your supply voltage is down in the 300v range, it might be a bit less than that.

   

• OR, wire them in parallel for something closer to 10 or 12 watts. Something tells me that the parallel won't stay clean any better than a single SE tube.

The single EL84 won't be but ~4w (honest, clean output power). A parallel pair of EL84 could only be as much as 8w, then. You'd need to switch output transformers (or otherwise halve the primary impedance) to get the doubling of output power to 8w when you add the 2nd EL84.

Of course, you'd still need to buy another output transformer to switch to the push-pull arrangement, so maybe that's not a barrier. Just as long as you realize adding the extra tube won't do the job on its own.

 

• OR, swap the socket for an octal socket for a SE EL34, 6L6, or KT88 build, which would help keep some of the clean headroom.

Yes, but the new output transformer case still applies. And does your existing power transformer have enough 6.3vac current to light the extra EL84 or the EL32/6L6/KT88? Does its high voltage winding have extra current reserve to deliver the doubled current that would have to be drawn for double-output-power?

Just reiterating that changing/adding tubes is not the only issue to consider.

I hadn't really looked at the layout, as the schematic may change. And the schematic's indication of a SE EL84 and a push-pull stage with a long-tail inverter just seemed like the choice hadn't been made.

Have you tried the Lonestar's Presence control before and know you like it? It looks like a glorified treble-cut to me; it might be hard to know in advance if you really want/need it.[/list]

[Gainzilla]

Hi HotBluePlates,


I don't need it too terribly loud. Loud enough to put a mic on it is ok. I don't need to keep up with a drummer. I would probably try and coax some cranked overdriven sounds for recording out of it though... Seems to do those pretty well too. :D


So if it's a champ, and I'm trying to build a single ended lodestar, should I consider adding some negative feedback? It's a little ragged cranked wide open... maybe even blocking distortion territory.


I'm actually using the Hammond 125E, which is a P/P Tranny that can handle SE as well. Here is a link to the specs: http://angela.com/hammondoutputtransformer125e.aspx
For Power, I'm using a Hammond 269Ex: http://www.mouser.com/ProductDetail/Hammond-Manufacturing/269EX/?qs=%252bHhoWzUJg4JFeqMvbuL8ng==]
 
Any thoughts, given my current transformer set?


[EDIT] Forgot to mention, no I haven't tried the presence yet. Should be easy enought to wire in though and test. I don't have the loop yet either. Not really sure if i should go for the full enchelada or just go passive. My Analog delay has a switchable buffer, which might be enough to keep the signal intact. I've never actually used a passive loop before so I have no Idea what to expect.


Cheers!

[HotBluePlates]

I'm actually using the Hammond 125E ... For Power, I'm using a Hammond 269Ex ...
 
Any thoughts, given my current transformer set?

EL84's draw 0.76A of heater current per tube. Your 269EX has 2.5A available.

2x EL84's need 1.56A to light the tubes, which leaves 0.98A. The 12A_7 tubes draw 300mA each to light them, so you have enough for 3x of those (you can probably add a 4th without concern, even though it technically exceeds the rating). That means you're good-to-go for push-pull EL84's and the preamp & add-ons you've chosen.

A 6V6 draws only 0.45A of heater current per tube, but a 6L6 takes 0.9A, an EL34 takes 1.5A, and a KT88 takes 1.6A. It looks like with a KT88, you still have enough heater current left for 3 12A_7's.

You'd ideally idle a single-ended tube at 100% dissipation; then an ideal load impedance would let the tube swing up to double-idle current and down to just touch zero current, in order to get maximum power output.

Your 269EX is 380v center-tapped, which is the same as saying 190-0-190v. If you use a conventional full-wave rectifier the most d.c. output voltage you can get is 190v * 1.414 = ~268vdc.

The top of the 2nd page of the KT88 data sheet shows a condition with 250v plate and screen where idle plate current is 140mA, which is 250v * 0.14A = 35w idle dissipation (100% of Design Max, but less than Absolute Max). But the 269EX is rated for 71mA from the high voltage winding, and the 125E is rated for 80mA maximum d.c. So you'd need bigger transformers for both the power and output unless you idle the KT88 at 50-60% of its dissipation rating, which will seriously zap its power output.

Likewise, the EL34 data sheet shows 100mA of idle plate current at 250v plate and screen. The 6L6GC data sheet looks promising with a Class A condition at 250v plate and screen which draws 72mA of idle plate current, until you motice that 72mA times 250v is 18w, that the condition was copied almost exactly from the original 19w metal 6L6 from 1937 and that you "30w 6L6GC" is only giving you 6.5w of output under these conditions.

I haven't made a bunch of big-output-tube SE amps, so I don't have a memorized set of good pairings of power and output transformers. If you were dead-set on using the big tubes, the approach would be to look at available output transformers which can handle the likely power output you'd wind up with and see what amount of idle current they can handle (this is a key rating for a SE output transformer). If the current limit is low, you will need a little higher supply voltage (power = voltage * current; you're shooting for a set maximum dissipation). Then you look for a power transformer which will deliver the voltage and current dictated by your choice of output tube and available output transformer.

If available power and output transformers allow, it is always best to choose the lowest supply voltage you can use. When supply voltage is low, your bias voltage will be lower for a target current. When your bias voltage is small, it makes the tube easier for the preamp to drive, and leaves you more options. Since you will be running SE and therefore Class A, you are best off using cathode bias and a small bias voltage means less reduction of the supply voltage left across your tube. And with low supply voltage, your ideal load impedance will stay low and more likely to be available in off-the-shelf transformers.

We haven't even gotten into selecting a load impedance, but I think you'll see the big issues to settle first are "what tube?" & "what transformers are out there?". Or maybe the question is whether it's worthwhile to change...

Hammond's Transformers (since they have the broadest selection with full published specs)

(NOTE: I've made mention of "ideal load" and "max power output" but an amp will work even if you have neither of these. But if you're changing the amp because you want more output and are paying for bigger tubes (and if you wind up needing to pay for bigger transformers), you might as well get everything you can for your investment.)

[Gainzilla]

First I'd just like to thank you (and eleventeen) for taking the time out of your life to help. That was the most insightful bit of explanation I've read on tube current draw... Most of it actually made sense. ;) the rest will provide some good study!


Ok, so I've read that some builders (Bruce from Mission amps to be specific) has said the 125E is overbuilt and can reliably take significantly more than it's rated for. (http://archive.ampage.org/threads/1/gadc/039055/6550_SE_Stereo-2.html#040199) That said, I get now how it's probably only a good idea (if even that) to push it in both PT and OT (ideally you would probably over spec). I know there is a SE el34 High Octane on the AX84 site, but after reading your response, it's clearly not running at its full potential.


I looked at a few data sheets... The kt66 looks borderline at 250v/85ma. That's clearly over the rating of both, but arguably still doable at a lesser % of its dissipation rating (not sure how to calculate that, or even how many SE watts I could expect).


In looking at P/P, it looks like 6v6 is the better option , both in terms of voltage, and clean tone (I've not used them really so I don't have a baseline. I think I've read that el84s are quicker to breakup?). That being the case, is what I'd be building be less of a smaller lonestar, and more of a Princeton FX (fx loop instead of reverb?)


Any thoughts on the kt66, and the 6v6 vs el34 debate? I suppose the option of running parallel el84s exists as well? If so, how many watts would that get me (that's the simplest to implement short of staying put).


Thanks again!

[HotBluePlates]

Ok, so I've read that some builders (Bruce from Mission amps to be specific) has said the 125E is overbuilt and can reliably take significantly more than it's rated for. (http://archive.ampage.org/threads/1/gadc/039055/6550_SE_Stereo-2.html#040199)

It may not be obvious, but when they mention Bruce in their discussion some details imply Bruce used the 125E as-designed in push-pull operation.

Single-ended transformers are made differently than push-pull, and have different demands. Comparing apples-to-apples, look at the push-pull 125 OT's vs the single-ended 125SE OT's and compare specifically the 15w 125E to the 15w 125ESE. The SE transformer is twice the weight, meaning twice the core-size. Push-pull operation balances the d.c. pulled through the primary and allows a smaller core size for the same power through-put.

"But I've been using a push-pull OT for a single-ended circuit, and it's worked fine." Yes, it will work. It some ways it may sound cool. But if you have a serious power target you need to hit and you want full bass response, you'll need a purpose-built SE transformer. You can drive a V8 with 6 spark plugs pulled, 4 flat tires and only gears 1 & 2 but it won't deliver what it should be able to do. The difference with a car is you know from long experience what to expect.

In looking at P/P, it looks like 6v6 is the better option , both in terms of voltage, and clean tone (I've not used them really so I don't have a baseline. I think I've read that el84s are quicker to breakup?).

The 6V6 and EL84 both have a 12w plate dissipation rating (late data sheets upate the 6V6 to 14w, but the rationale for doing so would also allow you to uprate an EL84 to 14w). The available power output will be some percentage of the tube plate dissipation rating, so you should use that as a guide for selecting the tube. A Champ is a "4w amp" so you might say SE 6V6 gives 4w (from a 12w tube), so we might say SE Class A yields about 1/3 plate dissipation as output power. That means SE EL84 also gives 4w.

Class A has a theoretical efficiency of 50%, so a 12w tube should give 6w output power. But reality is 4w. When you move to push-pull Class A, you get a bit more output power (maybe in the 40% range), so a reasonable expectation for 6V6's or EL84's is 12w * 2 * 0.4 = ~10w. "But a Deluxe Reverb is a 22w amp. What gives?" Rules change when you move to Class AB. Class B has a theoretical efficiency around 70%, and AB is something between A and B, so there's a wide range the power output could fall. It also depends on the specific load impedance and supply voltage used, so the same tubes in Class AB could give differing power outputs in different amps.

About the "EL84's are quicker to beak up" bit:
EL84's have higher transconductance than 6V6's (common data sheet Gm #'s are 11.3mA/V for EL84, 4.1mA/V for 6V6) which means for a given plate current change it takes less control grid voltage change for the EL84. A practical result of this is that for the same idle condition (say both tubes single-ended with 250v plate and screen) the EL84 requires less bias voltage to land on the target current (~9.3v for 250 plate & screen vs 12.5v for the 6V6). The input voltage to drive an output tube to its maximum clean output power happens when the input signal drives the control grid momentarily to 0v or a little before reaching this point. So using the numbers given before, a 9.3v peak input signal drives the EL84 to maximum clean output while it takes 12.5v peak to drive the 6V6 to the same point.

Since both tubes give the same output power, it sounds like the EL84 "breaks up quicker" or the 6V6 "stays clean longer" only because you have to turn the volume control to a higher setting to get the 6V6 to distortion.

That said, the 2 types have different shapes to their characteristic curves, so it is still true the EL84 sounds different from the 6V6. But aside from the different tube socket, bias considerations and that the 6V6 needs a bigger drive signal, the tubes are interchangeable in a given amp.

[HotBluePlates]

I looked at a few data sheets... The kt66 looks borderline at 250v/85ma. That's clearly over the rating of both, but arguably still doable at a lesser % of its dissipation rating (not sure how to calculate that, or even how many SE watts I could expect).

Well, the MO Valve data sheet that condition comes from (towards the top of page 2) says the plate dissipation rating (Pa max) is 25w. Power = Voltage * Current, so 250v * 0.085A = 21.25w. They're idling a bit under maximum dissipation.

Assuming a perfect tube, maximum power output for class A happens when you divide supply voltage by plate current, or 250v/0.085A = ~2.9kΩ. But tubes are perfect devices, and Marconi found a happy condition listed here with 2.2kΩ. Notice the bias listed is -15v; If you use cathode bias, this is the voltage across the cathode resistor and the tube feels the supply voltage minus the bias voltage. This implies 250v + 15v = 265v, and you can get almost 268v with a solid-state rectifier, so this condition looks perfect for your power supply.

Yes, the current is higher than the rating of the transformer, but you'll draw less than full heater current and that may keep the voltage from sagging too much. Don't bother with a rectifier tube, as you're not gonna get useful sag from it in a small class A amp.The only downside is the power output is only 7.25w at 9% distortion, which will approach double-power but may not be enough for you (or maybe it's just perfect).

2.2kΩ is not available from the 125E, but if you have a 4Ω speaker load, 2.1kΩ is and would be ideal. You could get a 4Ω load by having 2x12" 8Ω speakers (maybe mix & match 2 Brit-style speakers?).The data sheet says idle current is 85mA for the plate and 6.3mA for the screen, so 15v/0.913A = 164Ω, and Marconi says use a 160Ω cathode resistor. 15v * 0.0913A = ~1.4w, so a 3w or higher resistor works well. The sheet doesn't say, but you'll want a 50uF bypass cap or higher (but only need a 25v rating).

So it looks like you picked a winner. 7.25w might seem small, but using 2 speakers will pump up the loudness for that power level.

[Gainzilla]

Apologies for the slow reply, as I've been busy with "real" life. lol


Thanks again for your help. Sadly, I only have 8 and 16 ohm speakers. I have to say, in my gut I feel like I should bite the bullet and swap over to PP 6V6. Both for less harsh tonality and for more clean headroom, as well as lower power requirements. So sounds like I'm building the Princeton Reverb after all.


If I do go that route, which do you recommend? Cathode or fixed bias. Fixed seems like the easier path, but is there a trade off tonally?


Thanks again!

[HotBluePlates]

... Sadly, I only have 8 and 16 ohm speakers. ...

A pair of 8Ω in parallel = 4Ω (total) speaker load.
 

I feel like I should bite the bullet and swap over to PP 6V6. Both for less harsh tonality and for more clean headroom, as well as lower power requirements.

6V6's can sound different than EL84's, but they deliver exactly the same (clean) power output. "More headroom" with a 6V6 is an illusion because although you dial the volume control higher to get the same amount of distortion (making it seem like EL84's distort sooner), the clean output power at the onset of distortion is almost identical.
 
So EL84's are easier to drive (because they have higher transconductance) but they don't have any less "real, acoustic" headroom.
 

So sounds like I'm building the Princeton Reverb after all. If I do go that route, which do you recommend? Cathode or fixed bias. Fixed seems like the easier path, but is there a trade off tonally?

Cathode bias makes sense with amps that largely stay in class A, which probably implies they have lower supply voltage. The Princeton Reverb is up around 430-450vdc, while a 5E3 Deluxe might be more like 360-380vdc (more or less), and it's still likely a class AB amp when pushed. Preamps are entirely different between these amps (as are rectifier tubes), so there's no direct tonal comparison based on supply voltage and bias method.
 
If you had a class A amp where the bias was exactly the same but you switched back-n-forth from fixed to cathode bias, you would hear essentially no sound difference. Much of what people report to hear as sonic differences are when they take an amp that has no business being anything but fixed-bias (running deep into class AB) and they try to jam a cathode resistor under it (to reduce power or whatever). In reality, they're not hearing each method as it was intended to be used.
 
So it's your call: you want a bigger amp? Then pick a model that suits you; you'll probably be buying transformers anyway. You want to stick with what you have and swap 6V6 for the existing EL84? Cool, just get a new socket and a new cathode bias resistor.

[Gainzilla]

A pair of 8Ω in parallel = 4Ω (total) speaker load.

I guess I could pillage the speakers from my Marshall... I think of that as kind of a last resort though.

So EL84's are easier to drive (because they have higher transconductance) but they don't have any less "real, acoustic" headroom.

Ok, I hear you. They break up at different points on the dial but I won't really see any increased headroom with el84s/6v6s (i.e. louder cleans).

Re: 6v6 voltage: My PT can't supply anywhere near the voltages you mentioned. And you're right, I originally put this amp together to be a low watt bedroom amp. For practice/songwriting/recording, it sounds really good. Maybe the right call here is to just wrap it up. If there is a lesson to be learned here it's likely that its better to get the collection of components that will enable the amp you want, rather than trying to make what you want fit into a given set of components. 

Anyway, I really appreciate your help. I'll post some clips soon.

Cheers,

[HotBluePlates]

... If there is a lesson to be learned here it's likely that its better to get the collection of components that will enable the amp you want, rather than trying to make what you want fit into a given set of components.

That is always a wise choice.

If you have a 4x12, have you just tried plugging your single-EL84 amp into it? You may get the extra volume you want right there...

[Gainzilla]

Made a quick sound clip today of some clean noodling (really quick. made no effort at mic placement. 1 take, no edits). Consider this the "before" recording. I've got some parts on the way and will gut/rebuild soon. It'll be interesting to see what I get on the other side.


Thoughts?


https://soundcloud.com/gainzilla/lonestar-se-amp-test#t=0:03


Cheers!