Back to TDA1558Q and its noise problems:
There is no reasonably clean ground for audio input provided by either the chip or datasheet.
Therefore I'm thinking of using a rail splitter much like the popular CMOY Headphone Amplifier. With the TDA1558Q in bridged configuration, there's no heavy current needed for "virtual ground" and so it could simply be a pair of ordinary 1/4 1% resistors, possibly a 2k value.
Any ideas or help would be sincerely appreciated.
Pin 4, the internal rail splitter is emitting:
+6.28
-5.81
(the power supply in use has 12.1v)
There's a 0.47v error in the internal rail splitter, so I don't think its useful, unless, perhaps stabilizing it can reduce the noise?
Reading materials:
Philips/NXP document showing that the company won't help you apply their products: http://www.nxp.com/acrobat_download2/various/AUDIOIC.pdf
The CMOY rail splitter: http://tangentsoft.net/audio/cmoy-tutorial/misc/cmoy-tangent-sch.pdf and maybe it can make a solution for getting a reliable input circuit ground on a bridged TDA1558Q?
There is no reasonably clean ground for audio input provided by either the chip or datasheet.
Therefore I'm thinking of using a rail splitter much like the popular CMOY Headphone Amplifier. With the TDA1558Q in bridged configuration, there's no heavy current needed for "virtual ground" and so it could simply be a pair of ordinary 1/4 1% resistors, possibly a 2k value.
Any ideas or help would be sincerely appreciated.
Pin 4, the internal rail splitter is emitting:
+6.28
-5.81
(the power supply in use has 12.1v)
There's a 0.47v error in the internal rail splitter, so I don't think its useful, unless, perhaps stabilizing it can reduce the noise?
Reading materials:
Philips/NXP document showing that the company won't help you apply their products: http://www.nxp.com/acrobat_download2/various/AUDIOIC.pdf
The CMOY rail splitter: http://tangentsoft.net/audio/cmoy-tutorial/misc/cmoy-tangent-sch.pdf and maybe it can make a solution for getting a reliable input circuit ground on a bridged TDA1558Q?
Compare the 16w amplifier with the sound you want to crank up -versus- the 64w amplifier with the sound you'd never want to hear again. Only one of these options is useful, so lets make clean amplifiers.
Yes, I am well aware how wattage is the most overstated factor of a quality amp. It seems to be the one that gets all the casual listener's ears to perk up. Them and the mobile audio crowd love them watts. Plus, I can't play music very loud due to neighbors.
I will probably snag a few of those TDA8561Q's. I think it also said there's some "switch on/off pop" protection. Which saves me from extending myself further into an area I am brand new to. The less I have to do the better. In addition there's better THD. Win Win.
I didn't exactly mean it that way. The goal isn't to do as little as possible, exactly. More so strike a nice balance, while trying to achieve the best product possible. But I want to build something myself.
I just realized I screwed up. I thought the TDA8561Q had 2 less wpc than the TDA1558Q, not the other way around like it is. Hence the giving up watts comment. I'm stupid.
The specs for the Velleman aren't for 12V single supply. I wonder how much worse that wattage gets running it that way. Is the THD for the TDA1521Q as good as the Velleman suggests, because I am having a harder time understanding that spec sheet?
Its not exactly for a 12v power supply. It uses the very common 12-0-12vac center tap transformer, which is a 24vct transformer. The power supply circuit is onboard with the amplifier itself, so you don't have to build a power supply. The reputation of the product is MUCH different that what we're looking at (and that's a good thing).
I think the datasheet is interesting because its SO very similar to the amplifiers we're looking at, and comparing it "may" show how to use a split-rail power supply with a bridged TDA1558Q, via a simple rail splitter (2 resistors like the CMOY) to get a reliable and noise-free input circuit.
here a photo of the test circuit using the 1558Q....
it did work at first start...
you can see that there s four electrolytic caps...
two are connected to the supply rails...
the two other are connected like this :
the negative side is connected to earth, and the
positive side is connected through a dual switch to
a half of each bridge, thus reducing the power
delivered by the non relevant amplifiers ,
so i can add a bass box connected from the non inverting
output of higher channel to the inverting input of the lower channel,
so the amp become a 2 X 4 W + 20 W for the boom box...
i m preparing a pic to help explain..
it did work at first start...
you can see that there s four electrolytic caps...
two are connected to the supply rails...
the two other are connected like this :
the negative side is connected to earth, and the
positive side is connected through a dual switch to
a half of each bridge, thus reducing the power
delivered by the non relevant amplifiers ,
so i can add a bass box connected from the non inverting
output of higher channel to the inverting input of the lower channel,
so the amp become a 2 X 4 W + 20 W for the boom box...
i m preparing a pic to help explain..
8561Q is a little better than the 1558Q, due to reduced gain that
did allow to reduce a little the distorsion...
power output capability is slightly better, in the range of
10% at equal supply voltage...
i can only warn about ground loops if smps are used,
as the negative is connected to earth if there s a safety
earth cable...
power supply with no safety earth, i.E, class 2 insulated in
european norms, don t suffer from these ground loops that
create buzz and hums even all things look perfectly connected..
the ground loop start from the amp which have its ground connected
to negative rail of the power supply..this latter will make
this ground connected to earth..
if the source signal has also a ground connection, then there
will be a ground loop...
suppressing the earth connection of one of the side, amp or
source generally resolve the problem but create another one as
the safety earth is no more adequatly connected...
This is a badly dangerous problem in the UK because outlets there are split-rail AC, much like U.S. Air-Conditioner, Water Heater, Clothes Drier, and Cook Stove 220 outlets--but yet somehow even more dangerous.
This is also a problem in the U.S. if the audio equipment is plugged into more than one outlet. Its really hazardous to musicians if the mixer board happens to be plugged into an outlet (at some distance away) that operates on the "other" rail, and shocks (especially to guitarists) are commonplace here in the U.S. Recently, this problem also causes "mooing" subwoofers for home theater installations.
So, I think that the plastic-bodied (absolutely no touchable metal parts) "wall wart" or "laptop cord" is a really good idea for safer starter projects. Its not totally safe; however, the potential output is considerably decreased, thus increasing safety (decreasing risk) by that same amount.
Ah, well, I suppose I'll go fight with this little amp some more. Gosh, I don't just want to give up on it; however, DIY'ers the world over have had very bad luck with TDA1554/5/8 family of chips. Some get it to work nicely but there are no support circuit or pcb examples to help anyone else.
Last edited:
Could you make a sketch or a drawing of the circuit?
All of my attempts, datasheet, extra load, split rail, single rail, etc. . . make a very annoying buzz! In fact, there's a wide variety of awful noises possible, along with some music that's quite good.
here the exact configuration i used..
as you can see, it s the minimal one...
pin3 is connected with the relevant ground pins...
a cap is added to pin4....
the switches allow the left/right channel to work
in non bridge mode, as a bass box can be added..
somewhat tricky, but it works very well...
of course, this particular implementation is of no
use if only two speakers are used...
i can only remind everybody to use a potentiometer
at the input..
also; be carefull with those plastic smps, as the negative side
is often connected to earth..
check if there s 3 wires in the ac input; and measure
the resistance between earth and negative output (the thing being
unplugged, of course !!)....
as you can see, it s the minimal one...
pin3 is connected with the relevant ground pins...
a cap is added to pin4....
the switches allow the left/right channel to work
in non bridge mode, as a bass box can be added..
somewhat tricky, but it works very well...
of course, this particular implementation is of no
use if only two speakers are used...
i can only remind everybody to use a potentiometer
at the input..
also; be carefull with those plastic smps, as the negative side
is often connected to earth..
check if there s 3 wires in the ac input; and measure
the resistance between earth and negative output (the thing being
unplugged, of course !!)....
Well, I'm trying that exact same circuit (in bridge mode).
My SMPS has only a 2 conductor ac cord.
I have a 10k potentiometer.
Without anything plugged into the inputs, there's a buzzzzzzz (guitar amp like noise).
Does yours do that when you use bridge mode?
So far, the 17th redo involved:
Replace the TDA1558Q
Replace all resistors
Replace all caps
Replace all wires
Replace the power supply
Exactly the same! BZZZZZZZ!!!
Help?
P.S.
I also get high dc offset when the amp is first turned on and it goes to normal after 2 seconds. But, this is very bad for speakers.
This is a brute force method to "turn down the noise" by readjusting the gain.
supplementary feedback resistors
Supplies:
4 of 2.2k resistors
Location:
2.2k from pin 15 to pin 12
2.2k from pin 16_17 to pin 10
2.2k from pin 15 to pin 6
2.2k from pin 1_2 to pin 8
It is still far more than enough gain for the MP3 player.
EDIT:
Working on it still, and testing, but this bit is nice progress. Finally, near silent at idle!
This mod also seems to stop the wide variety of misbehaviors and racket.
1.9k and 1.5k values are even better (smaller figures = less noise).
The RCA Pearl MP3 player has no trouble pushing this amp to clipping with the 1.5k supplementary feedback resistors.
supplementary feedback resistors
Supplies:
4 of 2.2k resistors
Location:
2.2k from pin 15 to pin 12
2.2k from pin 16_17 to pin 10
2.2k from pin 15 to pin 6
2.2k from pin 1_2 to pin 8
It is still far more than enough gain for the MP3 player.
EDIT:
Working on it still, and testing, but this bit is nice progress. Finally, near silent at idle!
This mod also seems to stop the wide variety of misbehaviors and racket.
1.9k and 1.5k values are even better (smaller figures = less noise).
The RCA Pearl MP3 player has no trouble pushing this amp to clipping with the 1.5k supplementary feedback resistors.
Last edited:
pin15 is a virtual ground in ac mode..
you cant reduce gain by injecting feedback to this pin...
what you have done is to reduce the gain of the non
inverting amps , making the bridge unbalanced, as the inverters
amps have not effectively reduced gain..
so a half of the bridge provide more output voltage than the
other half...
if there s still noise with a 10K potentiometer in minimal position,
then noise is due to a bad layout , as you said that changing the
power supply didn t change nothing..
a photo of you proto would help, if possible..
Well, I didn't test in bridge mode until now. The proportions are lopsided. Feedback from pin 15 is only half as effective as feedback from the input. Darnit! No wonder they discontinued this one and made a new one with more reasonable gain setting.
sorry that it didn t work according to your wills...
since i didn t test the 8561Q, i m in the waiting of your
experiences, as i surely will go with this one next time..
could it be that i was just lucky with the old fashion ic ?..
generally, ic s work perfectly at first shot, if i except
some that were badly designed like the horrible TDA2005,
the STM concurrent of the 1558Q....
since i didn t test the 8561Q, i m in the waiting of your
experiences, as i surely will go with this one next time..
could it be that i was just lucky with the old fashion ic ?..
generally, ic s work perfectly at first shot, if i except
some that were badly designed like the horrible TDA2005,
the STM concurrent of the 1558Q....
Wahab,
Could you clarify these issues with the TDA2040 diode schematic you posted?
Considering Daniel's issues with the TDA1558... I think I will focus my efforts on the TDA2040 until I get the TDA8561Q.
How do the THD and gain values compare between the TDA8561Q and TDA2040?
Thanks
For TDA2030, TDA2040, TDA2050, LM1875. . .
You set the gain you want.
On a non-inverting amplifier. . .
If the feedback resistor is 27k (inverting input to speaker) and its partner is 820R (0.82k) (inverting input to signal ground) then you can divide the two with the calculator to find the gain (albeit there's a more exact formula and a simple division will be off by "1")
A voltage divider:
Its like a potentiometer, pot, l-pad, except there's no dial. You can divide a voltage divider's two resistor values to find its effect. What it does is divide. You can too.
You set the gain you want.
On a non-inverting amplifier. . .
If the feedback resistor is 27k (inverting input to speaker) and its partner is 820R (0.82k) (inverting input to signal ground) then you can divide the two with the calculator to find the gain (albeit there's a more exact formula and a simple division will be off by "1")
A voltage divider:
Its like a potentiometer, pot, l-pad, except there's no dial. You can divide a voltage divider's two resistor values to find its effect. What it does is divide. You can too.
I've used the tda2003, tda2005, etc. . . family of chips in bridged design. The efficiency is rather horrible; however, audio performance depends on your design. Being op amps, they'll do almost exactly as you ask; however, much of what you ask for will probably turn to heat instead of speaker output, and we don't want that extra expense.
The TDA1558Q--I do have to report that its not ever going to qualify for high fidelity class audio. The "disqualification" is the noise output, due to its locked in setting of wildly inappropriate gain.
My experience with TDA1558Q was rather cruel, in fact. Its presentation is desirable and to the extent that most amplifiers of any price cannot match it.
It does indeed do the "music takes flight" experience (VERY RARE), along with noise. And that is a heartbreaking combination. Normally, it takes at least three of full bandwidth speakers to make the presentation that this amp can do with one or two. Since the problem is "locked" inside the chip, I have to tell that one goodbye.
Remaining options:
With TDA8561Q, I'll look forward to seeing if Philips, NXP has mastered their symmetry chip technology. After it comes in, its going to face some stiff competition in the form of LA4628 (which is a great performer but heatsink will cost slightly more). Fail: I've already tested Toshiba and KEC's similar offerings and have to report that they don't even meet mid-fi standards. More research: We also need to find out what similar chips are offered by the major players, ST and Natsemi.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.