PEBKAC warning: When measuring the frequency vs. distortion, I had the THD+N box checked resulting in slightly higher distortion measurements ( my concern at 8:50 ). I can't perform reliable THD+N tests unless I shield the equipment so must stick with THD.
Both UTC and ST had same schematics in datasheets. It seems also JSMSemi is also manufacturing those. Maybe same layout or improved. Checked price from lcsc because didn't find from mouser or digikey anymore. Those were about 0.35 usd each in 5 or 10 quantity. Edit: my favorite chip TDA7294V is also on lcsc from ST only and around 2.34 usd each from 1+ quantity. (about 100W chip)
It's interesting to see the counterfeit audio amp IC performance mirroring my experience with counterfeit linear regulators. I recognize the wonky "ST" or "57" logo from counterfeit L7805CV regulators I keep encountering in broken devices and on the usual sites. From my experience and anecdotes from others, the fake L7805CV seems to perform perfectly fine at the median of the datasheet values, but runs way too hot and fails prematurely above 1A, often failing short circuit, while the genuine article is reliable at its rated 1.5A. The counterfeit also seems to have a worse dropout characteristic, becoming unstable around 8.4v input, while the genuine is close to 7v. Interestingly, others' testing seems to show the "ST"/"57" L7805CV could be a relabeling of a Chinese-market vendor's L7805 with a different suffix, one that has a datasheet indicating a 1A current limit.
I'm quite enjoying this series as I've been a big fan of the TDA2050 for some time now, though I use boards with blocking caps on the output that require only a single rail supply as a decent 24V power supply can be had quite cheaply. I'm curious as to the difference of performance between these two setups...
Hi! I would like to inform you that I work with one of the largest importers of electronic components in Argentina. They are so significant in the business that every year, the people from ST semiconductors pay for their plane tickets, along with his wife's, to attend the annual meeting in Silicon Valley. As for the TDA2050 and TDA2052, not all "fake" ones are actually fake. Many are manufactured by ST, but come out flawed. The wafers have some defects on the edges that lead to lower performance, overheating, or low gain and offset voltage. These are sold in bulk as "bulk trash" and are usually bought by some Chinese companies to finish encapsulating them, engrave their codes, and sell them. This means that many of these TDAs, which seem fake because they don't meet their nominal specifications, are actually "real" but defective products that ST would never sell to their loyal customers. Instead, they resell them to second-tier companies in China to offset manufacturing costs and reduce the price impact on the original ST semiconductor products.
The secret to use Chinese chips is to convert the values from Chinese units: 1 Chinese amp is 0.5 western ampere; 1 Chinese volt is 0.6 western volts; and 1 Chinese watt is 0.3 western watts. If you derate the components accordingly , you'll be fine. Also, short circuit protection means it will protect itself by burning on short circuit. 😂
Maybe the QA is measuring THD+N and displaying this on the graph but it's subtracting the noise from the THD and giving you a low number? I don't know, never owned or used a QA. On my stand alone distortion meters, noise can be eliminated with filters to give a more accurate THD measurement.
These ICs are really just an op-amp with chunky output transistors; and because they are _inside_ the feedback loop, any errors they introduce are minimised. The fact of everything being in good thermal contact will also be working in your favour.
Great video. Can the audio analyser output its data in CSV format or any other similar raw data (text) format so you could post it or look at it in Excel or some other analysis s/w? If that was spossible then you'd have a confidence check for what it's showing you on its plots..
Yes, it can save data. I'll need to research the issue further. All the other measurement "anomalies" were my fault, but this one seems like a bug. There are a lot of settings under various menus that must be made before setting up the sweep menu for the test.
@@JohnAudioTech You might try to manually measure THD around inflection point where distorsion rises and replot in excel/libre.calc? Also you might check ARTA or holm impulse...
I actually waited for this so bad (writing at uhh 0:00 ) :D i have built a stereo Fake tda 2050 on a diy pertinax board, and i think i would not gonna bombard it with +-20 something volts, only with +-15V :D I was like at the last video, i was like wait what, utc isnt counterfeit, then the test came similar with the real ST TDA2050, and yeah. Now its gonna be a fun video :DD Because i remember i had an UTC TDA2003 and it was a pretty good amplifier, it was also sounded pretty good, it died due the screw unscrewed itself somehow that i havent realized and it cooked itself unfortunatelly, but i loved that chip, it was the amplifier that i was growing up on. It was like uhh 5years ago, i was pretty much a child, wanted back my UTC2003 and i got fake tda2003, and they exploded after 5-10mins on 12volts. Yeah... I tried once i had a near 22Volt supply and the UTC TDA2003 worked so damn good. but when i tried my fake TDA2003 that i currently have, it started just to block off the output, but when i reaplied the 12Volts it worked again. Currently the tda2003 i have is some kind of "yw TDA2003" and also i tried that with only 19.5Volts instaed of the UTC that i was using a different supply, which was 22Volt DC. And it was rated for 10Watts, and 13.8Watts came out to 10% THD so i think that was pretty good, almost 14watts from a 9-10watts rated chip. and of course thats on 8ohms. Pretty good chip, its just the counterfeits of 2003 are junk.
The red line going back a little bit is just strange. I think there is some stability issue going on, maybe high frequency oscillation is starting at some point. Would be interesting to look at it with a oscilloscope to see what is going wrong. Of course the chip is not as good as the original one but usually those cheap PCBs layout and circuit are not great when it is about stability and
Could be the action of the limiter circuit at higher signal levels, but then it starts letting more power through as the signal increases more. I agree, the scope may reveal what's going on. Those boards have a nice ground layout and bypassing.
I have received an amplifier technics su- x 302 it is svi 3102 ic based amp but problem is that Ic is not available in my area so Iam thinking of modifing amp with 2 tda 2050 . Tda 2050 can take 50 volt in single supply. So I want to use +42 volt from dc section of to operate ic. I would like to know whether tda 2050 support dc 50 volt in single supply.
It should work, but I would rather use TDA7293. Those should handle 8 ohm speakers at +/- 42V. Eventhough I think japanese powersupply will sag anyhow so this 42V will drop quite a lot when loaded.
Even the very fake ones are actually good amplifiers. If you need more than 10 watts and want to keep em clean , slapping a voltage follower stage after em will improve the output power significantly. The holly grail is to have em bridged , and add output transistors. If its well set up , then you have a mighty decent amplifier.
Can someone suggest me the best chip for hifi sound, 2x15watt is ok, a fake TDA2050 do the job? Is there something better? Now I have TDA7297, I think midrange sound not good. Is the TDA2050(or TDA2030) be better than that? I saw a version TDA2030 with dc12V, and a another with ac dual12v, are they have sound diference?
Watching the graph line go backwards was a looney tunes moment. 😆 I wonder if the "fake" chip is a QA reject? Might explain why some of the numbers hit the mark. 🤷♂👍
It's not so much the specs on these counterfeits, of course they are going to closely get all the data properties some what correct, but if you want your viewers to really see what makes them generic you need to open them up. That's where you will see the difference in quality of the B/E leads from the chip showing how much thinner the leads are and how they are attached from the chip itself and how these knockoffs have it built together under the case with so much less quality compared to the legitimate transistors .
Do you think the factory does not sell these after sorting them ? 2050/2040/2030 the same chip probably, rejects of 2050 labeled 2040 and so on. Only faulty chips are thrown away.
@@modorangeorge4991 I think a lower speced chip may be mislabelled now and then and they will be "rejects" that can't be sold on the regular market however ebay sellers etc may scoop them up at bargain prices.
@@modorangeorge4991 Decapping and examining the die would reveal a different circuit. Lots of similarities, but the TDA2050 came out years after the other ones. The main differences are the higher supply voltage rating and current handling capacity of the 2050.
PEBKAC warning: When measuring the frequency vs. distortion, I had the THD+N box checked resulting in slightly higher distortion measurements ( my concern at 8:50 ). I can't perform reliable THD+N tests unless I shield the equipment so must stick with THD.
Hey John,
What mods did you do to the board to accommodate the TDA instead of the LM.
Thanks
@@MelbourneLife-xj2ee I changed the value of the resistor and capacitor that is connected from the output to ground (see the TDA2050 data sheet).
Both UTC and ST had same schematics in datasheets. It seems also JSMSemi is also manufacturing those. Maybe same layout or improved.
Checked price from lcsc because didn't find from mouser or digikey anymore. Those were about 0.35 usd each in 5 or 10 quantity.
Edit: my favorite chip TDA7294V is also on lcsc from ST only and around 2.34 usd each from 1+ quantity. (about 100W chip)
It's interesting to see the counterfeit audio amp IC performance mirroring my experience with counterfeit linear regulators. I recognize the wonky "ST" or "57" logo from counterfeit L7805CV regulators I keep encountering in broken devices and on the usual sites. From my experience and anecdotes from others, the fake L7805CV seems to perform perfectly fine at the median of the datasheet values, but runs way too hot and fails prematurely above 1A, often failing short circuit, while the genuine article is reliable at its rated 1.5A. The counterfeit also seems to have a worse dropout characteristic, becoming unstable around 8.4v input, while the genuine is close to 7v. Interestingly, others' testing seems to show the "ST"/"57" L7805CV could be a relabeling of a Chinese-market vendor's L7805 with a different suffix, one that has a datasheet indicating a 1A current limit.
Amazing they would bother counterfeiting a nearly 50 year old design given the chip die size would be fairly small and cheap to produce anyway.
Seemes that this fake TDA2050 is actually some TDA2030. This chip is meant to be run ca +/- 16V @ 4 ohms.
Sounds right, i thought the same...
These little ICs (even the fakes) are great for small projects that need a couple of watts of audio!
Thanks you so much for these tests , they are truely very informative and very helpful,🥰🥰😊😊
Now you have a new suscriber due to add multilanguaje audio tracks.
The TDA2050 is the GOAT. Have a Happy New Years John.
Happy New Year to you as well! Another year down already. Life is like a roll of toilet paper. The closer you get to the end, the faster it goes.
@@JohnAudioTech yeah 2025 sounds like science fiction. Where's my flying car?
@@JohnAudioTech LOL ! John !
@@JohnAudioTech And the less shit you can take
GOAT? It's a BISON!
Thanks 👍
Very nice video
I'm quite enjoying this series as I've been a big fan of the TDA2050 for some time now, though I use boards with blocking caps on the output that require only a single rail supply as a decent 24V power supply can be had quite cheaply. I'm curious as to the difference of performance between these two setups...
Hi! I would like to inform you that I work with one of the largest importers of electronic components in Argentina. They are so significant in the business that every year, the people from ST semiconductors pay for their plane tickets, along with his wife's, to attend the annual meeting in Silicon Valley. As for the TDA2050 and TDA2052, not all "fake" ones are actually fake. Many are manufactured by ST, but come out flawed. The wafers have some defects on the edges that lead to lower performance, overheating, or low gain and offset voltage. These are sold in bulk as "bulk trash" and are usually bought by some Chinese companies to finish encapsulating them, engrave their codes, and sell them. This means that many of these TDAs, which seem fake because they don't meet their nominal specifications, are actually "real" but defective products that ST would never sell to their loyal customers. Instead, they resell them to second-tier companies in China to offset manufacturing costs and reduce the price impact on the original ST semiconductor products.
Thanks John😊
The secret to use Chinese chips is to convert the values from Chinese units: 1 Chinese amp is 0.5 western ampere; 1 Chinese volt is 0.6 western volts; and 1 Chinese watt is 0.3 western watts. If you derate the components accordingly , you'll be fine.
Also, short circuit protection means it will protect itself by burning on short circuit. 😂
Maybe the QA is measuring THD+N and displaying this on the graph but it's subtracting the noise from the THD and giving you a low number?
I don't know, never owned or used a QA. On my stand alone distortion meters, noise can be eliminated with filters to give a more accurate THD measurement.
It has a separate mode for THD+N. It displays the data as it takes measurements. I'll have to investigate more.
You are right! Went back to the menu for the sweep and the THD+N box was checked.
@@JohnAudioTech Well that's odd ! but you found it ! I look forward to the next vid (as always !) Happy New Year !
Maybe sweeping the frequency on the distortion plot has an effect on distortion, as in a stable sine wave causes it to settle down.
That's pretty surprising performance for such a cheap counterfeit amp. Thanks John.
Yes, the distortion performance was surprising.
These ICs are really just an op-amp with chunky output transistors; and because they are _inside_ the feedback loop, any errors they introduce are minimised. The fact of everything being in good thermal contact will also be working in your favour.
Who writes 10 as 1E+01. silly software devs not paying attention?
@@Krmpfpks Same same
Could it be that fake chips are TDA2030 marked as TDA2050? In any case fake chip is very good for projects at lower PSU voltages.
could be, though they would also take any random TO220-5 packaged device and remark it as well.
Great video.
Can the audio analyser output its data in CSV format or any other similar raw data (text) format so you could post it or look at it in Excel or some other analysis s/w?
If that was spossible then you'd have a confidence check for what it's showing you on its plots..
Yes, it can save data. I'll need to research the issue further. All the other measurement "anomalies" were my fault, but this one seems like a bug. There are a lot of settings under various menus that must be made before setting up the sweep menu for the test.
@@JohnAudioTech You might try to manually measure THD around inflection point where distorsion rises and replot in excel/libre.calc? Also you might check ARTA or holm impulse...
Simple screw up. I had the THD-N box checked. I can't perform a proper THD+noise test without good shielding.
Interesting! Whats the analyzer you're using?
I am building 4 channel test amp with UTC TDA2050L chips. I would use it for measuring speaker drivers and active crossover tuning experiments.
They are not bad anyways. I usually use them under +-20v at 8 ohms. Not that bad anyways.
Thanks so much John!!
I actually waited for this so bad (writing at uhh 0:00 ) :D i have built a stereo Fake tda 2050 on a diy pertinax board, and i think i would not gonna bombard it with +-20 something volts, only with +-15V :D
I was like at the last video, i was like wait what, utc isnt counterfeit, then the test came similar with the real ST TDA2050, and yeah.
Now its gonna be a fun video :DD
Because i remember i had an UTC TDA2003 and it was a pretty good amplifier, it was also sounded pretty good, it died due the screw unscrewed itself somehow that i havent realized and it cooked itself unfortunatelly, but i loved that chip, it was the amplifier that i was growing up on.
It was like uhh 5years ago, i was pretty much a child, wanted back my UTC2003 and i got fake tda2003, and they exploded after 5-10mins on 12volts. Yeah...
I tried once i had a near 22Volt supply and the UTC TDA2003 worked so damn good. but when i tried my fake TDA2003 that i currently have, it started just to block off the output, but when i reaplied the 12Volts it worked again. Currently the tda2003 i have is some kind of "yw TDA2003" and also i tried that with only 19.5Volts instaed of the UTC that i was using a different supply, which was 22Volt DC. And it was rated for 10Watts, and 13.8Watts came out to 10% THD so i think that was pretty good, almost 14watts from a 9-10watts rated chip.
and of course thats on 8ohms.
Pretty good chip, its just the counterfeits of 2003 are junk.
Unisonic still produces all the goodies from the past - even TDA7375. I used their TDA2822 chip to good results.
Some old design car amplifier IC-s has a same bumped distortion cruve at low rfeq.
The case and legs are looks similar with an TDA2003.
Have you ever tested the aca amp ? Its a class amp ,sold as pcb or as a kit.
The red line going back a little bit is just strange. I think there is some stability issue going on, maybe high frequency oscillation is starting at some point. Would be interesting to look at it with a oscilloscope to see what is going wrong. Of course the chip is not as good as the original one but usually those cheap PCBs layout and circuit are not great when it is about stability and
Could be the action of the limiter circuit at higher signal levels, but then it starts letting more power through as the signal increases more. I agree, the scope may reveal what's going on. Those boards have a nice ground layout and bypassing.
Hey John, i am a kid and it it hard for me to get ic but my unc game me alot of tda2050 with a brand called HLF. Is that brand ok or its just very bad
I have no experience with that brand. I would try to get the UTC ones.
It seems even more better than lm1875
Better distortion but the fake can't handle the current.
It's such a simple circuit, it's probably harder to do a bad job than a good one.
I have received an amplifier technics su- x 302 it is svi 3102 ic based amp but problem is that Ic is not available in my area so Iam thinking of modifing amp with 2 tda 2050 . Tda 2050 can take 50 volt in single supply. So I want to use +42 volt from dc section of to operate ic. I would like to know whether tda 2050 support dc 50 volt in single supply.
It should work, but I would rather use TDA7293. Those should handle 8 ohm speakers at +/- 42V. Eventhough I think japanese powersupply will sag anyhow so this 42V will drop quite a lot when loaded.
Max voltage across the supply pins is 50v so you are good in single supply mode or +/- 25v with a split supply.
@@Tegelane5 Thank you
@@JohnAudioTech Thank you 😊
8:53 maybe it does some integration when it plots.
I later found I had the THD+N box checked.
Even the very fake ones are actually good amplifiers. If you need more than 10 watts and want to keep em clean , slapping a voltage follower stage after em will improve the output power significantly.
The holly grail is to have em bridged , and add output transistors. If its well set up , then you have a mighty decent amplifier.
Can you stack these chips to get pro output power e.g. 500 to 1000 Watts per channel continuous?
Not high enough voltage capability
UTC 2050L is good. I use them.
Smaller die has less parasitic capacitance, maybe that is why good THD, but also naturally wimpy for power especially at 4Ω
Can someone suggest me the best chip for hifi sound, 2x15watt is ok, a fake TDA2050 do the job? Is there something better? Now I have TDA7297, I think midrange sound not good. Is the TDA2050(or TDA2030) be better than that? I saw a version TDA2030 with dc12V, and a another with ac dual12v, are they have sound diference?
Assuming its implemented correctly, they are all going to sound the same.
Watching the graph line go backwards was a looney tunes moment. 😆 I wonder if the "fake" chip is a QA reject? Might explain why some of the numbers hit the mark. 🤷♂👍
It's not so much the specs on these counterfeits, of course they are going to closely get all the data properties some what correct, but if you want your viewers to really see what makes them generic you need to open them up. That's where you will see the difference in quality of the B/E leads from the chip showing how much thinner the leads are and how they are attached from the chip itself and how these knockoffs have it built together under the case with so much less quality compared to the legitimate transistors .
MAybe they are not fakes but rather factory rejects that don't quite meet the specs.
They are fakes. Decapping the IC reveals a smaller die inside.
Do you think the factory does not sell these after sorting them ? 2050/2040/2030 the same chip probably, rejects of 2050 labeled 2040 and so on. Only faulty chips are thrown away.
@@modorangeorge4991 I think a lower speced chip may be mislabelled now and then and they will be "rejects" that can't be sold on the regular market however ebay sellers etc may scoop them up at bargain prices.
@@modorangeorge4991 Decapping and examining the die would reveal a different circuit. Lots of similarities, but the TDA2050 came out years after the other ones. The main differences are the higher supply voltage rating and current handling capacity of the 2050.
Good ole ebay mystery meat.