Mmmmmm..that's about what I paid for my 3116. I have always liked the look of those sleek inductors. And nice to see the Orange drops. If I needed that much power I would definitely give this a try.
the cheap 7492 version is available with orange drops:
Tda7492 hoch macht leistungsstarken Verstrker bord amplificador digitale 50 w* 2/100 w se mit heizkrper hei in Funktion
er Verwendung von hohen- Geschwindigkeit smt-bestckung maschine, bertemperatur reflow, die Leistung ist sehr aus integrierte
less than 6€ shipped.
i do not need, but it's too cheap to not get it.
i ordered that board more than a year ago and it never arrived.
now it is only 13€ shipped ..... looks like another good deal.
Tda7492 hoch macht leistungsstarken Verstrker bord amplificador digitale 50 w* 2/100 w se mit heizkrper hei in Funktion
er Verwendung von hohen- Geschwindigkeit smt-bestckung maschine, bertemperatur reflow, die Leistung ist sehr aus integrierteless than 6€ shipped.
i do not need, but it's too cheap to not get it.
i ordered that board more than a year ago and it never arrived.
now it is only 13€ shipped ..... looks like another good deal.
Is adifferential snubber the pre-lc zobel i've.tried? if so, i certainly noticed a positive difference in listening tests.
yes, have you tried the "bootstrap snubbers", 4 of them, double the parts, if you want equal values you double the capacitor and half the resistor, but I suggest to use TI values
No. the bootstrap snubber is on list of to do's, and was going to use datasheet values.
the difference here is that the differential snubber doesnt go to GND.
because of the info i read, and using a pllxo, i was going to use across speaker. when i read xrk review of 7492 sound, and it having the boucherot cell I thought worth investigating. i definately heard a positive improvement side by side without the bootstrap snubber to GND yet fitted.
the difference here is that the differential snubber doesnt go to GND.
because of the info i read, and using a pllxo, i was going to use across speaker. when i read xrk review of 7492 sound, and it having the boucherot cell I thought worth investigating. i definately heard a positive improvement side by side without the bootstrap snubber to GND yet fitted.
Just posted the photo of the mocked up boucherot cell on the 3116 thread.
After my comparison, I personally do think it must be contributing to the 7492 sounding more precise, and clear. I certainly heard that difference on the 3116.
As I mentioned, I used LTspice before physical mod. I added the 3116 output circuits after my pllxo (This is a 4th order butterworth xo on HP) circuit (using 15uH/470nf LC and the 3116 datasheet values of bootstrap caps/bootstrap snubber/outputfilters)
the results were phase allover the place 2 gentle rolling slopes, wavey lines. 0db did not align with 0degrees, nor did the phase or voltage line up with any value in particular. in fact, the phase for both +/- started aligned with each other until around 300-400hz then one slowly rolled off. When it was about 90degrees below the other, the top phase gently rolled off to kind of follow the bottom slope and eventually they levelled out at 180degrees apart around the 20kHz.
Adding the Boucherot cell as per 7492 schematic resulted in perfectly flat phase, which had exactly 180degrees between +/- throughout 20-20k frequency, and when the phases did switch they did so in a verticle fashion. none of this gentle rolling slopes of phase, starting at 0degrees and ending up at 180degrees when reaching 20kHz.
It also put the 0hz at exactly +/-40db with the 0degrees at 0db on the spice graph. and left at 20kHz 360degrees away still at 180degrees apart, with the voltage hitting the 0degrees at 0db perfectly.
I would say, and I am a novice just piecing together information put forward online by knowledgable speaker/XO/Amp builders, that the dramatic difference this circuit makes to the phase is what makes the difference.
I read an article by Ludwig about crossovers which included sections on phase etc. in that he discusses that effects of phase are very audible.
I read about how drivers have particular optimal min/max Hz range according to size (e.g i.e for my 8inch woofer it would be 180Hz to 1715Hz, between 3.2 & 3.5 octaves, and when sounds are played simultainiously from different areas of spectrum the driver will not necessary cope to well, and he determined this caused what I call a "muddled" sound. Of course, he achnowledged some of it is down to quality, and design of drivers, but he also demonstrated a good arguement for phase being responsible too.
So after reading that, and testing the Boucherot cell both in spice graphs and reality, I am judging that it corrects the phase issues which cause a more compressed or "muddled" sound...thus a cleaner, clearer, more open, higher detailed sound. NOT increased highs at the detriment to bass. infact, allowing more detail of the bass to come out....which my mentioning of the bass "strike" was about. I could clearly hear the instant the bass note was initiated in both drums and electronically generated dance music, and I still heard the deep belly of the bass the note created.
I personally think the circuit correcting the phase to dead straight, verticle and 180degrees apart allows the drivers to open up sounds and harmonics start to shine through.
just my impressions, and guessing.
Last edited:
i read that, too.
but both 7492 and 7498 say 60k is the 'typical' value.
i just wanted to gt knowledge of what that 'typical' means.
i revived i big 7498 board that i got a while back from 'aliexpress-sale'.
it woke up after replacing the 3.3V voltage regulator.
i do not like that 749x boards need that thing as an external part.
during the process i removed the power caps and put one big KZN and two small PSG chemicons in place.
i replaced the signal input caps with 2.2uF film caps a had in my stock.
i don't know if that changes anything, but they fitted perfectly into the spaces and the stock caps were just big and had different prints on them....looked dubious all in all.
i removed the volume pot.
the board has a nice feature:
it holes prepared 'big' and 'small' pots.
so i was able to put an input-connector across the pot-outputs.
now i can level both channels indepently.
i bought that board because of the easier to mod through hole components.
btw i tested it without the heatsink.
no problems for short runs.
maybe the big inductors do their job quite well....?
groe tda7498 100w+100w class d audio verstrker bord dc32v in groe tda7498 100w+100w class-d audio-verstrker bord dc32vArbeit: Klasse d Ruhestrom: 50mA Effizienz aus Verstrker auf AliExpress.com | Alibaba Group
it woke up after replacing the 3.3V voltage regulator.
i do not like that 749x boards need that thing as an external part.
during the process i removed the power caps and put one big KZN and two small PSG chemicons in place.
i replaced the signal input caps with 2.2uF film caps a had in my stock.
i don't know if that changes anything, but they fitted perfectly into the spaces and the stock caps were just big and had different prints on them....looked dubious all in all.
i removed the volume pot.
the board has a nice feature:
it holes prepared 'big' and 'small' pots.
so i was able to put an input-connector across the pot-outputs.
now i can level both channels indepently.
i bought that board because of the easier to mod through hole components.
btw i tested it without the heatsink.
no problems for short runs.
maybe the big inductors do their job quite well....?
groe tda7498 100w+100w class d audio verstrker bord dc32v in groe tda7498 100w+100w class-d audio-verstrker bord dc32vArbeit: Klasse d Ruhestrom: 50mA Effizienz aus Verstrker auf AliExpress.com | Alibaba Group
It does say "typical". But it also says " minimum 48K". And you say 30K SE, which isn't in revision 5 datasheet. This might cause confusion in peoples choice of input caps!!!
I am a newbie. Really don't understand what your getting at? How does the difference you've noted explain why the zobel/Boucherot cell doesn't belong there on BD modulation?
Does it? Single ended is signal and gnd, differential is positive signal, gnd, negative signal. Picture that as 3 parallel lines. Inputimpedance differential input STM is 60k, that would be positive to gnd 30k, gnd to negative 30k, total across positive and negative 60k. That would give 24k to 30k inputimpedance SE, signal only sees half the differential input.
Sloa119b doesn't make it any clearer.
The way I see it.....
The 3116 bootstrap snubber filters out high frequencies to GND, which as the 119b points out prevents IC shutdown as well as improving sound and performanc3.
My suggested LC snubber does not pass anything to GND. It does not show any potential across it. The effect of a Boucherot cell across speaker terminals presents the speaker to amp outputs with constant resistance, which it does through controling the "potential storage of energy" within the voice coil, increasing response. For similar reasons the zobel is in the 3116 output filter, and helps control emi.
Placing the snubber across the P&N L inputs has nothing to do with removing frequencies as with bootstrap snubber. It is there to control the potential storage of energy within the L that magnetic flux can create, and works with current passing through coil in either direction.
Doing this stabilises the resistance created by the L so that the IC output legs don't see varying resistance.
The Boucherot cell in this instance couples the IC legs to L with constant resistance, and because the L is not "storing potential energy" it creates through inductance, it must pass energy straight through and out the other side.
Overall this all combines to increase LC response, just as it does when placed across speaker terminals.
The increased response is evident in the LTspice graph changes to phase lines.
Just try it on LTspice within the 3116 output circuit, you'll see what I mean.
The way I see it.....
The 3116 bootstrap snubber filters out high frequencies to GND, which as the 119b points out prevents IC shutdown as well as improving sound and performanc3.
My suggested LC snubber does not pass anything to GND. It does not show any potential across it. The effect of a Boucherot cell across speaker terminals presents the speaker to amp outputs with constant resistance, which it does through controling the "potential storage of energy" within the voice coil, increasing response. For similar reasons the zobel is in the 3116 output filter, and helps control emi.
Placing the snubber across the P&N L inputs has nothing to do with removing frequencies as with bootstrap snubber. It is there to control the potential storage of energy within the L that magnetic flux can create, and works with current passing through coil in either direction.
Doing this stabilises the resistance created by the L so that the IC output legs don't see varying resistance.
The Boucherot cell in this instance couples the IC legs to L with constant resistance, and because the L is not "storing potential energy" it creates through inductance, it must pass energy straight through and out the other side.
Overall this all combines to increase LC response, just as it does when placed across speaker terminals.
The increased response is evident in the LTspice graph changes to phase lines.
Just try it on LTspice within the 3116 output circuit, you'll see what I mean.
Oh, no
ebay.
aliexpress is much better when those kind of problems occur.
got no refund from the ebay seller.