The NOS Tungsol 6550 will cost a fortune, but in my experience will run at over 500V without problems. (Did in my amplifiers.)
That said I built a number of 6550 PPP amps running off of 525V regulated supplies and the ones I keep track of are running >10yr old Svetlana 6550C (the originals I delivered with the amps when new) without failure so far. They idle at 50mA each and are UL connected. These tubes are still available under the Winged C moniker.
The 6550C is essentially their take on a 6550A which is significantly uprated from the original 6550.. Last time I purchased these they were still quite good. No idea what they are like today.
I would avoid the current Tungsol 6550, my experience with them in relatively unstressed service was poor.
http://www.mif.pg.gda.pl/homepages/frank/sheets/135/6/6550A.pdf
Caveat is the only tube I know that meets that spec is the 6550C and some KT88 - modern ones aren't that much more than a 6550 and sound pretty good in my experience.
Thanx Kevin. Of the four not heavily used '65 Tung-Sols I bought (for right about $500), two look and test pretty much NOS, the other two seem close.
Two of them are within 1mA at 400Vp and -48V on my Maxi Matcher but who knows where they'll be at 525Vp and about 60-65% Pd. The other two are higher current and not close.
So I'll have separate bias pots for each tube.
I have run old well used Tung-Sols (not reissues) at 725Vp in '74/5 Ampeg SVTs. At recommended 24mA of course. No problems, they just wore out eventually.
The reissue Tung-Sols I have heard don't survive even 500Vp too well but are OK at 450V. The Svets I have heard can handle the high voltage but I doubt they sound as good as the originals, even the '60s 3-hole greyplates like I have, never mind the no-hole '50s black plates, which are reputedly better but I've not heard them. I have a few Svets from the late '90s.
The highest power 6550 I have had were the Sovtek 6550D or E, they would do 350W in a sextet in the '70s Ampeg SVT at 695V. Made a very reasonable modern replacement for the worn-out Tung-Sols in that amp. I doubt they would compare in audio service though.
Two of them are within 1mA at 400Vp and -48V on my Maxi Matcher but who knows where they'll be at 525Vp and about 60-65% Pd. The other two are higher current and not close.
So I'll have separate bias pots for each tube.
I have run old well used Tung-Sols (not reissues) at 725Vp in '74/5 Ampeg SVTs. At recommended 24mA of course. No problems, they just wore out eventually.
The reissue Tung-Sols I have heard don't survive even 500Vp too well but are OK at 450V. The Svets I have heard can handle the high voltage but I doubt they sound as good as the originals, even the '60s 3-hole greyplates like I have, never mind the no-hole '50s black plates, which are reputedly better but I've not heard them. I have a few Svets from the late '90s.
The highest power 6550 I have had were the Sovtek 6550D or E, they would do 350W in a sextet in the '70s Ampeg SVT at 695V. Made a very reasonable modern replacement for the worn-out Tung-Sols in that amp. I doubt they would compare in audio service though.
Last edited:
If you're still considering the Edcor OPT, I asked them about max voltage and they said 500V was the design rating.
New issue Tungsols 6550 ...I toned them down to 450V and not to go over /equiv 35W total quiescent dissipation. Any higher and they runaway fast. The best I got was max 2000 hrs life before getters went grey. Outgassing metals ? All the todays versions aren't stiched up for the heavy duty jobs, when compared to the famed A version of yester-year where many designs ran them way above 500V..
richy
richy
Thanx zelgall, that's useful to know. I have run 100W Edcor OTs above 500V before, at about 530V, in one with the same PT I'm using on this build (One Electron BFT-1B). No evidence of stress noticed.
Rich, your info confirms what Lord Valve told me about the reissue Tung-Sol 6550 when they first came out, with the 2000 hour life being something new. Thank you.
Rich, your info confirms what Lord Valve told me about the reissue Tung-Sol 6550 when they first came out, with the 2000 hour life being something new. Thank you.
Hi Ian,
This "extra screen proctection" you mentioned by upping the stopper values somewhat - how does it work? Certainly not by the meagre few additional Volts being dropped across the upped stoppers, I suppose?
Regards, Tom
This "extra screen proctection" you mentioned by upping the stopper values somewhat - how does it work? Certainly not by the meagre few additional Volts being dropped across the upped stoppers, I suppose?
Regards, Tom
The function of the elevated value screen resistors is current limiting I believe. Not sure how the math works though.
Hi Fleghorn,
Looking at the GE datasheet of 6550A, for UL mode @ 450V Ea and Eg2 and Eg1 = -48V a maximum screen current of 38mA for a PP pair is claimed (means Ig2,max = 19mA for a single tube in such a PP pair).
Further looking at the Ig2 plots in that data sheet, one can see the Ig2 skyrocketing point due to massive Ia-Ig2 takeover moves towards Ea = 0V for Eg1 getting more an more negative.
The most negative value Eg1 value curve plotted for Ig2 shown there is at Eg1 = -20V and the Ig2 skyrocketing point is below Ea = 10V already (actually it is not even shown plotted in the graph). For even more negative values of Eg1 this takeover point would move even more towards Ea = 0V.
Honestly I don't see how this "current limiting" should work if (iddle state) Eg1 is reasonably negative (f.e Eg1 more negative than say -40V).
The scheme of considerably upping screen stoppers for "current limiting" may work for tetrode / BPT mode and relatively moderate values of Eg1, but for UL mode with unusally high Ea=Eg2 values and unsually negative iddle Eg1 values, I doubt it. At least that is what I derive from the datasheet plots.
Regards, Tom
Looking at the GE datasheet of 6550A, for UL mode @ 450V Ea and Eg2 and Eg1 = -48V a maximum screen current of 38mA for a PP pair is claimed (means Ig2,max = 19mA for a single tube in such a PP pair).
Further looking at the Ig2 plots in that data sheet, one can see the Ig2 skyrocketing point due to massive Ia-Ig2 takeover moves towards Ea = 0V for Eg1 getting more an more negative.
The most negative value Eg1 value curve plotted for Ig2 shown there is at Eg1 = -20V and the Ig2 skyrocketing point is below Ea = 10V already (actually it is not even shown plotted in the graph). For even more negative values of Eg1 this takeover point would move even more towards Ea = 0V.
Honestly I don't see how this "current limiting" should work if (iddle state) Eg1 is reasonably negative (f.e Eg1 more negative than say -40V).
The scheme of considerably upping screen stoppers for "current limiting" may work for tetrode / BPT mode and relatively moderate values of Eg1, but for UL mode with unusally high Ea=Eg2 values and unsually negative iddle Eg1 values, I doubt it. At least that is what I derive from the datasheet plots.
Regards, Tom
Good point, I was thinking I guess of pentode connection not UL. The UL screen grid stoppers in the circuit I'm using are 220r only, and Patrick Turner used only 220 or 330r in his UL tetrode/pentode mode tubes.
The grid resistors have nothing to do with grid DC bias or operating conditions at idle in this case.
If they are too large, it may happen -during operation, especially when you play loud- that the output tubes go into some kind of thermal runaway: the grid doesn't discharge as fast as it should, plate current stays high, anode glows red.
If you haven't observe something like this, then probably the grid res. values are o.k. It the AC behaviour that is affected,not the DC conditions.
This has only remotely to do with longevity, some manufacturers specify lower grid resistors if Va exeeds some predefined value.
I'm sorry but it's not true. The thermal runaway can be during the idle conditions (power ON, no signal applied). This resulted by the grid1 leakage current. This current is increasing as the temperature increase, and this current reduce the bias voltage, as it drops voltage over the grid resistor. To avoid this problem the manufacturers apply gold plated 1st grid (Svetlana do it for sure).
Sajti
Maybe, if the cook the tube at idle, reaching very high temps, as you say. I have never observed something like this though.
To avoid the problem, manufacturers recomend a maximum value for the grid leak resistor, which is by the way much higher under cathode bias compared to fixed. The cathode resistor counteracts the current increase to a certain extend.
Not every tube has gold plated grids
You are lucky! I had this issue with JJ KT88, even I used the recommended grid resistor. The early JJ tubes had quality issues sometimes.
Yes. If You reduce this value, You can maybe violate the maximum voltage ratings carefully. I used 6550 UL with 500V, but with direct coupled cathode follower. The series resitance of the 1st grid was no more than 100ohms. It was rock solid bias.
Yes! The oroginal 6550 from GE has the maximum grid resistor value is 50kohms. The common types (such as Sovtek) has 100kohms, but the winged C Svetlana has 220kohms! I use winged C, but I've never go over 100k, to stay on the safe side. It works perfectly with 450V HT, triode connected.
Sajti
- Status
- Not open for further replies.