Very nice.
With no music playing, do you hear any hum? Even with you ear close to the speaker cones?
It might be possible to reduce the 60Hz power supply harmonics with a bit of modification to the grounding. Do you have a schematic for the power supply and maybe some pictures of your wiring.
With no music playing, do you hear any hum? Even with you ear close to the speaker cones?
It might be possible to reduce the 60Hz power supply harmonics with a bit of modification to the grounding. Do you have a schematic for the power supply and maybe some pictures of your wiring.
There might be the tiniest bit of hum, but my room is pretty noisy right now. I'll be able to tell in the morning.
The schematic for the PSU boards is attached. GND is tied at the DC output end of the board only. 18kuF caps in all positions. 4 x 0R47 => 0R118.
Essentially 36kuF => 0R18 36kuF => CRC. I don't remember exactly which diodes I used for rectification.
Picture of overall wiring / layout is below. One thing I'm reasonably certain would reduce a bit of the noise / 60Hz harmonics is to move the DC wiring from being so close to the back of the chassis as it leaves the filter boards. It's right up next to the PEM, and I haven't found a way to twist it at that point b/c the V- and V+ supply attachment points are on the far "left" and far "right". The secondaries are not as close to the DC supply as they might look in the photo, but I can likely move them closer to center. I use the toroids in a lot of projects, so... I rarely if ever properly cut the primaries / secondaries to their ideal shortest lengths for a specific build.
I've also got some other PSU boards with a different layout that have produced good results. I have a few extra sets of 4320 regulator boards / separate filter boards also. What's most likely to happen is that I'll move that PSU / filter board back somehow or reverse it. It might even fit vertically in this chassis. That will happen most likely when I get the "R21" boards built and tested.
Any tips and tricks are always welcome.
Thanks again!
The schematic for the PSU boards is attached. GND is tied at the DC output end of the board only. 18kuF caps in all positions. 4 x 0R47 => 0R118.
Essentially 36kuF => 0R18 36kuF => CRC. I don't remember exactly which diodes I used for rectification.
Picture of overall wiring / layout is below. One thing I'm reasonably certain would reduce a bit of the noise / 60Hz harmonics is to move the DC wiring from being so close to the back of the chassis as it leaves the filter boards. It's right up next to the PEM, and I haven't found a way to twist it at that point b/c the V- and V+ supply attachment points are on the far "left" and far "right". The secondaries are not as close to the DC supply as they might look in the photo, but I can likely move them closer to center. I use the toroids in a lot of projects, so... I rarely if ever properly cut the primaries / secondaries to their ideal shortest lengths for a specific build.
I've also got some other PSU boards with a different layout that have produced good results. I have a few extra sets of 4320 regulator boards / separate filter boards also. What's most likely to happen is that I'll move that PSU / filter board back somehow or reverse it. It might even fit vertically in this chassis. That will happen most likely when I get the "R21" boards built and tested.
Any tips and tricks are always welcome.
Thanks again!
It appears that your PS has two totally separate diode bridges and CRC filters, as with mono blocks.
This is a case I did not discuss in the Build Guide. Try eliminating the ground loop breaker by connecting the RCA GND and In- to (PCB) GND as it is with the channel. And each PS GND should have a thermistor to chassis GND.
This is a case I did not discuss in the Build Guide. Try eliminating the ground loop breaker by connecting the RCA GND and In- to (PCB) GND as it is with the channel. And each PS GND should have a thermistor to chassis GND.
Is there any connection between the output grounds to the two sides of the board?
believe there are through for jumpering the grounds together, but I do not see any jumpers in the image.
The secondaries are independent. It appears that the two halves can float relative to each other.
What am I missing?
believe there are through for jumpering the grounds together, but I do not see any jumpers in the image.
The secondaries are independent. It appears that the two halves can float relative to each other.
What am I missing?
No. It might appear that way, but that is not the case. The V+ and V- supply boards are tied together only between GND2.2 and GND3.3.
Yes, two traces, one on each board connected as described above.
Maybe the jumper creating the common GND between the boards circled below was not clear in the photo. This is what I meant by: "GND is tied at the DC output end of the board only."
It is also much easier to see here how all the DC wiring is very close to the PEM. I have a few ideas re: how to change the layout to remove that particular issue / situation.
Hmmm…my interest is peaked! I missed out on the GB marathon but I’ll need to procure a set of boards or order from JLPCB.
Nice work Patrick, thanks for the QA403 measurements - puts it in perspective. That’s very low THD performance at 1W! Is this the XA25 for the masses? Perhaps!
Thanks Lynn @lhquam for your contributions…
Best,
Anand.
Nice work Patrick, thanks for the QA403 measurements - puts it in perspective. That’s very low THD performance at 1W! Is this the XA25 for the masses? Perhaps!
Thanks Lynn @lhquam for your contributions…
Best,
Anand.
I don't have the best hearing, but I'd say the tiniest bit of hum with my ear against the grill. The speakers hooked up at the moment (Tekton DI) are rated at 98dB / 2V83. Nominal 4R. My gut says that's exaggerated a bit.
I see the largest component of the hum at 60Hz, and that's around -80dbV (the QA is calibrated) at 60Hz. So, that component is ~0V0001rms output, I think. That translates to ~50dBSPLrms at 0.1m.
That seems a tad off unless my hearing is more suspect than I thought. It's likely I did an error putting something into one of the spreadsheets.
I never rely on my math, I use online calculators and/or a spreadsheet that someone else made.
Mainly... it's perfectly wonderful, and from a practical standpoint, I'd consider it silent at listening distance.
------------------------------------------------------
On to listening.
Early morning listening is a highlight of my day. I get up about 4 hours earlier than my wife on average, and "waking the bear" is ill-advised. I'd say over 50% of my listening time is <70dBSPL and very, very rarely in the "sweet spot". Right now, I'm, sitting about 2 ft from my left speaker and 12 ft from the right. It's important to me what with a little balance tweak I have a reasonable experience. It's also important that it sounds alright when I'm at the bench 5m away.
I have background music on ALL the time. It's on now. I think I've "critically listened" for maybe 5 or 6 hours this year. I've had music on for 100s and 100s of hours this year.
I also listen to "crap" recordings 98% of the time. I'm all digital, with 90% of that being streaming and the other 10% CDs. Most of the masters are not fabulous. My system needs to work well ... for me.
I'm also one of those folks that uses emotion to evaluate the system. I don't hear well enough to know if this or that is better or worse over time unless there's a pretty exaggerated difference. However, there is something about certain combinations of things that brings goosebumps or tears or gets the toes to start tapping involuntarily. As a reasonable rule of thumb, something that will immediately put me off is if people don't sound like ... well... what I think people sound like. I don't listen to enough live music to know exactly what certain instruments sound like in real life. I can pick out guitar tones pretty well. Percussion can be a tell-tale for me.
You've seen the measurements. So, for those that like listening impressions, to me, it's just as important to ask about the listener.
What I think so far...
It's a great amplifier, obviously. I put it in just after an Aleph J variant I'd been listening to for a couple days.
Everything is subtle, but I'd say (to me) these are pretty different presentations.
Biggest differences to me -
The AJ to me has a way of taking an edge off that can be really welcome with certain songs / recordings. This amp decidedly has a little less of that. I find myself turning certain songs down because... well... something in my brain says turn that crap down. Note - I do have volume leveling turned on through Roon / Qobuz, so it's not necessarily a jarring increase in overall dB b/c of loudness wars era music vs. older recordings. Some things just weren't clicking with me. That could be my mood or whatever. I'm only on day 2.
The tradeoff... DETAIL retrieval, particularly with percussion. It's an odd thing to call out, I guess, but it's what I hear. Weird little things jump out at me. Example, the Rascal's "Groovin'" just came up. The tambourine bells are prominent. The rimshots? sticks? stand out. The skin of the bongos is 'real'. This type of thing has held true through a lot of songs. I was looking for words to describe it... then this song just came on. I find myself turning my head when a familiar song comes on b/c I hear something that sounds a bit different, and very, very nice.
(One of) the litmus tests for can it stay - Alisson Krauss - A Hundred Miles or More is one of my "test recordings" if I really have any. Her voice (to me) has a way of indicating whether something works in that critical vocal range. Also, do tears well up with certain songs, not just because of the emotion of the lyrics, but the presentation. I adore that recording. For those that like that type of music, get it. For those that want a bigger treat get the video though PBS.
Enough -
OK... it works. It works REALLY, REALLY, REALLY well. I'll be listening to a few special recordings and sitting back down in the sweet spot a bit more over the coming days when I get time. I can trade turning down an intentionally streamed recording of Adele's 19 (I know it's crap and I use it to see if I cringe) for having something that sounds (to me) truly different to any amp I recall having put into the system for the last year or so. It makes me want to listen to more music to "see what I hear". Ain't that what we all want?
If anything particularly fun comes up, I'll certainly post. I'll also share what PSU changes I wind up making and/or wiring layout. I fully admit that I could have taken more time to not just "put it in a chassis and go". Until then, you're probably all wishing I'd quit typing.
Lynn... this is a wonderful learning experience. I'm also going to use this to learn a little bit more about how the bias current varies with amplitude as you've written. Maybe I just lucked out, but when I was doing the test measurements, I kept the DMMs hooked up. I saw no more than a few mV change with sustained test tones. To me... with "real" music, it just won't matter, but I'm still learning to measure things, and that sounds like a fun thing to see if I may want to "optimize".
Anyone on the fence - just do it. Ya know you want to. I'll certainly be curious to hear what others think.
Credit to -
- https://sengpielaudio.com/calculator-db-volt.htm (This site is fantastic IMO)
- DIYer fireanimal for creating the spreadsheet I use for dBSPL at listening distance when inputting the speaker sensitivity, voltage input, and distance.
Last edited:
Anand, Thank you! Also... and I think you've done it a few times. After a few days, I will be swapping the speaker cables to see how it sounds with a 'negative 2nd'. With distortion this low, will it matter? I have no clue, but I know it'll eat at me until I do it. Heck, even if it has a psychological impact to just know that the warmness of the music will cover me like a soft blanket... 
Also, given the 'flatness' of the THD vs. power curve up to clipping... I am going against my original thoughts. I actually will be using this to try and "move a little more air" this weekend. I want to see if it can be that perfect hybrid of early morning grooves and also use it when I want to have the neighbors come say "Hi". That's also why the DIs are hooked up.
Also, given the 'flatness' of the THD vs. power curve up to clipping... I am going against my original thoughts. I actually will be using this to try and "move a little more air" this weekend. I want to see if it can be that perfect hybrid of early morning grooves and also use it when I want to have the neighbors come say "Hi". That's also why the DIs are hooked up.
Always fun to see a good build come together.
It is easy to have boards fabricated from the JLC gerber files. You will end up with 5 PCBs, so enough for some experimentation. The trick will be selecting which output devices to use. I am thinking of using my existing M2x chassis with its dual-mono supply; it provides a solid +/– 24V @ 1.5A. I used the Mouser format text files in the ZIP package to formulate two BOMs and ordered the parts list, which arrived yesterday.
I am still tinkering a bit with my F5m, which has the same type FETs that I intend to use with the ZD25: FQA28N15 & IXTQ26P20P. I also have some IXTQ36P15P to try. The F5m likes running on 26V rails and 1.5A bias current.
It is easy to have boards fabricated from the JLC gerber files. You will end up with 5 PCBs, so enough for some experimentation. The trick will be selecting which output devices to use. I am thinking of using my existing M2x chassis with its dual-mono supply; it provides a solid +/– 24V @ 1.5A. I used the Mouser format text files in the ZIP package to formulate two BOMs and ordered the parts list, which arrived yesterday.
I am still tinkering a bit with my F5m, which has the same type FETs that I intend to use with the ZD25: FQA28N15 & IXTQ26P20P. I also have some IXTQ36P15P to try. The F5m likes running on 26V rails and 1.5A bias current.
^ Awesome suggestions. I, for one, have taken some notes from above. I don't personally (yet) have the skills to select alternate actives for circuits. Heck even with a slight deviation from the 'norm' using the actives provided kindly by Lynn, there was a modification needed that I'd have never figured out on my own.
Whenever someone like you is kind enough to mention a few actives to try, I always see what's available or put them on a list along with the suggested circuits and notes for other circuit mods if they're not "drop in replacements". Your suggestions for the ACA and F6 come immediately to mind as things I've tried. Plus... I am pretty sure we share at least one set of speakers in common (3As). So, when you say something may sound a certain way etc. room acoustics / brains / ears etc. notwithstanding, yours are some of the first comments I look for. Plus, I have that super-cool PSU you helped tremendously with.
It's projects / platforms like this that help me (and maybe a few others) continue to learn about what's important in a circuit and what parameters we get to play around with.
tl;dr- IXTQ26P20P and IXTQ36P15P added to my "list o' stuff to order / find".
Whenever someone like you is kind enough to mention a few actives to try, I always see what's available or put them on a list along with the suggested circuits and notes for other circuit mods if they're not "drop in replacements". Your suggestions for the ACA and F6 come immediately to mind as things I've tried. Plus... I am pretty sure we share at least one set of speakers in common (3As). So, when you say something may sound a certain way etc. room acoustics / brains / ears etc. notwithstanding, yours are some of the first comments I look for. Plus, I have that super-cool PSU you helped tremendously with.
It's projects / platforms like this that help me (and maybe a few others) continue to learn about what's important in a circuit and what parameters we get to play around with.
tl;dr- IXTQ26P20P and IXTQ36P15P added to my "list o' stuff to order / find".
Thank you ItAllInMyHead, you are the champion of the Holy Grail. First build of all GB. Hat off!
Your measurements and listening experiences are very impressive.
And watching your PSU image I would add all connections between the two sides of the board with thick wires to make one low-impedance common ground. One thin jumper may be not enough.
Your measurements and listening experiences are very impressive.
Yes, this what I am looking for all my life too.
And watching your PSU image I would add all connections between the two sides of the board with thick wires to make one low-impedance common ground. One thin jumper may be not enough.
