diyAudio (
-   Chip Amps (
-   -   Another TDA2050 amp built and tested.............. (

ashok 28th June 2007 06:07 PM

Another TDA2050 amp built and tested..............
2 Attachment(s)
After seeing so much written about the TDA series of chip amps, I couldn't resist testing the TDA2050 which is so cheap over here.
All parts are standard and the circuit is off the application note. Only the gain has been increased slightly due to component value changes. Input R = 47K and input C = 1uF ( Philips film )
Feed back is 47K and 1K with a 47uF electrolytic to ground.
Supply is +/- 25 volts dc from a 50 VA(!) torroid . I didn't have anything else.

The heat sink is too small for loud listening levels. The temperature was about 50 deg C I guess. I couldn't touch it for more than a second. Will check that tomorrow.

The sound was..... well...... incredible for a $1/- chip ! The first few minutes it seemed to be a bit rough in parts but as time went by it got better . Will have to let this play for a while before coming to conclusions. It looks like voice will be fine, deep bass is definitely there. The sound is tight and treble is clean .
However at times things are not OK as I seem to hear background trash. Mostly with complicated sounds. Might be some RF hash . Might need some compensation caps . Right now there are no compensation caps anywhere.
Supply cap is only 4700uF per rail but seems to be fine.

Never expected this to sound so good. Background noise is inaudible . Will check that again tomorrow. I'm going to get some more boards made and build a few more. Good to have some low cost good sounding rugged amps around for experimentation.

ashok 29th June 2007 01:55 PM

The next day !
I did a listening test with a well known $ +1K amp. No names will be mentioned !

Surprise, surprise ! The TDA2050 beat it by a good margin in some areas !
The TDA was
1. Cleaner sounding but noisy vocals sound a bit rough.
( this could be a speaker problem (?)).
2. Bass was deeper and more distinct.
3. Midrange was slightly more forward ( certainly more clear cut ).
4. Generally, instruments had more air around each other.
5. Upper mids could sometimes be a bit brash.
6. Sound was more dynamic. The ref amp sounded compressed.
This was quite surprising . The ref amp sounded veiled in comparison ! Note that the listening levels were matched within 0.5db .

I think the LM3886 is better . Will also know how it compares with the TDA7293/94 in a short while.

For guys who are forced to stay on a very tight budget ( or do not want to spend more ) , the TDA2050 must be one of the best options available.

Some observations:
Running it at full blast .....just clipping least 30 % of the time, it sounds good. Test signal was pop music off a CD player.
Heat sink temp rose to 60 deg C at the chips.
The 50VA torroid just got a bit warm.I was expecting it to get hot.
DC off set was less than 3mV on both channels.
A bigger torroid with 10,000uF caps might improve the sound some more. My speakers are nominally 8 ohms . They dip to about 5.8 ohms in the lower mids I think .

With some tweaks I think it will sound even better . As it is , it is great. I'm happy I tried it. Will have to refine the pcb a bit.
Great $1/- chip .
Tweaking it too much would be expensive and therefore pointless because the higher powered chips are also not very much more expensive and offer better performance.Over 90% of the cost goes towards the rest of the amp.


ashok 30th June 2007 10:55 AM

Day two !
2 Attachment(s)
This is the second day of testing and so far 125 page views but no comments !

As mentioned earlier , the transformer is currently a 50 VA torroid as that was all I had at hand. It’s too small but sounds surprisingly good and does not heat up much playing into 8 ohm speakers at full blast . I will be ordering a 100 or 120 VA 18-0-18 V unit .
The heat sink is also a bit small. I managed to reach 68 deg C playing loud on 88db/watt speakers. No harm done. In a proper case with fins on the outside it should be OK for music.

I changed the speakers and the occasional roughness in mid range mellowed down a lot. Then I changed the input cap to a Solen film cap. Now it really sounds very good. It might just sound as good as a LM3886 . I cannot vouch for that until I try it out in an A/B test.

What makes this amp so interesting is the heart of it cost next to nothing and it sounds so good with a respectable 20 watts per channel! But then that is true only for a DIY amp. This is also not for people who like to rock the house down.

I made a pie chart to compare the cost of various parts of the amp. Looks interesing.
Resistors are less than 1% of the cost ! The ‘power devices’ are just 3% of the cost !
The casing ( Aluminum box ,RCA sockets and speaker binding posts, IEC socket, mains filter ) is the most expensive part.

ashok 30th June 2007 11:01 AM

2 Attachment(s)
Here is a picture of the board and the transformer I used.

consort_ee_um 30th June 2007 01:50 PM

You need a zobel network on the output to stop any high frequency oscillation. I like the idea of the power supply components being on the board. I have used it myself but you need to have a star ground layout at the storage capacitors.
Is it stereo or bridged mono?

ashok 30th June 2007 05:11 PM

The board does carry a Zobel network as per the application note. 2.2 ohms ( vertically mounted ) and 0.47uF . Yes the board is reasonably star grounded .
I neglected the earth of the second channel and connected it the wrong way by a few mm and it had 100Hz supply ripple on the output ! I reconnected it to a different location and the 100Hz was gone. The board layout has to be done quite carefully.

The board will have to be modified slightly. This is a stereo board and with the existing supply it goes pretty loud.
I think the supply needs to be stiffer........bigger transformer and at least 10,000uF per rail.

I later added a 100pF capacitor at the input to filter out all RF trash . It didn't seem to matter audibly.
The amp sounds very good. I can't get over the fact that a $1 chip can sound so good.
Remember the 810 chip ....

ashok 1st July 2007 05:01 PM

On second thought...............
I watched the power supply rails when playing music. It dipped quite a bit on loud music. However the ripple levels did not go up too much though the average dc levels did drop more than it should.
So a bigger transformer might be more effective than increasing psu capacitance. So a 100 or 120 VA trafo with a minimum of 4,700uF per rail should do well. I'm targeting a 4 ohm load ( about 35 watts per channel ).

scaesic 4th July 2007 05:00 PM

what kind of slow blow fuse did you use?

i was planning on using 2000uf caps on power supply rails but i might bump it up to 4000uf after reading this.

ashok 5th July 2007 08:55 AM

2 Attachment(s)
I've attached the circuit I've used with the component values shown on it.
The supply and supply fuse is shared between both channels.
I used 5 A slow blow because I had it . I've just changed it to 3A slow blow and it's working fine. I've used it about 4 hours now with the 3 A fuse and playing it pretty loud ( ocassional clipping).
This is a non inverting scheme and gain is +48 ( +33.6dB ).

So assuming a maximum theoretical 35 watts into 4 ohms , the input required would be just under 0.25 Volts.

Output is at pin 4 . The small +/- supply indicators near the opamp is part of the software package. You can ignore that.

Pin 5 is the +ve supply and pin 3 is the -ve supply.


ashok 5th July 2007 09:02 AM

The bridge rectifier is a BR1010 without any snubbers.

All times are GMT. The time now is 12:13 AM.

Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 18.75%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio