>

Biasing A 40 Watt Amp With El34 Tubes

Discussion in 'The Workbench' started by matttornado, Sep 18, 2018.

  1. Pete Farrington

    Pete Farrington Well-Known Member

    Joined:
    Jan 18, 2021
    Messages:
    573
    Likes Received:
    556
    I’m confused - where’s the 30 watt rating come from?
     
    Matopotato likes this.
  2. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    I think @myersbw may have been looking at the KT66 plate absolute rating of 30W instead of the design rating of 25W.
    Do not use the "absolute" rating for reference on nominal bias set.
     
    myersbw likes this.
  3. Pete Farrington

    Pete Farrington Well-Known Member

    Joined:
    Jan 18, 2021
    Messages:
    573
    Likes Received:
    556
    Yikes, yes, ‘absolute max’ ratings are an absolute swizz. As are, to a lesser extent, design max. Because they don’t factor in component tolerance (eg OT and PT winding turns), valve tolerance (abs max), and mains voltage variation.
    As I see it, if the goal is to avoid the valve’s anode exceeding its dissipation limit under any reasonably foreseeable combination of scenarios, then a design centre rating (or an equivalent thereof) is the data point that needs to be used for the calculations.
    To derive a valve type's design centre rating from a design max, the industry standard seems to be to multiply the design max figure by about 0.87.
    Which brings the KT66 anode dissipation limit to about 22W. So in fixed bias AB operation, idle them below 16W.
    My suspicion is that the only reason that its 30W absolute max limit was ever published was to make it seem competitive with 6L6GC, ie a marketing con.

    Here's a bit of background info regarding valve rating systems from RC20 p65 http://tubebooks.org/tubedata/RC20.pdf
    [​IMG]

    Here's the RCA application note justifying / introducing the design max system to the industry http://www.one-electron.com/Archives/RCA/RCA-AppNotes/RCA 1958 AN-174 Design-Maximum System for Rating Electron Tubes.pdf
     
    Last edited: Jul 11, 2021
    myersbw likes this.
  4. matttornado

    matttornado Well-Known Member

    Joined:
    Mar 14, 2012
    Messages:
    1,523
    Likes Received:
    1,030
    I recently put a set of Svetlana EL34s in and with the bias adjustment knobs cranked, i’m at 34 mA each. i did not open the amp up to measure the plate V again. Amp sounds good so i’m happy with it.
     
    myersbw likes this.
  5. Pete Farrington

    Pete Farrington Well-Known Member

    Joined:
    Jan 18, 2021
    Messages:
    573
    Likes Received:
    556
    And that’s fine. Once you know what it is for a given set of conditions (principally the mains voltage and the output valve idle current), how can it change?
    FYI when you get chassis out next, the HT voltage supplied to the OT’s CT is a better metric. Because it eliminates the possibility of inducing oscillation, which can easily occur when attempting to measure the anode voltage of output valves.
     
    matttornado and tschrama like this.
  6. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    I know all of that but am sure others appreciate the information. :D

    The plate wattage rating is only for the power tube plate itself.
    We and I have been preaching here that all these amplifiers have design specifications which should be adhered and the power tubes intended use/operation needs to fall into those parameters.

    Also the KT66 came out way before 6L6GC because it was an answer to and in competition with the 6L6's of 1936/1937.
     
    Last edited: Jul 11, 2021
    Pete Farrington likes this.
  7. myersbw

    myersbw Well-Known Member

    Joined:
    Oct 22, 2015
    Messages:
    1,295
    Likes Received:
    776
    Location:
    SW Ohio
    Ha! Surely y'all know how we like to abuse these tubes for squeezing 150 watts audio out of an 18 watt Marshall! :band: (But, the adapters needed are awkward!)

    All kidding aside, go take at the Tung-Sol data sheet. It gets interesting how obscure data sheets can relay data. Tung-Sol will give just one value to plate current..."not more than" 30 watts. Which, isn't that normally what we expect to see in a data sheet? Usually, you would think a "rating" is absolute max unless specifically stated so, but many don't relate the info in that way. But, it was interesting to see how the plate current "rose to the occasion"...literally. 25 watts is certainly a conservative rating for the Gold Lions (as I'd wish all manufacturers would do....and agree on specific reporting, but they don't).

    Yep...and I wonder if that's why Tung-Sol and others like to hint at a 30 watt plate dissipation. In regard to wattage, it's nice to see the conservative side with manufacturers like Bad Cat and a few others. You get a comfortable plus side in regard to their amp rated output. I suppose I'd best leave the power rating topic for another post. Given some wildly posted peak-to-peak, peak, rms numbers given in specs with the intent to mislead the consumer.
     
  8. myersbw

    myersbw Well-Known Member

    Joined:
    Oct 22, 2015
    Messages:
    1,295
    Likes Received:
    776
    Location:
    SW Ohio
    Agreed. And, this is a real danger when we assume the OP's of posts understand snippets of information given in this forum. I know there are a handful (of regular posters anyways) that have a good handle on safety procedures, ability to analyze a given circuit correctly. I try to be as overly detailed as possible (and longwinded at that). But, better that than hearing someone ended up dead as a result of what isn't said and where assumptions about a skillset is made.

    Now, one fault I do have with the Gold Lion data is it lists "maximum signal"...I don't read that as "maximum rating". Consider the fact we often exceed a "maximum signal" in many a Marshall head and you have to acknowledge how often we abuse the tubes beyond their "rating"...and, they continue to work. I've accepted the obscurity in how they differ language, but it would be nice to see a consistency of language that's more absolute.

    It's also one reason I really love the old RCA manuals' data. It wasn't written with just data...it was almost a taught class in itself!
     
  9. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    I am going to say that "not more than" is Russian for "absolute" in this case. :D

    The other datasheets usually indicate 30W "Absolute" and 25W "Design" I am pretty sure Tung-Sol falls in line with that.
     
    myersbw likes this.
  10. Pete Farrington

    Pete Farrington Well-Known Member

    Joined:
    Jan 18, 2021
    Messages:
    573
    Likes Received:
    556
    Indeed, however publication of KT66 absolute max anode limits seems to have been a ploy that GEC decided to use in the mid 70s.
    GEC M-O KT66 April 1977 https://frank.pocnet.net/sheets/086/k/KT66.pdf
    Perhaps to give a false impression that the KT66 was competitive with the 6L6GC, and the KT88 with the 6550A (the latter seems to have been introduced in the early 1970s).
    Sorry, I meant to include that in my last post, to show what I was referring to.
    KT66 valve info published prior to the 70s mentions only a max anode rating, rating system unspecified, possibly an in-house Marconi system, as the design max system wasn't officially introduced (to North America, at least) until 1957.
    Marconi KT66 June 1956 https://frank.pocnet.net/sheets/126/k/KT66.pdf
    Actually the earliest published type info I can find doesn't even mention an anode dissipation rating!
    Marconi KT66 supposedly 1938 https://frank.pocnet.net/sheets/179/k/KT66.pdf

    Tung Sol Drg #GT00014 KT66 https://tungsol.com/specs/kt66-tung-sol.pdf
    These days, the conformance of valves to their type norm seems to be totally unregulated, as the above data doesn't conform to KT66 type nominal, almost nothing seems to align.
    The 'cathode current - no more than 80mA' is ridiculous, standard KT66 circuits require cathode currents far higher than that, see 'Typical Operation' sections in the above vintage info.
    And the nominal heater current of 0.9A is some way below that of the 1.27A of KT66. Perhaps the lower heater power explains the lower cathode capability?
    And the cap it all, the anode pentode characteristic curves don't note the screen grid voltage they apply to
    Anyway, whatever the marketing etc states, that valve is not a KT66, the evidence of that is clearly presented in its published info.

    Even when valves were essential, that didn't happen, so fat chance of it now, especially as the only manufacturers and engineers still in operation have never trained / operated under Western systems :)

    New Sensor Corporation Genalex KT66 https://www.ehx.com/wp-content/uploads/2020/11/kt66-genalex.pdf
    Surely the key issue with above is that is a crudely amended version of info previously published by GEC in the 1950s. By cynically presenting vintage publications as their own, New Sensor would seem to be displaying a complete disregard of truth and accuracy, and contempt for valve purchasers. Such behavior disgusts me.
    As they're prepared to engage in such practices, how can any info from them be considered credible?
    I suspect that their business model is predicated on the basis of operating in an unregulated seller's market, which fundamentally is grateful that valve manufacture is still ongoing.
    Which, though it may stick in the craw to admit it, we have to be.

    EDIT - Oh dear, here's some British shysters doing the exact same thing, ie amending vintage valve info and passing it off as their own. The difference being that this lot are just rebranding valves made elsewhere.
    https://brimaruk.com/BrimarUK/wp-content/uploads/Brimar-KT66-DataSheet.pdf

    Assuming that you're reading that from the 'Typical Operation' section of the info, I agree, eg
    [​IMG]

    The above is showing a set of operating conditions, and the results that are achievable. Very usefully, it includes conditions for both idle ('no signal') and max unclipped signal power output P out ('max signal'), so we can see how stiff the HT supplies need to be.
    The above are not ratings / limiting values, but rather demonstrate how the valve might be used to achieve high audio power outputs, without exceeding the valve's limiting values.
     
    Last edited: Jul 12, 2021
    myersbw likes this.
  11. myersbw

    myersbw Well-Known Member

    Joined:
    Oct 22, 2015
    Messages:
    1,295
    Likes Received:
    776
    Location:
    SW Ohio
    LOL...I agree! I think I need what Pete is drinking. Actually, forget the pint...bourbon is my preference these days!

    And, yes, @Pete Farrington , how can you trust what you see? So long as we're schooled enough to know how to do the math properly and "read between the lines" when data is obscure...we're fine. I'm still concerned about the many want-to-be techs that venture outside of some basic maintenance at times.

    I suppose all must be good thus far. I have yet to hear of anyone succumbing to chaotic service attempts due to misleading info in here. But, then again...how would we know?
     
    Pete Farrington likes this.
  12. Chris-in-LA

    Chris-in-LA Well-Known Member

    Joined:
    Jul 19, 2018
    Messages:
    2,737
    Likes Received:
    2,589
    Location:
    Los Angeles
    If I understand you, using the 1R method adds 3 to 4 Mv to the reading? If so, I should be setting the bias a little hotter as read by my meter. Does this hold true with other tubes as well? 6l6? Kt66?
     
  13. Pete Farrington

    Pete Farrington Well-Known Member

    Joined:
    Jan 18, 2021
    Messages:
    573
    Likes Received:
    556
    As cathode current includes anode and screen grid current, that’s an option.
    But we’re dealing with valve guitar amps, not lab equipment / guided missiles, and so it seems reasonable to ignore the discrepancy. Particularly as it’s a ‘safe side’ error.
    Suppressor grid type pentodes (eg EL34) will have a higher proportion of screen grid current than beam type pentodes (eg 6L6, KT66).
    With typical ohm screen grid resistor values, it’s easy to get the screen grid current, just put voltmeter probes on either leg, eg 1VDC across 1k means 1mA current.
     
  14. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    Like @Pete Farrington mentiond the cathode current includes the screen cuurent.
    The value is usually about 10% and that does depend on the power tube type but is also associated with the plate voltage.
    So for cathode current measurements 6L6/KT66/5881/7581 type plate at 30mA would have screen at about 3mA for total of 33mA and 6CA7/KT77/EL34 type plate at 40mA would have screen at about 4mA for total of 44mA and so on.
    Still it is a guestament and you should try your amplifier with that set of tubes at different volumes and see were you like idle bias best for your playing.
     
    Last edited: Jul 15, 2021
    Chris-in-LA and Pete Farrington like this.
  15. Matopotato

    Matopotato Member

    Joined:
    Mar 3, 2021
    Messages:
    45
    Likes Received:
    38
    @mickeydg5
    I swapped the power tubes between themselves and the fluctuating swapped to the other side. So it seems tube related.
    Lubricated with some special spray for that purpose. Not sure there was any difference, but no harm, so I take it as a good thing.
    Tried to dial in but hard to hear what sounds good and not. 22mV sounded weakish and 42mV a bit too hot. Will see how 37 or 38 feels.
    Main issue remaining is that clean channel sounds quite thin or low. Definite higher output from the other three. Tweakin gain, vol, master but same result.
    Is there any of the preamp tubes that is more dedicated to the clean channel?
     
  16. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    If the bias fluctuation follows the power tube then it is a tube issue.
    You need to stop the fluctuation before deciding on a final bias number but if 38mA suits you then that is the ticket.

    Is the CLEAN channel low on both CLASSIC and CRUNCH?
     
    Matopotato and Pete Farrington like this.
  17. Matopotato

    Matopotato Member

    Joined:
    Mar 3, 2021
    Messages:
    45
    Likes Received:
    38
    Thanks @mickeydg5
    OK, tube sounds likely then.
    Not sure how to stop the fluctuation, since it is always there?
    When trying to measure it flows and is very sensitive to the touch compared with the other tube.
    When left alone for a while, the steady one keeps its biasing but the fluctuating one could have gone off on some tangent. Pretty random.
    It's a DSL40CR, so the channels are Classic gain, green (the one I call CLEAN). This is lower than the 3 others (Classic Red = CRUNCH, Ultra Gain Red & Green = OD1&OD2). Those three are somewhat within a similar range.
    Turning the gain around has effect of course, but keeping at the same and the Clean quieter vs 3 others louder stands out. I keep it around 9-12.
     
  18. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    As for the power tubes, make sure the pins on the power tubes are cleaned. Hit those pins with a cleaner/lubricant and try again. If the fluctuation persist then another set of tubes are in order.
    If that set is new and under warranty you know what to do.
     
  19. mickeydg5

    mickeydg5 Well-Known Member

    Joined:
    Sep 30, 2011
    Messages:
    26,709
    Likes Received:
    13,887
    Location:
    united States of America
    As for the CLEAN being knocked down in volume it makes me wonder.
    Do you have jumper wires? Like wires with alligator clips on each end or something similar to use?
     
  20. Matopotato

    Matopotato Member

    Joined:
    Mar 3, 2021
    Messages:
    45
    Likes Received:
    38
    For the multimeter? No I use the sort of "fat needle" probes. I can get one to stick in its biasing input. The ground can touch anywhere on the chassis.
    Or should i use alligator clips for some other measurement? Like the various pins on the tube?
    If there is anything with the clean, and the other 3 being ok, i would have guessed either of the preamp tubes. But not sure which one or why...
     

Share This Page