Very low MID-bass in TDA7293

okay guys,
i did tda on a veroboard today. there is less than 20mv DC in the output. i used the circuit picked up from Rod's site, with minor modifications

1)added 2uf mkp input capacitor as he doest have one, which is necessary.:D
On the basis of following page, made other few changes
Avoiding TDA7293 pitfalls
2)changed the bootstrap capacitor (as suggested in pitfalls of tda) to 100uf
3)changed the NFB cap to 100uf.

NOTE-the decoupling caps are smaller in comparison to datasheet on the rod eliotts site, and i have used ceramic 100nf ones alongwith them.

The problem is, i did a side by side comparison to tda2050 on 18v rails. but! i am really disappointed by the bass tda7293 puts out on my speakers.
can someone lend me an insight?:shhh:
 

Attachments

  • p127-f1.gif
    p127-f1.gif
    25 KB · Views: 647
Moderator
Joined 2002
Paid Member
What were you expecting, and what are you used to?

18V rails is very low for the 7293/94 and in general the 20xx series requires much smaller power supplies, so it's not really surprising that you are not hearing what you want to.

Can you share the PSU schematic and some pictures?

In general increasing the bootstrap cap has its own consequences. I would try a smaller cap first, 22 to 47uF is a good starting point. The capacitor needs to be charged and discharged through the bootstrap pins, so a smaller cap would help. You can bootstrap it from the output with the consequence of slightly lower operating voltage limit - but that's 100V as opposed to 120V, so I wouldn't worry.

Bootstrapping (electronics) - Wikipedia, the free encyclopedia

Just switching a chip without considering the entire system is a bit like painting your car and complaining it won't go any faster.
 
i am going to have my transformer rewound so running it on 25-0-25 is going to take some time.
i am using 20000uf per rail,bypassed with 0.1uf polyester.

i compared with tda2050 not because i wantes more power and not getting it. its about side by side comparison of the sonics.
i want some more bass. i actually read somewhere on the forum that feedback cap has a corespoinding value to the feedback resistor. since i changed feedback cap without any calculation blindly based on the suggestions i believe there is a mistake there.
if anyone can help me find out what exactly is it?
thanks for your time,sangram ji. appreciated.
 
Moderator
Joined 2002
Paid Member
Feedback cap is calculated by f= 1/(2*pi*R*C) where f is usually about three octaves below the lowest frequency you want to get out of the system.

Optionally, eliminate that cap by using a servo. Cheap capacitors in that position seriously degrade the sound. So does a servo, but it is better than a cheap cap.

I doubt that is the cause of your woes. 22 to 100uF is a step in the right direction, but only if you used a good capacitor. What is available in the Indian markets is usually very poor quality, usually the wrong value and with high parasitic inductance. To get a slightly better result you can parallel that with a small 100nF polyester cap, that should take care of the problem to some extent.

You cannot compare side by side the way you are doing it. Starving a chip of power is not going to do it or you any favours. Optimise your power supply for the amplifier, and then compare.

Bass is a double whammy because speaker efficiency drops rapidly at low frequencies, and so does your ear's sensitivity to it. Power amplifiers have a much harder time creating a convincing bass presentation because of these reasons. 18V rails means you get about 13V into the load, a shade over 20 watts. At 20Hz, this is a molecule in the drop in the ocean.

Good Luck with your search.
 
Dear sangram ji,
Sorry for the late reply, i was doing just the same with the feedback cap. i have put a cap in parallel. the bass has improved a bit.
but there is a funny thing that i observed but i am not able to find theoretical meaning to it.
when i use the same transformer,speaker and heat sink, the tda2050 chip has heavier bass,better sonics,and heats up the transformer and heat sink a bit(shows the load is being drawn,dissipated) whereas when i use tda 7293, the treble is great but the transformer and heatsink runs cool and there is very less bass.
i understand these chips are ment to be near their voltages i am going to fix one of my transformers (or maybe just fix the secondary cables) to +-25v rated @ 4amps.
would it be sufficient for two channels, kindly consider that i am going to use 6ohm speakers with 20mf capacitance per rail?
 
Last edited:
Moderator
Joined 2002
Paid Member
The TDA20xx has bipolar output stage and the 72xx is nMOS stage on the output. The nMOS is slightly tougher to get right, as you will see by the number of blown products. Also not sure if you have genuine chips, the China effect is staggering.

I don't think too much of the ST chips. I have no respect for them and anyway they are all obsolete products. Heavy bass could simply be an audible effect of lower damping factor of the arguably much weaker TDA20xx output stages, or lower treble output. There's no way to tell without measurements or a good reference point. If you don't have either (for example, a decent commercial or DIY amp with proper specs and some measurements) there's no way to be sure what you're hearing.
 
okay guys,
i did tda on a veroboard today. there is less than 20mv DC in the output. i used the circuit picked up from Rod's site, with minor modifications

1)added 2uf mkp input capacitor as he doest have one, which is necessary.:D
On the basis of following page, made other few changes
Avoiding TDA7293 pitfalls
2)changed the bootstrap capacitor (as suggested in pitfalls of tda) to 100uf
3)changed the NFB cap to 100uf.

NOTE-the decoupling caps are smaller in comparison to datasheet on the rod eliotts site, and i have used ceramic 100nf ones alongwith them.

The problem is, i did a side by side comparison to tda2050 on 18v rails. but! i am really disappointed by the bass tda7293 puts out on my speakers.
can someone lend me an insight?:shhh:

Hi!!! Increase the input capacitor to 10uf.
 
i believe the chips are genuine, my uncle stocks them and he tells me beforehand i admit the china effect is staggering..even cheap crappy la4400 have duplicates. and the infamous sc5200/1943 toshiba pairs.

i havent done measurement on my amps as i am a beginner. i have very few export quality amps no wait.. exported amps original pioneer(onlyworks on 110volts) sx-2600 reciever and kenwood rse-7 amplifier.

i know these chips or receivers are no where near the audiophile grade, but we shouldnt forget we are living in india where most of the people dont even have bread to eat. these cheap chips and crappy components are the only way for the most of us to enjoy music(atleast students like me) no harm in making an amp, say for beginners, is it?
i am sure, these chips will atleast perform better than the la4400 chips that do duty on multimedia speakers and amps that are sold at great prices in the market.
thanks for the useful insight, really appreciate it
 
Member
Joined 2019
Paid Member
Hi
i found this

Why are there only disastrous PCB Versions for TDA7293/TDA7294/LM3886 etc. available?


chris

Chris: a lot of discussions about disastrous 7293 builds revolve around the not-so-great documentation sheet that exists for the TFA7293. Unlike the detailed doc for the LM3886 (along with solid schematics presented therein) the ones presented in the 7293 doc seemed not to have been rigorously tested. Take the boot loop capacitor value presented in the sheet (22uF?) turns out to be way too low and may cause oscillations (as far as I read) and everyone now just puts 100uF cap there. Also, people are confused over whether anything greater than 220uF caps on the boards are really safe to use (even the designers of the Xmas amp said 1000uF caps on the board caused oscillations way up in the extreme upper frequency range) but no one really knows if that’s (potentially) disastrous for our boards.

ClaudeG sent me a PM as we discussed why we can’t put at least 1000uF caps on the board and his take on viewing the disastrous history of builds with this chip was people just didn’t seem to know what they were doing and just thought they could handle the 7293 like all the other chipamp chips were in Vogue at the time. And how people didn’t know to insulate the back of the 7293 before they mounted it on the heatsinks is beyond me.

Had rigorous designer/builders adopted this chip we have lots more great builds these days but the LM1875 and LM3886 chip builds flourished as the circuitry design seemed to make them more robust (or the documentation for them was just superb). The 47 Labs Gaincard amp using the lm3886 is probably a big reason why the 7293 chip was left in the dust.

The Xmas amp is a nice, stable build.

Just my two cents.

Pete
 
i
i know these chips or receivers are no where near the audiophile grade, but we shouldnt forget we are living in india where most of the people dont even have bread to eat. these cheap chips and crappy components are the only way for the most of us to enjoy music(atleast students like me) no harm in making an amp, say for beginners, is it

Hi mate,
dont you think 10uf of audio grade capacitor will cost me more than the whole board :p ?

Then try scavenging from used equipment instead of trying to buy from your “local markets”. Things that used to work might have higher quality components in them than what they’re pushing on the street. I survived my lean college years by doing exactly that - I got more parts from junk than I ever purchased. And all of my design efforts centered around what it CAN GET, not buying specific stuff to follow someone’s exact plans or schematic. If you have limited options this is what is DO (and if you want to save 15% you switch to GEICO).
 
Chris: a....... Also, people are confused over whether anything greater than 220uF caps on the boards are really safe to use (even the designers of the Xmas amp said 1000uF caps on the board caused oscillations way up in the extreme upper frequency range) but no one really knows if that’s (potentially) disastrous for our boards.
...

Pete


Hi Pete


As i remember in the ebay LM1875 thread a lot of people wrote that 220µF are enough. i do not remember that somebody comment about oscillation. i remember that a lot of member wrote that the sound is much "better" = neutral with good quality , e.g. panasonic FC.
I am interested in why ? is it because of the ESR/ESL of the electrolytic cap?
would be a polymer better here?


chris
 
Member
Joined 2019
Paid Member
Hi Pete


As i remember in the ebay LM1875 thread a lot of people wrote that 220µF are enough. i do not remember that somebody comment about oscillation. i remember that a lot of member wrote that the sound is much "better" = neutral with good quality , e.g. panasonic FC.
I am interested in why ? is it because of the ESR/ESL of the electrolytic cap?
would be a polymer better here?


chris

Hi Chris: The 220uFs on the Xmas amp are certainly enough. There is no issue with getting a decent lower end response with the currently used Panasonic’s and, amazingly enough, putting the 100uF Silmic IIs on the negative rails (replacing two of the 220 Pannys) doesn’t seem to hurt the bass response any.

I’ve built several lm3886 gainclones in the past couple of years and have noticed that the use of the recommended 1000uF (or higher) seemed to muddy the sound more than help the bass response. Dropping to 470s or 220s made everything sound better and smoother. Granted most chipamp-type amplifiers suffered from the lack of solid bass (except maybe the My_Ref series from the late Mario P and now the Dario Fremen amp) using smaller caps led to a smoother sound. Someone, somewhere, stated that as audiophile electrolytic caps got bigger they get slower in response time (maybe lower ESR does play a role here) but until I started fooling around with Class D amps the Panny FCs were mostly recommended for placement around regulators and not as caps directly involved with signal transmission. Class D folk seem to embrace them however.

The use of polymers, at least for signal integrity/transmission, seems to suffer from a huge information void what I see. Any enlightening info you can pass my way I’d be grateful. Only place I’ve seen them used is in irs2092 amps you buy on EBay and then they usually are used for pre-amplification chip signal modification.

Cheers,

Pete