6DJ8 transconductance at different currents

the thermionic watercooler

6DJ8 transconductance at different currents

Postby Shannon Parks » Fri Sep 17, 2010 6:16 am

Some quick numbers I gathered this morning. Bypassed 150 ohm cathode resistor. Adjusted DC current by measuring 100 ohm resistor at the plate and using a variable HV DC supply. Put 100mV 1kHz signal at grid. Measured AC across said 100 ohm resistor. 6N1P is a variant, of course. ;) Actually surprised at how well it does at higher currents. Definitely needs rebiasing and not a drop in sub, fwiw. Note: these are just Gm tests. I kinda recall the EH being the most linear last time I monkeyed with these tubes and the 80's JAN as not-so-hot. I'll get back around to that as I'm tweaking a phono stage. Just was curious the drop in Gm below 10mA.

Obviously I'm violating some maxes here (nice plate dissipation on the 6N1P for example) so I wouldn't choose those as standard operating points. I was just swagging a fast test. I also want to retest some of these with +10% on the filament. Presume that that will be worth it.

- Shannon

6922 JAN Mar-86

54.5V 5mA 8710
78.8V 7.25mA 9660
103.2V 10mA 10480
125.6V 12.5mA 11090
147.8V 15mA 11790

JJ E88CC ~2004

46.4V 5mA 8740
68.9V 7.25mA 9740
90.8V 10mA 10460
112V 12.5mA 11060
133.5V 15mA 11530

EH 6922EH

43.6V 5mA 8380
64.6V 7.25mA 9180
84.9V 10mA 9800
103.6V 12.5mA 10360
124.2V 15mA 10780

6N1P 04 06 Svetlana Rebadge

108.1V 5mA 4940
142.9V 7.25mA 5570
176.2V 10mA 6050
207.2V 12.5mA 6450
238V 15mA 6820
User avatar
Shannon Parks
Site Admin
 
Posts: 3762
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby Geek » Fri Sep 17, 2010 3:47 pm

Hi Shannon,

Great work, thank you! :))

Can I repost this info on my GeeK ZonE?

Cheers!
-= Gregg =-
Fine wine comes in glass bottles, not plastic sacks. Therefore the finer electrons are also found in glass bottles.
User avatar
Geek
KT88
 
Posts: 3585
Joined: Sun Oct 21, 2007 3:01 am
Location: Chilliwack, British Columbia

Postby Shannon Parks » Sat Sep 18, 2010 2:52 pm

Of course! I did finally find the Gm curve charts I was looking for here:
http://frank.pocnet.net/sheets/009/e/E88CC.pdf

Now testing 12AT7s, 12AV7s and 5965s for transconductance. The first two look just like the datasheets, but the RCA & GE 5965s (with the big plates) have much more Gm at 10mA - very cool as I already loved this tube. I like all these due to the abundance of NOS and higher Mu. Hopefully get my audio analyzer test rig running this afternoon.

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3762
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby Geek » Sat Sep 18, 2010 3:57 pm

Thanks!

The 12AT7 info would be really handy [:)

Cheers!
-= Gregg =-
Fine wine comes in glass bottles, not plastic sacks. Therefore the finer electrons are also found in glass bottles.
User avatar
Geek
KT88
 
Posts: 3585
Joined: Sun Oct 21, 2007 3:01 am
Location: Chilliwack, British Columbia

Postby EWBrown » Sat Sep 18, 2010 7:41 pm

The 6DJ8s were originally designed as an RF tube, and have some of the variable-mu remote cutoff characteristics of many of the RF "front end" tubes, This was useful in order to implement AGC in the FM tuner's RF amplifier and converter/mixer stages.

Ditto for the 12AT7, and it's siblings 6DT8 and 12DT8, and "half-brother "6AB4.

I've also found the JAN 6992s to be somewhat less desirable, at least in the phono stage, and other high-gain applications. These were most likely intended for RF service in some military "boat anchor" rather than in high-fidelity audio. [:) That's probably why they are less $$$ than the good 6DJ8s and ECC88s. JAN doesn't always mean "the best tubes available".

Unlike some other "RF" twin triodes, these tubes can work very well in audio service. 6BZ7s, 6BQ7s, 12BZ7s, etc, will work as VAs or LTPIs, but they can be of variable quality, and some of them are just downright nasty sounding. (sick)

I really like the 5965s in SRPP mode, I run them in my latest 300B SET - the one built up on the ST70 chassis.

The 6414 and the somewhat obscure 0528 are very similar in operating characteristics to the 5965, but these have the taller bottle and longer plates like 12BH7As. They also work nicely in the "ST70" 300B amp.

Another interesting ,somewhat obscure twin triode is the 6350, these are similar to somewhat beefier 12BH7s, but they have a different pinout - the grids and cathodes are "swapped", so they are not plug-in interchangeable..

/ed B
Last edited by EWBrown on Sun Sep 19, 2010 1:23 pm, edited 1 time in total.
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby Geek » Sun Sep 19, 2010 4:42 am

EWBrown wrote:Another interesting ,somewhat obscure twin triode is the 6350, these are similar to somewhat beefier 12BH7s, but they have a different pinout - the grids and cathodes are "swapped", so they are not plug-in interchangeable..


They are also ranging from $5 to over $50 on fleabay $)

Cheers!
-= Gregg =-
Fine wine comes in glass bottles, not plastic sacks. Therefore the finer electrons are also found in glass bottles.
User avatar
Geek
KT88
 
Posts: 3585
Joined: Sun Oct 21, 2007 3:01 am
Location: Chilliwack, British Columbia

Postby EWBrown » Sun Sep 19, 2010 1:06 pm

The ridiculously high flea-bay 6350 prices are most likely a result of VTL and Manley Labs - which are essentially the same company - using them for the LTPIs on some of their more powerfu (and very expensive)l monoblock amps. Their smaller (and stereo power) amps use 12BH7s, 6464s, abd even 12AT7s as LTPIs,

I lucked onto a very large quantity of GE 6350s several years ago at a NH hamfest, they were "pulls" from the MIT computer lab, and they were really, really dirt cheap, - two xerox copy paper boxes of assorted tubes (mostly twin triodes) for the princely price of $15, and at least a quarter of them were 6350s and 6414s (and the oddly numbered 0528s, which were end-user marked 6414s). Best of all, more than 90% of them tested "like new". I still have over 100 of them which I kept for my own use.

I've gotten other surplus MIT CL tubes from other hamfests and fleas, and the best (for free) was a large ziplock freezer bag full of 12AX7s 12AU7s and 12AT7s. There were a couple of broken ones, and after testing them, I found only a few bad ones. Free is ALWAYS a good thing...
Frank, who gave them to me had literally thousands more in "egg-carton" bulk packaging, which he was selling in bulk lots for some pretty steep $$$. $)

The only relatively major expense was buying enough of the white tube boxes, even in large quantities, they out-cost the tubes by a long run :|


I only wish I knew about this " tubular bounty" back in the early to mid 1970s when I lived in Cambridge Mass, and worked as a tech, less than a mile away from MIT, and I had friends who worked in the labs and other maintenance functions...


Apparently the preventitive maintenance program on those ancient huge tubed computers, was to swap them out after only 50 to 100 hours' operation.

The MIT (and Draper) Labs must have had these tubes by the boxcar load, as they still show up at occasional NE fleas and swapmeets.

I figure that the 6350s have definite possibilities for use in small PSET or PPP amplifiers,

THen there was the "Hosstraders" hamfest, in Deerfield NH, back in the mid 1980s, when some guys from the NE Mass AT+T and WE factories were sent to the 'fest to sell off large quantities of surplus NIB, NOS 300Bs for $15 each. They also had massive amounts of 6080s and 6AS7s, and other "odd" WE tubes, most of which I knew nothing about.
Back then, "telephone repeater tubes" for $15 each was considered kinda steep...

But back then, I didn't know what WE 300Bs were really worth, or what they even were good for. My tube interests back then were more ham radio related, and big triodes didn't fit into my needs back then. :'( :'(

/ed B
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby Shannon Parks » Fri Sep 24, 2010 9:52 am

Interesting. This morning I tested 5965's paralleled and will up the data later. Basically, they were around 12000Gm at 10mA - just a touch better than the single triode 6922s at 10mA. I like this operating point for them (and want to use them in this phono stage!). It is exactly 10x the Gm of a 12AX7 and just edges out a 6922. I did try to test the 6922 and 6N1P paralleled, but at 10mA+ they oscillated madly, probably from the jumper wires I soldered on. Dang RF tubes. (shakes head)
User avatar
Shannon Parks
Site Admin
 
Posts: 3762
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby EWBrown » Fri Sep 24, 2010 2:48 pm

I've also found that WE396As, 5670s and 2C51s (all essentially identical) to also be oscillators-in-waiting. I copied Gary Kaufman's "Baby Blue" 407A linestage http://www.the-planet.org/linestage.html and it thought that it was a wide-spectrum oscillators, and the 396As were microphonic as (666) . It took careful grid and plate stoppering, a sprinkling of ferrite beads, and "biasing" the filaments at a point midway between the cathode potentials (this is a VA/CF circuit) to tame them down, and then only some of the dozen 396As which I tried, behaved nicely.

This linestage does have a LOT of gain, far more than is really needed for most LS applications, and 12AU7s, 6CG7s and 6SN7s are a far more suitable tube for this purpose. But then, I just HAD to try it... Just to stay with the "fancy tube" theme, I used a Bendix "Red Bank" 6754 for the rectifier.

If you parallel 6922s, 6DJ8s, 6N1Ps, etc, you really need separate grid stoppers, and e plate stoppers for each triode section, to keep these hyperactive little tubes from going into RF "wobbulator" mode.

All of this will be good experience should you ever decide to try using the Russian 6C45s, which are very high transconductance pulse radar triodes, and they just love to oscillate. Taming them into behaving properly in the audio realm can be a real test of one's patience >:o

I've also tried using small ferrite beads as "stoppers" with varying levels of success, this all depensd on the ferrite material's permeability, useful frequency ranges, etc, and can bexome a complex, involved procedure.

Since my FB collection is mostly of random samples and stuff which I salvaged from various electronic carcii, I just have ot wing it, as I have no idea of what the ferrite material really is.

/ed B
Last edited by EWBrown on Fri Sep 24, 2010 3:00 pm, edited 3 times in total.
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby EWBrown » Fri Sep 24, 2010 2:48 pm

If you parallel 6922s, 6DJ8s, 6N1Ps, etc, you really need separate grid stoppers, and even plate stoppers for each section, to keep these hyperactive little tubes from going into RF "wobbulator" mode. ANd try to solder the stopper resistor's body as close to the tube socket as reasonably possible.

All of this will be good experience should you ever decide to try the Russian 6C45s, which are very high transconductance pulse radar triodes, and they just love to oscillate. Taming them into behaving properly in the audio realm can be a real test of one's patience >:o

I've also tried using small ferrite beads as "stoppers" with varying levels of success, this all depensd on the ferrite material's permeability, useful frequency ranges, etc, and can bexome a complex, involved procedure.

Since my FB collection is mostly of random samples and stuff which I salvaged from various electronic carcii, I just have ot wing it, as I have no idea of what the ferrite material really is.

/ed B
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby Shannon Parks » Fri Sep 24, 2010 3:34 pm

EWBrown wrote:If you parallel 6922s, 6DJ8s, 6N1Ps, etc, you really need separate grid stoppers, and e plate stoppers for each triode section, to keep these hyperactive little tubes from going into RF "wobbulator" mode.


I figured as much! And just one more reason to continue with my friendly 5965s (which I have a shoe box full). These are the *big* plate GEs and RCAs, not the wimpy 12AV7 plates (look about the size of 12AT7s). Obviously, those huge plates are the reason we have 3pF Cgp. 6GK5s just aren't as sexy, either.

--Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3762
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby EWBrown » Fri Sep 24, 2010 8:51 pm

The 6414s and IBM house-numbered "0528"s are like heftier versions of the 5965, and usually cheap when found at flea markets and hamfests.
They have the same pinout, and relatively close operating characteristics, close enough to be readily interchangeable, especially in SRPP service (like with SorenJ's 6AV5GA amp he built a couple years ago). So far I've been able to find 5965s for cheap on e-bay and at fleas. Just more of those "nasty computer tubes" to those who are not in the know (???)

I've never had any runaway RF oscillator or microphonic problems with any of these three tube types.

IIRC, VTL / Manley Labs use 6414s as well as 6350s, in their LTPIs, so that fact, alone, brings up the price from the usual big tube vendors. Otherwise they are generally considered as being in the "plinker" category by most.


/ed
Real Radios Glow in the Dark
User avatar
EWBrown
Insulator & Iron Magnate
 
Posts: 6389
Joined: Wed Mar 19, 2003 6:03 am
Location: Now located in Clay County, NC !

Postby Shannon Parks » Mon Sep 27, 2010 6:43 am

I finally did run the Sovtek, JJ and JAN 6922s in my phono stage with the input GC paralleled running at 10mA. No issues after first power-up, and then I put a grid stopper in there anyhow. Had been doing full bandwidth tests, fwiw.

I was able to tune the preamp for the RCA 5965s and then use 6922s +/-1dB RIAA, though I wasn't able to do the opposite. It was cool to roll all these tubes in the same preamp with +/-1dB and zero tuning changes (just barely though):
RCA 5965
GE 5965
IBM 5965 (clear top) manufacturer?
RCA 12AV7
Sovtek 6922
JJ E88CC
JAN 6922 (80's)

I can tweak any of those tubes to +/-0.15dB RIAA, but I like the 'tube rolling' tuning. Still working on a super slick method for tuning. Did order some K71-7 Russkie caps last night: polystyrene, 250V rated and +/-0.5% tolerance. Suhweet.

--Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3762
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Postby lynxx » Thu Sep 30, 2010 12:50 pm

Shannon,
Have you ever tried the 6C45pi tubes for a RIAA. I know they are not dual triode but just curious on the performance.
Thanks
Bryan
Downtown Orange Park, Florida
User avatar
lynxx
 
Posts: 38
Joined: Mon Feb 14, 2005 11:22 am

Postby Shannon Parks » Fri Oct 01, 2010 3:05 pm

I don't have any of those, but I'll probably get some for testing sooner than later. Tore my preamp apart (again) after realizing the Miller capacitance of my paralleled 5965s would load the carts too much. Now testing a hybrid cascode - 5965 with 2SK170 ala Allen Wright's phono stages. Had to wire this up from scratch using one of my ST70 boards. Hope to test it this weekend.

--Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3762
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

Next

Return to diy hifi

Who is online

Users browsing this forum: No registered users and 0 guests