I am still hoping to try OPA1622 without the buffer. Would the values in the attached schematic be OK with 10k pot? I guess putting the DC blocking capacitor (C1) between the OPA1622 and the volume pot wiper would make impedance matching around OPA1622 inputs at DC impossible or it is still an option?
Regards,
Oleg
Regards,
Oleg
Attachments
Yesterday I have finally received my new measurement tool Focusrite Scarlett 2i2 (2nd gen.) and performed some initial tests (attached). These are my first measurements of audio spectrum ever, so I might have done something wrong. Please let me know if something looks suspicious.
Anyways, what I realized is that there is a sweet spot for output and input levels which results in the lowest THD. Go a little bit higher with input or output level and 2nd and 3rd harmonics start to rise quickly. The noise floor seems to dominate the THD+N level at any settings if I interpret it right. Interestingly, both right and left channels have remarkably similar performance so I can use either of the two for measurements.
So far I only managed to use the interface in 48kHz mode, thus spectrum only gets to 24kHz. I know 2i2 is capable of up to 192kHz sampling rate but so far I haven't found a way to set it so high on linux. The only settings that REW let's me choose are 44.1kHz and 48kHz.
In any case, such a performance should be sufficient for my basic measurement needs.
Anyways, what I realized is that there is a sweet spot for output and input levels which results in the lowest THD. Go a little bit higher with input or output level and 2nd and 3rd harmonics start to rise quickly. The noise floor seems to dominate the THD+N level at any settings if I interpret it right. Interestingly, both right and left channels have remarkably similar performance so I can use either of the two for measurements.
So far I only managed to use the interface in 48kHz mode, thus spectrum only gets to 24kHz. I know 2i2 is capable of up to 192kHz sampling rate but so far I haven't found a way to set it so high on linux. The only settings that REW let's me choose are 44.1kHz and 48kHz.
In any case, such a performance should be sufficient for my basic measurement needs.
Attachments
It's working!!! Finished the headamp today and listened to some youtube music using the laptop output, so no quality comments for now. Most important is that there are no obvious problems so far. Tomorrow will connect it to Subbu DAC and listen to some music from my limited collection.
There are a bit more wires in the box than I would like (mostly due to adding a buffer board) but the amp is dead silent with and without the source attached. My DMM measures 0.000 VDC and 0.000VAC with input floating, same when it is grounded. Since my DMM resolution is limited to 1mV I assume that the actual DC offset is below 1mV.
So far I'm satisfied with the fact that it did not blow
I turned it ON all at once. Before I only tested the PSU under load. The PSU when feeding 66 Ohm dummy load (from each rail to ground) shows -5.999 and +6.002 VDC (regs set to +-6 VDC by 0.1% resistors).
So far all is as designed. Next are the measurements, but first I have to make probes for the Focusrite 2i2. I only have some random cables which do not suit my needs. I'll post measurements when done.
Regards,
Oleg
There are a bit more wires in the box than I would like (mostly due to adding a buffer board) but the amp is dead silent with and without the source attached. My DMM measures 0.000 VDC and 0.000VAC with input floating, same when it is grounded. Since my DMM resolution is limited to 1mV I assume that the actual DC offset is below 1mV.
So far I'm satisfied with the fact that it did not blow
I turned it ON all at once. Before I only tested the PSU under load. The PSU when feeding 66 Ohm dummy load (from each rail to ground) shows -5.999 and +6.002 VDC (regs set to +-6 VDC by 0.1% resistors). So far all is as designed. Next are the measurements, but first I have to make probes for the Focusrite 2i2. I only have some random cables which do not suit my needs. I'll post measurements when done.
Regards,
Oleg
Attachments
Just listened to some music using the head amp. It sounds really good. I have nothing to compare it with except the Focusrite 2i2 but then I would have different source so it won't be a true head amp comparison, but rather DAC+headamp combination.
I also have a strange issue. In all my gear I use safety GND lift circuit (attached). If I use it in the head amp and connect it to my DAC which also has such ground lift inside I get really loud mains hum. The volume of hum is independent of the volume pot setting. Disconnecting one channel does remove the hum from that channel only but the other is still humming. Only if I bypass the GND lift and connect the PSU ground straight to the headamp chassis the hum is totally gone... Now I wander where the problem can be? That same DAC works nicely with the dual mono MyRef FE amp where each channel has safety ground lift. The amp had mains hum problem without the GND lift regardless if amps GND was floating or tied to chassis. I am really puzzled now. May be someone has an idea where the problem is? The GNDs of the two channels first meet at the buffer PCB inside of the headamp after they split at the DAC end. The small headamp bypass PCB (near the RCA sockets) switches both signal and signal return using two relays for maximum isolation when the headamp is OFF.
I also have a strange issue. In all my gear I use safety GND lift circuit (attached). If I use it in the head amp and connect it to my DAC which also has such ground lift inside I get really loud mains hum. The volume of hum is independent of the volume pot setting. Disconnecting one channel does remove the hum from that channel only but the other is still humming. Only if I bypass the GND lift and connect the PSU ground straight to the headamp chassis the hum is totally gone... Now I wander where the problem can be? That same DAC works nicely with the dual mono MyRef FE amp where each channel has safety ground lift. The amp had mains hum problem without the GND lift regardless if amps GND was floating or tied to chassis. I am really puzzled now. May be someone has an idea where the problem is? The GNDs of the two channels first meet at the buffer PCB inside of the headamp after they split at the DAC end. The small headamp bypass PCB (near the RCA sockets) switches both signal and signal return using two relays for maximum isolation when the headamp is OFF.
False alarm. Problem solved. I just had stray connection to chassis from an auxiliary power supply negative terminal via the standoffs. Now it is isolated and the headamp is dead silent at any condition.
Next is to assemble a couple of probes to measure the headamp. Missing connectors/clips are ordered.
Next is to assemble a couple of probes to measure the headamp. Missing connectors/clips are ordered.
Finally received the cables and connectors for the probes. Before to cut the wires I hooked up the headamp to the Focusrite 2i2. The headamp drives mostly its own feedback network (1.5kOhm) in parallel to 52kOhm input impedance of the 2i2 via long cable (~3m). The result is that I can't measure any difference between the loopback and the headamp in the loop which means that the headamp has distortion at least an order of magnitude lower than the audio interface and that the OPA1622 has no trouble driving long cables without the isolation resistor at its output. The spectra I get are attached. There is a tiny bit higher noise floor and distortion compared to the previous loopback results (see post #182) which I guess is due to single-ended connection instead of balanced. There is also some mains hum pick up which is very sensitive to interconnects physical placement. By moving the cables around I managed to reduce the mains component by about 10 dB, so I guess it is mostly external to the headamp.
In the next days I'll complete the probes and will try to measure the headamp under real load.
Regards,
Oleg
In the next days I'll complete the probes and will try to measure the headamp under real load.
Regards,
Oleg
Attachments
