Sure Electronics New Tripath Board tc2000+tp2050

[sendler]

Inductance added?

It's not a sim. It's a real measurement from a circuit that has basically the same topology as the Tripath data sheet.


Maybe it's not, but my understanding from those who were doing the testing is that it resulted in audible sibilence. Can't verify it myself...

I was going to say, something else more than just the resistance has changed for the one peaking response. That is what I see from adding some inductance to a purely resistive model of the speaker in my sims. Maybe they are showing 2-8ohm test resistors versus one real 6 ohm speaker? Or one different filter with smaller coils and too much caps for comparison? Either way it is a good example of the draw back of class D. The required output filter is fussy to specify depending on what speaker will follow it and a purely resistive model won't really predict what types of responses will be found in the field with different speakers. I am leaning toward a filter that is allowed to peak as they show much better consistency 4 ohm versus 8 ohm at 20K and only the peak at 70K changes which can be stabilized by adjusting the resistor in a Zobel.

 

[Neil Davis]

I was going to say, something else more than just the resistance has changed for the one peaking response. That is what I see from adding some inductance to a purely resistive model of the speaker in my sims. Maybe they are showing 2-8ohm test resistors versus one real 6 ohm speaker? Or one different filter with smaller coils and too much caps for comparison?

The actual circuit and testing that was done is proprietary info...but I didn't think there would be a problem publishing the graphs from one test result so that you get an idea of how these amps behave.

Either way it is a good example of the draw back of class D. The required output filter is fussy to specify depending on what speaker will follow it and a purely resistive model won't really predict what types of responses will be found in the field with different speakers.

Enclosing the output filter in a feedback loop such as what is used in the Ucd amps will get around this problem. However, the Tripath/Apogee/TI modulators don't allow that option. Zetex is the only vendor I know of with a digital input and feedback from the output. There will be more in the future...

I am leaning toward a filter that is allowed to peak as they show much better consistency 4 ohm versus 8 ohm at 20K and only the peak at 70K changes which can be stabilized by adjusting the resistor in a Zobel

You might want to model the output filter recommended by Apogee, and tailor it for the Tripath switching frequency. It uses additional poles that are placed at a higher frequency than the main filter to ensure adequate EMI reduction. See the discussion at: Understanding output filters for Class-D amplifiers | Audio DesignLine

I've got a couple of circuit boards that use this filter--I could sell you one at cost if you want to experiment.

 

[sendler]

Been there

The actual circuit and testing that was done is proprietary info...but I didn't think there would be a problem publishing the graphs from one test result so that you get an idea of how these amps behave.



Enclosing the output filter in a feedback loop such as what is used in the Ucd amps will get around this problem. However, the Tripath/Apogee/TI modulators don't allow that option. Zetex is the only vendor I know of with a digital input and feedback from the output. There will be more in the future...



You might want to model the output filter recommended by Apogee, and tailor it for the Tripath switching frequency. It uses additional poles that are placed at a higher frequency than the main filter to ensure adequate EMI reduction. See the discussion at: Understanding output filters for Class-D amplifiers | Audio DesignLine

I've got a couple of circuit boards that use this filter--I could sell you one at cost if you want to experiment.

It so happens that the first filter I tried was that one from the "Understanding" site. I prefer the ubiquitous common/ differential cap filter plus zobel that Tripath and others recommend. The more you work with the "Understanding" filter, the more you keep reducing the resistors down to 0 where it starts to turn into the other filter more or less.

 

[Ouroboros]

I use a couple of these amplifier boards to bi-amp a pair of home made two-way speakers. On both the mid/bass unit and the tweeter I have added a series C-R Zobel network to flatten out the impedance rise of the units at HF. (I have an HP35665 analyser so it's easy for me to check the load impedance up to 100kHz). I'd rather modify the actual speaker loads to get a fairly flat impedance curve, and then use a "text book" second order filter rather than simply change the filter to roll-off at a higher frequency.
As others have said, post-filter NFB topologies (such as UCD) suffer much less than the Tripath amps, but even with UCD, there are advantages in correcting the speaker impedance at high frequencies as a highly reactive load on a UCD amplifier can adversely affect amplifier stability. (I know only too well, having designed a number of UCD amplifiers for driving step up output transformers for 100V speaker driving. These amplifiers have to be stable with ony the OPT attached,and no speakers plugged in.)

 

[kd5eet]

Has anyone tested the 2*100 or 4*100 on 12 volts and found power before clipping?

 

[robob]

hello,
I am thinking of buying this amp board, I would like to know : I'm also thinking of buying a module DAC kit with Analog output specifcation 2V rms. @75 ohms. My knowledge in electronics are as good as my English : can i connect directly the dac to two of the amp input ?

 

[dr_vega]

hello,
I am thinking of buying this amp board, I would like to know : I'm also thinking of buying a module DAC kit with Analog output specifcation 2V rms. @75 ohms. My knowledge in electronics are as good as my English : can i connect directly the dac to two of the amp input ?

Yes, you can connect them directly. But if your DAC doesn't have a volume control, neither does the Sure. Things will be loud.

-dr_vega

 

[robob]

hello,
I shall use my computer to control the volume.
Thank you for answer.

 

[wwenze]

Which side?

I'm going to solder on an extra 2200uF cap for each channel and am not sure which side is nearer to the amp chip electrically-wise, the side with the film cap acting as bypass for the 330uF caps or the side next to the output filters?

Also, have read about the side effects of parallelling different caps, can anyone point me to somewhere?

 

[dr_vega]

I'm going to solder on an extra 2200uF cap for each channel and am not sure which side is nearer to the amp chip electrically-wise, the side with the film cap acting as bypass for the 330uF caps or the side next to the output filters?

Also, have read about the side effects of parallelling different caps, can anyone point me to somewhere?

The side next to the output filters is closer to the power amp chips.

-dr_vega

 

[chaos_maniac]

Hi there, just registered to this very interesting forum.

I'm currently modding both my Lepai TA-2020 aswell as my Sure TK2050 2x100W. I've come to the point where I replace the filter inductors, but stumbled upon an issue I hope you can help me with.

Background:

I bought a lot of 10pcs 10uH toroids, and seeing as newbie-ish I still am with class D-amps I bought cheap ones. What I realised now afterwards was that I see people all over using red iron powder toroids, while the ones I bought seem to be light green ( #12 material according to iron powder material properties ).

Now with background in place, to my question: Will the TK2050 or TA2020 work well with these inductors or is it like asking for trouble? My knowledge in inductors is very limited and I don't wish to ruin neither my amps nor my speakers.

Link to the item I bought: Inductor 10uH 10 uH +/- 10% 5 pcs - eBay (item 260512722675 end time Feb-23-10 01:11:57 PST)

I'm thinking of keeping the original output filter capacitors, for the TK2050 that should give me at least an extra 20kHz to the output I suppose just by replacing the original inductors.

Thanks in advance, I don't dare solder these toroids to the boards before I know it's a harmless procedure.

 

[soldersmoker]

You may want to try the output filter swap before you give up on it as it is not any harder to do than what you have done already and will make a bigger improvement. The stock coils are much too high in value and too whimpy in current.

Okay, I finally swapped out the output filters as you suggested, Sendler. The first thing I noticed was that perceived loudness is greater than before. I have to turn down the volume pot on my preamp quite a bit compared to pre-swap.

For several weeks I have been listening exclusively to my ST-70 rewired in triode mode, and I have been really liking what I have been hearing. I've been spinning vinyl 95% of the time. So that's been my frame of reference. I'm chastened and pleased to admit that I'm liking the Sure amp more at first blush. But as I've found with all things hi-fi, it's best to live with the sound for a while before passing judgment, as sometimes an impressive initial hearing later reveals previously overlooked deficiencies.

That said, this f'in amp is knocking my socks off! Delicate, powerful, refined, smooth, musical, it's got it all as I sit here listening.

That is all for now. Thanks to all of you who blazed the trail.

 

[wwenze]

Strange hiss and failed soft-start

Double whammy, oh well...

I shorted the 1uF input capacitors via bare wires under the PCB, yet this gives me a hiss both times I tried. The hiss is there even without anything else connected to the input - just the shorting of the input capacitor, and goes away after I remove the short. What could go wrong here?

Also, I made a manual soft-start with a 270 ohm resistor (the only high-power value I had) between the PSU and the amp, it seems to take forever to charge - the voltage across the amp terminals never goes past 5.3V - with or without extra capacitors. Likely too high a resistance. I'm going to tackle this another way with a DPDT switch disconnecting the amp from the PSU and the capacitor in off position so there's nothing to disturb the capacitor's charging.

But is there any math to tell what value should be used just in case? I'm guessing ard 20~ish for 15000uF considering the PSU trips with over 4400uF normally.

Or I should just heck and just use 4400uF.

 

[chaos_maniac]

wwenze, I suppose that your hiss is due to the bare wire acting as an antenna picking up RF. Why are you shorting the 1uf input cap and just not remove it instead?

 

[sendler]

Short

Double whammy, oh well...

I shorted the 1uF input capacitors via bare wires under the PCB, yet this gives me a hiss both times I tried. The hiss is there even without anything else connected to the input - just the shorting of the input capacitor, and goes away after I remove the short. What could go wrong here?

You are shorting the amp's 2.5v off set trimming circuit to ground through R14. You can remove R14 if you plan to remove the spike suppressor that is under the RCA connector but it still might short back through your volume control depending on it's layout. Because of the single rail power supply and the 2.5v trimming voltage, I think these amps will always need an input cap.

 

[sendler]

New favorite coil

I have a new favorite coil and I think my search is over. The 7443310390 from Wurth. I was very happy with the 7447709003 but then checked their web site before ordering a bulk quantity and found that they had a new coil on offer. It is their newest design and does sound a bit better than the 7447709004 or 7447709003 that I was using and they come in the perfect value right in between those two at 3.9uH. They don't look shielded but Wurth says it is a shielded bobbin and they measure with good shielding. They sound very transparent and holographic, and are the perfect size to drop right in.
Digi-Key - 732-2125-1-ND (Manufacturer - 7443310390)

 

[wwenze]

You are shorting the amp's 2.5v off set trimming circuit to ground through R14. You can remove R14 if you plan to remove the spike suppressor that is under the RCA connector but it still might short back through your volume control depending on it's layout. Because of the single rail power supply and the 2.5v trimming voltage, I think these amps will always need an input cap.

Thanks a lot. I see, I missed this when reading the schematics. Didn't spend much time reading it though.

 

[soldersmoker]

They sound very transparent and holographic, and are the perfect size to drop right in.

"Drop in," eh? You mentioned how easy I'd find swapping out the coils, but I found replacing the coils difficult, since they're surface mount and the PCB isn't a through-hole design. I ended up soldering up some short lengths of wire to connect the coils to the board. Is this standard practice or am I missing something?

My music sounds good, so I'm happy for now.

 

[sendler]

All of the 12X12 Wurth coils drop in

"Drop in," eh? You mentioned how easy I'd find swapping out the coils, but I found replacing the coils difficult, since they're surface mount and the PCB isn't a through-hole design. I ended up soldering up some short lengths of wire to connect the coils to the board. Is this standard practice or am I missing something?

My music sounds good, so I'm happy for now.

Are you using the Wurth coils? They are 12X12mm the same as the ones you are taking out. They are easy to fit right back in. Getting the old ones out can be tough. I think I am going to buy another cheap soldering iron and have someone help me desolder both side at once. Otherwise the stock coils can be desoldered on one side while prying up which breaks the housing on the other side allowing the first contact to rise clear of the pad. Then the other side can be desoldered normally.

 

[sendler]

Redo on favorite coil

I have a new favorite coil and I think my search is over. The 7443310390 from Wurth. I was very happy with the 7447709003 but then checked their web site before ordering a bulk quantity and found that they had a new coil on offer. It is their newest design and does sound a bit better than the 7447709004 or 7447709003 that I was using and they come in the perfect value right in between those two at 3.9uH. They don't look shielded but Wurth says it is a shielded bobbin and they measure with good shielding. They sound very transparent and holographic, and are the perfect size to drop right in.
Digi-Key - 732-2125-1-ND (Manufacturer - 7443310390)

These three coils are all pretty close. I had another chance to really crank it up and am reversing my preference back to the 7447709003. It is a little more aggressive sounding and is therefore fussier about nailing the Zobel but also a bit more spacious. They are all about the same price.
Digi-Key - 732-1238-1-ND (Manufacturer - 7447709003)