EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: kenta_hirono on August 24, 2022, 05:29:13 am
-
Hi to all I'm new here,
I'm about to make a couple of DIY bookshelf speakers powered by some 19V laptop bricks laying around.
I opted for some A/B class IC's for car audio as those usually works with relatively low voltage and support 4ohm speakers like the ones I have.
I made some simulations with LTSpice about the powerboard based on an LD1084 LDO regulator, but I don't get if I made some mistakes, stability analysis looks no so good and testing the example circuit for the LDO it looks good but output caps got no esr, with recommended esr then go crazy as well.
I attach diagrams and simulations screenshoots, thanks for your help.
-
The regulator has internal feedback from the output, and tries to maintain it 1.25V above ADJ. By not breaking that internal loop but instead applying some voltage to ADJ, you are testing the regulator in closed-loop unity follower configuration. Also, observe that AC feedback to ADJ through your voltage setting divider is killed by the filtering capacitor (or attenuated by a fairly high fixed ratio when Xc < ESR). In this circuit, the regulator really is nothing more than a voltage follower with +1.25V offset, as far as AC is concerned.
I think you are overthinking this. You are not doing anything extraordinary with this regulator. Simply apply the capacitance they ask for and it should work.
-
yep ld1084's datasheet say very few, I based on the LT1084/LM1084 documentation, but
- the amplifier IC needs a large cap (1200uF/35V esr .025ohm) and a smaller ceramic one
- the 150uF/35V caps I got are .13ohm esr
in the simulator (if I did it correctly) you can see a pole around 20kHz-40kHz that can be into audio-signal range
as the simulator give me a strange behaviour when testing the reference diagram with real components (the included example got caps with esr 0 and show no instability, but not with real like esr) and the c2 cap on adj pin should help but seems to worst the simulation further
I'm in doubt there is something I don't understand in this feedback-pole-related-amplification or the LTSpice model for the ldo is no working as supposed.
-
Why such an overkill?
Just place two 1N540x diodes in series with the supply voltage and you're within operating voltage range. Super-cheap and easy.
-
Your "base" simulation makes no sense because you have grounded the input and the output of the regulator.
The big one simulates something, but I'm not sure what it is.
It doesn't seem to be the open loop response of the regulator.
It surely isn't its output impedance.
-
The amplifier says absolute max input voltage 18v
You could just have three 3A diodes each drop ~0.7v and you have your voltage.
The amplifier chip isn't that great - you're looking at max 7w or so with 19v, if you want 1% thd.
You could just get a bunch of chinese TDA2050 or TDA2030 at around 50 cents each and keep it simple.
You get around 8-10w with both at 4 ohm 0.5-1% THD, powered directly from 19v , but they can run much better from a split power supply (ex up to +/- 18v or 36v for tda2030 (min. +/-6v or 12v) , up to +/- 25v or 50v for tda2050 (min. +/- 4.5v or 9v)
The tda2030 datasheet below even has example of using two such chips to make bridge circuit using split power supply so you don't need a big electrolytic on the output and to also get more output power.
TDA2050TB : https://lcsc.com/product-detail/Audio-Power-OpAmps_HGSEMI-TDA2050TB_C434516.html
datasheet : https://datasheet.lcsc.com/lcsc/1912111437_HGSEMI-TDA2050TB_C434516.pdf
TDA2030 : https://lcsc.com/product-detail/Audio-Power-OpAmps_HGSEMI-TDA2030_C2683154.html
datasheet : https://datasheet.lcsc.com/lcsc/2106070404_HGSEMI-TDA2030_C2683154.pdf
example of tda2050 amplifier project using split power supply : https://diyaudioprojects.com/Chip/DIY-TDA2050-Hi-Fi-Chip-Amplifier/
you could also easily get a couple toroidal transformers with two 12v/15v/18v/24v secondary windings for a split power supply design (or even a single winding) and keep it analogue and low frequency, so that you don't introduce switching noise into your sound.
-
@Benta I choose to add a linear regulator to avoid noise from the switching power source and get a bit of flexybility to replace it with other ac/dc adapter with different output voltage
@mariush the two amplifier I choose are 24W x2 @1% THD 16V, my speakers are 25W so no need more power, the ones you posted need negative voltage to operate correctly
@magic the "base" is the reference design, I put a load on the output and an AC source to test the stability of the ldo on load variations https://duckduckgo.com/?q=ldo+stability+testing (https://duckduckgo.com/?q=ldo+stability+testing) https://www.eeweb.com/esr-stability-and-the-ldo-regulator/ (https://www.eeweb.com/esr-stability-and-the-ldo-regulator/)
-
@Benta I choose to ad a linear regulator to avoid noise from the switching power source and get a bit of flexybility to replace it with other ac/dc adapter with different output voltage
Two comments on that:
1: all laptop bricks I've seen are 17, 18 or 19 V.
2: these are audio amplifiers designed for automotive use, which is the electrically noisiest and dirtiest environment you can imagine (OK, just in front of an airport radar would be worse). They are designed for dirty power.
Up to you. Two 8-cent (TME) diodes will do the job just as well.
-
You use TDA7375 and TDA7376 in your schematics
The datasheet is here for TDA7375v : https://www.st.com/resource/en/datasheet/tda7375v.pdf (https://www.st.com/resource/en/datasheet/tda7375v.pdf)
The datasheet is here for TDA7376 : https://www.st.com/resource/en/datasheet/tda7376b.pdf (https://www.st.com/resource/en/datasheet/tda7376b.pdf)
For the first, you have on page 6 figure 12 and 13 : 7w per channel on 4 ohm, and 16v input , 1% thd .. 24w in BTL mode
TDA2050 and TDA2030 can run in either split power supply OR single power supply mode... split power supply mode makes it possible to avoid a big electrolytic capacitor on the output.
Take the TDA2050 datasheet : https://datasheet.lcsc.com/lcsc/1912111437_HGSEMI-TDA2050TB_C434516.pdf (https://datasheet.lcsc.com/lcsc/1912111437_HGSEMI-TDA2050TB_C434516.pdf)
Go to page 4 and you have the SPLIT power supply typical application.
On page 6 you have the SINGLE supply example. You need an extra electrolytic for AC decoupling but that's pretty much it.
With 19v input and 4 ohm speakers, the TDA2050 can do around 8w with THD 0.5% at 19v - using +/-10v line on the figure 7 on page 8.
But it can do much more with just more voltage... ex a 36v 66 watts wallwart is 11$ : https://www.digikey.com/en/products/detail/tri-mag-llc/ICM30-360/10445552 (https://www.digikey.com/en/products/detail/tri-mag-llc/ICM30-360/10445552)
With 36v (Equivalent of +/- 18v), the tda2050 can do up to around 20w and the tda2030 will do close to 30w
Though at around 70% efficiency, the 66w psu above would give you around 40w of audio in total I guess.
-
Change the output voltage of the laptop power supply to 18V, by replacing a resistor on the TL431 feedback circuit.
Failing that, I know I'll probably get flamed for this, but I'd just run it straight off 19V. Those audio amplifier ICs can take 28V. Fair enough, they're only guaranteed to work up to 18V, but you'll probably be able to push it up by 1V, without any problems. They're tough and designed for a much harsher environment, than your home, so I wouldn't worry about the slightly higher voltage.
-
Change the output voltage of the laptop power supply to 18V, by replacing a resistor on the TL431 feedback circuit.
Failing that, I know I'll probably get flamed for this, but I'd just run it straight off 19V.
1: do you own a laptop? The bricks don't have screws these days, they're practically impossible to open. You suggested this in another thread where I posted about LED driving using bricks. The idea is dead in the water.
But running the parts at 19 V I second. Not that I like running devices outside their specs, but these are rugged and for automotive, it'll probably work.
-
@magic the "base" is the reference design, I put a load on the output and an AC source to test the stability of the ldo on load variations https://duckduckgo.com/?q=ldo+stability+testing (https://duckduckgo.com/?q=ldo+stability+testing) https://www.eeweb.com/esr-stability-and-the-ldo-regulator/ (https://www.eeweb.com/esr-stability-and-the-ldo-regulator/)
It's not the reference design because there is no power to the LDO input.
-
Change the output voltage of the laptop power supply to 18V, by replacing a resistor on the TL431 feedback circuit.
1: do you own a laptop? The bricks don't have screws these days, they're practically impossible to open. You suggested this in another thread where I posted about LED driving using bricks. The idea is dead in the water.
It's just ultrasonically welded plastic, not diamond case.
You can use some adjustable pliers or a vice to squeeze the plastic case near the seam until it cracks open, and then you can use a flathead screwdriver to extend the break until both parts are apart.
Or you can use a dremel.
I've also resorted to heating a very thin knife or a blade on the stove until red hot, and then drag it along the seam to melt the plastic enough to insert a flathead screwdriver as a crowbar of sorts to break the seam along the edge.
-
Well, yes, and then you can't put it back together ;)
-
There are plenty of amplifier chips that can run off the 19 V without a regulator. They may need a little more votlage overhead, but with 19 V as opposed to some 14-16 V with an LDO you have that extra headroom. If it is just to get 18 V one could use 2 diodes in series to get some 1-1.5 V of drop.
-
Change the output voltage of the laptop power supply to 18V, by replacing a resistor on the TL431 feedback circuit.
1: do you own a laptop? The bricks don't have screws these days, they're practically impossible to open. You suggested this in another thread where I posted about LED driving using bricks. The idea is dead in the water.
It's just ultrasonically welded plastic, not diamond case.
You can use some adjustable pliers or a vice to squeeze the plastic case near the seam until it cracks open, and then you can use a flathead screwdriver to extend the break until both parts are apart.
Or you can use a dremel.
I've also resorted to heating a very thin knife or a blade on the stove until red hot, and then drag it along the seam to melt the plastic enough to insert a flathead screwdriver as a crowbar of sorts to break the seam along the edge.
Well, yes, and then you can't put it back together ;)
Don't bother. Transfer the PSU board into the speaker enclosure, which would need to be earthed, if it's metal. If the PSU enclosure isn't that badly damaged, put it back together with tape and keep it in the speaker enclosure.
-
There are plenty of amplifier chips that can run off the 19 V without a regulator.
For that voltage, I would suggest a TPA3116 module, they're cheap and perform very well. Another option is the MA12070 but that's significantly more expensive.
-
@magic the "base" is the reference design, I put a load on the output and an AC source to test the stability of the ldo on load variations https://duckduckgo.com/?q=ldo+stability+testing (https://duckduckgo.com/?q=ldo+stability+testing) https://www.eeweb.com/esr-stability-and-the-ldo-regulator/ (https://www.eeweb.com/esr-stability-and-the-ldo-regulator/)
It's not the reference design because there is no power to the LDO input.
Oh cr4p!!!
this has sense now.
I remade some simulations, with few edits, looks like some components didn't work has I expected (like 1n5819 diodes leaks a bit to much)
Seems stable enough with some 0.1~0.2 resistors to increase caps esr, but dumping is a bit affected too duh
I tried to simulate somehow the load of a push-pull/btl
-
@magic the "base" is the reference design, I put a load on the output and an AC source to test the stability of the ldo on load variations https://duckduckgo.com/?q=ldo+stability+testing (https://duckduckgo.com/?q=ldo+stability+testing) https://www.eeweb.com/esr-stability-and-the-ldo-regulator/ (https://www.eeweb.com/esr-stability-and-the-ldo-regulator/)
It's not the reference design because there is no power to the LDO input.
Oh cr4p!!!
this has sense now.
I remade some simulations, with few edits, looks like some components didn't work has I expected (like 1n5819 diodes leaks a bit to much)
Seems stable enough with some 0.1~0.2 resistors to increase caps esr, but dumping is a bit affected too duh
I tried to simulate somehow the load of a push-pull/btl
You're making this much more difficult than it needs to be.
Have you tried powering it straight from the 19V PSU?
At the very most you need a capacitance multiplier, which consists of a transistor, resistor and a couple of capacitors, is all that's required.