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

472 473 474 475 476 477 478 479 480 481
  1. DrivenKeys
    Cool, I'll try that 1st. Thank you. Any idea about that loose cap? I'm trying to determine if it's a sign to replace it.

    Also, I was asking the community if I should reduce pot impedance while I'm at it. That's why I detailed the issue is with 250 ohm cans, not just ask about theory. I hope those who have gone through this experimentation already could perhaps save me a few hours and expense of my aimless troubleshooting (admittedly breaking things as I go). I wish I had the time and cash to just try whatever comes to mind, not there right now.

    Thank you, anybody who can help.
    Last edited: Jul 22, 2019
  2. DrivenKeys
    Well, now that I'm looking at it, I wonder about the resistors directly after the pot. EDIT Nevermind, I finally re-read the very informative post by yourself explaining the optional parts and grid stop. You are very patient and helpful. I'm still wondering what happens if the resistor values are changed.

    Thank you for your patience as I'm learning. I search all I can, but this has become a very, very long thread.
    Last edited: Jul 22, 2019
  3. tomb
    I'm not sure what you're asking when you say, "I'm still wondering what happens if the resistor values are changed."

    Every volume pot is designed for zero volume at the bottom of the range, and zero-resistance/maximum volume at the end of the range. The only question becomes, "Where in that travel does your comfortable listening volume take place?" As mentioned many times previously, the stereo volume pot used in the Starving Student is not the best (low cost and the "starving student" theme applies). Cheaper volume pots, among other symptoms, exhibit poor channel matching, often at the extremes of the volume travel - both at the bottom of the range and the top of the range. If your comfortable listening volume happens to fall in either of those ranges where the channel matching is bad in the volume pot, then your listening experience is full of irritation.

    What determines where the comfortable listening volume falls? Three variables apply: 1) the gain of the amp, 2) the efficiency (sensitivity) of the headphones, and 3) your personal listening preference. There is also the recorded gain setting of the music, but let's assume you've adjusted that through your media player, or that the differences are typically small enough that the change in volume pot setting is trivial. You can't change #2 unless you change the headphones. You can't change #3 unless you train yourself to listen at different volumes (could be ear damage as a result). So, you're left with doing something with the amp.

    If the Starving Student were a typical solid-state amp with feedback, you could change the resistors forming the voltage divider ratio in the feedback loop. However, the Starving Student is not a typical solid-state amp. Neither does it have feedback. The basic gain is provided by the tubes and is not variable. Unfortunately, the basic gain is quite high with the tubes and circuit in the Starving Student, meaning that for most headphones and sources, the comfortable listening range for most people occurs within the first quarter turn of the volume knob. This is a bad place for comfortable listening. It does not give you much precision in adjusting the volume and the channel matching in that area with a cheap volume pot sucks.

    In order to remedy that, we began the practice of applying resistors to the signal input of the circuit. Why? Well, the goal is to put the comfortable listening range further in the volume travel than sooner. In other words, we want to "push" the comfortable listening range of the volume pot from 9 o'clock to 12 o'clock. How do we do that? Well, continuing to turn the volume pot does what? It decreases the resistance. So, if we add resistors to the input, we increase the resistance in the circuit before it even gets to the volume pot. Done perfectly, the input resistors push the range to 12 o'clock, where once the comfortable listening range was at 9 o'clock, .

    This is not rocket science, but the actual variable values are hard to determine. Every headphone is slightly different and so is your hearing. We supplied (or suggested) resistors in the 50K, 100K, and 150K range to take care of the differences. If you lower that input resistance or remove it altogether, it will have the effect of moving the comfortable listening range down to the bottom of the dial (with most headphones). However, if you happen to have an uncommonly high headphone impedance (affects how easily the amp drives the headphones) or low efficiency headphones, you may find yourself at the end of volume pot travel, yet still not receiving sufficient volume. In that case, lowering/removing the input resistance may mean that you end up in the 12 o'clock range, instead.*

    * At some point, the power the amplifier can supply enters into the mix, but this has usually not been a problem with the Starving Student.
    Last edited: Jul 23, 2019
  4. DrivenKeys
    Thank you, I really appreciate your detailed response, and continue to learn.

    Please forgive my lack of detail. Before I posted at all, I understood the basics of the input resistors before the pot, but just needed to confirm I hadn't missed anything. I've been working with speakers my whole life, so I understand a basic series circuit. Since there's a lot I don't understand here, I tread lightly and question everything I think I know. Even if I feel stupid asking, a question is never stupid if it prevents a stupid mistake.

    I was wondering about changing the values of the grid stop resistors, just out of curiosity and conversation. Please don't think I expect you to spend the time educating me on this whole circuit. I don't really plan to change the grid stop resistors, but I don't yet understand the purpose. That's ok, I haven't even learned what the tube does yet, and I'm not asking you to type that out, unless you really want to.

    I'm in a fairly desperate work situation right now, and don't have the time to experiment like I'd like. I'm really just trying to get some basic answers from those with experience, so I can stop opening this thing up and just enjoy it. Basic beginners stuff that most ppl wish they knew before learning the hard way. I like learning the hard way, but must throw this in a box if I can't just wrap it up. Basically, I only have the time to take it apart once more, so I need to get the most out of that one time. Replacing a few affordable parts all at once is an acceptable expense, especially when I know that leaving it in one piece prevents me from breaking anything.

    If a cylindrical capacitor soldered to a board, stops working, but works again when pressed on and wiggled, is it typically damaged? Will it continue reliably, or should I pre-emptively replace it next time I remove the board from the case? I believe it's been knocked around and the leads have been bent back and forth many times as I've worked near it. It's not that expensive to replace. Should I just replace it while I have it open?

    From everything I've found through search, I'm under the impression that my 100k pot without signal resistors will allow me to comfortably drive 600 ohm cans. I've only gathered this anecdotally, and wanted to know if I have a realistic expectation, or if I should move to a lower impedance pot. Anybody who has played with this idea could save me a lot of time.

    Thanks again for all your help, I'm very happy for this community and hobby.
  5. tomb
    OK - it was early in the morning and combined with your earlier comments about lack of volume control with 250 ohm Beyers, it was a reasonable assumption to explain the input resistors - especially since the grid-stop resistors have absolutely nothing to do with that.

    If you've read somewhere that grid stopper resistors attenuate gain, that's absolutely wrong. They are there to increase the stability of the tube performance and stop potential oscillation, among other things. The only thing they attenuate are frequencies well above the audible range. Maybe you should think of them as a safety valve for a tube. :)

    Are they needed for the Starving Student? Probably not and Pete's original schematic didn't include them. However, among other things, they could address some of the comments in this huge thread where Starving Students were susceptible to cell phone noise or even radio signals.
  6. DrivenKeys
    Ahhhh, thank you. I'm glad to have them, lotsa potential rf probs around here.

    I'm excited to finally be back in this hobby. I'm thirsty for knowledge, wanna learn this thing. Looking forward to reading all about tubes, etc.
  7. onesojourner
    My pot is on the way from mouser. Hopefully I will have time to get it installed this weekend.
  8. DrivenKeys
    Nice, take your time! Which pot did you get?
  9. onesojourner
    I went with the one from the parts list.

    Audio D-Shaft 50K
  10. bdbell
    I built a SSMH several years ago, but never used it much because all of my headphones at the time were low impedance - but recently got a pair of 250 ohm DT990 and none of my DAP's could really drive them well enough so I pulled out the SSMH - there was too much noise and hum so I decided to mill a new PCB for it on my CNC. I did up the design in eagle and started milling my PCB, took several tries because I haven't done many PCB's on it for a while...finally have a PCB that works and has everything in the proper location to fit into the enclosure I already had built (aluminum). But when I got it all installed inside the case and tried it last night there was a very annoying noise mostly in right ear even with volume at 0. Did some searching on this thread and found I should connect the ground from the first power capacitor directly to the DC jack instead of to the ground plane - that fixed the noise problem. I now have no hum or noise at 0 with the DT990's, as the volume is increased I get hum and a little noise - but with lower impedance earphones it is much more noticeable...I will really only be using it with the DT990's so the hum is not really noticeable with them when playing music, but...want to try to get it as good as possible as I will be using it a lot.

    All the grounds are connected to the ground plane on my single sided PCB - should connect grounds of all of the capacitors to the DC jack, or should I connect all of the grounds for everything to the DC jack, or just the two big 680 uF caps? I would like to get it right before I put it back into the case so would appreciate any advice on how best to deal with it.
  11. onesojourner
    Alright I got the new pot and got it soldered in. I get a hum and nothing else. There is some popping when turning it on but other than that just a steady hum. Turning the pot does not change the volume of the hum.
  12. tomb
    We're not actually sure the grounding does much through the power supply. The Cisco power supply may actually be isolated from ground inside its circuitry. It's one of the reasons we theorized that static charges can develop on the RCA jacks (killing some DIY DACs that output directly from a DAC chip). The primary grounding in the SSMH takes place through the casework. Did you scrape the anodizing away around the inside surface of the back plate at the power jack and RCA jacks? The RCA jacks will work better if they are not isolated with insulating washers.

    This pic may help:

    Note that it's still important to provide a ground between the DC jack and the back plate.

    The other possibilities are the tubes and the power supply. If either one are bad, they can cause some bad humming.
  13. tomb
    I'm sorry that replacing the pot doesn't appear to fix it.

    I just reviewed your photos again. Let me ask a question (don't hit me if this was obvious) - are those MOSFETs connected directly to the copper of the heat sink? If so, they are directly shorted to ground. The Drain pin (center pin) is the same as the tab on the MOSFET. The tabs must be insulated from ground.

    If they are isolated, then sorry for the suggestion, but I can't see any insulating washers or pads in the photos.
  14. onesojourner
    They are screwed directly to the heatsink with a bit of thermal paste. and that means the drains are both tied together I guess? I have some of that cheap thermal tape that seems like it would insulate ok. Should I try that?

    :edit: I decided to just unscrew the fets from the heatsink (breaking the connection between the tab and the heatsink) and turn it on for a second. There was no noticeable change.
    Last edited: Jul 26, 2019
  15. tomb
    I guess I was thinking about the "normal" situation where you have a metal case and the heat sinks are bolted to the case, effectively grounding them. In any event, the MOSFET drains are supposed to be tied together, if you look at the schematic. The difference is that both drains are then connected to the "+" portion of the power supply. If your heat sink is not grounded to anything, then maybe it's OK.

    Well, I'll keep looking when I get a chance. At least this suggestion didn't cost you any money. :wink:
472 473 474 475 476 477 478 479 480 481

Share This Page