That's interesting. To be quite honest I've not done any more than glanced at the TPA3116 datasheet so I wasn't aware of the 1SPW option.
I did take a quick look (see attached) and I could probably make a TPA3116 fit on the board. I considered doing this but looking at my usual suppliers (Farnell, Digikey & Mouser) the TPA3116 is around double the price of the 3132 for small quantities, which is a surprise. I wonder why?
I've had a look around for 41Hz audio gear and there's an awful lot about it but it seems they have now disappeared? I might add that it seems odd to supply small size SMD designs as kits, but regardless they are clearly popular. Also saves on assembly overheads I suppose.
I did take a quick look (see attached) and I could probably make a TPA3116 fit on the board. I considered doing this but looking at my usual suppliers (Farnell, Digikey & Mouser) the TPA3116 is around double the price of the 3132 for small quantities, which is a surprise. I wonder why?
I've had a look around for 41Hz audio gear and there's an awful lot about it but it seems they have now disappeared? I might add that it seems odd to supply small size SMD designs as kits, but regardless they are clearly popular. Also saves on assembly overheads I suppose.
Very nice job! I love how small they are.
Now sit back and see how long it takes YJ or one of the other Chinese manufacturers to take your design and sell it in mass. That is how we ended up with the famous YJ blue/black TPA3116D2 amp designed by member Danzz.
Thanks! Although I suppose it's a sign that you have got somewhere in life when China copies your products 😀
To take the last bit first. Yes, they were supplied as kits but the SMD parts were of physically largest kind that could fit in order to make it possible for even reasonably novice builders to DIY.
As you can see from this picture below the TA2021 is much larger than the TPA3116 (similar size to STmicro HSSOP36 size) so it is certainly possible. And it had full LC output filter as was required for that chip as it did not feature a filterfree operation mode like is standard on the TPA3116 (and the only option on TPA3132).
The size of the chip especially it's thickness allowed other components to mounted on the same side without interfering with the direct heat slug to cabinet bolting. Just like a TPA3116 a regular heat sink isn't really needed. To do the same with the TPA3116 you'd probably need some heat transfer material in order to fit components on the same side as the chip as the TPA3116 is too thin to allow this otherwise.
It's important to note that if you want to experiment with the 1SWP modulation mode on the TPA3116 that is it not a filterfree modulation mode. You must use full LC output filtering which leads me back to one of the most interesting parts on the amp32, the custom made dual output inductors. The only writing on them is "100 QY-" and I have no idea who made/makes them but they are very similar to the ones referred to below. It's the same incredibly compact physical package and most of the characteristics are the same as well.
http://www.xfmrs.com/pdf/XF100910DL-xxxx.pdf
Last edited:
Those inductors are rather interesting. A look on the XFMRS website showed no results, so I can only assume that the particular series in question is out of production. I'll drop them an email and see what they say.
As for heatsinking, the TPA3132 has the pad down of course, and with a fair bit of via stitching to the other side of the board I've had no issues with heat. Into a 4 ohm load at max input voltage it gets hot, probably about 60-70C. But at that power it distorts like hell anyway, and during all normal use it just rises in temperature, but doesn't get warm.
Out of interest how much did the amp32 kit cost? It looks like they used a thick board and decent components, a guess from here would be about $10 BOM for small batches.
Also, I went to a bit of effort to source nice Camdenboss screw terminals and fit them on the board. Without the terminals I could decrease the size, although this would remove the plug and play capability. What are your thoughts on this?
As for heatsinking, the TPA3132 has the pad down of course, and with a fair bit of via stitching to the other side of the board I've had no issues with heat. Into a 4 ohm load at max input voltage it gets hot, probably about 60-70C. But at that power it distorts like hell anyway, and during all normal use it just rises in temperature, but doesn't get warm.
Out of interest how much did the amp32 kit cost? It looks like they used a thick board and decent components, a guess from here would be about $10 BOM for small batches.
Also, I went to a bit of effort to source nice Camdenboss screw terminals and fit them on the board. Without the terminals I could decrease the size, although this would remove the plug and play capability. What are your thoughts on this?
Last edited:
If you're using the TPA3116 it's probably advisable to use an inductor capable of the full current even though it would hardly ever be used but it ensures you're reasonably safe even with low impedance loads. Coilcraft HA4158-EL (and similar) are still compact enough to fit but has even better data.
The idea behind using the TA2021 which is slug up (similar to TPA3116) instead of the TA2024 which is slug down (similar to TPA3118) is that you can route power plane directly under the chip and have ground plane exclusively on the reverse side. This makes routing far more optimal at the expense of requiring some kind of heat spreader, usually the casing itself.
I prefer (vertical) pin terminals for such compact boards. Screw terminals really takes up a lot of room, and anyone using such compact boards probably don't mind the small extra hassle. There's a lot of options of larger boards with screw terminals around for the multi-thumbed people.
I think the kit was €25 shipped. Board was quadruple thickness copper. All components were hand picked from an engineering point of view rather than "audiophile" preferences which usually gives the best end result.
The idea behind using the TA2021 which is slug up (similar to TPA3116) instead of the TA2024 which is slug down (similar to TPA3118) is that you can route power plane directly under the chip and have ground plane exclusively on the reverse side. This makes routing far more optimal at the expense of requiring some kind of heat spreader, usually the casing itself.
I prefer (vertical) pin terminals for such compact boards. Screw terminals really takes up a lot of room, and anyone using such compact boards probably don't mind the small extra hassle. There's a lot of options of larger boards with screw terminals around for the multi-thumbed people.
I think the kit was €25 shipped. Board was quadruple thickness copper. All components were hand picked from an engineering point of view rather than "audiophile" preferences which usually gives the best end result.
IO390, can you find out how much shipping would cost to Mumbai, India? Slipped into just a padded envelope.
They look like nice inductors. Not cheap though, to use such parts would increase the cost of the board by about £4-5, so if I was going to invest in a batch I'd have to be sure I'd have interest. Still, £22 or so isn't too bad... Or is it?
Well if I could do away with the screw terminals then I'd be able to have a much more convenient layout. Some beefy 100th spacing pin terminals would be much easier.
Is increased copper thickness necessary for only a 25W device? Careful routing with 1oz copper is fine for something of this power IMO, although it is certainly a possibility to have increased thickness. This is getting expensive haha
And yes, I've never been a follower of the audiophile approach to component selection. I'm already using decent components on this, given how cheap most SMD chip parts are there is really no excuse to use crappy components.
Keith,
An amplifier including shipping via Royal Mail International Standard (5-7 days to India) would be £22.
Probably not needed but I think it might have something to do with better quality vias with increased copper thickness. We are also talking about an almost 10 year old design. A lot of things have changed and been improved since.
No problem.
Indeed. Though if I ever need greater current capacity on vias, I just bundle a load together (plenty of that on these boards). But yes, PCB manufacturing has come a long way (not least in price)...
Nice. Yup. The 2nd type certainly looks exactly like those used on the amp32 (and amp6).
SMD Inductors for Class D 7G09B | SAGAMI ELEC CO., LTD.
Thanks for the links gentlemen.
I've sent off for a few quotes, so let's see what they come back with.
I've sent off for a few quotes, so let's see what they come back with.
Please note that you should only do a TPA3116 version for your own interest to begin with at least. What you already have built fills the niche left and already ticks off all the boxes plus some.
I am still planning to use TPA3132, I'm simply looking at using full LC output filters. This would enable the amp to be used in a normal system without having to worry about short speaker wires etc... Anyway I ordered a couple items from Coilcraft and Ill modify one of the existing boards and see how it goes.
I was thinking I could perhaps have the existing version plus the LC filter model at a slightly higher price perhaps (still £25 or less).
Also, you mentioned how this design "ticks off all the boxes plus some". May I ask, what are the "boxes"? Is this by comparison to an amp32 replacement? I thought it would be useful to know for when for when I make future improvements.
Well I stumbled upon an improvement. The two 4.7uF coupling caps sticking up on the component side of the board were going from the neutral inputs on the amp to ground (single ended input on this amp). There's no signal going in here so the 4.7uF caps were quite unnecessary - replaced with some 100nF ceramics and it has a much lower profile. I'll mod this on the existing boards that are here, the two that have been dispatched have electrolytics.
Should have same impedance as positive inputs, so closer to 6.8uF if you use classII ceramics I would think.
edit: your probably right, impedance ceramic will be lower than combined positive input impedance. I would replace the electrolytics on input if you want to improve, and maybe put the small DC decoupling ceramics closer to chip than the bigger value?
Last edited:
- Status
- Not open for further replies.