Tube datasheets are only casual suggestions, right?

[imaradiostar]

100th post! It'll probably get ignored because it's too long but I'll post anyway because I'm excited about my projects.

I have a ton of iron I'm trying to use up while I get better at building amps. I'm trying to build some stuff for family and friends and give or sell it to them cheap and chalk it up for experience. You could say I'm trying to "hone my craft" so I'm building like crazy- I have a mini EL84 plexi, a jtm45, a SE el34 amp, a pentode into 7591a pushpull amp, a dumble clone, a stereo blackface champ with reverb, a super lead with 4 miniature beam tube outputs and any of a number of other amps partially completed and stacked up next to my bench.

So this brings me to my question- how loosely should I interpret the tube datasheets?

I don't know enough about how to use a load line but I assume you can run a tube far above its rated Zp.

Example 1: EL34 single ended amp, plate volts around 460, screen around 325. I would usually use lower plate voltage and run the screen and plates at almost the same voltage. I've read in other threads that pulling back screen voltages will allow a single-ended output section to saturate more symmetrically- makes sense, is it true? (I haven't scoped it yet to compare) Assuming I'm using close to twice the normal SE EL34 plate voltage I'm going to use a Hammond 125 ESE at 5k. Does this make sense? There is no info on the datasheets to tell me if this is right or not. I don't know enough about load lines to know if I can calculate load this way.

Example 2: EL34's or KT66's with a low voltage push-pull output, probably cathode biased. I have a PT that originally ran two 6bm8 output sections- plate voltage was high and the 6bm8's cooked on modern voltages and they were originally run over the datasheet values but not by much. The amp had a ton of tubes (preamp) so there is an abundance of filament current but not a ton of plate current. IIRC unloaded AC plate voltage is around 285-300 per leg, DCR was about 60 ohms per side. I wanna run a pair of KT66's or EL34's and per Duncan's PSU calc and practical experience with the PT it'll make about 330 volts at around 140ma. This is about all I feel safe pulling from it at rest. The kt66 data says 4k at 250 volts and 8k at 390 with much lower screens. I would probably have the screen close to the plate volts- 325ish. Could I run a push-pull cathode biased output at somewhere between 6k and 8k load to lower the full load current draw? I assume this would prevent excessive current draw under load. Same applies for EL34's- I rarely see them run at higher plate loads unless plate voltage is over 500. If they're biased up to 80-100% of dissipation will they sound decent in spite of the mismatch? I'm trying to use higher power tubes but I don't wanna make high power!

Example 3: 7189's at 425 volts or so...screen voltages closer to 300 volts. Could I crank the Zpri up to prevent overloading the tubes? I have a nice 9k hifi output that would make a neat amp with fixed bias high voltage el84's and some sort of Marshally front end.

Example 4: "JAN" military tubes. I have 5686's (12ax7 sized 8w beam power tube), 6005w's (6aq5) and a few others that are rated for rather low plate voltages- under 300 volts. Is it safe to assume the Army had them overbuilt in the interests of pilot's radios and intercoms still working in extreme conditions? If I were to treat them as the above example and exceed the rather low plate voltage maximums would I have to lower screen voltage substantially (ie: beam tubes, less screen current needed) to keep them from blowing up or would higher Zpri be enough?

I know this has been a mouthful- I just wanted to hear practical examples and reflection from those that know!

thanks,

jamie

[Steve Conner]

Hi Jamie

Indeed, the datasheet is just like the "Serving suggestion" picture on the front of a packet of cookie dough, that shows a plate of perfect cookies.

The same tube had different datasheets from its different manufacturers. If you look at the Philips EL34 datasheet, there are about 20 sets of example conditions all worked out for you, covering everything from a single tube Class A at 10 watts, to a push-pull pair cranking 100W off of 800V B+. It tells you what output transformer Z to use, what percentage distortion you'll get and so on, for each set of conditions. (The example you gave with 460 plate and 325 screen is one of them, IIRC.)

Plate voltage is more of a guideline than a deadline. In the olden days there was no distinction in datasheets between recommended operating conditions, and absolute maximum OMG-Do-Not-Exceed values. I've seen people run 2 or 4 times the maximum plate voltage it says in the datasheet.

OTOH, screen power dissipation is one of the OMGDNE's, and somewhere in some datasheet, it says that the OMGDNE plate voltage for the EL34 is 2kV peak.

For tubes that weren't intended for the kind of service you intend to put them into, the maker won't have bothered to characterize them in that way, so you have to make up the "recipe" as you go along. Tubes are forgiving enough that you can just mess around until you get something that "tastes good", and if you don't see any parts of the tube glowing red that shouldn't, it'll probably be fine. It's no big deal if your cookies don't look like the picture on the packet.

[6267]

Most tubes can take a higher plate voltage without arcing over, but they become harder to bias, requiring higher and higher load impedances and closer to class B operation. The screens have gain, essentially, so some tubes need fewer volts here for the same effective power out as other tubes. They will melt with too high voltages, you'll see them glowing red or gold through any holes in the plate that may exist. If you are using tubes with big differences between screen and plate maximum voltages, it's in your best interest to use a separate screen supply.

[imaradiostar]

You mention el34 data close to what I wanna try- I've been unable to find class A single ended data over 300 volts. Would limiting screen voltage to around 300 and using plate voltage around 450 be the right way to get an EL34 to clip more symmetrically in SE use?

I've looked at your Toaster power supply. I don't suppose you know of a simple way to generate a regulated screen supply- maybe some sort of resistor/zener combo establishing a screen voltage then a large resistor and mosfet following it to do the heavy lifting?

On my single ended el34 design (not yet running) I intend to use to separate sets of diodes- one into a cap for plate supply, one into a choke input filter for screen and preamp supply. I'm excited to see how it sounds because it sounds great in my mind! I previously built a similar amp for my little brother but used about 360 volts for the plate and screens and it didn't sound quite right when driven hard- I'm assuming this is because it was clipping unevenly because of the highish plate and screen voltages for SE use.

jamie

[Steve Conner]

Oh, single ended. That's a different story: I've not seen data for SE operation of an EL34 much over 300v either. All of the higher voltages are for push-pull pairs in Class-AB.

The reason for this is what Mr. 6267 said. As you use more plate voltage, you have to run the tube at a lower and lower current, to keep the plate dissipation below burning-up point. That would make it clip less symmetrically, not more, in SE service, as the tube can now deliver much higher peak currents in relation to its idle current. And symmetrical clipping demands that it peak out at twice the idle current.

Unless you adjusted the screen voltage for symmetrical clipping, I guess, then you could use any plate voltage you wanted. The screen voltage sets the maximum current the tube can deliver. With 460 on the plate, I'd guess maybe 250 on the screen.

Oh, and if your screens are getting hot, a higher primary Z will just make them worse. The classic example is the open-circuit speaker load resulting in instant screen death.

[Old Tele man]

...if you're so inclined to treat Doctor's Prescription Instructions as "...only casual suggestions", well maybe current tube datasheets could be construed as "casual suggestions" too.

...the original manufacturers (not current importers) published the "limiting" and "recommended maximum" values for people (engineers) to go by when they designed circuits to use those tubes. Moreover, most of the example circuits (Class A, Class-AB1, Class-B, Class-C, etc.) were laboratory-proven designs that 'showcased' the capabilities of their tubes...and were usually optimized for one aspect or another.

...so, what TODAY might be 'casual,' back in the HEY-DAYs was 'gospel' information to engineers, technicians, and TV-repairmen/women.

[Enzo]

When RCA or someone published specs for their tubes back when, it was with the idea that your designs would be things like hifis, radios, TVs. The specs were published so that if you followed them, you could expect long reliable life from their tubes. Mom and Dad would not be happy having to put a new pair of output tubes in their TV every few months. When we design guitar amps, we are expecting to wear out the tubes in short order, and we are putting the tubes into overdriven situations that create distortion. The specs were published thinking designers would want to minimize distortion.

Casual is maybe the wrong term. When Leo Fender built his amps with an extra 100v on his 6V6s, he wasn't building a table radio.

Think of it like a car. Your mom buys a nice average car to drive around town in. If she follows the specs in the owners manual she should get long reliable service from that car. Now imagine you bought that same car, but you plan to race it at the track. YOu won't be following the manual.

[imaradiostar]

I think it's cool that the Dr. Z Route 66 is lifetime guaranteed...because he was careful to follow the guidelines for reliable KT66 usage outlined in the datasheet.

I think part of what bugs me is how difficult it is to duplicate the exact conditions of the RCA manual or other manufacturer info. Most datasheets call for screen voltages a hundred or more volts lower than the plate. Most amps seem to ignore that and have the screens at or above plate voltage. Did they expect use to use regulated screen supplies or huge dropping resistors or what?

jamie

[Enzo]

They don't have to be regulated. You can drop to them or even use a separate supply. What amp am I thinking of, Music man maybe, with super high B+ on the plates and the screens are running on something like 350v. They are runing off separate supplies. Most of your amps already have a string of drops along the B+ line. The screens would be no more odd downstream than say the reverb stage.

[Steve Conner]

Ah, screen voltages. It seems to me that the screen voltage is the most commonly ignored limiting value.

When Jim Marshall designed his first amps, he completely ripped off the Fender designs that used 6L6s. Now, 6L6GCs are quite happy running with a high screen voltage, so Fender hooked the screen supply up to almost the same voltage as the plates, with just a 470 ohm resistor. Marshall started out with KT66s, and those seemed to handle the high screen voltage all right, too. But the EL34s that Jim ended up using couldn't take it, and tended to blow out their screens. Legend has it that people who played them cranked, including Hendrix, had to change to 6550s to make their amps hold together for the length of a gig.

Ever since, people consider the plexi to be the holy grail and copy the circuit in spite of this shortcoming.

The "bible", RDH4, says that the limiting screen voltage printed in the datasheet can be exceeded, provided that you use a screen resistor big enough to sag the instantaneous screen voltage below the datasheet limit on current peaks. This works because it's not excessive screen voltage that damages the tubes, it's excessive screen power dissipation.

This is what the majority of guitar amps do: 1.5k seems to work for EL34s, and 470 ohms for 6L6 type tubes. I think they use this method simply because it's the cheapest way of dropping the screen voltage. However, I've heard tell of amps that use gas tube regulators, voltage dividers, and so on, for the screen supply.

The Toaster power supply I designed could be simplified still further, to a string of zeners, a MOSFET, and a couple of resistors. I used the TL431 cascoded with a HV MOSFET instead of zeners, because I wanted the voltage to be adjustable.

[imaradiostar]

I have a bunch of 12 volt 5 watt zeners- any reason why I couldn't use them to drop the voltage for the screen supply on a small amp? Maybe with a resistor to ground to make sure there is enough current to make them function as they should?

jamie