Simple Hifi Home Stereo Amp Builderino

Discussion in 'Shock Brother's DIY Amps' started by kleydejong, Feb 8, 2019.

  1. Billnchristy

    Billnchristy Tele-Afflicted

    Posts:
    1,109
    Joined:
    Sep 21, 2017
    Location:
    Georgia
    I agree with an asterisk. I have used 3w SET amps on 87db efficiency speakers and 90db is adequate volume but the dynamics aren’t there because when the music needs 8w, you just can’t deliver.

    You could jam some folk music with a watt or two.
     
  2. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    Hmm, I did just go through my tube stash and have a few old Baldwin 12AU7's that might be perfect for this project. I like this idea.

    [​IMG]

    I have some 1N4007 diodes on hand for rectification.

    The spec sheet I found for my PT indicates 120ma on the B+. Probably not enough for 6L6GC's. JJ 6V6S?
     
    FenderLover likes this.
  3. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    My current setup is nearfield in a home office.

    Good idea, I'll see if I can rig up the inputs and outputs to test it out.
     
  4. jsnwhite619

    jsnwhite619 Friend of Leo's

    Age:
    34
    Posts:
    2,264
    Joined:
    Sep 10, 2013
    Location:
    Georgia
    1/8" to 1/4" adapter is all it needs. Won't be stereo through an amp combo, but nothing fancy just to hear it.

    Sent from my XT1585 using Tapatalk
     
  5. FenderLover

    FenderLover Friend of Leo's

    Posts:
    4,556
    Joined:
    Jun 11, 2009
    Location:
    Minnesota
    Near field, like you said, that would work fine with 6V6. There's nothing really lost because if you did double the power the volume is not appreciably greater, and speaker efficiency is where you can realize gains if you need it.
     
  6. David Barnett

    David Barnett Doctor of Teleocity

    Age:
    62
    Posts:
    10,459
    Joined:
    Mar 17, 2003
    Location:
    The Far-Flung Isles of Langerhans
    Overnight sensations are likely less than 87dB/1W/1M. I can't find a sensitivity spec on Paul Carmody's site, but a pulled-from-my-ass guess would put them in the 84-85 range - they're 4.5 liter boxes.
     
    Billnchristy likes this.
  7. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    Guess I should put a new speaker build on the project list! I think I'll give it a shot and go from there.
     
  8. Billnchristy

    Billnchristy Tele-Afflicted

    Posts:
    1,109
    Joined:
    Sep 21, 2017
    Location:
    Georgia
    I have built a couple pair of Jordan single driver full range speakers but they aren’t super sensitive. I’ve always wanted to do some fostex but they are so dang ugly!

    I had a friend that had fostex 12” full range driven by a 1 watt set amp, he loved it.
     
  9. shortcircuit

    shortcircuit Tele-Meister

    Posts:
    175
    Joined:
    Sep 9, 2017
    Location:
    North Carolina
    I built one based on the Lacewood amp a while ago. I listen at a low to modest level in my office, and the amp is plenty loud. My schematic and pics:

    Hifi.jpg outside.jpg guts.jpg

    Instead of a choke, I used hefty caps to filter the power supply. The amp is dead quiet and sounds great. What I'd do differently though is instead of a 6SN7 for the driver, I'd use a 6SL7 for more gain, no cathode bypass caps for the driver, and global negative feedback to the cathode of the driver.
     
    Last edited: Feb 9, 2019
    kleydejong likes this.
  10. FenderLover

    FenderLover Friend of Leo's

    Posts:
    4,556
    Joined:
    Jun 11, 2009
    Location:
    Minnesota
    Cool build Shortcircuit. I can't imagine your gain is too low. Or maybe you are saying you would trade more gain to be reduced by NFB and no bypass. Kleydejong, that would be one way to use your stash of 12AX7's.

    err... that would be 680 ohm cathode for center bias with 12AU7 at your voltages.... sorry.
     
  11. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    That's extremely helpful, thank you! Great looking build, very much what I'd like to make here.

    Regarding the power supply - just confirming some details so I understand them. Please correct any mistakes:

    Because you used diodes you can pretty much use as large of a capacitor as you want. You wouldn't necessarily want to run the first B+ node before the 500r/50w resistor because we're dealing with a hifi amp and the extra filtering is needed whereas a guitar amp you can kind of get away with it being a bit more noisy and unrefined. The larger the filter cap the more filtering that occurs.

    The 500r/50w resistor must be 50w because there may be upwards of 130ma of current going to the four plates at 330v, therefore 330*130 = 42.9 watts. In most power supplies I've built for guitar amps the heavy current load going to the output transformer / power tube plates exits before any dropping resistors - hence the 2w-ish dropping resistors.

    Is the value of the 500r important? Other than to pinpoint your B+ for the 6V6 plates? I may have an 8r 50w resistor in the parts bin as a backup for my resistive load box.
     
  12. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    I have a big multi pack of almost every resistor a guy could ever want at 1/4w - so I'm okay to tweak the plate / cathode resistors a bit to work well with whatever tube and voltage setup I can get.

    To determine that center bias are you drawing load lines? I really need to wrap my brain around that...
     
  13. FenderLover

    FenderLover Friend of Leo's

    Posts:
    4,556
    Joined:
    Jun 11, 2009
    Location:
    Minnesota
    Load lines work well, but I use a program called TubeCad that I have found to be quick and very accurate. It's available from GlassWare Audio, but unfortunately it does not work on win 7 and newer OS.

    If you dropped in a 12AX7, 100K plate and 1500 cathode is center biased. With no cathode bypass and a little negative feedback like Shortcircuit suggested you would have better linearity and tamp down the gain a bit.
     
    kleydejong likes this.
  14. shortcircuit

    shortcircuit Tele-Meister

    Posts:
    175
    Joined:
    Sep 9, 2017
    Location:
    North Carolina
    Yep.

    Exactly, because of the diodes, I can use large capacitors, and I want the extra filtering. The big caps and the resistor are cheaper than a big choke.:D

    I'm using the 500R resistor to drop the voltage from 333V to 290V. The voltage drop is 43V, so the current across the 500R resistor is 86ma. The dissipation (voltage drop x current) is 3.7W. As you can see, a 50W rated resistor is totally overkill. I used the 50W resistor because I wanted to bolt it to the chassis. It was only a couple of bucks. The outside of the chassis where the resistor is bolted does get kind of warm though.

    The reason I wanted 290V going to the output transformer is that it will drop to around 275V at the plates of the 6V6 with my OT's. The guy who designed the Lacewood amplifier did optimization tests on the 6V6 tubes and reported there is point of diminishing returns on power vs THD above 275V on the plates. https://www.cascadetubes.com/optimization-of-the-6v6-se-ul-amp/ Can I tell the difference with my crummy speakers? No. But with the bias set at just above 9W and the low plate voltage, the 6V6's will last forever.
     
    D'tar and kleydejong like this.
  15. shortcircuit

    shortcircuit Tele-Meister

    Posts:
    175
    Joined:
    Sep 9, 2017
    Location:
    North Carolina
    BTW, in the picture, I have a set of NOS Sylvania tubes, but I am running Russian 6N8S (6SN7) and 6P6S (6V6). I didn't like the Russian tubes for guitar, but dang, they sound really good for hi-fi.
     
  16. bermuda

    bermuda TDPRI Member

    Age:
    54
    Posts:
    50
    Joined:
    May 15, 2017
    Location:
    Florida
  17. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    Needed to save some pennies to order some parts - including a pair of Edcor GXSE10-5K output transformers. I also have some small parts I need to order. In doing so I was reviewing my current iteration schematic.

    [​IMG]

    I feel like the B+ step down circuit could use some improvement. I could probably just copy a power supply designed by somebody who actually knows what they're doing. But I'm going to attempt to understand what's going on here.

    B+ Filtering

    Reviewing Blencowe: http://www.valvewizard.co.uk/smoothing.html

    A typical guitar amp like an AB763 Deluxe Reverb would feed the B+1 stage leading to the plates of the Power Tubes directly off the first filter cap. As stated earlier, this is fine for a guitar amp, perhaps not so much for a hifi amp. In my above schematic I have B+1 feeding off the second filter cap - but I don't have any resistor in between.

    As I understand Blencowe, the smoothing occurs with the RC low pass filter. The goal is to achieve filtering of all AC signals past 0hz which = DC. In theory one could chain a number of RC filters together in series to achieve really good B+ smoothing. This would require sufficiently high B+ and at some point the voltage drop across each resistor would become a problem. But that's the idea I'm working from.

    He offers a 22uf cap + 4.7k resistor producing a 1.5hz cutoff frequency. Just to check that for my own understanding. He offers the equation:

    f = 1 / (2 pi R C)

    So f = 1/(2*3.14*4700*22)

    f = 0.000001539


    Likely did the maths wrong in there - but I do see 1.5 as the result if you move the decimal over 6 places.

    Obviously I could just use an extra choke between those first two filter capacitors, but I'm kind of a cheap idiot and want to use what I've got on hand. Chokes seem quite optimal as they provide great ripple reduction and don't reduce the B+ as severely.

    So my idea is to swap the GZ34 to diodes. Then I'll add a resistor between C4 and C5.

    1hz frequency is the goal. Let's play with some C and R values.

    C = 32uf

    R = 4.7k

    0.0000010 = 1/(2*3.14*4700*32)

    I believe that would be approximately 1hz.

    Votlage Drop

    Next I want to ensure that the 4.7k isn't providing to strong a voltage drop.

    Using a lot of estimations here. The AB763 schematic says this power transformer + GZ34 rectifier will make 420vdc to the plates of the 6V6's. Going from a GZ34 to diodes should bring that up maybe 30vdc to about 450-460vdc.

    This will produce 111ma of current due to:

    6V6 = 45ma
    12AU7 = 10.5ma

    45 + 45 + 10.5 + 10.5 = 111ma

    If I take 450vdc * 111ma of current and run it through a 4.7k resistor = 521.

    Uh oh...

    A voltage drop of 521 is obviously not acceptable. Did I do that math right?

    Is this the practical reason why guitar amps never place an RC filter before feeding the plates of the power tubes? And the reason why a choke would be so helpful for a hifi circuit?

    Try Again

    As I'm trying to absorb this problem I reviewed some of the schematics above. I need to get the resistance down - like way down. I'm thinking like 250r to 500r - which can work as I have a few extra 10w cathode bias resistors on hand in that range. To keep the frequency point low I then need to compensate with higher capacitance - which I can do as I am now running diode rectification.

    If R = 330r

    and C = 300uf

    0.0000016 = 1/(2*3.14*330*300)

    That's quite close to what I had before with the 4.7k resistor + 22uf capacitor.

    I may give a little and try two 100uf capacitors in parallel for 200uf. R = 500r and C = 200uf.

    0.0000015 = 1/(2*3.14*500*200)

    Resistor Wattage

    Can the 500r / 10w resistor handle the power at this stage?

    V = IR

    55.5 = 111ma * 500r

    .111 amps * 55 voltage drop = 6.1 watts.

    I think you generally want to double this to be safe, so I think I'm close-ish at 10w of power handling.

    Schematic updated:

    [​IMG]
     
  18. David Barnett

    David Barnett Doctor of Teleocity

    Age:
    62
    Posts:
    10,459
    Joined:
    Mar 17, 2003
    Location:
    The Far-Flung Isles of Langerhans
    I have a couple of questions about htis circuit - (1) what's with the 220K feedback resistor around the 6V6? Shouldn't there be a small resistor between this and the plate of V1A? And there's no grid leak on V2, to load C2, is this intentional?
     
    kleydejong likes this.
  19. kleydejong

    kleydejong Tele-Holic

    Posts:
    660
    Joined:
    Aug 30, 2010
    Location:
    Orange City, IA
    I stole the 220k feedback resistor from this build:



    He describes it as a shade feedback resistor or an anode follower.

    I do believe there should also be a grid leak and a grid stopper on the power tube, thanks for pointing that out.
     
  20. Tom Kamphuys

    Tom Kamphuys Tele-Meister

    Age:
    39
    Posts:
    162
    Joined:
    Sep 18, 2018
    Location:
    The Netherlands
    uF is 1e-6F. If you use that in your equation, your frequency will have the decimal moved over 6 places.
     
    kleydejong likes this.
IMPORTANT: Treat everyone here with respect, no matter how difficult!
No sex, drug, political, religion or hate discussion permitted here.


  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.