HiFiTOY log: I've moved from CM6631A to CM6642 USB IC which is so crazy compact and need almost no surrounding parts that I found possible to add even more aluminum caps on the 33V rail. There are 11pcs 1000u/35V KZH NCC 15-17mOhm ESR of each, so totally it is 11000uF 1.5mOhm tank + 7pcs 1210 1u 100V Murata's ESR=5mOhm! Near to ideal power rail, I believe. Furthermore, now BT stream isn't optional but always onboard because I got a free room for BT module. I also ordered a bit customized MIDEN inductors for a better THD and EMI performance.
PS: I spent a lot of time with C-Media and its distributor to get a simple, if not say trivial, application info regarding 6642. I think about 1 month I have thrown for it ))
PS: I spent a lot of time with C-Media and its distributor to get a simple, if not say trivial, application info regarding 6642. I think about 1 month I have thrown for it ))
Attachments
Back to 3251 test, so far with no heatsink and 1.4W THD+N is 3 times higher than claimed in SLAA788A )) THD itself is close to SLAA788A THD+N result for TPA3251 i.e. .002% but my THD+N is .006%.. S/N with preamp 1VRMS sensitivity -110db(A), I think I lost somewhere 3db at least.
Attachments
Last edited:
The debugging of the stealth panel is an interesting process. I've printed front part of the heatsink, painted by cheap black spray(I hate cheap paints, they never do decent looking surface but I had no time to order good one), and here we go! YouTube
Attachments
I've found the way to improve S/N for TPA3251 amp, now S/N < -112db(A) at max volume, and < -114db(A) at middle volume setting. Still, have some clicking noise during amp's Off procedure, and ugly clipping behavior at speaker load and no load(seems phase margin too low). That self-oscillating noise around clipping is audible, hence not acceptable, funny that nobody cares about it and about Off click-noise as well in the Ti and their e2e forum, they are deaf? ))
After some modification I've got not too bad clipping with tpa3251 in the post-fb mode, however, THD+N looks higher than original Ti's app-note(at speaker load THD, within measured range are better than THD+N up to 3 times). Signal/Noise ratio is still good -112db(A) at max gain(sensitivity 1VRMS). Finally, due to non-ideal clipping behavior of the tpa3251 in post-fb mode and 4ohm rated speaker load, the amp losing 15W of 1% THD power i.e. good clipped TAS5624 in the same conditions will produce about 130W. I personally don't hear a harshness after modification but probably somebody will. Frequency response at 20kHz 4 Ohm vs no load <.5db, hence post-fb doing its work 
Attachments
Last edited:
Hi there! No, I didn't give up the project yet but currently busy with some HPA stuff Next one is HiFiTOY HiFiTOY | index
Next one is HiFiTOY HiFiTOY | index
You forgot the y-scale on thsi one...
https://static.wixstatic.com/media/...3417f9f1afe02939aa31b~mv2_d_2557_1397_s_2.png
//
https://static.wixstatic.com/media/...3417f9f1afe02939aa31b~mv2_d_2557_1397_s_2.png
//
Is there a final version of this amp or whatever we want to call it? Is it possible to connect two channels together to achieve twice the performance of the woofer?
I plan to activate a 3-way box Blue Note 2 (BlueNote - Ausgabe 6 2007 - Lautsprecherbau-Magazin 2007 | Lautsprecherbau - Lautsprecher selber bauen).
Can these elements (4 in total) be served centrally and via Wifi or similar?
Ps .. hypex FA253 DIYclassd.com is known
English by Google translator
I plan to activate a 3-way box Blue Note 2 (BlueNote - Ausgabe 6 2007 - Lautsprecherbau-Magazin 2007 | Lautsprecherbau - Lautsprecher selber bauen).
Can these elements (4 in total) be served centrally and via Wifi or similar?
Ps .. hypex FA253 DIYclassd.com is known
English by Google translator
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.