Thresholds Stasis 3 PSU redesign

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I am fan of well designed (and well built) PSUs.
The goal is to reduce 60hz (conducted and EM coupled), increase channels separation and provide a big reservoir of current.

Solution1:
I was thinking about moving the PSU (Transformer, rectifier, filter) into a secondary chassis and replacing the capacitor filter with a Pi filter. In this way, only clean(er) DC would enter inside the amplifier chassis.

Solution2:
New chassis + toroidal transformer + Pi filter + voltage regulator.

Is 1 or 2 worth? Am I going crazy with my PSU obsession?
 
"Am I going crazy with my PSU obsession? "

If it's worth for you to investigate much time and effort - why not? That's obvioulsy similar to the way which Mr. Pass has gone when designing his Xs-power amps which are famous for their superb sound.

I would not go for a regulated supply, a couple of highest quality elcaps, may be combined with a Pi-filter, will do a perfect job.
 
"Am I going crazy with my PSU obsession? "

If it's worth for you to investigate much time and effort - why not? That's obvioulsy similar to the way which Mr. Pass has gone when designing his Xs-power amps which are famous for their superb sound.

I would not go for a regulated supply, a couple of highest quality elcaps, may be combined with a Pi-filter, will do a perfect job.


I was thinking to use 15mF, 1mH (38mOhm), 22mF for the C-LC filter (I already have good quality parts in house), that should filter much better than the standard 15mF capacitor..
 

Attachments

  • psu.JPG
    psu.JPG
    15.8 KB · Views: 326
I have sent you a PM

Thank you for your PM, I am looking into it.


Low volume sessions are actually the moment when I care more about low PSU noise, so I was also considering to use a couple of e-bike battery packs to replace the transformer for low power listening sessions.

I need to measure the energy use during a low volume session, and do some math about the sizing of the batteries..
 
Is this the Stasis 3 in pic?
Why not add more caps in parallel to the 2 main caps in an external enclosure, since there's not much room inside?
What is the PS ripple now (VAC)? What is your goal? My goal for PS ripple in an amp is 2mV.
 

Attachments

  • Threshold Stasis 3 inside.jpg
    Threshold Stasis 3 inside.jpg
    153.1 KB · Views: 268
Yeah, that's the guy.

I evaluated the idea of adding external caps in parallel, but I am concerned about two main aspects:
1. the stray resistance of the wires wold degrade the performances
2. The AC current flowing trough the wires would generate a lot of EM noise

I'd rather build the whole PSU (transformer, rectifier and filter) into a separate enclosure, so I would gain room and reduce EM inducted noise too. I am quite sure that those red/blue wires are partially concatenated the magnetic flux of the transformer.

Where did you measure the ripple, at the PSU output or at the amplifier output?
I haven't performed any measurement with the scope yet but, according to the simulations I'd expect to see at least 2 orders of magnitude more ripple than the number you are talking about, with the current filtering.

I don't have a goal in my mind yet in term of ripple. I would need to do a performance/price analysis in order to set a goal. What's your approach to obtain such a low ripple level?
 
Putting some distance from that transformer would be wise.


I check ripple at the last cap in the PS or at the pcb.
If there's space then I would use inductors for a CLC, but otherwise my approach is just the most caps that fit, and try to place a cap near the output transistor. 100mF per output transistor is a good start is what Papa whispered to me once.
I just added an LC (4mH+ 2x68mF) section to a friend's DIY F5, so it went from 'normal' CRC (30mF-0.1Ohm-30mF) to cRcLCC. Ripple went from 50mV to 2mV (but lost 2VDC at rails). So clean, and now can't go back...
 

Attachments

  • Mark F5 crcLCC 1.jpg
    Mark F5 crcLCC 1.jpg
    370.9 KB · Views: 230
Last edited:
2mV is quite impressive! I am gonna start with a CLC and go from there..I feel quite confident that it'll be quite a good improvement compared to the actual 15mF capacitor only.
I think the Stasis 3 is quite of a fine amplifier with an ok-ish PSU that need improvements.
I am not particularly worried about loosing a few % of voltage at the rails, as long as the ripple is reduced, the channels are well decouple and the rails voltage is stable during the peaks of current.

I guess Papa suggested to install capacity closer to the output transistors in order to improve the slew rate and for decoupling purpose..
 
Today I assembled a prototype version (cabling need to be redone properly + fuse and switch need to be installed).
Oh Lord, the amp sounds amazing. I am not just talking about the 60Hz ripple that totally disappeared, but also about an unexpectedly massive overall quality improvement. The soundstagr is now so wide that is seems like my room has no walls, the percussions are incredibly punchy, and the details are so clean and defined.
I didn't expect such a massive improvement from a PSU update, definitely the stock PSU was holding back this marvellous amplifier. With 300$ of components, my amplifier now sounds like it belongs to another category. Wow!
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.