I had the idea of building a little amplifier/speaker module to monitor the modular output based on the classic TDA2030 amplifier in the pentawatt package (of course, @Stef already beat me to it). I am working with a 12V single supply (off an old 5A laptop switching PSU) and looking to build something similar to the datasheet (edited image off the datasheet below). I have a couple of questions for those who have played around with the 2030:
- Am I okay with electrolytic capacitors rated at 25V, especially C7?
- R6/C6 right before the speaker looks like a Zobel network. I am working with a salvaged 3Ω 75W speaker. Do I need a resistor rated at 1W instead of the typical 1/4W ones? How critical are the R6/C6 values for the kind of application I am looking for?
25V for the capacitors should be more than enough at 12V input voltage.
And for the Zobel network - yes, this is one - I used a 2W resistor for my input voltage of 19V, and its not getting very warm, but I think 1/4W would be a little bit too weak.
That’s great, thanks! At 1W I have 0.57Ω and 68Ω. How much difference the <1Ω will make for a 3Ω speaker?
The correct values can be calculated using the formulas on these website: Zobel network calculator - Impedance equalization circuit - Step by step
Maybe its interesting
Of course you can use more robust resistors,
Of course you can use more robust resistors,
Or put several 1/4W in parallel.
(To add the wattage they must have the same value, or else it’s a bit more complicated to compute the resultant wattage)
Unfortunately, it’s not possible to use that calculator as I need the inductance of the voice coil of the speaker. I tried locating the speaker driver datasheet but had no luck.
hm, ok, but in this case - if you cant measure the inductance - you could use the values found in some speakers datasheets. I’ve found values of around 1.5mH for 75W/4Ohm speakers, and 0.7mH for 8Ohm.
So I think with a resistor of 2Ohm/2W (or 3 of your 0.68 Ohm in series) your should be close enough to the perfect value.
Thanks for the suggestion! A quick correction, the speaker is an Eastech at 7.5W, not 75W as I wrote by mistake. I checked datasheets of similar Eastech speaker drivers and the specs for the inductance were all over the place, which is not very helpful. I might just put a socket for a couple of resistors in series and try a couple of values around 1Ω as specified in the TDA2030 datasheet. If I ever find more information, I can replace them easily.
Here’s what I came up with, which unfortunately is not working. When I turn this on, I get plenty of noise and oscillation, and the circuit gets overwhelmed when I inject a signal at the input. I got no signal at all, just oscillation.
The specs of the build:
- I used a, quite likely fake, TDA2030A desoldered from a little amplifier module that you can get cheaply online.
- The IC is mounted at a 45-degree angle with a large heatsink.
- The circuit was built on stripboard as shown below. All soldering is fine, no bridges etc.
- Electrolytic capacitors are rated at 25 or 50V, the rest are film box capacitors, except the 100nF which is ceramic. The diodes are 1N4007 and I have a socket for the 1Ω 1W resistor (tried 0.57Ω, 1.6 and 5Ω).
- Power comes from a 12V 5A switching laptop kind of brick.
- Speaker, input, and supply grounds are connected with crocodile clips.
- Tried both 3Ω 7.5W and 8Ω 0.5W speakers.
This had me scratching my head for a while. Unless I made a mistake in the stripboard layout (which is always likely), I am going to say that this is not likely to work because of the chip quality and/or limitations of stripboard design (i.e. jumpers and whatnot).
Even if the IC is original they are very sensible to pcb design and ground connections. I would not try this on stripboard. With an etched pcb with the design following the example in the datasheet I had no problems, nearly no noise and no oscillation.
But I will have a deaper look at your stripboard design.
This IC follows the rule #1 in oscillator and amplfifier design: ampflifiers always oscillate, and oscillators never.
ok, your stripboard design looks fine, only - in my opinion - too much hopping for the ground lane. Maybe you could try to change the orientation of the stripboard, from 17 rows/12 columns to 12 rows/17 columns, so that you could place all ground connections on one row without the need to place them all over the board.
If you can etch your own pcbs I could send you the kicad project.
Or maybe try to use the manhattan style, if you have some copper plated pcb material. Use a big copper plate for ground and glue small strips on it (just an example image):
In my case I bought an TD2030AV from UTC at the german shop reichelt.de for 0.60€.
Thanks for the suggestions! I figured that the ground jumpers might have something to do with it, but I did not expect things to be that bad. I mean, there is no signal whatsoever, just the nastiest motorboating and whining.
I could try another stripboard layout, but I suspect that it might be a huge difference. Or etching a board, as you suggested, but I am really trying to avoid the hassle of doing that. I looked to see if there is a board I could order from JLCPCB, but couldn’t find any on github, so I just ordered a couple of boards as kits. I also have a TDA2040 that I might try in a point-to-point build until the kits arrive, as you suggested.
Not abandoning this project yet!