Jenghis, did you read through the thread I linked? Among the things mentioned is setting the sockets up so that both the 6BQ5/EL84 and the Russian 6Π15Π (6p15p) work. 
A quick check of EBone shows several attractive listings for excellent EB (ev) suffix tubes.
There's plenty of chassis space for 2X new O/P trafos and a 0A2 gas discharge regulator that makes 5 WPC full pentode mode "finals" reasonably linear, open loop. The GNFB loop yields an adequate damping factor and linearizes the modest cost Edcor O/P "iron".
A quick check of EBone shows several attractive listings for excellent EB (ev) suffix tubes.There's plenty of chassis space for 2X new O/P trafos and a 0A2 gas discharge regulator that makes 5 WPC full pentode mode "finals" reasonably linear, open loop. The GNFB loop yields an adequate damping factor and linearizes the modest cost Edcor O/P "iron".
The 6P15P-EV that Eli D refers to in the previous post is a great tube, but has lower maximum anode and 2nd grid voltages at 330 and 330 Volts respectively. The 6P14P-EV is basically the Russian version of the EL84M, and is useable at 400+ Volts. If you keep your voltages below the 330 Volt maximum of the 6P15P-EV, it is hard to beat. This is especially true at its very low cost of around $3 per tube online. This is the tube that Decware ships it’s SE84 amps with.
The key behavioral difference between the 6Π15Π and the EL84 is screen grid fragility. Notice that 150 V. g2 limit. Don't wire the 6Π15Π in "normal" ultra-linear (UL) mode.
Max. open loop linearity in full pentode mode is obtained by regulating g2 B+ at a fraction of anode B+, regardless of how tough g2 is. The "150" V. 0A2 gas discharge regulator combines very well with 2X (1/channel) "12" W. multi-grid power O/P tubes' screen grids. In particular, it or another form of regulation is crucial, when the 6Π15Π is employed.
When g2 B+ is regulated, some liberty can be taken with anode B+ limits. The key to reliability and longevity is RUTHLESSLY enforcing the anode dissipation limit.
Heat, not mere voltage, is what causes damage and, ultimately, failure. In the thread previously linked, I added additional modifications to a console amp that had been modified by Mr. Gillespie. An important feature Mr. Gillespie provided is adjustable self bias. The exact voltage a PSU produces can only be determined by bench measurements. Within reason, the specific value is less important than knowing it. Applying P = VI to the voltage on the plate and the cathode current yields a favorably "fudged" value for the plate dissipation. 
Remember, g2 contributes to the cathode current. Adjust each O/P tube's bias to obtain a "fudged" 12 W. or slightly lower dissipation.
Max. open loop linearity in full pentode mode is obtained by regulating g2 B+ at a fraction of anode B+, regardless of how tough g2 is. The "150" V. 0A2 gas discharge regulator combines very well with 2X (1/channel) "12" W. multi-grid power O/P tubes' screen grids. In particular, it or another form of regulation is crucial, when the 6Π15Π is employed.
When g2 B+ is regulated, some liberty can be taken with anode B+ limits. The key to reliability and longevity is RUTHLESSLY enforcing the anode dissipation limit.
Heat, not mere voltage, is what causes damage and, ultimately, failure. In the thread previously linked, I added additional modifications to a console amp that had been modified by Mr. Gillespie. An important feature Mr. Gillespie provided is adjustable self bias. The exact voltage a PSU produces can only be determined by bench measurements. Within reason, the specific value is less important than knowing it. Applying P = VI to the voltage on the plate and the cathode current yields a favorably "fudged" value for the plate dissipation. Remember, g2 contributes to the cathode current. Adjust each O/P tube's bias to obtain a "fudged" 12 W. or slightly lower dissipation.
Yes, the power supply already has a rectifier tube.
Wading through Mr McShane's semi-random listings for tube rollers is a way to spend a pleasant few minutes or hours.
Shipping costs into Canada from the US are very high; even within Canada it makes a huge difference if there's a local supplier.
The Tubestore in Ontario has the JJ tube for $12USD
www.thetubestore.com - 6CA4 / EZ81 Tube Types
Long&McQuade is a national music store 'here'; not the first place I'd shop for tubes, but since it's local to many people the savings on shipping are a plus..
I believe that the OP also is local to Pacific TV (John Albion)- another good tube supplier.
Pacific T.V. Tube Catalog
Last edited:
Jenghis, did you read through the thread I linked? Among the things mentioned is setting the sockets up so that both the 6BQ5/EL84 and the Russian 6Π15Π (6p15p) work.
Yep, I had a read of it, some very helpful points. For now I think I'll leave the tube sockets alone and just try to build a functioning amp. I'm assuming they can be modified down the line if needed
Indeed I am. I talked to John Albion several years ago in fact.
I think what Eli is referring to is the use of Pin 1.
In the EL84/6BQ5 pin 1 isn't connected to anything inside the tube. Sometimes amp builders will use those 'unconnected' lugs on the tube socket as connection points for joining components. So if there is any chance of ever installing a 6P15P tube in the 6BQ5 socket, you want to make sure the #1 lug on the tube socket is unused.
Attachments
It looks like the Pin 1 lugs aren't being used, so that's good.
You will notice that the chassis is being used as a circuit ground connection - the metal lugs on the tube socket attachment are used as connection points.
(That's why those lugs were there on older sockets... ).
Nowadays most builders separate the circuit ground from the chassis in a more deliberate way.
Get a good soldering station/iron with a medium chisel tip, lots of solder and solder wick for the disassembly steps!
It's excellent practice.
You will notice that the chassis is being used as a circuit ground connection - the metal lugs on the tube socket attachment are used as connection points.
(That's why those lugs were there on older sockets...
).Nowadays most builders separate the circuit ground from the chassis in a more deliberate way.
Get a good soldering station/iron with a medium chisel tip, lots of solder and solder wick for the disassembly steps!
It's excellent practice.
Attachments
Last edited:
Ah that makes sense, I thought 6P15P would require a modified socket. I'll reread the thread Eli linked. The EH EL84 tubes I have now sound fine, and are almost brand new, so I'll use them for initial tests.
It looks like the Pin 1 lugs aren't being used, so that's good.
You will notice that the chassis is being used as a circuit ground connection - the metal lugs on the tube socket attachment are used as connection points.
(That's why those lugs were there on older sockets...
Nowadays most builders separate the circuit ground from the chassis in a more deliberate way.
Get a good soldering station/iron with a medium chisel tip, lots of solder and solder wick for the disassembly steps!
It's excellent practice.
I don't know if I like the idea of the chassis itself used as a circuit ground, especially since the ground wire isn't directly coupled to the chassis. I think I'll use the terminal strips to join all the circuit grounds to a common safety ground, wtfamps has a good article on properly grounding an amplifier.
It'll probably be almost as much work to disassemble the amp as it will be to reassemble the new version
Not only must pins 1 and 6 be kept clear. One of those pins has to be jumpered to pin 3 (the cathode). Unlike the EL84/6BQ5, the 6Π15Π's suppressor grid (g3) is not internally connected to the cathode.
Sockets correctly configured for the 6Π15Π accept EL84s, without incident. The other way around does not work. The 6Π15Π-EB (6p15p-ev) is a very nice tube and (at least for now) favorably priced. JMO, having flexibility at your disposal is a good thing.
So if I jumper pin 1 or pin 6 to pin 3 on the socket, does that make the 6P15P a drop in replacement for the EL84? I'll have a closer read of the thread you linked and see if I can find the reference to the 6P15P.
You'll need to pay attention to the lug strips since often one or more of the lugs are connected to the attachment bolts. I usually put a 'reminder' over the lug if I'm likely to use it without thinking about the chassis connection.
You will find lots of (often conflicting) info on where that chassis connection should be for the circuit ground, and how it should be made. Brace yourself!
In the Boyuu thread it went on for a while...battle of the experts...
It's a good opportunity to see how it was done originally ('properly') and also a chance to judge the quality of the 'mod' work done subsequently.It'll probably be almost as much work to disassemble the amp as it will be to reassemble the new version
Simply installing a proper jumper is insufficient. Remember, g2 in the 6Π15Π is fragile. In the linked thread, full pentode mode "finals" are employed in combination with g2 B+ regulated at "150" V. A 0A2 gas discharge tube was selected as the regulator, for both convenience and cosmetics. Your chassis currently contains no SS and gas discharge regulation will keep things that way. BTW, I'm not anti-SS, but, some times, all tubes is nice.
The adjustable self bias idea that originates with Mr. Gillespie allows the amp to be correctly configured for the specific O/P tube specimens that are employed.
I'm very impressed with the "El Mighty Cacahuate", which I hadn't seen before. The whole page is well thought out, well written and essentially error-free, and that combination pretty much *never* happens in our hobby. And, it's an elegant engineering design for homebrewers. I intend to recommend it to anybody who'll listen.
There's a 3dB peak output increase by using the "ultra-linear" tap, which is more important in these fleawatts than normal (which is: no importance at all). Rules out the One Electron OPT's but others exist if that becomes your final direction. Personally, I'd just go triode and sit closer.
In any and all cases, good fortune,
Chris
There's a 3dB peak output increase by using the "ultra-linear" tap, which is more important in these fleawatts than normal (which is: no importance at all). Rules out the One Electron OPT's but others exist if that becomes your final direction. Personally, I'd just go triode and sit closer.
In any and all cases, good fortune,
Chris
Last edited:
Well since this is my first build, I should probably keep things simple. Though a gas discharge tube would look pretty cool...
The speakers I intend to use them with are somewhat low efficiency, but my listening space is pretty small. If 12 watts Class D does it for me currently then ~3 watts of triode should be nearly enough for most listening. The ultra-linear option should help for the rare time I really need to crank it.
Also, I've started drawing up the power supply circuit the amp currently uses, and it is different from the one in the manual pic I posted earlier. The cap values are the same, but the resistor values are different. hopefully I'll finish it tomorrow, the wiring in there is not well organized.
Last edited:
An important issue for fleawatts (and really for all amplifiers if we want to optimize them) is minimizing the "hang time" after overdrive. It will happen and so should be optimized.
Valve amplifiers in class A or AB1 clip when the output valve's grid is driven positive (above zero, referred to the cathode). Here the grid conducts (more positive than the cathode, so imagines itself an anode - the affrontery!) and changes the charge on the coupling cap feeding it. Until this bleeds back the output valve is mis-biased. Stop me if you've heard this one.
Bleed down time constant is just the R's and C, and can be set anywhere you'd like. If it's really too small it will roll off the very low bass too soon. If it's too big, clipping is much more audible. A good number for big output transformers and big speakers is commonly considered to be about 1/4 second. Smaller, and especially single-ended, output transformers benefit by shading this down some. Maybe even more than a factor of three in special cases.
Easy to do, important, but unfortunately usually inadequately discussed IMO.
The trade-offs between triode and "ultra-linear" are better covered in the tube-o-sphere, so I'll shut up.
All good fortune,
Chris
Last edited:
Using the Cacahuate as an example, where does it stand in this respect, given the posted schematic?
Please forgive me, because I've used my personal shortcuts publicly, and misleadingly. I don't think in time constants, rather in Hertz, so I let the 2pi difference slip into conversation. Actually (and you can trust me this time Charlie Brown) the usually accepted optimum choice is 4Hz, not 1/4 second time constant. An actual 1/4 second time constant is quite audible and way too large. Again, I'm sorry for my sloppiness.
WRT the El Mighty Cacahuate (gotta luv it) the driver is so stiff that it can be ignored and the relevant time constant is between the 0.1uF coupling cap and the 0.47M Ohm grid leak resistor. Its RC time constant is 0.1 exp -6 times 0.47 exp +6 equals 0.047 or a frequency knee of 3.4 Hz. This should be considered the largest time constant, and therefore the lowest frequency, for an optimum design of *any* amplifier with an output transformer. Amplifiers with loop feedback around the output transformer need a much more conservative number, and amplifiers with standing current in the primaries (like a single-ended design) or with smaller output transformers or with smaller speakers, would benefit from something more appropriate also.
My personal choice would be to cut the grid resistor in half or so (the driver is plenty capable and this is more forgiving of imperfect vacuum in the output valves) and to adjust the frequency with capacitor value. With a good solid 50 Hy or more (with current) primary and bookshelf / roughly 60 Hz F3 speakers, I'd look in my stash for some nice 716P's of 0.33 or 0.22uF to start. This choice is fairly audible in real usage, so you can experiment if interested.
Again, I apologize for my sloppiness.
All good fortune,
Chris
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.