He's not measuring input impedance. He is simply checking the DC resistance from the grid pin of one side of the 12AT7. For a build that does not have a volume control, this is a good method to check that you've got the right value for the grid leak resistor (220K). It is easy to unplug the amp, pull the 12AT7, and meter between pin 2 (or 7) of the socket and ground.
Since he has a volume control, it will be in parallel with this measurement. You can check with the wiper is all the way at the "top", in which case you have the resistance of the pot in parallel with 220K. Or, you can turn the pot the other way until the wiper is at the "bottom" (ground). In this case, all you measure is the grid stopper on the 12AT7 - i.e., 100 ohms.
Arh. That makes more sense. That method would also find any solder bridges between the grid pin and GND.
One suggestion: Rather than using music as a test source, I suggest using a tone generator. There're plenty of software packages that'll turn your PC into a tone generator. I think even the free demo version of TrueRTA will do it. That makes measurements much easier and predictable.
But let's verify the DC levels first. With 330R, the 10M45 gives about 8 mA if I recall correctly. So the voltage at pins 3 and 8 (cathode) should be 220R * 0.008 = 1.76 V. According to the 12AT7 data sheet, this should put the anodes (pins 1, 6) at around 225 V. The DC voltage at the grid (pins 2, 7) should be 0 V. Also, the voltage across R13 should be about 2.65 V.
I forget which power transformer you said you were using but I'm assuming you have some 300~400 VDC on B+. I'm hand waiving on the datasheet, but I'm thinking that should land you around 35~40 mA of cathode current in the 6L6 output tube. So you should have some 20-ish volt across the 560R cathode resistor - R17.
If all the DC levels seem reasonable, I suggest the following AC tests:
The 12AT7 has a mu of about 60 and it's loaded by a current source. So the AC voltage (RMS) should be roughly 60 times higher at the anode (pins 1, 6) than the AC voltage at the grid (pins 2, 7). In reality, you'll probably have around 45~50x the voltage. Hook up a tone generator (1 kHz, 100 mV RMS) and verify that this is the case. If not, try pulling the output tube and measure again.
If the input stage checks out, the problem must be in the output stage. Disable the cathode feedback (i.e. ground the bottom plate of C12).
You should be able to perform these tests using a PC-based tone generator (if you don't have a hardware one) and a basic multimeter.
~Tom
One suggestion: Rather than using music as a test source, I suggest using a tone generator. There're plenty of software packages that'll turn your PC into a tone generator. I think even the free demo version of TrueRTA will do it. That makes measurements much easier and predictable.
But let's verify the DC levels first. With 330R, the 10M45 gives about 8 mA if I recall correctly. So the voltage at pins 3 and 8 (cathode) should be 220R * 0.008 = 1.76 V. According to the 12AT7 data sheet, this should put the anodes (pins 1, 6) at around 225 V. The DC voltage at the grid (pins 2, 7) should be 0 V. Also, the voltage across R13 should be about 2.65 V.
I forget which power transformer you said you were using but I'm assuming you have some 300~400 VDC on B+. I'm hand waiving on the datasheet, but I'm thinking that should land you around 35~40 mA of cathode current in the 6L6 output tube. So you should have some 20-ish volt across the 560R cathode resistor - R17.
If all the DC levels seem reasonable, I suggest the following AC tests:
The 12AT7 has a mu of about 60 and it's loaded by a current source. So the AC voltage (RMS) should be roughly 60 times higher at the anode (pins 1, 6) than the AC voltage at the grid (pins 2, 7). In reality, you'll probably have around 45~50x the voltage. Hook up a tone generator (1 kHz, 100 mV RMS) and verify that this is the case. If not, try pulling the output tube and measure again.
If the input stage checks out, the problem must be in the output stage. Disable the cathode feedback (i.e. ground the bottom plate of C12).
You should be able to perform these tests using a PC-based tone generator (if you don't have a hardware one) and a basic multimeter.
~Tom
I assume that the sound is not very loud under these conditions. I usually sit about 1 meter from my 87db speakers but at full tilt it is plenty loud. To me "very low volume" means that you can berely hear it under these conditions. Is this correct? Is the weak sound clear or distorted?
Was the volume low when the amp was bench tested? Are both channels equally low? Assuming the amp had low volume when bench tested, were the input switching and output connectors used? I usually connect the CD player directly to the board, and the speakers directly to the OPT's using clip leads. If the amp had low volume under these conditions the problem is on the PC board or tye OPT's and wiring.
With the amp off measure the resistance from the coupling caps to ground (each end). The readings should be higher than 100K. I vaguely remember a board where R15 and R16 were switched causing very low volume. This would cause a low reading.
As has been suggested, voltage readings may help us narrow things down a bit. Since the backs of the tube sockets are exposed, measure the voltage on each pin of the output tubes and the 12AT7's with the meter on DC volts and the negative lead connected to ground. I would expect the DC readings to be normal if the sound isn't distorted.
At this moment I've tried to connect the amp in triode mode (basic, with choke and power supply cap), and the sound is a lot more clear and without distortion, then I tried to connect the amp with UL and CFB but without the switches and it just sounds like heaven, so the problem is with the switches.
---
David
---
David
I'm kind of surprised that nobody has suggested it but, man, please neaten up your lead dress.
Shorten the wires to only necessary lengths.
Excessive wiring can only lead to noise and poor performance.
Not saying this has anything to do with your low output but it would be a lot easier to trace the signal paths and power wires if they were neater.
Just look at some other builds to see how they wired theirs.
Good luck on finding your error.
Shorten the wires to only necessary lengths.
Excessive wiring can only lead to noise and poor performance.
Not saying this has anything to do with your low output but it would be a lot easier to trace the signal paths and power wires if they were neater.
Just look at some other builds to see how they wired theirs.
Good luck on finding your error.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.