Tho"ri*um, n. [NL. See {Thorite}.] (Chem.) 
A radioactive metallic element found in certain rare minerals, as thorite, pyrochlore, monazite, etc., and isolated as an infusible gray metallic powder which burns in the air and forms thoria; -- formerly called also {thorinum}. Symbol Th. Atomic weight 232.0.


Someone I knew once told me he was challenged to find a pair of 842 tubes. At the time I thought "what's the deal, it's just another transmitter tube". A search through the books revealed this to be a cousin of the 10 family. Along with the 841, the 842 is identical in construction to the 10Y with exception of the grid. The same stunt was pulled on the 845/211 and 811A/812A: take a tube, change the grid winding pitch, and you have yourself a new tube type. The same is true with the 841 and 842, they have a different grid giving them a new set of parameters. Where the 841 now has a mu of 30 and plate impedance to match (30K), the 842 was taken in the opposite direction and mu was lowered to 3. As can be expected this meant that Rp went down to 2K5. Holding the same anode voltage and current, the bias would change to around -100V making this tube a bit harder to drive but also yielding more power in return. Looking at different sources for my tubes I could not find a single reference to anyone ever having had these tubes for sale. After several months of searching, the challenge proposed at the time seems to have been a real one; even I couldn't find this tube.

Almost six months ago I found the first 842, an RCA ST16 shaped tube testing as new. Darn, where to find a second one. E-bay paid off for once, it seemed not many people are looking for this tube and I got lucky. I found someone offering a globe 842 (WOW!) in used but good condition. Couldn't resist, but still don't have a pair :((((  I had kind of given up hope on getting a matching second. The reason why this tube is so scarce leaves me guessing. The globe and ST I have tells that this tube was made for a long time (late 20's to late 40's), so there have to be quite a few. Either there is a stash somewhere waiting to be found, they were made only in small numbers or people were throwing them all away not knowing what to do with them...


RCA 842 (ST16) and nameless 842 globe (GE?)




When you're not looking for something, you always seem to keep running into that certain something over and over (a.k.a. irony). I ran into two more 842's last week, RCA ST that tested like new on my AVO. One or two more and I will have a backup pair. Circuit design in progress, see below.


What did I tell you... give up looking and they pop up all over the place. Scored two more 842's yesterday plus seven (!) more today. My "stash" is guaranteed and I can safely build an amp with this tube and listen to it for some time to come. I should have put money on that bet ;))


What's in a name?
The name thorium should be obvious :)   For those shrugging and scratching their heads: Thorium is the element added to the tungsten filaments to increase electron emission. Hence the filaments are Thoriated Tungsten. In case of this amp, both tubes have Thoriated filaments.



The driver tube will be the 10Y, it has already proven to be an excellent driver in my Exodus amplifiers, both sonically and technically. Along with the fact that it runs on the same anode voltage and it has the same appearance of shape and filament (hey, the eye wants something as well !!!), makes it an excellent candidate in my book. 

The 10 will be auto biased by means of a cathode resistor. The bias is chosen at 350V/15mA and results in a grid (bias) voltage of around -27V. This is by no means carved in stone. As with all tubes, there is always a deviation to the curves plotted in the manuals. 

As you might be used to by now, my circuit is based around an inter-stage loaded driver into the output tube. Well, hate to be a bore, but the same goes here. I find the IT is the best solution for driving output tubes due to the ability to offer a low impedance to the grid of the power tube which comes in handy when driving some current in there. Second, it replaces the anode load (choke or resistor), a capacitor and a resistor/choke to ground (one part instead of three). Sound doesn't suffer either from using this approach. The 1:1 ratio insures the best possible bandwidth and coupling between primary and secondary. As I don't need extra gain, I do not see a reason to chose for a step-up or step-down ratio.



The output tube is the long lost cousin of the 10, namely the 842. This tube is a rare one as I have not found any since this year. The 842 is identical in construction to the 10, with exception of the grid pitch. This grid gives the 842 a mu of 3 and Rp of 2K5. I've biased it at 425V/24mA using fixed bias. This is still nagging me, so I might be changing this at a later date. The fact that I now have a 20K pot in the grid circuit does not give me peace of mind. It is a trade-off to the 3K9 resistor that I would otherwise have to stick in the cathode circuit.

The load for the 842 is a 12K OPT made to operate over 80Hz. This means the induction is a bit lower than a standard full-range OPT giving it a benefit of fewer windings and a smaller core. HF is increased a bit due to this, something I am after for my Oris horns. Bass is handled by seperate amps, so I do not need to concern this amp with frequencies below 50-80Hz.


A problem I ran into was with the B+ supply of the 842. As the bias voltage is quite high at -100V, this means that I would need a B+ of around 540V using auto biasing and a high value cathode resistor. I don't like using a cathode resistor (3K9) that is higher than the plate impedance of the tube (2K5). The option that is left is to use a fixed bias supply.  I haven't used fixed bias for quite a while now, the reason being is that I found the sound of the auto bias better to my liking (rounder, sweeter, bla bla). Another dislike towards fixed bias was that if the bias supply died while playing, it would most likely cost me the output tube. This is not a real issue when playing with a Sovtek 2A3, but it IS when I have my 80 year old, new testing UX210's plugged in. Why use it with the impossible to find 842 then... argh, I'd better stop thinking about it before I change my mind...

I'm looking around to see if I can't make some sort of protection on the amount of anode current.


Each filament supply will have it's own power transformer to minimize any coupling between the different filaments and will have taps to be used in combination with a current source or a passive supply setup. The supply will use two 15mH filter chokes (ferrite core as seen in speaker cross-overs) as choke-input filter connected together with a 22.000uF cap in an 18dB filter. The last choke will act as current-source for the filament so to speak.

For the 10Y I chose a current regulated supply to get it as quiet as possible. The components were chosen so I can easily change between passive DC and current source.


I may wind air core chokes later on, this will only require a lump of isolated wire and a bobbin. RIght now I am playing with a new LCR meter to see what I will require in terms of length and windings. I just made a 20mH choke and it looks very nice at 100Hz and 1kHz (same values).


More work to be done here soon....

B+ supply for the 842

Jim de Kort