Having been a big TPA3316D2 user and fan for over 2 years now, I am just starting to use the very low-cost TDA7492 amp boards available from China. I recently bought a pair from Aliexpress for about $8 ea shipping included. They arrived in a few weeks and I have since set them up, and even started modifying them in a way similar to the TPA3116D2 with regards to power rail caps and snubbers. I think this amp deserves its own thread to keep the 3116 thread on topic...
Datasheet here: http://www.st.com/web/en/catalog/sense_power/FM125/CL1503/SC983/PF214773
For the purposes of comparison of sound quality between the TPA3116D2 and the TDA7492's (both amps have same mods), I am using rather nice speakers that I just finished building - the ScanSpeak 10F/8424 & Dayton RS225-8 Ref Monitor shown here:
This is the current monitor's measured response and acoustic XO:
Group delay for this monitor (sealed Limkwitz transform bass) is +/- 2.5ms above 33Hz:
Here are the amps now connected to my miniDSP powering my ScanSpeak 10F/8424 & Dayton RS225-8 FAST reference speakers (http://www.diyaudio.com/forums/full-range/273524-10f-8424-rs225-8-fast-ref-monitor.html):
Having shown you the performance of the reference monitors, I can say that using these speakers, compared to the 3116, these amps sounded OK stock, but seemed a bit harsh and edgy on the top end, and lacked some forcefulness in the bass department. So I proceeded to add a pair of Panasonic OSCON 330uF 25V SEPF caps and a Panasonic 560uF 35V FM cap to the single stock 2200uF 25V cap that was there. I also added 330pF and 10R snubbers at the inputs of the inductors. With the mods in place, this amp sounds very very good! I think it has better high frequency resolution that the 3116, and the extra snubber smoothed out the harshness. The OSCON upgrade always helps - a lot better bass authority and the sound stage and imaging improved.
Here is how I added 330pF and 10R snubbers to the input side of the inductors, basically same location where the "bootstrap" snubbers would go on the 3116. Note that the TDA7492 does not have external bootstap caps.
I pulled the massive heatsink off of the board to reveal a Feixiang pcb - here is a high res layout photo. It looks like the Feixiang board pretty much follows the suggested circuit but as you can see from the photo, they added some stuff like a protection diode on Vcc, and a pair of small SMT by-pass caps after the electrolytics very close to the pins for the best noise suppression - a nice touch. Input and output filter caps are EPCOS films. The input caps have been upgraded from factory suggested 470nF to 1uF EPCOS film caps. The output filters seem to be using a compromise arrangement of 22uH and 220nF and 100nF film caps. Usually this would be with 33uH inductors. :
The amps run warmer than the 3116, but one of my amps runs hotter than the other and one of the inductors (far right) runs hot to the touch which is why I am looking at it closely to see what the problem is. It looks like the ferrite shield casing is chipped or cracked.
Here is the datasheet suggested schematic:
Here are the datasheet pinouts:
You can get these for circa $7 shipping included:
http://www.aliexpress.com/item/2015-Digital-Amplifier-AMP-Board-TDA7492-High-Power-Powerful-with-Radiator-Hot/32331617990.html
An externally hosted image should be here but it was not working when we last tested it.
Datasheet here: http://www.st.com/web/en/catalog/sense_power/FM125/CL1503/SC983/PF214773
For the purposes of comparison of sound quality between the TPA3116D2 and the TDA7492's (both amps have same mods), I am using rather nice speakers that I just finished building - the ScanSpeak 10F/8424 & Dayton RS225-8 Ref Monitor shown here:
This is the current monitor's measured response and acoustic XO:
Group delay for this monitor (sealed Limkwitz transform bass) is +/- 2.5ms above 33Hz:
Here are the amps now connected to my miniDSP powering my ScanSpeak 10F/8424 & Dayton RS225-8 FAST reference speakers (http://www.diyaudio.com/forums/full-range/273524-10f-8424-rs225-8-fast-ref-monitor.html):
Having shown you the performance of the reference monitors, I can say that using these speakers, compared to the 3116, these amps sounded OK stock, but seemed a bit harsh and edgy on the top end, and lacked some forcefulness in the bass department. So I proceeded to add a pair of Panasonic OSCON 330uF 25V SEPF caps and a Panasonic 560uF 35V FM cap to the single stock 2200uF 25V cap that was there. I also added 330pF and 10R snubbers at the inputs of the inductors. With the mods in place, this amp sounds very very good! I think it has better high frequency resolution that the 3116, and the extra snubber smoothed out the harshness. The OSCON upgrade always helps - a lot better bass authority and the sound stage and imaging improved.
Here is how I added 330pF and 10R snubbers to the input side of the inductors, basically same location where the "bootstrap" snubbers would go on the 3116. Note that the TDA7492 does not have external bootstap caps.
I pulled the massive heatsink off of the board to reveal a Feixiang pcb - here is a high res layout photo. It looks like the Feixiang board pretty much follows the suggested circuit but as you can see from the photo, they added some stuff like a protection diode on Vcc, and a pair of small SMT by-pass caps after the electrolytics very close to the pins for the best noise suppression - a nice touch. Input and output filter caps are EPCOS films. The input caps have been upgraded from factory suggested 470nF to 1uF EPCOS film caps. The output filters seem to be using a compromise arrangement of 22uH and 220nF and 100nF film caps. Usually this would be with 33uH inductors. :
The amps run warmer than the 3116, but one of my amps runs hotter than the other and one of the inductors (far right) runs hot to the touch which is why I am looking at it closely to see what the problem is. It looks like the ferrite shield casing is chipped or cracked.
Here is the datasheet suggested schematic:
Here are the datasheet pinouts:
You can get these for circa $7 shipping included:
http://www.aliexpress.com/item/2015-Digital-Amplifier-AMP-Board-TDA7492-High-Power-Powerful-with-Radiator-Hot/32331617990.html
Attachments
Last edited:
Ok datasheet shows it is a little weaker than TI tpa3116, distortion sees steep rise above 20watt, TI above 30watt, so distortion is probably what makes the 7492 appear more powerfull. The output filterfrequency is much closer to audioband so peaking is more audible with 7492, I think you already found that. The filter used is in datasheet for 6 ohm, your SS10F monitor probably above that. Datasheet 7498 shows a little more advanced outputfilter with more damping of peaking probably.
I do not see higher distortion at the speaker frequency response and HD measurement. Speakers are generally limiting factor but HD seems comparable at the acoustic level to 3116.
Interesting that distortion makes it sound more powerful? The gain is higher than 26dB so things are just louder by about 6dB or so. However, turning input gain on miniDSP down by -6dB it still sounds very dynamic.
For $6 to $8 though, they are an absolute deal. How can you lose?
Why do they run hotter though even though same 90% quoted efficiency?
i subscribe this thread and suggest anyone who likes that board/chip to try its bigger brother the 7498.
the black 7498 board is a steal for only 10€ - 11€ shipped from china.
what i miss on those boards is the possibility to easily mod it to differential inputs.
would be super easy with only throgh hole caps on the negative inputs, too.
the black 7498 board is a steal for only 10€ - 11€ shipped from china.
what i miss on those boards is the possibility to easily mod it to differential inputs.
would be super easy with only throgh hole caps on the negative inputs, too.
The snubber used on the input of the inductor for the TPA3116 was 330pF and 10R. I just used same one because that was what I had on hand. I don't even know if it is right for the 7492. It does seem to have similar values for the inductor and output capacitor values.
Do you see 22R used somewhere? Oh I see what you are saying as there is a 330pF and 22R connecting the two inductor inputs together. I don't know if that is a snubber as it doesn't go to ground - just between the positive and negative legs.
Do you see 22R used somewhere? Oh I see what you are saying as there is a 330pF and 22R connecting the two inductor inputs together. I don't know if that is a snubber as it doesn't go to ground - just between the positive and negative legs.
Last edited:
I doubt that higher distortion on a sinewave is the reason that an amp sounds more powerful. In my experience its dynamics that are associated with the perception of more power - so I'd reckon that the TDA has a more constant noise floor, less dynamic variation with signal.
I've not tried this chip yet but its possible it shares some heritage with NXP's classD offerings and the TDA8932 was definitely subjectively more dynamic in the bass to my ears than the TPA3116.
If the chip's running hot, I'd firstly look at the output inductors - perhaps they're too lossy at the switching frequency. Certainly they look rather small ones.....
Last edited:
Yes. You are right. This part is absent in the TPS3116 schematic. In the TPA3116, there are pins from the chip (BSNX) that are specifically for connecting to a snubber, that I do not see in the TPA chip. Anyway, comparing the schematics side by side, it seems that the out filter sections of the two chips are different. I need to look up the TI paper on Class D amp output filter design to get more details.
Regards,
Sound clip with 10F ref speaker
http://www.diyaudio.com/forums/atta...5-8-fast-ref-monitor-10f-fast-7492-amp-01.asc
Change extension to .mp3 to listen.
http://www.diyaudio.com/forums/atta...5-8-fast-ref-monitor-10f-fast-7492-amp-01.asc
Change extension to .mp3 to listen.
i found a way to solder my favorit connector to the blue 7492.
and i changed the power cap for a chemicon kzn 2700uF / 35V.
it only fitted from the bottom side.
the three holes are a nice feature for different lead spacings.
maybe i'll add a polymer cap from the topside if i change back to lower voltage.
let's see how that thing acts with more voltage...
and i changed the power cap for a chemicon kzn 2700uF / 35V.
it only fitted from the bottom side.
the three holes are a nice feature for different lead spacings.
maybe i'll add a polymer cap from the topside if i change back to lower voltage.
let's see how that thing acts with more voltage...
I just ordered two of the 7498's. At $11 ea how can you not ? 
What are you using for power supply? They recommend 35v smps for $32 with shipping!
I will try two 19v bricks in series for 38v. Won't get the high amps but voltage is there for high peak transients.
19v x 4.65 amp smps bricks are $6 from Amazon.
I cut an old RCA cable in half and solder the shielded cable directly to the exposed pins on the bottom of the input connector. This way I keep it shielded to reduce noise pickup. The connector cable it came with has pin but is not shielded.
Look at bottom photo to see what I did.
What are you using for power supply? They recommend 35v smps for $32 with shipping!
I will try two 19v bricks in series for 38v. Won't get the high amps but voltage is there for high peak transients.
19v x 4.65 amp smps bricks are $6 from Amazon.
I cut an old RCA cable in half and solder the shielded cable directly to the exposed pins on the bottom of the input connector. This way I keep it shielded to reduce noise pickup. The connector cable it came with has pin but is not shielded.
Look at bottom photo to see what I did.
Last edited:
i use sla batteries for an outdoor party machinery.
i have those cheap 160watt voltage boost converters.
they work surprisingly good.
i have a little trouble with two cubo12 which are driven by one of the black boards.
the sound gets heavily distorted at higher volumes.
i want to check if i can get rid of that with higher voltages.
cubos are better tested outside.
those things so loud just before the distortion starts, it's too much for an indoor test.
...i just finished university and now have six audio projects running at the same time.....
i have those cheap 160watt voltage boost converters.
they work surprisingly good.
i have a little trouble with two cubo12 which are driven by one of the black boards.
the sound gets heavily distorted at higher volumes.
i want to check if i can get rid of that with higher voltages.
cubos are better tested outside.
those things so loud just before the distortion starts, it's too much for an indoor test.
...i just finished university and now have six audio projects running at the same time.....
i have that psu, too, and is a satisfying deal.
but the batteries are just better.
24V DC 6.5A 150W Schaltnetzteil Treiber für LED Streifen Strip Netzteil Neu | eBay
Yes for fun measure SPL between the 100watt 7498 and tiny Sure3110 with low 4 ohm load, zobelled fullrange "rated 4 ohm" chassis (I have small TangBand I use as TVspeakers, a little brighter than TC9 Vifa's).
I just ordered two of the 7498's. At $11 ea how can you not ?
What are you using for power supply? They recommend 35v smps for $32 with shipping!
I found this one... 160w + 160w. It's about $60 but comes with power brick, and case.
I'm thinking I could power up a couple of subs to go with my TPA3116D2 50w + 50w.
ÅS¤Ñ©ç½æ-¥xÆW NO.1 ©ç½æºô¯¸
(there's some English on that page)