Optimizing TDA7294 Output

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Well, amplifiers can sound different from one another; however, the adjustable schematic at the bottom of post#474 can be set for system symmetry or you could dial it in to imitate a different amplifier if you wanted to.

The parallel amp can use a larger input cap and should use a larger bootstrap cap; but otherwise, the solo and parallel (master chip) can use that same schematic.

I'm actually removing LM3886's from my integrated receiver and installing Parallel TDA7293 boards. Indeed the LM3886's are clearer at the cost of more forwards (great for analytics, voicecomm and movie soundtracks) and have the name brand hard clipper, SPIKE system noise as well. Personally, I don't need those television amplifier features in my integrated receiver. And, I believe that the Parallel TDA7293 boards are more significant for rocking the house with music, because that amp can run to the rails and maintain a palatable tone in all conditions. I think that since the usage is so much different, the LM3886 and TDA7293 are not really 1:1 comparable. The ST part, on usage, is most likely to direct-replace STK chips and inexpensive tube amplifiers, for music replay, but definitely not for the clearest voice track on your television. Natsemi markets the LM3886 for television movie soundtracks and they were wise to do so because it excels at replaying extra intelligible voicecomm (at some cost to tone).

If we want both of the above features simultaneously from 1 chip amplifier, the makers of LM3886 (National Semiconductor division of Texas Instruments) suggests that we upgrade to LME49830 if music replay is desired. However, ST suggests that we take the Parallel TDA7293 board, set BOTH chips to slave mode for zero gain (input is then pin#11, and the feedback loop is relocated) and drive these rather inexpensive lateral fet packs ($1.90 ea @10 units each of which contain 2 latfets) with a replacement voltage amplifier stage, which can be made of tubes like Akido, or BJT, like the SSA, or small signal chips.
Natsemi/TI makes highly favorable made-for-music (awesome "made for using" devices that are far better quality than "made for selling" devices) LME chip amplifiers, and ST makes inexpensive latfet packs that are unlikely to have an availability shortage. So, of course I'm curious if a "shotgun wedding" can be arranged?

Instead of a have your cake and eat it too scenario, what I currently have is some LM1875's for analytical purposes and relaxed SPL listening at superb quality. Although they may be frail, the LM1875's are getting the job done for my analytic purposes.

After the LM3886's are evicted from the receiver, I may attempt a fine tuning on those because they are indeed capable of superior imaging. They can beat the LM1875's on imaging but not on tone (and the basic "resolution" appears to be nearly break even). In that comparison, the LM1875's are more suited to mixer desk and near field needs, but the LM3886's are certainly better at distance. However, that receiver has been unplugged for some years on account of unfriendly tone when cranked up loud with the LM3886's. This has made the volume control some odd 75% useless, since it is fantastic at low volume but almost broken bone painful when turned up loud. Well, of course, I'm absolutely certain that the TDA7293 Parallel boards can beat the LM3886's in that one performance aspect (when small venue and high SPL meet, aka "rockin the house").

SO, to try for an answer to your question:
The ideal usage for the TDA7293 versus the LM3886 varies so greatly that it is like apples and oranges. It would be possible to say which is better if given a specific application, but otherwise they're impossible to compare.
On the second part of your post, most chip amplifiers show exactly the same 22k vs 1k or 22k vs 680R with some odd random caps for generic datasheet schematic. . . despite the fact that the chips differ in internal design. Be careful! That sort of generic schematic is for showing where the parts should go. But the component values are nonspecific/generic, and that may cause overheat, sound like a caffeinated Basset Hound, install a severe bass blocker, cause audible distortion or even cause breakage. I would not be able to explain this thoroughly, since I'm not an engineer, so here's Eddie on the topic of changing the generic schematic to something more specific that supports the chips:
Well I got it out because I intend to add a preamp (tone controls etc) to it so it is a truly integrated amp. . . and found out that I have TDA7296 chips running off a 62 volt split supply.[a 31+31vdc split rail supply]. . .I used the same board that was in the "no-fi" tabletop stereo. I removed the ripple caps (2200uF per rail is all it had) and installed 330 uF caps on the board. I used a capacitor bank like Daniel's with 4x2200 uF per rail and it made a noticeable improvement. Then I swapped the bootstrap and feedback caps (it used the data sheet 22 uF caps with smaller feedback resistors - really terrible!) with 100 uF caps and it made a huge difference.. . .It works really great. . .I am truly inspired that such performance is available from such a generic chip by optimizing design and layout. . . .
Unabridged version is at post#354
Solo amp bootstrap cap range is 33u~47u.
Parallel amp bootstrap cap range is 68u~100u.
See also AndrewT's posts on the topic of NFB-Shunt cap sizing for clear quality bass.
Future plans for the TDA7293 Parallel involve begging Kean for some help on improving the imaging (indeed he can do that) because I'm a bit greedy when it comes to imaging; but, otherwise, and except for the need of clean power which was not difficult, the amp is excellent.

P.S.
That's about all the typing I'm good for today. My fingers are sore from the "house brand" of HTML and trying to outdo Google, which was, apparently, possible, albeit some effort to do it.
 
Last edited:
That adjustable schematic at the bottom of post#474 has turned out to be really quite pleasant. Just now, I tried it with that rather bad single layer board, and it worked appreciably (at somewhat near 40x before the darned thing was practical). Of course, I'd much prefer to use the TDA7293 Parallel dual layer compact board; however, the single layer solo TDA7294 board was more educational in the form of a challenge.
P.S.
I'm pretty much in "on break" status, except for refitting that one receiver for the purpose of rocking the house.
 
Let us know when you get it hooked up. I'm getting too many TDA* based amps around here myself - with too little time to work on them. :( Plus a move to CT coming up this Summer. Might grab a couple of these for fun!
I put together a ps with a 28-0-28 transformer I salvaged from and old SS amp and connected up the 7294. Absolutely dead silent.
I cannot detect any sonic advantage over the 7297 I was using except for it's annoying blue led.
I tried several speakers I have laying around.
RS Minimus 8, RS minimus 7, PSB 30 MKII (which I recently refoamed)
Dual CL172 (placed among the top in huge 1976 speaker soundoff in Japan) and my regular Tannoys.
All the speakers are getting on in years, if that makes a difference? but, to me, the bigger the speaker the better the sound.
I've switched back to the 7297, it runs cooler with just an excuse for a heatsink and ps.
Not being compentent enough to be modder, I wonder if changing a few periferal components on these little amps makes the kind of difference that speakers can make?
 

Attachments

  • BRIDGE TDA7294 SD 400%.JPG
    BRIDGE TDA7294 SD 400%.JPG
    254.2 KB · Views: 1,465
  • TDA7294 BRIDGE 400%.JPG
    TDA7294 BRIDGE 400%.JPG
    331 KB · Views: 1,451
What you say about this design?

APEXAUDIO

nice !!! I like the elco arrangement, very neat !

This is mine, a 3 channel amp using 3886 for high/mid and a 7293 for the unabridged lows :D as the 3886 spike doesn't like my woofer :rolleyes: It's based on a 100x80 mm doubleside board, is still under construction because next to a 3ch substractive filter, limiter and dc protection I also like to add some automatic on/off logic.

attachment.php


But the get back on topic : I've seen 7293 designs where the driver and final stage rails were separated using diodes, anybody anything to say on that ???
 

Attachments

  • 2013.01.jpg
    2013.01.jpg
    208.4 KB · Views: 3,113
Last edited:
But the get back on topic : I've seen 7293 designs where the driver and final stage rails were separated using diodes, anybody anything to say on that ???

This has been brought up so many times. There are also several replies on older threads. You should search for TDA7294/7293 and look for threads a few years old ! It was explained beautifully by ilimzn !
Search for a thread called 'TDA7293v' June 2002 ! Ilimzn posted sometime in April 2006 !
 
Well, amplifiers can sound different from one another; however, the adjustable schematic at the bottom of post#474 can be set for system symmetry or you could dial it in to imitate a different amplifier if you wanted to.

The parallel amp can use a larger input cap and should use a larger bootstrap cap; but otherwise, the solo and parallel (master chip) can use that same schematic.

I'm actually removing LM3886's from my integrated receiver and installing Parallel TDA7293 boards. Indeed the LM3886's are clearer at the cost of more forwards (great for analytics, voicecomm and movie soundtracks) and have the name brand hard clipper, SPIKE system noise as well. Personally, I don't need those television amplifier features in my integrated receiver. And, I believe that the Parallel TDA7293 boards are more significant for rocking the house with music, because that amp can run to the rails and maintain a palatable tone in all conditions. I think that since the usage is so much different, the LM3886 and TDA7293 are not really 1:1 comparable. The ST part, on usage, is most likely to direct-replace STK chips and inexpensive tube amplifiers, for music replay, but definitely not for the clearest voice track on your television. Natsemi markets the LM3886 for television movie soundtracks and they were wise to do so because it excels at replaying extra intelligible voicecomm (at some cost to tone).

If we want both of the above features simultaneously from 1 chip amplifier, the makers of LM3886 (National Semiconductor division of Texas Instruments) suggests that we upgrade to LME49830 if music replay is desired. However, ST suggests that we take the Parallel TDA7293 board, set BOTH chips to slave mode for zero gain (input is then pin#11, and the feedback loop is relocated) and drive these rather inexpensive lateral fet packs ($1.90 ea @10 units each of which contain 2 latfets) with a replacement voltage amplifier stage, which can be made of tubes like Akido, or BJT, like the SSA, or small signal chips.
Natsemi/TI makes highly favorable made-for-music (awesome "made for using" devices that are far better quality than "made for selling" devices) LME chip amplifiers, and ST makes inexpensive latfet packs that are unlikely to have an availability shortage. So, of course I'm curious if a "shotgun wedding" can be arranged?

Instead of a have your cake and eat it too scenario, what I currently have is some LM1875's for analytical purposes and relaxed SPL listening at superb quality. Although they may be frail, the LM1875's are getting the job done for my analytic purposes.

After the LM3886's are evicted from the receiver, I may attempt a fine tuning on those because they are indeed capable of superior imaging. They can beat the LM1875's on imaging but not on tone (and the basic "resolution" appears to be nearly break even). In that comparison, the LM1875's are more suited to mixer desk and near field needs, but the LM3886's are certainly better at distance. However, that receiver has been unplugged for some years on account of unfriendly tone when cranked up loud with the LM3886's. This has made the volume control some odd 75% useless, since it is fantastic at low volume but almost broken bone painful when turned up loud. Well, of course, I'm absolutely certain that the TDA7293 Parallel boards can beat the LM3886's in that one performance aspect (when small venue and high SPL meet, aka "rockin the house").

SO, to try for an answer to your question:
The ideal usage for the TDA7293 versus the LM3886 varies so greatly that it is like apples and oranges. It would be possible to say which is better if given a specific application, but otherwise they're impossible to compare.
On the second part of your post, most chip amplifiers show exactly the same 22k vs 1k or 22k vs 680R with some odd random caps for generic datasheet schematic. . . despite the fact that the chips differ in internal design. Be careful! That sort of generic schematic is for showing where the parts should go. But the component values are nonspecific/generic, and that may cause overheat, sound like a caffeinated Basset Hound, install a severe bass blocker, cause audible distortion or even cause breakage. I would not be able to explain this thoroughly, since I'm not an engineer, so here's Eddie on the topic of changing the generic schematic to something more specific that supports the chips:

Unabridged version is at post#354
Solo amp bootstrap cap range is 33u~47u.
Parallel amp bootstrap cap range is 68u~100u.
See also AndrewT's posts on the topic of NFB-Shunt cap sizing for clear quality bass.
Future plans for the TDA7293 Parallel involve begging Kean for some help on improving the imaging (indeed he can do that) because I'm a bit greedy when it comes to imaging; but, otherwise, and except for the need of clean power which was not difficult, the amp is excellent.

P.S.
That's about all the typing I'm good for today. My fingers are sore from the "house brand" of HTML and trying to outdo Google, which was, apparently, possible, albeit some effort to do it.

My question to this good eplanation are to read here (post #22):
http://www.diyaudio.com/forums/chip-amps/26661-lm3886-tda7293-tda1514-3.html#post3510969
thank you therefore.
 
Earlier I made an error when I said there was no need to remove the board from the Jim's Audio single layer TDA7294 kit. Indeed that was an error. The power, layout, of that particularly annoying board is quite bad sounding and I have not been able to remove the coloration. Well, not all of it. The best I could do with that board is reduce its contributions. But, that particular board does still manage to sing along with the music a little bit, and not correctly. The TDA7294 should not sound like that. Thus, I recommend to remove the board.

So, Bob, how do you feel about a point-to-point adventure?
Its either that or get the Parallel TDA7293 boards, which have only one coloration (easily fixed), at small signal, the routing of the nfb-shunt resistor, and that takes only about 2 minutes to fix, by sticking that ~730R resistor (or my resistor plus variable gain trimmer assembly) near the edge of the board, direct route from pin2 to the nfb-shunt cap. I like that one much better since the power layout doesn't make distortion. That is so much easier.
 
Last edited:
Any of you tried this excellent board?

TDA7293 X 4 BTL 350W Amplifer PCB Ultra Reliable | eBay

01ca3c62.jpg


Sounds really good. It is a parallel and bridged config. The sound will be influenced by the Opamp used but I use the OPA1612 (bipolar input) or OPA1642 (JFet input) which sound amazingly good!

You can also get rid of the opamp and use it as a stereo amp

I've built several of those and they're really good.

Ciao!
Do
 
Any of you tried this excellent board?

TDA7293 X 4 BTL 350W Amplifer PCB Ultra Reliable | eBay

01ca3c62.jpg


Sounds really good. It is a parallel and bridged config. The sound will be influenced by the Opamp used but I use the OPA1612 (bipolar input) or OPA1642 (JFet input) which sound amazingly good!

You can also get rid of the opamp and use it as a stereo amp

I've built several of those and they're really good.

Ciao!
Do
Gee thanks pinnocchio for pointing out to all of us another 7293 design! :D I'm up to my ears in 7293 variations and haven't had the time to start some of them, much less complete them. Too many projects.

I do recognize the branding of yuanjing though. I think he's got some other product designs out there too. Glad to hear this one is performing well for you. Unfortunately, my projects are on hold until I get moved to CT.

Thanks for sharing,
Rick

Edit: BTW, did you bought the full kit, or just the PCBs and supplied your own components? Did you do any parts swapping?
Edit 2: The eBay prices of these assembled amps seems to vary anywhere from $24USD to $42USD. That's a pretty wide swing, but shipping balances some of it out. I wonder which ones are more clones/fake than the others! :)
 
Last edited:
Gee thanks pinnocchio for pointing out to all of us another 7293 design! :D I'm up to my ears in 7293 variations and haven't had the time to start some of them, much less complete them. Too many projects.

I do recognize the branding of yuanjing though. I think he's got some other product designs out there too. Glad to hear this one is performing well for you. Unfortunately, my projects are on hold until I get moved to CT.

Thanks for sharing,
Rick

Edit: BTW, did you bought the full kit, or just the PCBs and supplied your own components? Did you do any parts swapping?
Edit 2: The eBay prices of these assembled amps seems to vary anywhere from $24USD to $42USD. That's a pretty wide swing, but shipping balances some of it out. I wonder which ones are more clones/fake than the others! :)

I actually bought both the PCB versions and the fully assembled modules all from Jims Audio. Unfortunately he does not have any left PCB of the stuffed version and his were close to 70$ each 2-3 years ago and sounded really good and I'm sure were genuine components. On the assembled ones the only thing I swapped was the Opamp since the NE5532 does not do it for me... I would not buy this assembled amp from another reseller at 24$-42$... Those are most probably fakes or cheap/fake components.

The PCB board I stuffed myself using high quality components sounded a little bit better but Jims assembled boards had some pretty good components on it already except resistors and some of the film caps.

Ciao!
Do
 
I actually bought both the PCB versions and the fully assembled modules all from Jims Audio. Unfortunately he does not have any left PCB of the stuffed version and his were close to 70$ each 2-3 years ago and sounded really good and I'm sure were genuine components. On the assembled ones the only thing I swapped was the Opamp since the NE5532 does not do it for me... I would not buy this assembled amp from another reseller at 24$-42$... Those are most probably fakes or cheap/fake components.

The PCB board I stuffed myself using high quality components sounded a little bit better but Jims assembled boards had some pretty good components on it already except resistors and some of the film caps.

Ciao!
Do
Hey, that's good to know info. I do have a China vendor that I've purchased from before, and generally find most of his assembled modules fairly good. His board is assembled and looks like it has genuine parts - but one never knows. The PCB is identical with the ground plane and all. His price is at the higher end at $42USD.

The board doesn't look like it contains any exotic parts and fairly easy to source most stuff. I wonder about the 7293 chip itself though. Are those all fakes and clones now too? Are some better than others? At least Jim's Audio has the bare PCBs available. Maybe I'll go that root for this build should I decide to do it. Thanks again. Rick
 
Dual mono TDA7293 parallel kit

See the modern, modular style paralleling without the ballast resistors?
An externally hosted image should be here but it was not working when we last tested it.

Favorite new style design: 2pack TDA7293 Parallel 170W Mono Amplifier Board Kits 29 | eBay
With the two pack of compact dual-layer boards in the link (ebay# 190862927141), it is possible to do higher quality builds like a pair of monoblocs, or dual-mono or virtual dual-mono with regs or virtual dual-mono with diodes.
And, when upgraded to fine tuned values, it can do very low heat output and good system symmetry.

Possible gain settings: 38x when using the on-chip vas.
Or here's a different idea: Unity gain in the case of setting both chips to slave mode, which disables the voltage amp, turns the board into a "latfet output pack" and can be used for a tube hybrid.

P.S.
What I haven't seen yet is a small signal op-amp added to do error correction mode feedback like this: http://www.diyaudio.com/forums/solid-state/238449-thd-loop-gain-enhancer.html Has anyone tried this with a chip amp? For example, with my ordinary schematic, the gain ends up at 38X. However, if we added the error correction/enhancer scheme that would turn the gain down for errors but not for music. Fascinating?
 
See the modern, modular style paralleling without the ballast resistors?
T2oguPXa8XXXXXXXXX_!!16989278.jpg

Favorite new style design: 2pack TDA7293 Parallel 170W Mono Amplifier Board Kits 29 | eBay
With the two pack of compact dual-layer boards in the link (ebay# 190862927141), it is possible to do higher quality builds like a pair of monoblocs, or dual-mono or virtual dual-mono with regs or virtual dual-mono with diodes.
And, when upgraded to fine tuned values, it can do very low heat output and good system symmetry.

Possible gain settings: 38x when using the on-chip vas.
Or here's a different idea: Unity gain in the case of setting both chips to slave mode, which disables the voltage amp, turns the board into a "latfet output pack" and can be used for a tube hybrid.

P.S.
What I haven't seen yet is a small signal op-amp added to do error correction mode feedback like this: http://www.diyaudio.com/forums/solid-state/238449-thd-loop-gain-enhancer.html Has anyone tried this with a chip amp? For example, with my ordinary schematic, the gain ends up at 38X. However, if we added the error correction/enhancer scheme that would turn the gain down for errors but not for music. Fascinating?



Hi, Daniel,

Have you tried out the corresponding 3-chips version by the same designer sold by the same EBay store? Thanks!

Regards,
 
Hi, Daniel, Have you tried out the corresponding 3-chips version by the same designer sold by the same EBay store? Thanks!
Regards,
No, I didn't need it yet. Looks fine though.
Either of those parallel boards could use this schematic, except that the 3-chip model would have a 100uF bootstrap cap.

My 70's receiver-amp refit project has a custom fit center tap transformer, previously powering 55 watts per channel with frail discrete amp, now powering 90 watts per channel with the durable 2-chip TDA7293 parallel boards. Due to the transformer amperage capacity, I didn't need the 3-chip boards at this time. And, now I find out that the smaller 2-chip parallel units easily exceed the capacity of my speakers.

P.S. I bought a box of TDA7293 chips to play with for parallel amp point to point projects, so it is unlikely that I'll need boards for some time. More powerful speakers probably comes first.
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.