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.

    Dismiss Notice

Millett "Starving Student" hybrid amp

First
 
Back
331 332 333 334 335 336 337 338 339 340
342 343 344 345 346 347 348 349 350 351
Next
 
Last
  1. hellomoto
    Quote:

    Originally Posted by tomb /img/forum/go_quote.gif
    There are some very nice Nichicon Fine Gold that will fit the bill that you can purchase from Handmade Electronics. I believe Mouser is carrying some now, too, but I'm not exactly sure on that.




    Couldn't find anything on Mouser. Is it okay to buy normal ones, and then if I want buy new ones to replace the standard ones?
    Quote:

    Originally Posted by tomb /img/forum/go_quote.gif
    As for the film caps - the sky's the limit. I'm not sure how much difference you're going to notice, though. The parts selection we made on the SSMH PCB website is about as good as it gets.[​IMG]



    The same as the Electrolytics.


    I've seen on your (very well made) build-thread that you added led. If i'm right, I just have to buy standards 3mm leds, and standard wire, and drill the sockets to put them in the AMP Right?

    Thanks you very very much for your help tomb!
     
  2. tomb
    Quote:

    Originally Posted by n_maher /img/forum/go_quote.gif
    Has anyone explained what exactly is going on that is so unique to the SSMH design that causes the issue? I read both your comments and Ti's and while there is anecdotal evidence to support that something is going on it seems like that something should have a more complete explanation. If you point me in the right direction I'll take it up with Pete.



    Cetoole actually had me doing several different tests, but to no avail. What happens, and this was first theorized by Dsavitsk (Colin was heading down the same road, too.), is that for some reason a charge builds up on the RCA jacks. This is first evident when the SSMH has been operating and for some reason, you disconnect the RCA jacks (like switching sources). A significant hum will develop until something discharges the jacks. In the case of the DACs frying, the DACs themselves are absorbing that charge, which burns them out.

    I remember distinctly doing this on an Alien DAC. My fumbling fingers accidently touched the signal portion of the RCA plug to the outside ground connection of the RCA jack on the SSMH. The DAC was fried in that channel, instaneously, and never worked again. I fried a couple of Bantams along the way, too. Ludoo fried a couple and there have been others, too. The one SSMH I listen to on a regular basis is plugged into my PC's sound card. There has never been any trouble with one of those. It seems more related to DAC's that have no intervening buffer or I/V circuitry between their outputs and the connections to an amp input.

    For awhile, I suspected the power supply created a ground that was at another potential from true ground, hence - some voltage present on the SSMH RCA jacks' ground. That's what Colin had me testing, but the tests were inconclusive. It's too intermittent, though, and more like Dsavitsk's theory: a charge that develops for some reason and is then grounded.

    It's either the power supply or the heater-biased MOSFETs. There's not much else in the amp that could do it.
     
  3. tomb
    Quote:

    Originally Posted by hellomoto /img/forum/go_quote.gif
    Couldn't find anything on Mouser. Is it okay to buy normal ones, and then if I want buy new ones to replace the standard ones?


    The same as the Electrolytics.




    Yes and yes.

    Quote:

    I've seen on your (very well made) build-thread that you added led. If i'm right, I just have to buy standards 3mm leds, and standard wire, and drill the sockets to put them in the AMP Right?



    Well, there's the matter of finding the right place to draw power for the LED's and sizing the resistor(s). The one on the SSMH PCB website BOM should suffice if you connect to the 48V portion of the circuit. Quote:


    Thanks you very very much for your help tomb!



    You're welcome!
     
  4. Juaquin
    Quote:

    Originally Posted by nullstring /img/forum/go_quote.gif
    IDK, my CBS tubes look very much different than my RCA tubes.
    CBS tubes have square getters and I prefer them over my RCA tubes.




    Different getters doesn't necessarily mean a different manufacturer. They (CBS, other brands) simply could have wanted their tubes to look different so when (if) they contracted RCA, they asked for a different getter design. I haven't done any research myself so I'm not making any claims, but the SSMH website suggests it and I've seen it repeated many times here. I believe at some point someone posted more specific reasoning behind it but I can't find that post. Hopefully someone else can clarify further, not that it really matters - all that matters is you like the sound of your tubes! [​IMG]
     
  5. JamesL
    I know this was discussed somewhere in the thread before but I can't seem to find it.

    There are several, albeit small nuisances with the pot that i want to fix.

    I tried putting my brain to use and came up with the following conclusion.. can someone tell me if it would be a good idea, or if its completely wacko?

    Quote:

    Put a series resistor at the input of the pot to:
    1. get better attenuation when completely counterclockwise
    2. better channel balance at normal listening volume
    3. Get more use out of the full rotation
    4. effectively truncate the the right portion of the pot's travel.

    Since the potentiometer follows a logarithmic curve, I would need about 10x the value of the potentiometer to effectively truncate the last 50% of the volume control.
    I should decrease the value of the pot first to maintain a reasonable input impedance



    currently, it doesn't completely mute when I turn down the knob, and i'd rather truncate than adjust the curve since there is quite a bit of noise at the far end of the pot, and I don't want to accidentally blow my ears out.
     
  6. tomb
    Quote:

    Originally Posted by Juaquin /img/forum/go_quote.gif
    Different getters doesn't necessarily mean a different manufacturer. They (CBS, other brands) simply could have wanted their tubes to look different so when (if) they contracted RCA, they asked for a different getter design. I haven't done any research myself so I'm not making any claims, but the SSMH website suggests it and I've seen it repeated many times here. I believe at some point someone posted more specific reasoning behind it but I can't find that post. Hopefully someone else can clarify further, not that it really matters - all that matters is you like the sound of your tubes! [​IMG]



    Yes. Based on my experience, Top Halo, Top Square, Side Square, Top D, Side D - they all appear on the major brands, including CBS. However, the gray-etched squished octagon tube designator - unique to RCA - appears on every tube. In some cases, it's been worn enough that it's almost invisible, but you can usually find evidence of it if you look hard enough.

    I've only gone through 400+ tubes, though, so there may be some different ones out there that I haven't seen yet.[​IMG]
     
  7. tomb
    Quote:

    Originally Posted by JamesL /img/forum/go_quote.gif
    I know this was discussed somewhere in the thread before but I can't seem to find it.

    There are several, albeit small nuisances with the pot that i want to fix.

    I tried putting my brain to use and came up with the following conclusion.. can someone tell me if it would be a good idea, or if its completely wacko?



    currently, it doesn't completely mute when I turn down the knob, and i'd rather truncate than adjust the curve since there is quite a bit of noise at the far end of the pot, and I don't want to accidentally blow my ears out.




    I agree that attempting to adjust the curve would be an exercise in futility, most likely. However, you should be able to attenuate the input with an optimum choice of resistors to get most music listening levels in a "good" range of the pot travel. Try 2X, 3X, even 4X times the pot impedance, depending on the impedance-sensitivity of your phones.
     
  8. hellomoto
    Thanks for your help tomb!

    All the stuff is now on the way!!!

    Bye,

    Antoine
     
  9. JamesL
    I just put in a 470k R in since I had one on hand. It did the trick perfectly. Volume range is perfect with my d1k's, but I'll probably swap it out with ~180k R so I get better range with my hd650's. =)

    Quote:

    Originally Posted by tomb /img/forum/go_quote.gif
    I agree that attempting to adjust the curve would be an exercise in futility, most likely. However, you should be able to attenuate the input with an optimum choice of resistors to get most music listening levels in a "good" range of the pot travel. Try 2X, 3X, even 4X times the pot impedance, depending on the impedance-sensitivity of your phones.



     
  10. the_equalizer
    As I posted a few weeks back, going through my tube manual I found a tube that seemed like a good substitute for the 19J6 in our beloved Millett "Starving Student" Hybrid.

    The tube in question is the 17EW8, a double triode in a 9-pin mini glass envelope with a 17.5 volt @ 150 mA heater; it's pinout the same as the 12A_7 tubes except for pin 9 which is unconnected (in the 12A_7 tubes pin 9 is the heater center tap)

    I went ahead and ordered a couple off of ebay ( USD $3.60 per tube !!) and received my package yesterday. Since the pinout is the same as the 12A_7, I decided to use my 12AU7 SSMH build and adapt it for the new tube.

    Unfortunately I haven't been able to find plate characteristic curves for the 17EW8 so, I've been unable to work on a load line that helps me choose good working values for the plate load and cathode resistors.

    Thus basically the only change I made to my 12AU7 circuit was adjusting the MOSFET gate biasing resistors to work with the heater voltage of the new tube. The heater voltage being close to the 19J6, I decided to use the same values as in the original circuit, that is R2, R4, R8, R10 = 220 Kohms.

    In all, it took about 15 minutes to modify the amp to run the 17EW8 tubes! Just desoldering a couple 390 Kohm resistors and soldering two new 220 Kohm ones. This leaves us with a circuit that is exactly the same as the original 19J6, but with a 9-pin tube socket with the 12A_7 pinout.

    For maximum versatility, a MOSFET gate bias switch could be wired so that the amp could be used with 12 volt and 19 volt heater tubes at the flick of said switch.

    [​IMG]

    I can happily report the new version works very well! I've been testing it for 3 straight hours and already I like it better than my 12AU7 version because the MOSFETs don't get as hot and the tubes glow a bit more. [​IMG] Soundwise, it seems to me that it has slightly more detailed treble and punchier mids than my new production JJ 12AU7 tubes.

    Here's the schematic and a couple pics. Good luck with your builds!

    [​IMG]

    [​IMG]

    [​IMG]

    cheers!
     
  11. timmyGCSE
    oh yea..thanks man, yet another excuse to do this project all over again!! haha naa just kidding great work! you're doing such a brilliant job in keeping this project alive. Maybe one for the future for me...
     
  12. ezzieyguywuf
    +1 at the_equalizer for keeping this thing alive! I'm still waiting for a chance to order and install a new pot. :-D
     
  13. dsavitsk
    Quote:

    Originally Posted by tomb /img/forum/go_quote.gif
    Cetoole actually had me doing several different tests, but to no avail. What happens, and this was first theorized by Dsavitsk (Colin was heading down the same road, too.), is that for some reason a charge builds up on the RCA jacks.



    To add a little more to this, here's my working theory. There is a measurable capacitance between the Cisco power supply's output ground terminal, and earth -- don't remember what it is, and can't find my Cisco to check right now -- I think it was around 50nF though. That is, there is no continuity between these two things. Something about the Cisco's design allows this capacitance to charge, and it seems to get discharged into a grounded source.

    The simplest fix is to simply stop using the Cisco and build a basic 48V supply instead. This could be done pretty easily and inexpensively. Indeed, one could use a Tread so long as the voltages are watched -- that is, the kit version won't work. Otherwise, some sort of protection network might work -- maybe some back to back zeners, or cap coupling the output and output ground of the dac? I'm not sure what would be most effective and least intrusive.

    But, the basic Starving Student design, a grounded cathode tube into a mosfet follower, is pretty harmless and is unlikely to be the source of the problem.
     
  14. dsavitsk
    Quote:

    Originally Posted by the_equalizer /img/forum/go_quote.gif
    The tube in question is the 17EW8, a double triode in a 9-pin mini glass envelope with a 17.5 volt @ 150 mA heater; it's pinout the same as the 12A_7 tubes except for pin 9 which is unconnected (in the 12A_7 tubes pin 9 is the heater center tap)



    It's a little less than ideal though not horrible. At a normal operating point, mu is 50, Gm is 4600 which puts rp at about 11K. With both sides in parallel, this is reduced to 5K5 which certainly works here. However, since we are running very low current through the tube, the reality is that rp is a lot higher, how high is hard to know. Also, mu of 50 is a lot of amplification, too high for many setups I'd think. But, it is a decent find and should be fun to play around with.
     
  15. nullstring
    Quote:

    Originally Posted by dsavitsk /img/forum/go_quote.gif
    To add a little more to this, here's my working theory. There is a measurable capacitance between the Cisco power supply's output ground terminal, and earth -- don't remember what it is, and can't find my Cisco to check right now -- I think it was around 50nF though. That is, there is no continuity between these two things. Something about the Cisco's design allows this capacitance to charge, and it seems to get discharged into a grounded source.



    What if we just attached ground to mains Earth?
     
First
 
Back
331 332 333 334 335 336 337 338 339 340
342 343 344 345 346 347 348 349 350 351
Next
 
Last

Share This Page