For a shortcut, you can try leaving the ceramic cap in circuit and then adding a bypass, a 4n7, 6n8 or 10n, little green polyester dip cap. Use the smallest size that works for some benefit.I remember a time when people complained about ceramic coupling caps, no one does anymore.
It arrived. And I knocked off R27.
I ran it with the 36db gain factory setting, initially, but then removed R27 to set 20db. The lower gain was a great improvement! Less blare and a little better resolution results with the lower gain.
After lowering the gain (as easy as removing R27), it is a rather good amplifier. The resolution isn't lacking, and the efficiency is top-notch.
There's plentiful output power, but the sonic signature is slightly loud at the lower-treble range, which can make the bass more difficult to hear, even though there's not actually any lack of bass. I suppose that either a power circuit mod and/or my BlareBuster input circuit, can probably straighten it out. Will try them soon...
I ran it with the 36db gain factory setting, initially, but then removed R27 to set 20db. The lower gain was a great improvement! Less blare and a little better resolution results with the lower gain.
After lowering the gain (as easy as removing R27), it is a rather good amplifier. The resolution isn't lacking, and the efficiency is top-notch.
There's plentiful output power, but the sonic signature is slightly loud at the lower-treble range, which can make the bass more difficult to hear, even though there's not actually any lack of bass. I suppose that either a power circuit mod and/or my BlareBuster input circuit, can probably straighten it out. Will try them soon...
Attachments
It seems there are many, many choices for flip up magnifiers from dirt cheap to really high dollar ones. Any suggestions what ones to consider?
Thanks! The range is near the 4khz, ear canal resonance. It isn't a huge amount. But it is enough to interfere slightly with hearing bass and baritone as well. In this frequency range, it seems like a power circuit error. Perhaps those 330u should have been 470u?
Patched it...
The resistor in the photo (1 ohm or higher) is optional, unnecessary, and makes the filter quit working when the amp is played quietly. Also optional is an FB43 ferrite on the line between the cap and diode.
Well, I can certainly hear the bass now. It doesn't blare as much, and it is able to play much, much louder.
Power decoupling caps are signal caps, so any change of size/value, will also change the tone. I decided not to change them. Anyway, I came up with a fast and easy patch, in the photo, attached. It is a series 1n540x diode (an ordinary standard diode) and then a 3300u cap added (most big value caps will do). Multi-channel applications would use a diode per each board, with the additional benefit that crosstalk is blocked (like dual-mono, except not expensive).
The resistor in the photo (1 ohm or higher) is optional, unnecessary, and makes the filter quit working when the amp is played quietly. Also optional is an FB43 ferrite on the line between the cap and diode.
Well, I can certainly hear the bass now. It doesn't blare as much, and it is able to play much, much louder.
Attachments
TPA3118D2
Same question here, additionally how does 330/470uF corresponds with 4kHz? And what about ESR?
At Daniel, what does the Diode+C construction in a single amp like shown on your picture? How does the diode fix the fact with the 330/470uF? I'm really trying to understand, but I actually don't get the mechanism behind.
If the bass is more "present" than before, dunno if you mean that, than I would blame the stock caps ESR and a poor power supply. But, even 3300uF won' help that much. What I could understand is, that the diode prevents affects from bus pumping back into the control loop of the power supply. But as this amp is BTL, there won't be such a thing, especially when a output filter is present.
So, I don't understand what's going on and what you fixed.
Same question here, additionally how does 330/470uF corresponds with 4kHz? And what about ESR?
At Daniel, what does the Diode+C construction in a single amp like shown on your picture? How does the diode fix the fact with the 330/470uF? I'm really trying to understand, but I actually don't get the mechanism behind.
If the bass is more "present" than before, dunno if you mean that, than I would blame the stock caps ESR and a poor power supply. But, even 3300uF won' help that much. What I could understand is, that the diode prevents affects from bus pumping back into the control loop of the power supply. But as this amp is BTL, there won't be such a thing, especially when a output filter is present.
So, I don't understand what's going on and what you fixed.
Last edited:
It doesn't correspond to a particular frequency, except that more thorough power decoupling can suppress sonic signature more thoroughly. If the decoupling caps are too large, both noise and audio are suppressed. That's why I didn't change them.
The 4x330u, caps in parallel is extremely low ESR. The cap that I added is Cornell LP series which has extremely low ESR.
4x330u, 1v drop, 3300u. That is CRC.
Diode was used for R, because the diode won't make rail sag fluctuations. This non-sag version was used because it is on a power amp that needs good current for bass--so that's why the "R" is a diode. The mechanism is still ordinary CRC filter.
It certainly was easy to do.
P.S.
In that photo, the bypass resistor is technically an error--an unnecessary option for making the filter work less well conditionally and dynamically. That latter bit, "dynamically," may provide some entertainment, so long as you didn't need the filter to work as well as possible.
I switched my bypass resistor to 2R, And finally hit the target--the amp is pleasant at full blast, but still interesting when playing quietly as well.
Last edited:
So I've got a bunch more measurements from various SMPS units I happen to have around. Rather than clutter up this thread anymore though I think I will start a new SMPS thread somewhere, so those interested keep an eye out this weekend. I will tell you that the little Chinese board I have is far from the worst, beating out a few Meanwell supplies even.
I am also curious about how the diode+cap helps the noise with the Sanwu boards. Making a half bridge rectifier to attempt to further rectify the noise? Seems just using a quieter supply would be a better bet.
I am also curious about how the diode+cap helps the noise with the Sanwu boards. Making a half bridge rectifier to attempt to further rectify the noise? Seems just using a quieter supply would be a better bet.
A whole lot less than a resistor would do (in comparison to a normal CRC).
The diode datasheets are informative though: High-current capacity + Low-voltage tolerance = least variance. The reverse voltage figure is not utilized. With my 8 ohm speaker, this amp is drawing low current, so I used a 1n5405 diode, and that's just fine.
Sure, it lost a volt, but it also lost that blare. Thereby, the proportion of useful output power, increased, considerably.