A couple of weekends ago I threw together a MONO-block Tubelab "Simple 45" amp, except I used its 5-pin cousin, the weird 2-grid #46 (from what I can tell, 45 plates and cathode/filaments in a 2A3 ST bottle with an extra grid).
Driver is the 5842 with the 10M45 CCS on top, pulling about 10 mA.
#46 filament is powered by one of Pete Millett's rectifier/regulator boards, spot-on 2.5 VDC, and grounded thru a 10 ohm resistor "fuse". Second grid is tied to the plate, as per the data sheet for Class A single-ended operation.
It's powered by an old Heathkit IP-32 regulated power supply; the 46 seems to like about 260-270 Vp and 33 mA (about -25 V on the grid thru a 47K grid resistor, IIRC), This operating point is a outside the specified limits for the #46, but folks seem to get good results with it, and there's no visible signs of distress.
OPT is one of the cheapie Edcor 5k > 8 ohms output transformers.
It's not exactly the sweetest 45 SET I've ever heard, just different; seems like it has more "sparkly" highs, super-clear mids, and more bass than other 45 SETs I've heard. I think I like it better than the #45, although time will tell. I suspect one reason it sounds good at this Op Point is that the plate resistance drops below about 2K ohms at 33mA plate current.
I have a pair of "Transcendar" 5K OPTs on the shelf, plus 2 different pair of bigger Edcors (CXSE25-8-5K and GXSE15-8-5K) in the mail. I'll try these later. I suspect that the big Edcors might have super bass for a 45-class SET amp.
HERE'S my question: the Edcor has an ultra-linear tap (which I use on all the pentode Simple SEs I've built). Could the second grid of the #46 be tied to the ultra-linear tap on the Edcor, like a quasi-pentode?
All the documentation that I've found on the #46 pre-dates the invention of the Ultra-Linear connection. The only options shown on the data sheets are:
(1) Class A: connect the second grid to the plate at the socket
(2) Class B: connect the two grids together at the socket
I'm thinking of trying it out, starting at the data sheet Class A operating point (-33 grid, 250 Vp, 22 mA plate current) and see what happens.
Famous last words: what could go wrong?
Driver is the 5842 with the 10M45 CCS on top, pulling about 10 mA.
#46 filament is powered by one of Pete Millett's rectifier/regulator boards, spot-on 2.5 VDC, and grounded thru a 10 ohm resistor "fuse". Second grid is tied to the plate, as per the data sheet for Class A single-ended operation.
It's powered by an old Heathkit IP-32 regulated power supply; the 46 seems to like about 260-270 Vp and 33 mA (about -25 V on the grid thru a 47K grid resistor, IIRC), This operating point is a outside the specified limits for the #46, but folks seem to get good results with it, and there's no visible signs of distress.
OPT is one of the cheapie Edcor 5k > 8 ohms output transformers.
It's not exactly the sweetest 45 SET I've ever heard, just different; seems like it has more "sparkly" highs, super-clear mids, and more bass than other 45 SETs I've heard. I think I like it better than the #45, although time will tell. I suspect one reason it sounds good at this Op Point is that the plate resistance drops below about 2K ohms at 33mA plate current.
I have a pair of "Transcendar" 5K OPTs on the shelf, plus 2 different pair of bigger Edcors (CXSE25-8-5K and GXSE15-8-5K) in the mail. I'll try these later. I suspect that the big Edcors might have super bass for a 45-class SET amp.
HERE'S my question: the Edcor has an ultra-linear tap (which I use on all the pentode Simple SEs I've built). Could the second grid of the #46 be tied to the ultra-linear tap on the Edcor, like a quasi-pentode?
All the documentation that I've found on the #46 pre-dates the invention of the Ultra-Linear connection. The only options shown on the data sheets are:
(1) Class A: connect the second grid to the plate at the socket
(2) Class B: connect the two grids together at the socket
I'm thinking of trying it out, starting at the data sheet Class A operating point (-33 grid, 250 Vp, 22 mA plate current) and see what happens.
Famous last words: what could go wrong?
The class A suggestion that G2 should be tied to the plate suggests voltage won't be a problem, but I've never seen a power rating for G2. 47 does have published G2 voltage ratings, so it might be a better choice.
I've used both as triodes in my TSE and liked them about as much as 45.
Win W5JAG
Sent from my SGH-i917 using Board Express
I've used both as triodes in my TSE and liked them about as much as 45.
Win W5JAG
Sent from my SGH-i917 using Board Express
I agree that G2 would probably take the voltage. The small potential difference between plate and G2 in uLinear also shouldn't matter. That is, I'm not terribly worried about damaging the #46, provided that I don't go all "TubeLab" on it. (how DO you spell "incandescence" anyway?)
My question probably boils down to this:
I know that "pure" pentodes (eg EL34) and beam power tetrodes (6L6 et al, the ones created to get around the seminal pentode patent; Mullard, was it?) will work and work very well in an Ultra-Linear circuit.
Will a pure tetrode work as well? It lacks the "beam" function of the third grid (pentode) or the beam guides (often also called G3) of the beam power tretrodes. Without that "focusing" function, will a different potential on the #46 tube just scatter the electrons all over heck? That is, does the ultralinear arrangement depend on beam focusing?
It's starting to dawn on me that either (a) somebody has already TRIED this, and it sounds like crap, or (b) I'm going to have to be the one to try it.
I can't find anybody to admit to (a), so I guess I'll just have to try (b).
Heck it's only one wire. I'll try it tonight and report back, assuming I'm not mutated into The Incredible Hulk's skinny cousin.
I have 6 or 7 #46 tubes and NO #47s, otherwise I might try that first.
de K5HAR
My question probably boils down to this:
I know that "pure" pentodes (eg EL34) and beam power tetrodes (6L6 et al, the ones created to get around the seminal pentode patent; Mullard, was it?) will work and work very well in an Ultra-Linear circuit.
Will a pure tetrode work as well? It lacks the "beam" function of the third grid (pentode) or the beam guides (often also called G3) of the beam power tretrodes. Without that "focusing" function, will a different potential on the #46 tube just scatter the electrons all over heck? That is, does the ultralinear arrangement depend on beam focusing?
It's starting to dawn on me that either (a) somebody has already TRIED this, and it sounds like crap, or (b) I'm going to have to be the one to try it.
I can't find anybody to admit to (a), so I guess I'll just have to try (b).
Heck it's only one wire. I'll try it tonight and report back, assuming I'm not mutated into The Incredible Hulk's skinny cousin.
I have 6 or 7 #46 tubes and NO #47s, otherwise I might try that first.
de K5HAR
No, the UL connection works with true pentodes as well as beam tetrodes. The third grid was added to reduce secondary emission which caused the "tetrode kink". I don't think it would make a big difference here. I have found at least one tube, the VT225/307A that doesn't work too well in SE UL. I haven't tried P-P yet since most of my P-P OPT's don't have the taps.
I never tried to glow a 45 but most of my tubes lack enough emission to get the plate red. I have gathered 10 or 20 45's over the years and they all work in SE with the usual 30 to 35 mA of plate current even at up to 325 volts, but I tried to run some in P-P AB2 and many lacked enough grunt to get to 20 watts out. Only the strong work there. That may be the reason for the recommendations with the 46's.
Hey, I would try it, but I don't think I ever did. I didn't find the 46's until a recent warehouse cleanout. I only have two. Next time I have a TSE on the bench.....
Yeah, I'd try it too ...
What didn't post up very well from my phone last night, is, how much power can that G2 in the 46 take?
We know voltage is no big deal because it's routinely run as a triode with G2 tied to either G1 or Plate ( Isn't there a thread somewhere on the relative merits of each connection? ). But I've never seen a published current rating for G2 in the 46. Wasn't 46 supposed to be a "do it all" tube? G2 was probably intended for a Local Oscillator connection.
47 does have published ratings for both voltage and currrent on G2.
But, you use what you got. Looking forward to the results.
Win W5JAG
What didn't post up very well from my phone last night, is, how much power can that G2 in the 46 take?
We know voltage is no big deal because it's routinely run as a triode with G2 tied to either G1 or Plate ( Isn't there a thread somewhere on the relative merits of each connection? ). But I've never seen a published current rating for G2 in the 46. Wasn't 46 supposed to be a "do it all" tube? G2 was probably intended for a Local Oscillator connection.
47 does have published ratings for both voltage and currrent on G2.
But, you use what you got. Looking forward to the results.
Win W5JAG
I remember a thread maybe 6 years ago about where to tie G3 in an EL34 for triode operation. I made some graphs with both connections and found that tieing G3 to the plate brings on the tetrode kink at low plate voltages. G3 was invented to cure the kink but sone "expert" disputed my findings since I took the time to hand plot the curves and I was told that "the graph was made in excel or similia, it is not a Sofia tester". I lost interest after that remark. If you have time to waste, read the thread and realize who the major players are.
http://www.diyaudio.com/forums/tube...mode-g3-where.html?highlight=g3+plate+cathode
I did experiment with wiring the grids together in various combinations. It seemed to me that wiring the EL34 up the way it was intended worked the best. A thread mentioned in post #18 is also worth looking into. It seems that RCA intended using at least one tube with G1 and G2 wired together. It also works with the 807, and maybe some sweep tubes (more experiments coming). Every tube works different when you try something unusual, so you just have to try it.
http://www.diyaudio.com/forums/tubes-valves/67632-rca-7094-tetrodes.html
!st Listening Results
A couple of solder tacks when I got home and the plate-to-G2 jumper is removed and G2 connected to the 40% tap on the cheapie Edcors (certainly no offense to Edcor; I love these OPTs with over-powered Sylvania 6AV5GAs).
A quick trip back to the "Music Room" and the mono LP system with the poor-boy open baffle (12 inch University "full range" with a cheap titanium-diaphragm horn; about 97-98 dBa IIRC):
Running conservatively at 250V B+, -30V bias for around 29-30 mA current (readings from the power supply; no meters hooked-up tonight), it sounds exactly like I expected from an uLinear set-up; extremely clean, very sharp bass, etc.
HOWEVER, it's probably TOO bright. If we were talking about a speaker, I would want to pad-down the tweeter. If we were talking about negative feedback (which I guess we ARE, just LOCAL not GLOBAL), I would want to reduce it by a few decibels.
Generally, I think there's something here, but it isn't perfect right out of the box using the Edcors and their 40% tap. If I didn't like the uLinear connection so much in general, I wouldn't screw with it any more, but I think we can save this puppy.
First thing I'm gonna do tomorrow is install a Triode-uLinear switch so I can get a direct comparison (and watch the power supply when I switch it).
Some ideas and questions:
(a) higher quality iron can't hurt; biggger Edcor are on the way, I'll try them this weekend.
(b) a different percentage tap may be called-for (which I think Edcor would provide for a few bucks); I wonder if a 20% Mullard-style tap would work better??
(c) I haven't put any components in the uL circuit yet at all (no grid resistors, no R-C filters, not nuthin').
I'm guessing that the Big Brightness of the 40% tap calls for some resistance, maybe 1K ohms or so (that is, beyond just a grid-stopper of 47 or 100 ohms).
Any thoughts on tamping down this local (uL) feedback brightness with Cs or Rs or R-Cs? That is, before I take the plunge and get custom uL taps...
I did find this nice article on uL characteristics (with some circuit ideas that look promising) but I haven't digested it yet.
http://www.oestex.com/tubes/1955%20Radiotronics.pdf
A couple of solder tacks when I got home and the plate-to-G2 jumper is removed and G2 connected to the 40% tap on the cheapie Edcors (certainly no offense to Edcor; I love these OPTs with over-powered Sylvania 6AV5GAs).
A quick trip back to the "Music Room" and the mono LP system with the poor-boy open baffle (12 inch University "full range" with a cheap titanium-diaphragm horn; about 97-98 dBa IIRC):
Running conservatively at 250V B+, -30V bias for around 29-30 mA current (readings from the power supply; no meters hooked-up tonight), it sounds exactly like I expected from an uLinear set-up; extremely clean, very sharp bass, etc.
HOWEVER, it's probably TOO bright. If we were talking about a speaker, I would want to pad-down the tweeter. If we were talking about negative feedback (which I guess we ARE, just LOCAL not GLOBAL), I would want to reduce it by a few decibels.
Generally, I think there's something here, but it isn't perfect right out of the box using the Edcors and their 40% tap. If I didn't like the uLinear connection so much in general, I wouldn't screw with it any more, but I think we can save this puppy.
First thing I'm gonna do tomorrow is install a Triode-uLinear switch so I can get a direct comparison (and watch the power supply when I switch it).
Some ideas and questions:
(a) higher quality iron can't hurt; biggger Edcor are on the way, I'll try them this weekend.
(b) a different percentage tap may be called-for (which I think Edcor would provide for a few bucks); I wonder if a 20% Mullard-style tap would work better??
(c) I haven't put any components in the uL circuit yet at all (no grid resistors, no R-C filters, not nuthin').
I'm guessing that the Big Brightness of the 40% tap calls for some resistance, maybe 1K ohms or so (that is, beyond just a grid-stopper of 47 or 100 ohms).
Any thoughts on tamping down this local (uL) feedback brightness with Cs or Rs or R-Cs? That is, before I take the plunge and get custom uL taps...
I did find this nice article on uL characteristics (with some circuit ideas that look promising) but I haven't digested it yet.
http://www.oestex.com/tubes/1955%20Radiotronics.pdf
Man, You guys are GOOD.
Just sat down after adjusting the P/S to around 34-35ma at 265V B+.
Brightness is nearly gone, at least significantly reduced, and it has that uLinear sound; like a good EL84 PP (say a Dynaco 35), but lots cleaner and crisper. Like an EL34 hooked-up with a 45, then ran away and their love-child was raised by Triodes.
Hooray, Tetrodes. Hooray, Beer.
Now, as this is the TubeLab Forum, after all, (good one) I need to find the practical limits.
These folks say they run 46s pretty hot (38ma at 272Vp, 10.3W dissipation)
Single Driver Website
whereas I'm at 34-35ma at say 260Vp (265V B+ minus the OPT voltage drop minus all of ~350mV across the 10ohm cathode resistor/fuse). That is, assuming the P/S meters are correct.
BTW here's a Cunningham data sheet on the 46 for reference. No actual data relevant to this issue (Hafler was still an infant, I think) but there might be clues...let me know if I missed something.
http://tubedata.tubes.se/sheets/029/4/46.pdf
As I should have expected, the power seems to be pretty darned good; where the 46 triode-strapped at data sheet op points makes LESS power than a 45 at it's recommended op points, I would guess this arrangement (ultralinear at ~9W dissipation) seems like it's making 2.5 to 3 W (is that possible??? 25-30% efficiency???). Probably a little less than that, but encouraging anyway, this Ultra-Linear Voo-Doo.
More later, maybe photos.
Raising the current generally lowerd the tubes plate resistance increasing the damping factor. I find this necessary with my big speakers to tighten up the bass and remove some "shout". I am using a SSE with KT88's and I cranked the current up to 100 mA per tube. I figured that I would replace the power transformer when it fried since I am sucking 220 mA from a 150 mA Hammond sourced Allied, but it has been playing this way for about 4 years. Can't touch the transformer after an hour of use though.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.