|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
69.141.96.215
At long last, retirement is not far off and I've started to seriously gather parts for a #46 DHT variation on the "El Cheapo" theme.As can be seen from the #46 data sheet, the type is a "natural" for near Class "B2" operating conditions. That means MOSFET drivers between the splitter and "finals". The MOSFET drivers are pretty well settled. FWIW, my initial "guestimate" for #46 "idle" current is 7 mA. However, questions remain and I seek input from the "inmates".
I'd appreciate recommendations for a winder of the 3.5 Kohm (end to end) PP O/P transformers. The "cast of characters" I know pretty darned well. I want good reasons for why I should prefer vendor "X" over vendor "Y". This amp will use GNFB and pooh/pooh remarks about NFB will be ignored.
Another undecided area is the tube to use as the splitter. I'm debating between the 'T7 shown in the original EC or a 6J6. The 6J6 will definitely require a line stage with gain. That's not a problem, as I have good stuff already on hand and I also have a pretty kick butt 6SN7 design to use. Plenty of "dodads" will be in this amp's enclosure and getting source selection out of the picture could easily be the best choice.
Eli D.
Edits: 03/12/17Follow Ups:
It's no longer 'el Cheapo' once you start stuffing rare and valuable tubes into such a circuit
You would be better off looking up the 6P23P which has 11 watt dissipation and mu of 8 (rather than mu of 5.6 for the 46) . It should drop straight in (apart from using lower value grid leaks) and costs a few dollars a piece . I would recommend a separate filament transformer though for the 6P23P . More power , cheaper and more repeatable than a 46 , surely worth a try :)
ps Avoid DN2540 cascodes , high impedence but not across the whole audio band as not particularly low capacitance . use a DN for the top but look into more suitable devices for the lower such as a JFET
George
The published works of Gary Pimm suggest drawing a different conclusion about the DN2540 MOSFET's when used in cascode.
Also, it is the upper one that could use improvement; a 10Mxxx chip has a lot more voltage gate-to-source( effectively; I realize they are multi-element chips ) and that exta voltage for the DN2540 to work under is useful in cutting capacitance...as well as offering up to 900V across it when the right model is chosen.
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
Eli just had Heyboer build me some monster 250 watt pp 5k transformers. They do everything better the Hammond and Edcors I have.
I realize that only 3 examples but Heyboer is a great company and the prices are great also.
Enjoy your retirement
Tom
The 3D21 has data for g1-g2 connected triode operation too...:)
cheers,
Douglas
Friend, I would not hurt thee for the world...but thou art standing where I am about to shoot.
RCA also specs out a driven g1/g2 scheme for 807. Essentially class B2 transformer coupled.
Edits: 03/14/17
I was wondering what to do with my 46s. Really good sounding tube.
Andy,Send me a PM, if you are willing to sell me a quartet.
Given what seem to be your tastes, you could try a trio of #46s, along with an interstage splitter trafo, as shown on the Cunningham data sheet I linked.
If you have "super/duper" efficient horn speakers, a #46 with g2 tied to the plate/channel running Class "A1" can be tried. Basically, you would be working with #45 circuitry.
Eli D.
Edits: 03/13/17
Eli:
Did you look at 46 prices lately?
Re. transformer, why do you need a custom one? Do you need special NFB winding(s)? Custom won't be cheapo.
Any insight why Class B?
I already have the quartet of #46s on hand. :> D I've got nice sockets too. ;> )I have a custom filament trafo on order from Heyboer, which is not particularly expensive. The main power trafo is a stock AnTek toroid, which (again) is not a "big ticket" item and already on hand.
The PP O/P "iron" needed is simple. 3.5 Kohms end to end, without UL taps. I'm going to go for at least 35 W. power handling. An "El Cheapo" uses "textbook" GNFB that includes the O/P trafo and that requires power handling capability above the power O/P. Core saturation due to a LF error correction signal SUCKS. Stability is pretty much a non-issue, as the only caps. in the signal path are between the LTP and driver FETs.
As for Class "B2", the #46 contains 2 grids, but is always used as a triode. Tie g2 to the plate and the #46 is much like the #45. OTOH, tie g1 and g2 together and the tube is nearly cut off, with 0 VDC of bias. Look at the data sheet. A near Class "B2" operating point ( slightly positive bias) to tame notching/crossover trouble is pretty (IMO) obvious, with 15+ WPC power O/P not being a stretch in the slightest.
Eli D.
Edits: 03/13/17
Power rating of 20 W is very conservative. 3K and 5K primary. Has a variety of secondaries, which is good for tweaking NFB. May seem pricey, but it sold for ~200 in today's dollars back then. It is highly interleaved, highly symmetrical, and has exceptional bandwidth. Cheap OPT may cause stability problems when inside NFB loop.BTW, I am hunting down 46s for use as drivers for 1624 triode-connected.
Edits: 03/13/17
LS-55
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: