more original s.e.x

[Paul Joppa]

The 6.3v winding of the transformer has two wires, with 6.3vAC between them. There are connected to tube socket terminals 7 and 8 respectively, so that the heater current (0.9 amps for each 6DN7) flows through the heater wire inside the tube - producing that dull orange glow.

There is a DC voltage difference between the heater wire and the cathodes. Only the tiniest current flows from heater to cathode. But the voltage difference can't be too large; most tubes are rated to withstand +/-100 volts maximum. Since one of the first-stage cathodes (B6) is at a high voltage (175v in the manual), the heater is biased halfway between the lowest cathode (A6 at 3v) and the highest (B6), or 90 volts.

There is other reason for this biasing; it reduces the possibility of hum being coupled from the AC-powered heater to the cathode. Traditionally the optimum way to do this is to place the bias at the midpoint of the 6.3v heater power, using either the 6.3v winding's center tap or a voltage divider (the two 100-ohm resistors that failed to show up on the schematic). The Rev 2.0 amp follows this traditional form of biasing the heater power even though it is only optimal for some 6DN7s, not all of them.

[Paul Joppa]

OK, thanks for the numbers. The voltages look basically good - some are further off that I'd like to see, but nothing indicates a wiring error - besides, it worked for a while!

So I'm going on guesswork here - I have an Original S.E.X. of my own, but I was not part of the design team back then, and more important not involved in tech support either.

However, based on the available data I suspect some instability of the tubes, specifically the power section which has an especially high transconductance - that makes tubes prone to unstable behavior throough tiny stray capacitances between wiring and the extremely tiny inductance of wire itself. The standard cure is what is called a "grid stopper" resistor, which damps the resonance before it can become an oscillation. So I dug out my own SEX amps and looked at the insides - sure enough, I had installed grid stoppers on mine at some point. I have no memory of doing this or of why I might have done it, but there they are. So I suggest you do the same.

Get four small resistors; any value from 100 ohms to 1000 ohms is suitable. Mine appear to be 100 ohms if these old eyes are reading the color code correctly, today we use 220 ohm stoppers in our products. Carbon composition is traditionally the best, but carbon film or metal film should work fine. On each tube socket, pin 1 is the offending grid. Unsolder whatever is connected there, and solder one end of the resistor as close to the terminal as possible and with the resistor vertical so there are no current flow kinks between the tube and the resistor. Kinks are inductive, and you want to separate the grid from all the parasitic inductances and capacitances with the resistor. Now re-connect whatever had been connected to that grid pin, to the other end of the resistor. Repeat on the remaining sockets.

[sl-15]

okay that makes sense, the only thing that was changed after swapping out the electrolytes on the tube sockets were the wires that are going from one tube to the other. i moved them around to get to the terminals and did not bend them back the way they were.

tonight i bent them back the way i had them originally and did not get any crazy sounds from the output transformer. then i was playing around with bending the wires with a wooden stick while the amp was on until the OT started to make high pitched sounds again. so the wires  definitely pick up something and i will get some resistors tomorrow and put them in.

is the high transconductance something that is a characteristic of this particular tube type (6DN7) or does one have it more than another like microphonics? is it something that gets worse the longer you use the tube?

anyways, thank you so much paul for taking the time to dig out your old amp and trouble shoot this. your help is very much appreciated. i will report back once i put the resistors in.

[Paul Joppa]

Thanks for doing the experiments, it pretty much confirms my guess about the problem. I'll add this experience to my notes on the Original SEX amp.

Transconductance is a property of the tube design - essentially, it's the current gain. It's a good thing, it just has some side issues. The power section of the 6DN7 has 7.7mA per volt transconductance, while the driver section is 2.5mA per volt. A 300B runs 5.5mA per volt, which was considered very high when it came out in the late thirties.

[Doc B.]

Guys, you are over-thinking this. That squeal that comes and goes when you move tube socket pins means you have a dirty tube socket or one with loose connections to the tube pins. Just crimp the socket sleeves a tiny bit to get a tighter fit on the pins. Might want to reheat the solder joints too.

[sl-15]

doc: interesting. the first thing i did when this happened is to deox the tube pins and sockets and i applied some jena labs contact enhancer. the tube sockets are still pretty tight but i can bend them even a little tighter. i will go ahead and reheat the tube pin connections as well.

paul: about the grid stoppers. you are saying to mount the resistor vertically pointing up from the tube pin and then connecting everything to the resistors other end. there is a 270K ohm resistor on A1 that goes to the ground bus and a cap that is connected to B1. do these also need to be connected to the grid stopper resistor or just the cable that connects A1 and B1? also what wattage rating should these resistors have.
thanks again

[Grainger49]

Stefan,

When I am trying to prove is a connection is good I use my meter.  I go from the component lead to something downstream of the solder joint.  In this case the tube pin is a good thing. 

[Paul Joppa]

...

paul: about the grid stoppers. you are saying to mount the resistor vertically pointing up from the tube pin and then connecting everything to the resistors other end. there is a 270K ohm resistor on A1 that goes to the ground bus and a cap that is connected to B1. do these also need to be connected to the grid stopper resistor or just the cable that connects A1 and B1? also what wattage rating should these resistors have.
thanks again

When I said "re-connect whatever had been connected to that grid pin, to the other end of the resistor" I meant everything. Only the grid stopper resistor goes to the grid pin; nothing else but the grid stopper goes to the grid pin.

The stopper carries no current, so it dissipates no power and can have any power rating.

[sl-15]

thanks for the clarification paul. i was further testing out the amps. i was trying to get them working without installing the grid stoppers which i am planning on doing eventually anyways though. just like i did the other night i was having them turned on, upside down, poking around with a wooden stick on the wires. this time with a multimeter connected to the output. with the meter i could tell very well which routing of the wires was the most quiet. the meter can also read frequency and every time the wires were in the wrong position the frequency would go up crazy high or i would see a ground hum at 60Hz. eventually i called it a day when i had the wires in such a way that the output was 5mV. it also says  in the manual that some wires need to be kept away from others.
i moved on to the second amp, which is the one with the failing cap that originally started this post. i was not able to get it"quiet" on the meter. i got a consistent higher output as well as a 120Hz hum on the meter. then to my surprise something started smoking and i easily could trace it back to the 5W wirewound resistor that is the last part of the power supply filter. i noticed already earlier that these get pretty warm but this time there must have been something wrong. it takes about 30 seconds after turning on and it starts to smoke. the good wirewound on the other amp gets about 130F hot and it measures 1Kohm in circuit. the smoking one measures 1.3kOhm in circuit.

i am wondering if the failure of the first filter cap somehow damaged the wirewound resistor and it took a while to show up.
 my question is if there is a different value i should put in place of the 5W. for example would a 10W be better and get less warm or is there a higher quality alternative to ceramic? if that is the case i would swap out the good one as well.  please let me know if you think there is anything else that i should swap out. although it is clear to me that the wirewound is the problem. it might even explain the hum of 120Hz. thanks

[Paul Joppa]

Yes, that 5-watt 1K resistor is under-rated by my current understanding - I recommend a 10-watt rating.

The "very high frequency" noise pretty much confirms that you have some oscillation going on. It may well be that the smoking amp is drawing more current as an oscillator than it would if working properly as an amplifier, which abuses that 5-watt resistor even more.

For what it's worth, here is a section of the "History of S.E.X." from the current draft - no quarantees that it's right, I'm still working on it, but this is the section specific to the Rev1.1 or Rev 2.0 with deYoung transformer:

======================================

The transformer and power circuit are satisfactory as is. However, if you are going to make any output upgrades, replace the 10uF/450v caps in the power supply with 47uF/450v or greater. I recommend 105 degrees C rating for longer life.

If you will retain the original output transformer, simply replace the 5-watt 1000-ohm resistor with a 10-watt 1000-ohm resistor, replace the 270-ohm 1-watt with a 270 ohm 5-watt part, and install grid stopper resistors at the power section grid, tube pin 1.

If you are not going to retain the original output transformer and circuit, remove the 1000 ohm 5 watt resistor and replace it with a piece of wire. Then change the cathode bias resistor from 270 ohms to 330 ohms 5 watt rating, and replace its bypass cap with one rated 25 or 35 volts. These changes will increase the available plate voltage, and reduce the plate current, for the best match to output transformers. At the same time, remove the feedback network (1200 or 2200 ohm resistor and 0.047uF capacitor) completely.  Install grid stopper resistors at the power section grid, tube pin 1.

With the above changes, the TFA-204 will perform very well.

If you want the benefits of parallel feed (recommended) then today (April 2013) you can use a Hammond 158L choke rated 15 henries, 75mADC with the Electronic Bottlehead OT-1 as used in Paramount. These will fit using some of the original  output transformer mounting hole plus one of the TFA-204 holes. Use the ones that get the iron farthest from the power transformer. An appropriate parafeed capacitor is 1.5 to 2.2uF. You should get somewhat deeper bass than the TFA-204.

Hopefully there will soon be a better alternative, with some new Electronic Tonalities iron. There will be an enclosed 60mA plate choke suitable for top-mounting, and an improved OT-1 as used in the Paramount, which can mount on the underside. These parts use the same mounting holes as the original output transformer and have greater bass extension and power handling.

The Magnequest BH-5 parafeed output transformer, or the similar TFA-2004Jr, is also suitable for installation under the chassis plate. The BH-7 plate choke has exposed terminals which would have the full high voltage on them, so it is not suitable for topside mounting.

[sl-15]

okay, quick update. i connected the working amp to the speakers again and had still some buzzing. more quiet this time but still noticeable. i then went ahead and put in the grid stopper resistors as you suggested. now the amp is dead quiet again. so the grid stoppers do make all the difference. next up is swapping out the wirewound resistors once i got new ones. thanks again paul for the technical support. very excited about your upcoming history documentation. her are a few pics. best, stefan