Greeting to all. I am a new member,a classical musician living in London.I first owned valve equiptment about 15 years ago and decided to teach myself valve electronics,with the aid of the superb Morgan Jones books, because I wanted to be able to be able to reproduce 78's as well as I could. However,as it happened,my first build was a microphone preamp because my commercial one failed during recording. It seemed natural to use differential pairs and I was frankly amazed that it worked first time.I have used this topology for my 3 subsequent preamps without problems. Recently I decided to have a go at some power amps so I began intending to put a balanced volume control at the input. I soon realised that if I wanted to stay differential then I absolutely needed a cathode follower stage if I was not to compromise frequency response.I duly built one; I had a 240-0-240 transformer available so I built a simple passive split supply,choke input on top and cap input with cascaded RC filtering for the negative. The mains is arranged to have the(toroidal) heater transformer come on upon plug in and the switch then turns on the HT. The circuit is a 6sn7 per channel sitting on a pair of Russian 6p15p (similar to el83) with 1k in their cathode circuits and the heater supply of these is elevated. Upon first power up, all seemed in order with about 34mA at 200 or so volts being drawn on top and the cathodes of the triodes sitting very steady at just over 5 volts. The negative supply was higher(or rather lower!) that I expected,partly because the transformer secondaries were 20V above nominal but also because the 4uF input cap was not reducing the DC as much as I anticipated. Nevertheless, about 260V is still within the rating of grid 2 (which is directly connected to 0V) for this valve. At any rate,I measured about 10V across each cathode resistor so all seemed well. The next morning I did the first audio test to be greeted by a deafening hum even at maximum volume attenuation although it did get louder as the volume was turned up. I checked my earthing and tried all the usual things although I knew in my bones that this would not solve the problem. Back on my workbench things had changed;the HT+ draw had more than doubled,the negative even more so,in fact I measured 30V across one cathode resistor before hurriedly switching off. I tested and cleared the supply with the valves removed but one curious result was measuring a 6V drop across the cathode resistor of one selectively removed pentode(!) One of the dual triodes was measuring 6V negative on its cathodes. I cannot understand what has gone wrong. Has anyone had this sort of problem? Is it possible that ceramic sockets could have internal shorts?-in this case one of the grid 3 pins is next to the anode pin but this seems a little farfetched. I would be most grateful for any help. Many thanks. N.A.
There is some confusion around about maximum rates for the 6P15P. I have seen some sites mentioning 330V, while other say it is 150V! Wavebourn (a member here) is convinced that the 6P15P is 'identical' to the 6P14P, which can handle 300V. Anyway, I have never experienced problems with a G2 and do not know what the consequences are, but maybe try a new pair of the 6P15P's and use some resistors in series with G2 (maybe 100R, to monitor G2 current and G2 dissipation).
Actually, I've found later that 6P15P is really a bit different from 6P14P. It's screen grid is denser, has more turns. It has more effect than in 6P14P and EL84 tubes. 6P15P is more linear tube made from 6P14P, for video amplifiers.
Thank you; I will try monitoring G2 current with a small resistor. However, the Soviet data sheets,which I would expect to be conservative, give 330V and 1.5W max for G2. Moreover I would have expected that as the voltage across the cathode resistor began to rise then this would counteract the rising current unless both G1 and G2 were fried and I was simply left with a diode. I think I will knock the B- down anyway as there is no need for it to be this high.
It would be nice to see you schematic.
For what you've explained, it more looks like tubes run away because of control grid currents caused by impurities in tubes.
For what you've explained, it more looks like tubes run away because of control grid currents caused by impurities in tubes.
Since I am new here and embarassingly unskilled in IT matters I have no idea yet how to post a schematic. It is, however, so simple that I can describe it in words; series resistors 100k from XLR pins(attenuator connected between them) to grids of dual 6sn7 cathode followers which are returned by 1M to 0V. Current sink 6p15p with G2 directly connected to 0V,each anode directly connected to each cathode and stopper resistor of 6sn7 and G3 and pentode cathode connected to -260V via 1k resistor.I repeat that all measured perfectly for several hours on first switch on so I suppose that eliminates any possibility of wiring error on my part. The 6p15p are ER grade with 2 OTK markings,the diamond mark, and the word "pyeryeprov" which my Russian dictionary tells me means "overproduction". Is it likely that such valves would be contaminated? The getter is pristine. Thank you.
You didn't mention g1 of the 6p15p - I assume this goes directly to -260V.
How is your soldering? A dry joint could leave electrodes floating. Too much flux left on the socket could cause leakage, which might only occur after a few weeks in a damp atmosphere (It has been warm and humid in London recently?). Try carefully inspecting your sockets. Internal shorts are unlikely, but poor pin contact can happen.
What about heater-cathode voltage? Do the valves share a heater winding? If the 6p15p heater is at 0v with the cathode at -260V, then it could act like another anode, injecting hum and stealing current from the real anode.
How is your soldering? A dry joint could leave electrodes floating. Too much flux left on the socket could cause leakage, which might only occur after a few weeks in a damp atmosphere (It has been warm and humid in London recently?). Try carefully inspecting your sockets. Internal shorts are unlikely, but poor pin contact can happen.
What about heater-cathode voltage? Do the valves share a heater winding? If the 6p15p heater is at 0v with the cathode at -260V, then it could act like another anode, injecting hum and stealing current from the real anode.
Thank you,yes, G1 directly to B-. I think my soldering is ok;I use silver low melting point solder,silverplated sockets which I clean with an eraser until bright before soldering so the solder easily flows on the pins-I am,in fact,almost paranoid about the possibility of dry joints,particularly on tube sockets and earth bus bars.As I said, the 6p15p heaters are elevated to between 2/3 and 3/4 of the way down the B-.It has indeed been humid in London recently; a delicious cake made by my sister and uncharacteristically not gobbled up immediately went mouldy in just over 2 days.However,I see no flux left on the sockets to absorb this moisture; I tend to keep the iron on the outside of the sockets so that any droplets of flux tend to fall on the chassis.
I think we need to see a circuit diagram. Click on the Post Reply button and you can attach a diagram via the full reply page. Or Go Advanced button.
Curiously,having reduced the B- to below 230V this strange current draw problem has disappeared. The loud hum remains although somewhat reduced. I shall try what DF96 has suggested and reference the heaters all the way down to the B-; I had been a little bit afraid of putting such a strain on the transformer secondary.
I assume you have separate heater secondaries for the two valves? If so, moving the pentode one down to B- should be OK, as modern transformers should have good insulation on all secondaries. By the way, B- (and B+) is an Americanism; we call it HT+ or HT- in Britain!
There is a possibility that this problem was caused by exceeding a heater/cathode voltage limit.
Thank you; I had checked the data sheet for this and it had said +/- 200V so I had thought that having the heaters about 60-70V more positive than the negative supply would be fine and maybe even beneficial. Referencing them all the way down to the negative supply has not cured the hum. At least, however, the DC conditions are now(for the time being, at least) completely stable. I think the AC heaters are going to have to go. It is possible,I suppose, that the hum is due to me trying to interface this balanced linestage with a commercial single ended amplifier and the problem may not be solved until I have the differential circuitry all the way through.I am certainly not about to start removing the earth connection from the plugs of the power amps. Nevertheless,listening through the hum,I certainly have had any lingering traces of prejudice against the cathode follower per se removed. There is no trace of deadening of the sound or any sonic characteristic at all that I can hear. If I can only kill the hum....
DC heaters have made no difference but the hum is clearly 50 Hz. Is it necessary to fit grid stoppers to pentode sinks? I never see them in schematics but I wonder if that is the problem...
IMO, grid stoppers are needed EVERYWHERE, and especially on hi gM tubes, which most pentodes are. Just add 1k's everywhere and see what happens. You can pull them out, one by one, if it makes no difference.
Refgards, Allen
Refgards, Allen
I am sure you are right; it is something I normally do but this is the first time I have built a cathode follower and the first time I have built a normal pentode sink; I usually just use transistor cascodes but I wanted to experiment. I am new to this forum,in fact I have never belonged to any internet forum before,but your name is very familiar to me as an expert and I would like your opinion on something exotic I tried this year in a differential microphone amplifier; I had read in one of the Terman books of the possibilty of turning a pentode into a high gain triode by using BOTH grid 1 and grid 2 as an imput so I put one connected like this in the tail of a differential pair and furthermore put a transistor in its cathode circuit hoping to achieve an enormous AC impedance without the penalty of pentode noise( although I have seen varying opinions as to whether or not this is a problem in this application).The result-little or no current seemed to flow in this outlandish set up despite about 500V from top to bottom. Maybe I made the wrong grid 3 connection. Many thanks.
Regards,Nicholas
Regards,Nicholas
This particular pseudotriode connection is mainly useful for making RF power amps, as it gives very high plate resistance and u-factors, like those of a "zero bias" RF power triode. Not so useful for audio applications.
My goodness! An immediate answer to something that has bothered me for months and is not easily accessible in textbooks. No wonder little current flowed; I had two of these pseudotriodes on top,one in the tail and then a transistor plus a few kohms underneath.I should probably not try such bizarre circuits. Many thanks.
OK, but how did you bias this "Super grid"?
If you treated it as a regular grid, that's the reason why no current flowed, small signal pentodes need some postitve bias on the screen, and even if wired your way, will still need some positive on both "grids".
The regular way of converting a pentode to a low noise triode is to connect the anode and the screen, and use the grid as a grid - this works very well.
I am a total fan of tube differential amplifiers in preamps, one of the few actually doing this in commercial practice, but I never go to the complexity of using a tube CCS. I find jfets do a far better job, far easier, and only need a circa -24 rail to operate. Quieter, more gM = higher OPZ, and far more reliable.
Check out the RTP3D circuit on my website: www.vacuumstate.com
Regards, Allen
Last edited:
Yes,this is how I normally triode connect pentodes with,additionally, grid 3 connected to the anode rather than the cathode. In this case,however,I connected grids 1 and 2 to the emitter of transistor and just used the usual arrangement with a zener from the base to the negative supply. It did not occur to me that the screen still knew it was the screen in this connection,if you know what I mean. I am sure you are right about jfets,however I suffer from what is probably a common fear in the relatively inexperienced in that I don't use fets because I am afraid of destroying them with static. I even go to the extremes of using a wristband with some fairly ordinary power BJT's or TL431C voltage references ! Many thanks.
Regards,Nicholas
Regards,Nicholas
- Status
- Not open for further replies.