1626 Tube Tests

arcane secrets and handshakes revealed!

1626 Tube Tests

Postby Thermion » Fri Jul 22, 2011 8:19 am

I have been designing a 1626 headphone amplifier based on Lloyd Peppards (Mapletree) ear+ amp and needed parameters for the 1626 tubes to input into TubeCAD. The available spec sheets for these tubes provide little information on mu, GM, RP and such. It is even conflicting between manufacturers. This is probably due to them not being designed or used in audio circuits. Anyway, I decided the best thing to do is measure the parameters on a sample of tubes myself at the operating point of interest.

I chose to test the 1626 tubes with centerpoint values of VP=200v and IP=15ma. I tested 4 1626 tubes that Shannon got from the Peoria Superfest last year. They are branded National Union and have identical markings and date codes. The averages for 4 tubes are as follows:
mu = 5.075
GM = 1671
RP = 3049

These tubes all tested very close to each other, so the standard deviations on the mean values are very small.

JT
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby EWBrown » Sun Jul 24, 2011 9:29 pm

Thanks! I would give your 1626 test data more credibility than the rather sparse information of which I have found, regarding its transconductance and plate resistance.

The only 1626 data I've seen which included gm and RP is from Tung-Sol, dated 1943, and that is just a very brief single page, which gives only a thumbnail description of both 1625 aand 1626 http://frank.pocnet.net/sheets/127/1/1625.pdf

And that shows the 1626 as having gm of 2500 umhos (2.5 mA/v) and rp of 2500. Which contradicts the usual givem mu of 5 (it would be 6.25 with the Tung Sol data).

thanx

/ed B in NC
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 Thermion » Mon Jul 25, 2011 8:10 pm

Thanks Ed! I was hoping someone would find the data useful. I am pretty confident in the measured values for that operating point. I also found the Tungsol datasheet and an couple RCA sheets giving a mu of 5. Definitely not consistent values. I think the Tungsol data gives the GM and RP for the max values of 250 volts and 25 ma. I haven't tested one yet, but that is my theory. TubeCAD simulations with my measured values are very close to prototype circuits. I only wish the 1626 had lower RP and higher GM. It really is a great looking tube.

I have a lot of ten 50's vintage RCA 1626's coming my way from Ebay. I'll post the results after I test the batch and see how they compare to the National Union.

The measurements are taken in a static not dynamic sense. I use a Heathkit SP-2717 high voltage regulated power supply to control the plate and grid voltages and measure the plate current through a 10 ohm resistor between the power supply and plate. Three multimeters allow me to measure VP, GV, and IP. I then hold GV constant and vary VP and record VP and IP. I then hold VP constant and vary GV and record GV and IP. The values then get entered into an excel spreadsheet that performs the linear regressions and spits out GM and RP as the slopes (or inverse slope) of the lines. Mu is then the product of the two slopes.

Regards,
Jonathan
Thermion
 
Posts: 181
Joined: Mon Jul 07, 2003 6:04 pm
Location: Fisher, IL

Postby EWBrown » Tue Jul 26, 2011 1:15 pm

I found that in the "classic" Darling SET design, that a 12B4A will operate at the same voltages and currents as will a 1626, but the output power will be about twice as much (1.5W rms vs 750 mW rms).

I plugged in a B+ voltage of 242VDC (240 to 245 is fine) and with an RK of 1K (bypassed) and any chosen 5K primary OPT, that the plate current is still between 25 and 27 mA, so it should be possible to make the amp "convertible" by parallel wiring two separate tube sovkets for the output section.

I chose 242 VDC (+/- 2 or 3V) because that is what I get out of a Hammond 369EX power trannie, with a couple of UF4007s, and the usual CLC filtering, loaded with two 12B4As and one 12DT8. That comes close enough to the original Danielek design requirements, close enough for government work, as they say (lol)

The driver tube could be the original 8352 / 6J4s, 5751, 12SL7 or even 12AT7 or 12DT8, all work well, though the latter four tubes work best with 100K plate and 1K cathode resistors.

An Antek 05-T200 toroidal 50VA power trannie will also produce very similar results. [:)

I suppose I could try out running a 1626 on my Hickok 539 tube tester and see how it compares. (???)

Not only is the 1626 a power triode, it also can serve as a rectifier, I have an old 50s vintage Heathkit capaccitor tester which actually uses a 1626 as a half-wave rectifier in the PSU. Now that's kinda weird... :/

But back then Heath made maximum use of the readily available, low cost post-WWII electronic surplus, which was flooding the market in the late 40s and early 50s, even more so than a decade or two later.

1626s are neatr tubes, and sometime I'll try them in a PSET or PP design, they are still relatively inexpensive, though not as dirt cheap as they were around 50 years ago =:o $) Ditto for 12B4As, they are still cheap and there must still be millions of them in the supply chain, so it seems...



/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 !


Return to test & measurement

Who is online

Users browsing this forum: No registered users and 8 guests