1. Time for some upgrades in server hardware and software to enhance security and take AK to the next level. Please contribute what you can to sales@audiokarma.org at PayPal.com - Thanks from the AK Team
    Dismiss Notice

Transformer selection for stereo EL34?

Discussion in 'Tube Audio' started by mbates14, Sep 9, 2018.

  1. mbates14

    mbates14 Well-Known Member

    Messages:
    768
    So continuing on with my project, I am selecting components to use in my BOM for the amp I want to build.

    I had another thread that I lost. I forget what it was and where it was. anyways.

    I am building a stereo HiFi amp with EL34s. the output power of each channel I am trying to get at least 40W. Id like to push it to 50 but I dont really need to be THAT much.

    So, I am wondering what power transformer should I use? I did some googling and 50W pair of EL34s in an amp are 120ma each tube. thats almost 500ma on a transformer. that would be huge.

    That explains why the transformer I have now gets too hot to touch after an hour or so of operation.
     

     

    Please register to disable this ad.

  2. triode17

    triode17 Super Member

    Messages:
    1,299
    Are you planning on Ultralinear? Or PP Pentode?
     
  3. mbates14

    mbates14 Well-Known Member

    Messages:
    768
    That I dont know. What difference is that going to make?

    If I chose to use my current OPTs, then they do not have the taps. But if I go with the hammonds then maybe?
     
  4. 6DZ7

    6DZ7 Super Member

    Messages:
    2,176
    Ya, googling... first mistake is to take one bit as the only answer.
     
  5. kward

    kward AK Subscriber Subscriber

    Messages:
    3,458
    Location:
    Utah
    You should start by defining the output stage. UL or Pentode? What primary impedance? What output power? What output tube? Class A or AB1? Fixed bias or cathode biased? These things will largely define your power supply requirements and ultimately the power transformer specs you'll need for the high voltage winding.

    If you can't find a better place to start, look at the power requirements of the Dynaco ST70. 35 watts per channel x2 on a stereo chassis, class AB1, EL34's, 4300 primary impedance (I believe).

    It will be easier to build it UL in my opinion. Less technical expertise needed in getting the power supply right, but all else being equal, UL will deliver a little less output power than pentode mode.
     
    Last edited: Sep 11, 2018
  6. gadget73

    gadget73 junk junkie Subscriber

    Messages:
    36,755
    Location:
    Southern NJ
    Not sure if they sell parts alone, but the power transformer used on the Latino ST-120 amp would be a good choice. For rectication, I'd go SS or twin vacuum rectifiers for 50wpc.

    Other option is either mono amps, or a "twin power" stereo amp, basically two mono amps on a single chassis, each with it's own power transformer. Two of whatever you're currently using would be plenty sufficient I expect.
     

     

    Please register to disable this ad.

  7. kward

    kward AK Subscriber Subscriber

    Messages:
    3,458
    Location:
    Utah
    Another source you could look at is the EL34 data sheet. From the Mullard spec, it says for two valves in push pull, pentode wired, fixed bias, 3.5KΩ primary, 54 watts plate power can be obtained with a 400V plate supply. (that is power developed at the plate--power at the speaker terminals will be something like 0.85 of that = 46 watts). The spec sheet says the stage will demand 220 mA of current at full power output. That's for a single channel. Double that for two channel stereo. That's the DC current requirement for the output stage. Add maybe 15 mA additional for the frontend tubes needed to drive the output stage per channel, 30 mA total. So total DC current required is 470 mA..

    Transformer specs for the high voltage winding are specified in terms of AC current. The conversion factor is approx 1.6.if you are using capacitor input filtering in your power supply So you would need a secondary high voltage winding that can source 752 mA AC. If you are using a center tapped type, each half of the secondary will conduct each half cycle, so per side, so you need 376 mA AC per side. You can probably reduce that to 350 mA AC per side and be fine.

    Now let's say you use silicon rectification. To deliver 400V to the plates using capacitor input filtering in your power supply, would require an AC winding of approximately 315V. So the transformer spec you need on the high voltage secondary winding is 315-0-315 at 350 mA, that is if using a 3.5KΩ output transformer that can handle 50 watts AC power.
     
    tcdriver and crispycircuit like this.
  8. mbates14

    mbates14 Well-Known Member

    Messages:
    768
    I figured my question was going to funnel down 10 million more questions back to me. Yikes! I hate that, but it is what it is.

    Its going to be (I think i mentioned) a stereo EL34 amplifier. so 2 channels, push pull, maybe 6XX7 pre-amp phase splitter tubes. 5U4 rectifier. The OPTs I have are NOT ultralinear they dont have those taps. The impedance of those transformers? I Dunno, they are pulls from a Conn organ chassis. I guess I can figure that out somehow later.

    Thats my target. So what does this funnel down to?
     
  9. gadget73

    gadget73 junk junkie Subscriber

    Messages:
    36,755
    Location:
    Southern NJ
    Kinda need to know the impedance of the transformers to figure out the operating points on the tube, which will determine power output and current requirements.

    Its not real hard to figure out though. Feed a known AC voltage from plate lead to plate lead. Measure the output voltage. Input voltage / output voltage is turns ratio. Turns ratio squared * speaker impedance = plate to plate impedance.
     
    mbates14 and tcdriver like this.
  10. 6DZ7

    6DZ7 Super Member

    Messages:
    2,176
    Do you have the outputs pin 1 grounded?
     
  11. maxhifi

    maxhifi AK Subscriber Subscriber

    Messages:
    2,961
    Location:
    Edmonton Alberta Canada
    The tubes should be biased around 50mA each, and then under peak load, draw more current. The amp should idle at about 200mA or so, and then need more power to achieve full output. Since amps are seldom operated at or near full output, this may not be a big deal.

    Some heating of the power transformer is normal in tube amps with vintage iron, unless it's hot enough to burn, don't worry about it too much.

    To answer your question another way, you can expect about 60% efficiency from a typical class AB amp, let's call it 50% to be safe here. So you want 50W/ch here, that means we need 200W of power transformer. Let's say you have 450V of B+, that's about 0.5A to get your 200W. So the thumb to the wind type "engineering", says you're not far off with a 500mA estimate. You can go lower too though, since it won't be putting out peak power all the time. All it really needs to do continuously is the bias current.

    To make things easier on yourself, just look up the specs of the transformers Dynaco used in the MK II, which was a 50W EL34 amp. Ok even easier, I'm attaching them - make sure to double it for stereo.

    "Rated 400-0-400V (800VCT) at 200 ma (actually, will deliver up to 300 ma continuous, albeit at a slightly lower voltage) , 5V at 3A, 6.3VCT at 5A, also has 60V bias tap. Primary 117VAC 60 Hz. "

    Ref: http://triodeelectronics.com/mq782.html
     

     

    Please register to disable this ad.

  12. shelly_d

    shelly_d Not An Audiophool

    Messages:
    6,605
    Location:
    Alameda CA
    Here is something to consider:

    http://www.antekinc.com/as-4t360-400va-360v-transformer/

    Twin 360 V secondaries will each provide .55 amps. Will give you about 500 to 510 volts DC out of a full wave rectified, capacitive filtered supply (solid state diodes). Their website also includes this statement: " In most of the cases, this transformer can be output 20% more power from its rating at 60Hz power source without any problem. "

    I am using one of their toroids in a Stereo 70 I have that needed a power supply rebuild. Runs nice and cool in that application. (I used a different transformer, 320 volts I think because I went with solid state diodes).

    Will deliver up to 8 amps at 6 volts for the filaments.

    You will need to set up a bias supply for this however. I managed to figure out how to do that using a grounded center tap power supply but I just no longer remember the circuit. I would have to search the forum here to find the circuit. (It has been a very long time)

    Good Luck

    Shelly_D
     
  13. kward

    kward AK Subscriber Subscriber

    Messages:
    3,458
    Location:
    Utah
    Right...power is the product of voltage and current. The Volt-Amp (VA) rating of the power transformer I suggested for the 3.5KΩ output transformer was 630 * 0.35 = 220 VA. Therefore if you are using a 400-0-400 CT unit, to support the same power output, you'd still need a 220VA power transformer, but you get it by increasing the voltage and decreasing the current. Current needed would be 220VA = 800 * X, where X is then 275 mA, not so far off from maxhifi's 300 mA estimation (or "right on the money" from the perspective of an engineer's guestimate :)).

    I haven't done the math, but I bet a 400-0-400 high voltage secondary would be pretty close for a 4.3K to 5K primary for 50 watts output.
     
    Last edited: Sep 11, 2018
  14. shelly_d

    shelly_d Not An Audiophool

    Messages:
    6,605
    Location:
    Alameda CA
  15. mbates14

    mbates14 Well-Known Member

    Messages:
    768
    I forgot to mention, I want to try to keep the PS in tube format. Trying to keep this thing all tube as possible, with the exception of "maybe" the front-end input switching, but I dont know yet.

    I love all the suggestions so far.

    However, the toroid option woudlnt really work out, because I would have to "de-uglyfy" it somehow. sits atop the chassis.

    As far as output transformer impedence, I am still trying to figure out how that is figured into the equation? Sorry for being math stupid, but I thought the current is drawn by the tube? not the transformer?

    Well, I know it still has to drive the speaker so I know there will be current drawn that way, but... ow my head hurts.

    To be honest, I am from solid-state digital land so going back to tubes is definitely presenting its challenges to me.
     
  16. mbates14

    mbates14 Well-Known Member

    Messages:
    768
    The chassis I am stealing the "iron" from is the one I am speaking about with that issue. Not sure what hot enough to burn is, but after a hour of good operation, maybe 2, the transformer is hot enough you cant keep your hand on it. soon as you touch it, it feels "burning" to the skin and have to move it away. like power tubes that have been on and biased fully for a couple minutes or so.

    This was the ultimate concern of mine which lead to this thread to see if I needed to upgrade the transformer.

    It originally powered 2 channels, one of 6L6GCs, and one of 6V6s. So I know with 2 channels of JJ E34Ls I may be taxing it a bit, but it works ok.

    Trying to save money on "iron", I wanted to part this guy out into a new amp design that is going to be my dream amp. Then again, I dont know if the OPTs on this guy can even handle 40 to 50w. But they do look exactly the same size as the A470s used in Dynaco so, maybe close?

    Who knows! maybe better off starting from scratch and selling the chassis as is. it is functional, technically. Also the amp is silent. no hiss, no hum, still has all of its original caps. Thats what I am trying to replicate in the new design is that quietness, but still good performance.

    I think the key to all that silence is the fact that they are not using the chassis as ground, they have a ground bus wire that starts at the pre-amp, input jacks, and snakes back through the phase splitter, the output tubes, and then the power supply ending there. Even the pre-amp tube filter caps are grounded at the other end of the bus.

    But I digress, I sort of ran this one into the same territory that my previous thread covered, so whoops on that one.
     
    Last edited: Sep 11, 2018

     

    Please register to disable this ad.

  17. kward

    kward AK Subscriber Subscriber

    Messages:
    3,458
    Location:
    Utah
    The impedance of the primary is the load the tube works into. Less impedance in the primary, the more current is allowed to flow. More impedance in the primary, less current is allowed to flow. So the impedance of the primary has a direct effect on how much current the power supply needs to provide.

    The easiest way to make some decisions here is to just pick your output transformers. The rest of the circuit can be tailored around them.
    Well I would recommend silicon rectification since this is a lot of current needed. But...if you really wanted to go with tube rectification, you will need dual 5AR4 rectifiers to support a 50+50 watt amp. The 5AR4 will drop about 35 to 40 volts DC just to run them. So you need to add 35 to 40V DC into the calculations so that when it is dropped across the rectifiers you are left with the plate voltage you want.

    If you go with the example I provided in post #7, the power transformer needed was a 315-0-315 secondary. But you need 40V DC additional so that when the 40V is dropped across the rectifier, you are left with 400V DC delivered to the plates. The voltage conversion factor (after transformer losses) is approximately 1.3. Doing some math: 40V DC / 1.3 = 31V AC extra needed from the high voltage secondary winding. Thus, with 5AR4 rectification, you need a 315 VAC + 31 VAC = 346 VAC secondary. That's close enough to 350 VAC, which is a quite common secondary, so you would go with a 350-0-350V secondary.

    Double checking the math: 350 VAC * 1.3 = 455V DC (with cap input filtering in the powers supply). Subtract 40V for the rectifier drop = 415V DC. Subtract 15V (an educated guess) to do some additional filtering through a choke) = 400V DC.

    Dual 5AR4 rectifiers will require 5 VAC at 3,8A filament current.
    4x EL34 tubes will require 6.3V at 6A filament current.
    Frontend tubes will require about 2A additional filament current. So you need a 6.3V secondary at 8A. You probably won't find one that big in an off-the-shelf unit, but you could go with dual 6.3V secondaries at 4A each

    Summing up for this example, you need a power transformer with the following specs (assuming stereo chassis build):
    • 350-0-350 center tapped at 350 mA
    • 5V at 3.8A
    • 2x 6.3V at 4A
    Edcor has a transformer that almost meets specs, the XPWR173. Its specs are:
    • 350-0-350 at 400 mA
    • 6.3V at 7A
    You would need an extra 5V transformer at 3.8A and an extra 6.3V transformer at 2A to round out your transformer needs.

    But this is just an example. You really need to know the primary impedance on the output transformers before you get too much further along here.
     
    Last edited: Sep 11, 2018
  18. kward

    kward AK Subscriber Subscriber

    Messages:
    3,458
    Location:
    Utah
    Nope...won't perform well for a dual channel stereo amp unless you have two identical output transformers. One channel running on 6L6's and the other on 6V6's definitely will not use the same output transformers....unless the engineers ignored common sense and used the same output transformers anyway. Possibly both are 8K or 10K primaries. Check the part numbers on those transformers. If they are the same, you might be able to proceed.

    But still need to know the primary impedances...

    One more thing to check--how many 6L6 sockets and how many 6V6 sockets? I expect either 2 or 4 per channel.
     
  19. shelly_d

    shelly_d Not An Audiophool

    Messages:
    6,605
    Location:
    Alameda CA
    Transformer cover:

    [​IMG]

    Here

    Shelly_D
     
  20. maxhifi

    maxhifi AK Subscriber Subscriber

    Messages:
    2,961
    Location:
    Edmonton Alberta Canada
    Agree with kward, no way are the transformers identical if one used 6V6 and the other 6L6. Why not save up and get some Dynaco transformers, would be a real shame to invest a lot of time and money, and be let down by transformers which aren't suitable for the job.
     
    crispycircuit likes this.

Share This Page