ppppp: I started reading the alterations suggested by ClaudeG in the "Aiyima TPA3251 Modification Build Thread !" mainly Post#s 105, 143 & 144 where he suggests numerous by-passing implementations. I’m going to try a few of the ones suggested in Post 144 on Lester’s board. I’ve ordered the parts so we will see.
Regards, Pete
D
Deleted member 148505
I suggest use the stock 1uF per slaa788a, deviating from the default PFFB values and configuration has a large impact on stability.
Putting 100nF bypass caps under large electros was being frowned upon (I remember AndrewT) because of possible ringing.
D
Deleted member 148505
D
Deleted member 148505
D
Deleted member 148505
Solder jumper SLAVE_JMP will pull the FREQ_ADJ to DVDD. We don't need to change resistors etc.
Also PFFB config is dependent on osc frequency, so don't change resistor for FREQ_ADJ. If you will change the frequency, you have to change the feedback components or disable PFFB because the amp will go unstable if the FB config is incorrect.
On the TPA3255EVM, moving the jumper to select Slave Mode disconnects FREQ_ADJ from ground and pulls it to 3.3V (I assume that is DVDD):
https://www.ti.com/lit/df/slar129a/slar129a.pdf?ts=1608646859138
On yours, by connecting a solder jumper to pull FREQ_ADJ to DVDD, won't it connect DVDD to ground (through whatever resistor you have - 30kR on TPA3255EVM)? Is that OK? I mean, it's not really a short because of the resistor, but the TPA3255EVM doesn't have any connection to ground at all once the slave jumper is in.
EDIT: 0.11 mA, I guess that's nothing 🙂
https://www.ti.com/lit/df/slar129a/slar129a.pdf?ts=1608646859138
On yours, by connecting a solder jumper to pull FREQ_ADJ to DVDD, won't it connect DVDD to ground (through whatever resistor you have - 30kR on TPA3255EVM)? Is that OK? I mean, it's not really a short because of the resistor, but the TPA3255EVM doesn't have any connection to ground at all once the slave jumper is in.
EDIT: 0.11 mA, I guess that's nothing 🙂
Last edited:
commstech, ppppp, asuslover & Lester, just let you tell you where I am with Lester’s 3255 amp.
It sounds really good now, sort of a silkiness with no harshness and nice detail within the music. As a reference, I’m driving the amp using 3e-Audio’s 48V SMPS (no longer sold) with the stock 4 output caps changed to four Panasonic’s 1000uF 63V. SE signal input. A number of the by-pass ideas are coming from ClaudeG’s post #144 in the Aiyima tpa3251 Modification forum.
1) I by-passed the 1800uF Panasonic FC caps with 1uF WIMA MKS (waiting on a new order since I’d rather use at least 2.2-3 uF for by-passing; approx a 1:500-600 by-pass).
2) the pre-tap 3255 caps were changed to 2.2uF Nichicon UES bipolars
3) I tried two different pre-1656 caps; either WIMA 2.2uF MKS or the suggested KEMET SMR 1uF caps. I really liked the WIMAs but pulled them only because they sounded really shrill on bright violin passages. The KEMET SMR quieted down the music a bit but I liked the silky calmness they imposed and details are still quite present.
4) I mounted a Elna Silmic II (10uF 35V) under the board across pins 4&8 of the op-amp DIP mount.
I think Lester’s amp sounds really good now. I might try two other by-pass options on the 1656’s pins 4&8 once I get the parts. I think that’s all I have for now.
Regards, Merry Christmas all
Pete
It sounds really good now, sort of a silkiness with no harshness and nice detail within the music. As a reference, I’m driving the amp using 3e-Audio’s 48V SMPS (no longer sold) with the stock 4 output caps changed to four Panasonic’s 1000uF 63V. SE signal input. A number of the by-pass ideas are coming from ClaudeG’s post #144 in the Aiyima tpa3251 Modification forum.
1) I by-passed the 1800uF Panasonic FC caps with 1uF WIMA MKS (waiting on a new order since I’d rather use at least 2.2-3 uF for by-passing; approx a 1:500-600 by-pass).
2) the pre-tap 3255 caps were changed to 2.2uF Nichicon UES bipolars
3) I tried two different pre-1656 caps; either WIMA 2.2uF MKS or the suggested KEMET SMR 1uF caps. I really liked the WIMAs but pulled them only because they sounded really shrill on bright violin passages. The KEMET SMR quieted down the music a bit but I liked the silky calmness they imposed and details are still quite present.
4) I mounted a Elna Silmic II (10uF 35V) under the board across pins 4&8 of the op-amp DIP mount.
I think Lester’s amp sounds really good now. I might try two other by-pass options on the 1656’s pins 4&8 once I get the parts. I think that’s all I have for now.
Regards, Merry Christmas all
Pete
Nice update Pete, thanks. I can't wait to jump in the game too, when I'll get all parts in january...
Just have question to you about pre-1656 caps. With 1uF input cap you created input highpass filter with slightly less bass, approximately -1dB at 100Hz and -2dB at 60Hz. Is it hearable? Ideally that cap should be of higher value to not attenuate low freqency that much. You can try to use 2x 1uF SMR in parallel.
Btw, if SMR sounds good, I'm thinking to use it also in output LC filter. 1uF value will fit nice in that place.
And Merry Christmas 🙂
Just have question to you about pre-1656 caps. With 1uF input cap you created input highpass filter with slightly less bass, approximately -1dB at 100Hz and -2dB at 60Hz. Is it hearable? Ideally that cap should be of higher value to not attenuate low freqency that much. You can try to use 2x 1uF SMR in parallel.
Btw, if SMR sounds good, I'm thinking to use it also in output LC filter. 1uF value will fit nice in that place.
And Merry Christmas 🙂
Depending on how much headroom you need down low (actual power output), you can just EQ any attenuation in the bass.
Well, I suppose it depends on what source, DSP, pre-amp, etc. you have. But that's an option, anyway.
Well, I suppose it depends on what source, DSP, pre-amp, etc. you have. But that's an option, anyway.
Of course it is. But we want to find out and "create" the pleasing sound using right components, so we should be aware of the sound characteristics of our selected components and their combination, right? IMO that's the point of this DIY hobby.
Asuslover, sure! I’d love to just have 4 to try them out if that’s a reasonable request. I’d rather not buy the 14 on EBay and then decide they didn’t add value to my builds.
Many thanks & Merry Christmas!
Pete
Merry Christmas to you as well ppppp!
I calculated a 1uF cap creates a cutoff at 16Hz or so while a 2 would give a 8Hz cutoff (2.2 = 7.2Hz cutoff). I may try paralleling another 1uF SMR under the board as a 2.2uF Kemet SMR cap becomes too big for easy mounting.
I did order 2.2uF Mundorf MKPs to try in the pre-opamp position as well. Waiting for them to arrive..mounting them up will be a challenge 😉
Pete
I'm curious about those Mundorfs. Personally I would love to use Mundorf Supreme as input caps, but there is a catch as 4x 10uF would be bigger than the whole Lester's board and add weight of 1kg (2.2 lbs) 😀
As to 1uF vs. 2.2uF ... Yes with 1uF and 10kOhm the cutoff is 16Hz, meaning that at 16Hz you will have half bass (-3dB). But that attenuation starting much sooner and from 100Hz down should be slightly hearable. But the more serious thing is that this C-R highpass filter also alter phase, and with 1uF you will get pretty big phase shift at audio band.
Try this calculator and compare the charts of 10kOhm + 1uF vs. 10kOhm + 10uF
RC High Pass Filter Calculator
As to 1uF vs. 2.2uF ... Yes with 1uF and 10kOhm the cutoff is 16Hz, meaning that at 16Hz you will have half bass (-3dB). But that attenuation starting much sooner and from 100Hz down should be slightly hearable. But the more serious thing is that this C-R highpass filter also alter phase, and with 1uF you will get pretty big phase shift at audio band.
Try this calculator and compare the charts of 10kOhm + 1uF vs. 10kOhm + 10uF
RC High Pass Filter Calculator
"Try this calculator and compare the charts of 10kOhm + 1uF vs. 10kOhm + 10uF
RC High Pass Filter Calculator"
ppppp: I usually just do freehand calculation for frequency low cutoff equation to examine the absolute number..
Fc=1/(2 x pi x R (ohms) x C (farads)), but it is important to know the roll offs I agree.
Unfortunately I don’t think KEMET SMRs will come in high enough of an AC voltage rating at 1uF to fit in the "C” of the FC space that is on Lester’s boards (with 15mm lead spacing).
Cheers, Pete
RC High Pass Filter Calculator"
ppppp: I usually just do freehand calculation for frequency low cutoff equation to examine the absolute number..
Fc=1/(2 x pi x R (ohms) x C (farads)), but it is important to know the roll offs I agree.
Unfortunately I don’t think KEMET SMRs will come in high enough of an AC voltage rating at 1uF to fit in the "C” of the FC space that is on Lester’s boards (with 15mm lead spacing).
Cheers, Pete
Wow Pete, thanks for coming back with these findings. I am currently happy with the sound, having only experimented with the op amp NE5532, LME49720, LT1364 and LME49860. My liking is more towards LME49860 now but i intend to order OPA1656 and some other op amp to try out. i shall be experimenting with the pre opamp caps and the pre Tp3255 Caps soon for different flavours.
D
Deleted member 148505
Correct, connection of 10K or 30K ohms from 3.3V to ground is nothing to worry about, no voltage divider since connection comes from low impedance 3.3VDC.
Attachments
Last edited by a moderator:
- Status
- Not open for further replies.