Hello.
I built a starving student 12SR7 amplifier. Basing on the post #9 in THIS thread I've changed the resistors values to fit new tubes.
My problem is that the amplifier is extremly quiet, even at maximum output.
When playing 1KHz from my ipod (about 0.9VAC) the output voltage without any load is over 3.5VAC. Once I put a 28Ohm load the voltage drops to something about 0.010VAC. I calculated the output impedance and it was something about 1150Ohm.
What have I done wrong?
I attached the schematic I've made and followed when building this amp. I've also connected pins 4 and 5 to the ground. I don't know if that matters but instead of 63v capacitors in C2 and C4 I've used 250V ones.
I built a starving student 12SR7 amplifier. Basing on the post #9 in THIS thread I've changed the resistors values to fit new tubes.
My problem is that the amplifier is extremly quiet, even at maximum output.
When playing 1KHz from my ipod (about 0.9VAC) the output voltage without any load is over 3.5VAC. Once I put a 28Ohm load the voltage drops to something about 0.010VAC. I calculated the output impedance and it was something about 1150Ohm.
What have I done wrong?
I attached the schematic I've made and followed when building this amp. I've also connected pins 4 and 5 to the ground. I don't know if that matters but instead of 63v capacitors in C2 and C4 I've used 250V ones.
Attachments
The IFR510 enhancement-mode MOSFET will drop about 5 volts. The resistor divider R2:R4 (and R8:R10) each divide 48V supply to about 16 volts. The source of each IFR510 therefore is about 5 volts less, or 11-12 volts, just enough to drive the heater of each 12SR7 tube. OK, got that part.
However ... the 12SR7 tubes are going to conduct something like 0.5 ma (if the data sheets can be extended to such low-voltage regime!), so there's a 20 volt drop on the anode resistor R1 (or R7). Maybe less. measure volts on pin 6 please of the 12SR7s. They should be nearly the same. Likewise as a double-check, measure volts at pin 3 of the 12SR7s. I'm expecting it to be somewhere between 1.2 and 2.0 volts.
OK, enough metrology. What's the problem? The amplifier voltage gain must come entirely from the 12SR7 for each channel. The output MOSFET devices only serve as impedance matchers, running in source-follower mode. This is cute, and a nice idea, since in source-follower mode, such MOSFET device characteristics are all but wiped out (they're NOT amplifying voltage).
So, the things that could be unfortunate are...
[1] the heaters of the 12SR7s not being hot enough. measure the voltage at pin 7 on each 12SR7.
[2] The bias-point is badly off, and the tube anodes are running substantially below 30 volts relative to ground. Much below 30 volts, and these 12SR7s aren't going to give much amplification at all. Its their nature.
[3] The 48 volt supply isn't really supplying 48 volts. Measure it!
[4] You have a particularly weak input "source". It should have the range of "line" or 600 millivolts nominal.
Finally - though not on the drawing, and probably not in the spirit of the "perfect headphone amplifier" - you certainly could try bypassing the cathode resistors R5 and R11 with some small electrolytic capacitors (16 to 68 uF, 12 volt, + lead to cathode-of-tube). This will result in substantially more gain from the 12SR7 tubes. They will be producing more 2nd order and even-order harmonics though. Bad? Maybe not!
POST FINAL: then again... you could have particularly insensitive headphones. Try some cheap "ear buds" from some left-over APPLE product. You know the pathetically "look-at-me! look-at-me! I buy apple!" white ones. They're quite sensitive. This is not a solution, but rather just another data point. If they work great ... then you definitely need to think about lowering the cathode resistor [method B] to a value closer to 1K or another half-dozen "fixes". Or the cathode-bypass caps... they do a good job increasing gain.
GoatGuy
However ... the 12SR7 tubes are going to conduct something like 0.5 ma (if the data sheets can be extended to such low-voltage regime!), so there's a 20 volt drop on the anode resistor R1 (or R7). Maybe less. measure volts on pin 6 please of the 12SR7s. They should be nearly the same. Likewise as a double-check, measure volts at pin 3 of the 12SR7s. I'm expecting it to be somewhere between 1.2 and 2.0 volts.
OK, enough metrology. What's the problem? The amplifier voltage gain must come entirely from the 12SR7 for each channel. The output MOSFET devices only serve as impedance matchers, running in source-follower mode. This is cute, and a nice idea, since in source-follower mode, such MOSFET device characteristics are all but wiped out (they're NOT amplifying voltage).
So, the things that could be unfortunate are...
[1] the heaters of the 12SR7s not being hot enough. measure the voltage at pin 7 on each 12SR7.
[2] The bias-point is badly off, and the tube anodes are running substantially below 30 volts relative to ground. Much below 30 volts, and these 12SR7s aren't going to give much amplification at all. Its their nature.
[3] The 48 volt supply isn't really supplying 48 volts. Measure it!
[4] You have a particularly weak input "source". It should have the range of "line" or 600 millivolts nominal.
Finally - though not on the drawing, and probably not in the spirit of the "perfect headphone amplifier" - you certainly could try bypassing the cathode resistors R5 and R11 with some small electrolytic capacitors (16 to 68 uF, 12 volt, + lead to cathode-of-tube). This will result in substantially more gain from the 12SR7 tubes. They will be producing more 2nd order and even-order harmonics though. Bad? Maybe not!
POST FINAL: then again... you could have particularly insensitive headphones. Try some cheap "ear buds" from some left-over APPLE product. You know the pathetically "look-at-me! look-at-me! I buy apple!" white ones. They're quite sensitive. This is not a solution, but rather just another data point. If they work great ... then you definitely need to think about lowering the cathode resistor [method B] to a value closer to 1K or another half-dozen "fixes". Or the cathode-bypass caps... they do a good job increasing gain.
GoatGuy
3. The power supply is acually about 50VDC so it's more that enough.
4. When playing 1KHz from my source it's about 0.911VAC.
I was trying to use both my cheap 32ohm headphones and my 300Ohm Sennheiser HD600.
I've also attached the measurements that I've made for you. I measured everything after tubes and mosfet's radiators got warm (about 1 minute after start up). Both channels have about the same voltages.
I would like to point out that I have very little experience with electronic but I am sure that I've connected everything properly.
So what I have to do now to get it to work?
4. When playing 1KHz from my source it's about 0.911VAC.
I was trying to use both my cheap 32ohm headphones and my 300Ohm Sennheiser HD600.
I've also attached the measurements that I've made for you. I measured everything after tubes and mosfet's radiators got warm (about 1 minute after start up). Both channels have about the same voltages.
I would like to point out that I have very little experience with electronic but I am sure that I've connected everything properly.
So what I have to do now to get it to work?
Attachments
Wow.
Well ... you certainly have come up with a good one! The voltages look fine. The problem "obviously" is that there is no meaningful amount of VAC on the output when the headphones (28Ω load) is in place. Yet, there is a nice (healthy!) 3.9V VAC on the output when there is no load (and being driven by input signal).
At first I was going to venture that the DRAIN of each MOSFET is attached to the wrong side of the 2K resistor that is in the V+ supply rail. But that wouldn't make sense - because the heaters would not glow. Wouldn't be enough current to drive them, and the poor 2K resistor would burn up.
So my next guess - again having the confirmation of there being proper voltage at the heater pins of the tubes to get them warm - is that C3 and C5 are maybe the cuprits. Are they really 150 uF?
IF YES, then they shouldn't be the problem.
IF NO - then you ought to be struck by only hearing treble in the earphones
That last part comes from Z = 1/(2×pi×F×C) ... where a C that is smaller makes Z to be larger, ... but F frequency also affects things. Basically - if the sound in the headphones sounds like its been passed through a very high-pass treble filter, then the 150 uF caps ... aren't 150 uF, but something smaller like 1.5 uF, or 150 nF (which wouldn't be surprising).
Because clearly the rest of the circuit is doing what its supposed to be doing. I've logiked my way to that conclusion.
You could go one step further in the confirmation process by feeding the output - without load - to the input of another amplifier, just to show that this thing is actually cleanly amplifying stuff, just not producing power. That would be cool, actually. If it were to, then it confirms basically that the output electrolytic capacitors (they are electrolytic, aren't they?) C5/C7 are the wrong value somehow.
GoatGuy
Well ... you certainly have come up with a good one! The voltages look fine. The problem "obviously" is that there is no meaningful amount of VAC on the output when the headphones (28Ω load) is in place. Yet, there is a nice (healthy!) 3.9V VAC on the output when there is no load (and being driven by input signal).
At first I was going to venture that the DRAIN of each MOSFET is attached to the wrong side of the 2K resistor that is in the V+ supply rail. But that wouldn't make sense - because the heaters would not glow. Wouldn't be enough current to drive them, and the poor 2K resistor would burn up.
So my next guess - again having the confirmation of there being proper voltage at the heater pins of the tubes to get them warm - is that C3 and C5 are maybe the cuprits. Are they really 150 uF?
IF YES, then they shouldn't be the problem.
IF NO - then you ought to be struck by only hearing treble in the earphones
That last part comes from Z = 1/(2×pi×F×C) ... where a C that is smaller makes Z to be larger, ... but F frequency also affects things. Basically - if the sound in the headphones sounds like its been passed through a very high-pass treble filter, then the 150 uF caps ... aren't 150 uF, but something smaller like 1.5 uF, or 150 nF (which wouldn't be surprising).
Because clearly the rest of the circuit is doing what its supposed to be doing. I've logiked my way to that conclusion.
You could go one step further in the confirmation process by feeding the output - without load - to the input of another amplifier, just to show that this thing is actually cleanly amplifying stuff, just not producing power. That would be cool, actually. If it were to, then it confirms basically that the output electrolytic capacitors (they are electrolytic, aren't they?) C5/C7 are the wrong value somehow.
GoatGuy
Well, I used 220uF 63V. Original specification says:
Sorry but I completly don't understand this. As I said, I have very little experience with electronic. I can hear a bass and treble but they're extremly quiet.
I can connect starving student's output into Objective2 Amp to see if it would work. Do you want me to do it?
I connected Starving student's output into O2's input but even at maximum volume I could only hear very quiet music and buzzing. But...
I accidentaly touched O2 and student's case at the same and... I almost became deaf... It played very loud (as it suppose to).
I made a quick investigation and it turned out that headphones ground wasn't connected properly (it acually was but it didn't had a good connection. It's hard to explain).
I connected headphone's ground in different way and now it plays as it suppose to.
Thanks for your help anyway.
I accidentaly touched O2 and student's case at the same and... I almost became deaf... It played very loud (as it suppose to).
I made a quick investigation and it turned out that headphones ground wasn't connected properly (it acually was but it didn't had a good connection. It's hard to explain).
I connected headphone's ground in different way and now it plays as it suppose to.
Thanks for your help anyway.
I'd say 7 times out of ten it's something simple but generally unpredictable like this. Probably was a good learning experience for OP, certainly was for me reading the post.
That's why all the tech support manuals usually start with questions like "Is your computer plugged in?" "Is the small red light just below the logo on the front panel glowing?", etc.
Happy listening.
Regards,
aerodoc
That's why all the tech support manuals usually start with questions like "Is your computer plugged in?" "Is the small red light just below the logo on the front panel glowing?", etc.
Happy listening.
Regards,
aerodoc
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.