I'm probably missing something obvious, but what is the impedance of the Alpair 7.3? I'll be powering mine with a tpa3116-based class d amp. The amp has an onboard LP filter for removing HF noise introduced by the amp's switching.
I looked at the Alpair 7.3 data sheet from the MA website. It's not dumbed-down enough for me. 🙂
I guess it's a more general question, as most commercial off-the-shelf speakers give a single number for their impedance, but I realize that in reality, most speakers have a variable impedance across their frequency range. So what is the single number that is quoted? The average? Or the minimum? Or something else?
Looking at the A7.3 impedance vs frequency curve, it looks like it never goes below 6 ohms, and is around 10 ohms at 20 kHz... what value should I use for the best-matched class-d LP filter?
Thanks!
I looked at the Alpair 7.3 data sheet from the MA website. It's not dumbed-down enough for me. 🙂
I guess it's a more general question, as most commercial off-the-shelf speakers give a single number for their impedance, but I realize that in reality, most speakers have a variable impedance across their frequency range. So what is the single number that is quoted? The average? Or the minimum? Or something else?
Looking at the A7.3 impedance vs frequency curve, it looks like it never goes below 6 ohms, and is around 10 ohms at 20 kHz... what value should I use for the best-matched class-d LP filter?
Thanks!
Also, a related question: according to the tpa3116 datasheet, it's technically possible to run the amp without any LP filter, and use simple ferrite beads for removing/reducing EMI. In this scenario, the speaker itself becomes the filter for the high-frequency switching noise. In my reading I've seen suggestions that this might damage some speakers.
Just curious if the Alpair 7.3 could potentially be damaged by sending it an unfiltered signal that includes class-d HF switching noise?
Just curious if the Alpair 7.3 could potentially be damaged by sending it an unfiltered signal that includes class-d HF switching noise?
The nominal voice coil DC resistance (ReVC) = 5.4ohms , whereas as you note the graph shows minimum of just over 6. Impedance is reactive with frequency, but keeping it simple, probably plan on the 6ohm value.
Experts in DIY class D amps should pipe in here, but I'd take the conservative approach. I'd guess that while not likely to be damaged by anything but a severely misbehaving amp, the Alpairs have sifficient HF resolution that any artifacts from power supply noise would not be camouflaged.
Experts in DIY class D amps should pipe in here, but I'd take the conservative approach. I'd guess that while not likely to be damaged by anything but a severely misbehaving amp, the Alpairs have sifficient HF resolution that any artifacts from power supply noise would not be camouflaged.
Thanks Chris! By the way, these A7.3s are currently in transit to me from Victoria, BC, along with a CGR dMar-Ken7.3 flat pack that I believe was made by you. 🙂
Matt: hope you enjoy them!
Dave should have some "suggested assembly" photos that might be helpful. While I never like to second guess any kit builder's skill set, and there are lots more experienced craftsman than I, these photos will show the process steps that I would use. Hint - if you plan on veneering or painting them, and have access to brad nailer, you'll need a lot fewer clamps.
I am noexpert, but...
A Class D amplifier or switching amplifier has a very high frequency carrier that needs to be filtered out. The LP filter is < 20 kHz, so you'd like to use that frequency. Since most impedance curves pnlu go to 20 kHz, that is as close as you are going to get. So i'd use the 20 ohm number.
The nominal impedance single number given is just a general guideline that puts speakers into "bins". This is usually the DCR +a little bit for the reactive impedance so A7.3 would be called an 8 ohm speaker.
dave
A Class D amplifier or switching amplifier has a very high frequency carrier that needs to be filtered out. The LP filter is < 20 kHz, so you'd like to use that frequency. Since most impedance curves pnlu go to 20 kHz, that is as close as you are going to get. So i'd use the 20 ohm number.
The nominal impedance single number given is just a general guideline that puts speakers into "bins". This is usually the DCR +a little bit for the reactive impedance so A7.3 would be called an 8 ohm speaker.
dave
I'm not an expert either, but I do know this: The main difference in the sound of different class d amps is the output filter Dave is talking about. Cheap amps have cheap, sometimes almost nonexistent filters. More expensive amps are likely to have decent filters with decent parts. Sure, you can mod a $20 amp into something useful, but you will spend more on parts than the original cost of the amp, and of course your time is free.
The Tripath amps are now technically old. There are some nice implementations starting at ~$200 and up. I have a nice one using the TA2022 with an analog power supply and name brand parts that sounds pretty good. Texas Instruments has a line of class d amps that start with TDA. There are a large number of demonstration boards that start at $20, no case, no power supply. Full implementations are starting to appear.
Now to answer your question, if you get a decent class d amp, you don't need to worry about the switching signal passing through to the audio range.
Bob
The Tripath amps are now technically old. There are some nice implementations starting at ~$200 and up. I have a nice one using the TA2022 with an analog power supply and name brand parts that sounds pretty good. Texas Instruments has a line of class d amps that start with TDA. There are a large number of demonstration boards that start at $20, no case, no power supply. Full implementations are starting to appear.
Now to answer your question, if you get a decent class d amp, you don't need to worry about the switching signal passing through to the audio range.
Bob
I'm likewise no expert, but based on my limited hobbyist experience: I suspect the sound quality of different class-d amps is a function of both the output filter and the chip itself. Like a chain, it's only as good as the weakest component. But, to be fair, I'm just making assumptions.
Just to be pedantic: the Texas Instruments line starts with TPA. TDA is from another company.
In fact, take a peek of this tpa3116 thread here on diyAudio if you're interested. Or you can check out the Wiki that I've been trying to maintain as a consolidated version of the crazy-long discussion thread.
This A7.3 discussion revolves around my desktop/nearfield setup. SPL is low, so I don't need a lot of power, and space is at a premium, so I want a relatively small amp, so class D is a good fit. Plus the efficiency/low power appeals to me... and I'm cheap and I like to tinker. 🙂
Previously I was using a HiFiMeDIY T2 amp, which is based on the TK2050 chip. It's a fairly well-regarded implementation.
But over on AudioCircle, in the "Cheap and Cheerful" forum, someone was talking up this $10 Sure TPA3110 board. For $10, I figured it was worth a try. I immediately noticed a substantial improvement in sound quality. I happily ran with that for a while, until I stumbled on that tpa3116 thread. The tpa3110 and 3116 are very similar, with the latter having more output power capability. This Yuan-Jing came out with a tpa3116 board that bears an uncanny resemblance to a suggested PCB layout that one of the forum members (Danzz) floated. The board uses mostly through-hole components so it's easy to modify. (It was only $20 when introduced, can still be had for that much off ebay.)
The general consensus is that the biggest improvement comes from upgrading the output filter inductors... and that's what brought me to bring up this question, since the ideal inductor is "tuned" to speaker impedance.
To me, the lower-power 3110 still has a slight edge over the 3116. But, the 3116 stock output filter is designed with 4 ohm speakers in mind; my current desktop speakers are 8 ohm. And interestingly, the 3110 is a filterless design.
Also, FWIW, my living room system uses yet another class d amp: it's an IRS2092-based amp, model "SDS-470" from the company Class D Audio. I have it powering what I consider fairly decent speakers (Salk Songtower). This is a fairly expensive amp (relative to the tpa311x amps anyway), and in my limited hifi experience, I consider it to be excellent. (Way too overpowered for desktop use though.) I used my stock tpa3116 build in it's place for a while, and didn't note any obvious differences, although I haven't had time for much critical listening on that system.
Anyway, just rambling about the TI TPA amps, as I've been having a lot of fun lately hacking on them and reading about others experiences.
There is a little something added after the LC filter most of times, not mentioned much, but it is parallel to speaker, TI calls a snubber.
Translated to single ended output parallel to speaker there is a 6.6 ohm resistor and a 5nF capacitor. How does impedance curve of alpair7.3 look with odd Zobellike parts of these values?
Translated to single ended output parallel to speaker there is a 6.6 ohm resistor and a 5nF capacitor. How does impedance curve of alpair7.3 look with odd Zobellike parts of these values?
The TPA amps and links have been mentioned to Bob before but he blew them off and went for the TA2022, which is surprising given that the 2022 is considered by some Tripath aficionados as the worst sounding of the lot IIRC.
I am betting the TPA will play with the 7.3 just fine, however I would suggest having 2 Elna Silmic II caps on hand for the TPA Power supply as their warm, smooth presentation may be a helpful fit for the Alpairs especially during break-in. Just something to think about.
Last edited:
No I did not. Reread my post! What I said was the Tripath amps are now obsolescent and the TI amps are coming on line. I don't do electronics. I am not going to buy a demo board, modify it and fit it in a box. When I got my 2022 amp, a TPA amp was not an option. When I stumble across a TPA amp that is fully fleshed out with decent parts, I will get one, and I think the time will be soon.
Bob
I have been playing the Red YJ TPA3116D2 amp with the A12P and it sounds pretty good. There is room for improvement, but even in stock, it's way too good for less than $20. Breaking in a pair of Alp 6.2m with the amp, and though not doing any serious listening during the break-in, I am not finding anything objectionable in the treble or mid-range yet.
I am thinking of getting a blue board soon...
I am thinking of getting a blue board soon...
Last edited:
As wonderful as this latest generation of digital amps may sound, I'm sure there's a lot of us on the same page as Bob - even if we ever had the skills, the urge to build electronic projects has long since waned. I've got a 12 yr old Jolida amp sitting on the bench for over 3 yrs now part way through a major rebuild - changed focus a bit, and then just lost interest.
Bob, Chris, if you guys are really interested in the TI TPA chips, but don't want to mess with DIY electronics, there are a few options:
Anyway, the point is, you can give these a try without even seeing a circuit board, and without much cash invested, if you're interested.
- SMSL SA-36A Pro. Uses the tpa3118, which is identical to the 3116 circuit-wise, only manufactured differently (different IC package). The 3118 is designed for use without a heatsink, so has a lower output spec. This is literally 100% finished, plug'n'play. $50. At least one person over on AudioCircle has one of these on order, should have a report on its performance soon.
- Yuan-Jing TPA3116 Finished Amp. This uses the same "blue/black" board that has been popular for modding in the big thread I mentioned earlier. Just like the one above, though, it's plug'n'play, ready to go. A little higher output power, although considerably more money. $120, not sure about shipping. But it's a known good performer.
- Sure TPA3110 Amp Board Only. No case, but immediately usable: has a 1/8" stereo input jack and 2.1 (or 2.5?) mm DC input jack... I would assume just about everyone has a spare wall-wart PSU handy that falls in the 10 to 16V range. For $10, what have you got to lose? Note you can also buy from Sure direct through their eBay store ($10 shipped from China). This is a pretty low-power chip, but plenty of power for nearfield applications and/or high-efficiency speakers. If you don't like it, drop it in the mail to me, I'll cover shipping. 🙂
- Yuan Jing tpa3116 board only. This isn't quite plug and play; it's the same board used in #2 above, and is the popular board that everyone enjoys modding. But it sounds great in stock form. If you're so inclined, however, you can improve it's performance with some modding. (And this is a tweaking forum, so even if it was perfect out of the box, someone around here would be compelled to mess with it.) Note the only work required to make this usable is to connect RCA inputs, binding posts, and a DC jack. I'd argue that's no more "electronics work" than wiring up DIY speakers.
Anyway, the point is, you can give these a try without even seeing a circuit board, and without much cash invested, if you're interested.
thanks for the links, Matt - those do look interesting - to think that I paid over $300 for my Kingrex T20 and PSU a few years back- Moore's law at work, I guess
edit; actually, the $88 kit version of 50w PC looks easy enough to kludge together.
Yuan Jing Audio - DIY KIT TPA3116 Class-D 2.0 Stereo Power Amplifier - $88.00 - DIY Kits
edit; actually, the $88 kit version of 50w PC looks easy enough to kludge together.
Yuan Jing Audio - DIY KIT TPA3116 Class-D 2.0 Stereo Power Amplifier - $88.00 - DIY Kits
Last edited:
I think option #2 (Yuan-Jing TPA3116 Finished Amp) is my best option. The SMSL is "rated" at 20W/c @8Ω, which means distortion skyrockets at ~7w/c. The Y-P is "rated" at 50w/c @4Ω, so I can probably 15w/c clean. I suppose I need to look at the data sheet and see, but I do know that a TA2020 into and A10P struggles to produce 80dB at 8ft.
Bob
Bob
I don't know room size nor its content so hard to tell🙂 I do know that all 311x amps I tried don't have trouble driving P70's very loud. Datasheet TI is for both 3118/3116, the skyrocketing is rather late I think http://www.ti.com/lit/ds/symlink/tpa3116d2.pdf
Thanks for doing what I was too lazy to do. Looks like I can get 20-30w at 8Ω before THD goes south, ok, north. That's about the same as the TA2022. Should work fine. The room is 15'x25'x8', but that is really irrelevant. The operative measurement is 8' from the speaker to the chair.
Bob
Bob
matt: going with 22uH .68uF ?
regarding looking at 4 options for these amps:
1. inputcap value could influence bass respons (HP filter) the first option smsl has 3.3uF epcos, the other YJ's 1uF. I think choice for 3.3uF is better regardless of gainsetting, most of times this will mean option 1 smsl has more extended bassrespons (less attenuated bassrespons)
2. the outputfilters have same values L&C, but again I would favor smsl here. At least you know where the inductors go wrong, 2.5-3Amp. YJ's inductors are ... who knows.
3. early in very long 3116 tread different voltages for powersupplies were tested by several people, 24V is not ideal, 21V and lower could be. With 24V there appeared to be some kind of stress, I believe in output part of chip, that negatively influenced SQ. Buying an tpa3116 amp that has integrated 24V powersupply...smsl comes with choice and again seems better in that respect.
Does the SMSL sound better? Too new to tell. Could be worse LOL
regarding looking at 4 options for these amps:
1. inputcap value could influence bass respons (HP filter) the first option smsl has 3.3uF epcos, the other YJ's 1uF. I think choice for 3.3uF is better regardless of gainsetting, most of times this will mean option 1 smsl has more extended bassrespons (less attenuated bassrespons)
2. the outputfilters have same values L&C, but again I would favor smsl here. At least you know where the inductors go wrong, 2.5-3Amp. YJ's inductors are ... who knows.
3. early in very long 3116 tread different voltages for powersupplies were tested by several people, 24V is not ideal, 21V and lower could be. With 24V there appeared to be some kind of stress, I believe in output part of chip, that negatively influenced SQ. Buying an tpa3116 amp that has integrated 24V powersupply...smsl comes with choice and again seems better in that respect.
Does the SMSL sound better? Too new to tell. Could be worse LOL
- Status
- Not open for further replies.