Peter,
Impressive construction of the monoblocs. Layout so neat and short, utmost compactness and really cool looks. Very clever indeed.
Regarding the bootstrap capacitor, I think you can (almost) safely remove it from the circuit - only beware some (100 -150 mV) DC offset on the output.
Proper grounding schemes for inverted and non-inverted amps differ indeed. Will it be audible in your particular layout is another question. But yeah, one usually does not want to share input and output grounds in non-inverting amps. Another thing to beware with non-inverting circuits is to keep input and output well apart from each other because they are in phase with each other and can interfere.
Concerning the sound differences between LM3875 and TDA7294, it’s my personal guess, that you will newer get those two sounding equal and that is because the different circuits and semiconductor types encapsulated in these chips. What you describe as (I am quoting you here) - It offers somewhat thick and veiled sound, as if I were looking for something through a thin mist, more like a haze. Obviously, it lacks involvement and I really don't know if this can improve after burn in time - in my personal experience this is a typical MosFet sound.Others may disagree on this.
Lowering the PS bypass caps may give you more similar sound to the LM3875 but degrades the overall bass and low mid performance IMO. Also PSRR of the TDA7294 is only 75dB compared to the 120dB of LM LM3875. Bypassing PS bypass caps with 100n film caps (very close or strait to the pins) is considered as a standard with all OpAmps in order to prevent them oscillating.
Cheers,
Argo
Impressive construction of the monoblocs. Layout so neat and short, utmost compactness and really cool looks. Very clever indeed.
Regarding the bootstrap capacitor, I think you can (almost) safely remove it from the circuit - only beware some (100 -150 mV) DC offset on the output.
Proper grounding schemes for inverted and non-inverted amps differ indeed. Will it be audible in your particular layout is another question. But yeah, one usually does not want to share input and output grounds in non-inverting amps. Another thing to beware with non-inverting circuits is to keep input and output well apart from each other because they are in phase with each other and can interfere.
Concerning the sound differences between LM3875 and TDA7294, it’s my personal guess, that you will newer get those two sounding equal and that is because the different circuits and semiconductor types encapsulated in these chips. What you describe as (I am quoting you here) - It offers somewhat thick and veiled sound, as if I were looking for something through a thin mist, more like a haze. Obviously, it lacks involvement and I really don't know if this can improve after burn in time - in my personal experience this is a typical MosFet sound.Others may disagree on this.
Lowering the PS bypass caps may give you more similar sound to the LM3875 but degrades the overall bass and low mid performance IMO. Also PSRR of the TDA7294 is only 75dB compared to the 120dB of LM LM3875. Bypassing PS bypass caps with 100n film caps (very close or strait to the pins) is considered as a standard with all OpAmps in order to prevent them oscillating.
Cheers,
Argo
Re: This boots ar made for???
Koinichiwa,
This is the case with some conventional discrete circuits where the load is a resistor. In the TDA7924 there is already a current source in the load section, so no, it does not raise OLG.
In theory this cap can be totally omited with this particular chip.
Sayonara
Koinichiwa,
Tube_Dude said:
This is the case with some conventional discrete circuits where the load is a resistor. In the TDA7924 there is already a current source in the load section, so no, it does not raise OLG.
In theory this cap can be totally omited with this particular chip.
Sayonara
argo said:
I was considering that, but how can you keep them apart when input and output are just pins on IC separated by few milimiters? Anyway, those amp is more resistant to oscillations than LM3875.
argo said:
I changed the input and feedback resistors to the same I use in my Aleph X, and indeed those amps are quite similar now in a sonic signature. TDA seems to have highs etched more and is slightly more forward, but I might have trouble pointing the right amp in a blind test
Attachments
I did some more listening today. Aleph X has definitely more air and is more refined. TDA based Gainclone is faster and more crisp sounding.
The similar impression with Zen, which came out pretty good today. Zen has nice depth and separation between instruments, but seems very slow compared to gainclone.
I did again comparison between OPA 549 and LM3875 and I prefer LM chip, it provides better presentation and is more musical. I'm converting OPA amp into LM based amp.
The similar impression with Zen, which came out pretty good today. Zen has nice depth and separation between instruments, but seems very slow compared to gainclone.
I did again comparison between OPA 549 and LM3875 and I prefer LM chip, it provides better presentation and is more musical. I'm converting OPA amp into LM based amp.
Peter,
It's not just your hearing either. I have the same feeling toward the OPA541 and OPA549 amps I tested. While I thought the OPAs had a flatter freq response and weren't as forward in the mids as the LM3875, I just enjoy listening to LM based amp more. I feel more involved in the music and don't feel like I am wanting for something with the LM3875 based amp when compared to the OPAs. I got goosebumps last night listening to Leo Delibes' Lakme Flower duet.
It's not just your hearing either. I have the same feeling toward the OPA541 and OPA549 amps I tested. While I thought the OPAs had a flatter freq response and weren't as forward in the mids as the LM3875, I just enjoy listening to LM based amp more. I feel more involved in the music and don't feel like I am wanting for something with the LM3875 based amp when compared to the OPAs. I got goosebumps last night listening to Leo Delibes' Lakme Flower duet.
fedde said:And what about:
Aleph X <-> LM3875 GC ???
Or do I interfere with your business now
Fedde
I still didn't do a formal testing between Aleph X and LM3875, but I will probably remove TDA7294 from monoblocks and put LM chip there and then compare both amps. But chances are it may be better than my Aleph X. Also, the Aleph is used above 120Hz only so this where it's compared.
I noticed that my Aleph X is very dependant on the feet I use. I changed recently for acrylic but somehow I don't like them much. It smears highs for some reason. I will go back to what I used before, because I think it was better.
As to the business part, I don't try to promote any product, whatever is better will be treated as such.
But any such comparisons have to be regarded as dependant on a given setup and are influenced by so many variables and tastes, that the outcome is always biased.
It the one based on second article in 1994 Audio Amateur. Ididn't use preamp in that setup and to control gain on Zen, connected the same pot as in GC at the input. I noticed before, that Zen sounds better with preamp.
Just to be sure that I have my setup right, I disconnected the transformer and diodes from TDA monoblocks and used the external supply I use with LM based gainclones as well as only 1000u per rail. My impression was very similar. Comparing TDA7294 and LM3875 chips, TDA has less air, is a bit darker sounding and the music is less involving.
I'm removing TDA chip from my monoblocks and installing LM3875
Just to be sure that I have my setup right, I disconnected the transformer and diodes from TDA monoblocks and used the external supply I use with LM based gainclones as well as only 1000u per rail. My impression was very similar. Comparing TDA7294 and LM3875 chips, TDA has less air, is a bit darker sounding and the music is less involving.
I'm removing TDA chip from my monoblocks and installing LM3875
before you remove the TDAs:
try with only 1000uF per rail
use a 22uF bootstrap cap
bypass the PS rails with 0.1uF
and if all of this doesn't work, try with a LM, but in the same chassis, same parts (transformer and resistors)
I saw you used another transformer, maybe this part is also sounding worse than the other ones you used
try with only 1000uF per rail
use a 22uF bootstrap cap
bypass the PS rails with 0.1uF
and if all of this doesn't work, try with a LM, but in the same chassis, same parts (transformer and resistors)
I saw you used another transformer, maybe this part is also sounding worse than the other ones you used
Bricolo had been having major problems with the IC 7294 earlier, oscillations seemed to be ruining his attempts to get the amp off the ground.
Peter, I guess you've already dismantled the TDAs. I would try actually using the muting resistors as per the AN and using a little more feedback - I think the TDA might be slightly happier with 100K in the feedback path, 56 might be the best. You can accordingly change the value of the input resistor, though I know it also changes a lot of other things and it may go against philosphy of using less feedback.
I would also try more capacitance, I would not be happy to run my TDAs below 3300 or maybe even 4700 uF per rail. The TDAs need a beefy supply, I find 4700 uF ideal.
However, I know there is a sonic difference - there is bound to be, the circuitry in both of them is completely different - but I find the TDA quite good personally.
One more thing, can you try 30 volts rails, I find the TDA does not sound good when above this level - 37 volt is a little high when the recommended spec. is 35 volt into 8 ohm, I like the 27 volts rails the best.
Peter, I guess you've already dismantled the TDAs. I would try actually using the muting resistors as per the AN and using a little more feedback - I think the TDA might be slightly happier with 100K in the feedback path, 56 might be the best. You can accordingly change the value of the input resistor, though I know it also changes a lot of other things and it may go against philosphy of using less feedback.
I would also try more capacitance, I would not be happy to run my TDAs below 3300 or maybe even 4700 uF per rail. The TDAs need a beefy supply, I find 4700 uF ideal.
However, I know there is a sonic difference - there is bound to be, the circuitry in both of them is completely different - but I find the TDA quite good personally.
One more thing, can you try 30 volts rails, I find the TDA does not sound good when above this level - 37 volt is a little high when the recommended spec. is 35 volt into 8 ohm, I like the 27 volts rails the best.
I still have 666kHz oscillations, bur 10mV peak to peak (but strangely, 10mV at volume pot=0 and pot=100, when I'm near 50%, the oscillation rises to 100mV!) I still can't understand
I can have oscillations, since I still have no casing
but variable oscillations, depending of the volume pot's position, that's really weird
I'll try with a 9.5kR+0.1uF between +in and gnd, to match the +in and -in impedances
I can have oscillations, since I still have no casing
but variable oscillations, depending of the volume pot's position, that's really weird
I'll try with a 9.5kR+0.1uF between +in and gnd, to match the +in and -in impedances
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.