tripath 2020 with low efficiency speakers

[john65b]

The McCormak is some serious bucks. As mentioned, I would do the UCD400 before dumpinmg that kinda cash. You may also want to look into the Panasonic SA-XR45, SA-XR55, SA-XR70. You may be surprised on how they can make the MMG's sing beautifully. And for a LOT less than other high end amps like the McCormak.

Currently have a XR45, getting the UCD180 and getting the XR55 (Bi-ampable, amoung other cool things). The Amp that sounds best to my (and my wifes) ears on the MMG's get permanent placement in my main music system and playtime, and the others get posted up on Audiogon/ebay.

I anticipate the best is UCD, then the XR55, and then the XR45.

We will see. At under $250 for each of these three for the big TEST, it won't kill my wallet...

[soufiej]

I have the Italian 2020 amp built from a kit and have tried several speakers with the amp. My most frequently used speakers during the last two years have been a pair of Rogers LS3/5a's (manufactured somewhere around 1979, so the original 15 Ohm version). Using both the SI T-amp and the Autocostruire amp, the 3/5a's never completely "light up" on all music. I also have used a pair of NSM 5s's and some inexpensive KEF Cresta 1's with these two amplifiers. The KEF's are the most efficient and the simplest load on the amplifiers. The 3/5a's might be a very tube friendly speaker, but the chip amps seem to have a different opinion of the BBC design. It isn't too surprising then to say the el-cheapo ($199 pair retail) KEF's are the best sounding speakers I have so far paired with these amplifiers. While they lack a bit of finesse and three dimensionality that the Rogers can impart to the music, the KEF's present the music in a very convincing fashion that is not that far off from the other two pairs of speakers. Possibly the fact the KEF's are ported and the Rogers and NSM's are sealed enclosures has a bit to do with the sound quality when paired with the chip amps. Both AS speakers have a bit of bloat in the bottom (as far down as they will go, at least) when used with the chip amplifiers. The lower efficiency of the AS designs also means the amplifiers are always closer to clipping and have less headroom to deal with the speakers' needs.

The Rogers amd the Autocostruire amplifier will fill my (21x17x9.5') room with an adequate volume level and I could be happy with the results on most music. As everyone else has discovered, the 2020 amplifier has a warmth that brings it much closer to my McIntosh tube amplifiers. But with the 3/5a's, the Macs, at about 90 watts X2, are the better choice. However, the 2020 amp will fill a desire for an everyday, almost 24/7 amplifier that will not use precious tube life. I haven't paired the Autocostruire amplifier with any of my pre amps as of yet, so there is still more experimentation to come.

I ordered a 12VDC, 5Ah VLRA which arrived yesterday and got placed in the system last night. (No switches, no cap.) The sound with the Rogers was immediately improved over the Duracell NiMH rechargeables I was using. Enough so to make me reconsider my plans and stay up another few hours listening. PRaT was improved to the point where my foot finally gave out. It really did sound like a difefrent amplifier had been placed in the system. The jump in performance from the stock $25 SI amp to the (slightly over $100 USD) Italian kit was one step forward. At this point, the change in power supply is another hurdle toward making the Autocostruire a bit more like the Macs.

So far the amplifier is stock and configured as instructed from Autocostruire. As I gather some more information regarding what I want to mod before the amplifier goes in an enclosure the next thing I'll do is to simply remove the output Zobel network and the diodes as suggested by Autocostruire in their "minimalist" version of the amplifier. This is a simple step I've not seen discussed in these pages. Have I just not found the information yet or does anyone have an opinion on the value ofretaining the above mentioned components?

[macpaulster]

I am using a Tripath 2020 amp made by HLLY electronics and available at http://www.p-macaudio.com

I am using some Klipsch bookshelf speakers (KSB 1.1) and the sound is absolutely wonderful. I've also tried these speakers with a Tripath 2024 amp, and I DO notice the difference in power, though slight (the 2024 amp also sounds wonderful with these little speakers).

My opinion is that in using these low power amps, speaker efficiency is very important. If you have some lower efficiency speakers (below 90db sensitivity), that you really like, you should certainly try them, but I think in most cases you'll get better results with high efficiency speakers.

[albin]

Quote: next thing I'll do is to simply remove the output Zobel network and the diodes as suggested by Autocostruire in their "minimalist" version of the amplifier.

hello
As far as I guess all these chips use the same die(internals).
The ta2024 datasheet shows only four output protection diodes instead of the eight shown for ta2020 and ta2021b,. this accounts for its power derating,
Be careful, or at least learn what these "Bits" are for before removing them
Same applies to the Zobel network.
regards

Max

[soundcheck]

Hi there.

Please consider that the distortion is extremely high at high volumes if you run low SPL speakers. The same for high R speakers, which will reduce the power capability.

The little thingy won't deliver the energy needed on the lower end on most setups.

Even on my high SPL setup it's not enough.

BTW: I ran Charlize (modded). Now I am running a T2021B type - which outperforms a 2020 in terms of delivering much more details.

I am running 98db/SPL speakers at 16R. ( The output filter needs to be adjusted accordingly)

I am only driving my fullrange speakers with it. For the low end I am using high wattage (icepower modules)

What I've also done. I put a small size cap (0,1uf - you can calculate your own value easily)
as coupling cap. This will free the amp ( and my fullrange speakers) from low frequencies. You'll see that it likes it. You also get high quality caps at low values for little money. Low value caps will always sound better then bigger caps.

I am running a Northstar battery (ESR 2mR) 90Ah buffered with 8*470 low esr type caps and a small filter.
The better (low ESR, low aging/degradation over time) and bigger ( high fill level is important) the battery the better the sound. That's what I figured.

Cheers

[macpaulster]

Soundcheck, great info! On the coupling caps how do you calculate the value, and where do they go on the board, prior to the input? Do you have pics of your setup? How is the power rating of the 2021B chip compared to the 2020?

[Saturnus]

Originally posted by macpaulster
On the coupling caps how do you calculate the value, and where do they go on the board, prior to the input?

C = 1/(2*pi*f*Zin)

C = Capacitance in Farads
f = Corner frequency in Hertz
Zin = input impedance in Ohms

It depends on the specific amp in question but you must replace the existing input cap. And if there's additional input filters like on the t-amp then you must bypass these. Note that corner frequency should generally be at least 2 octaves lower than the lowest frequency your speaker can produce, ie. the speakers -6dB point. If used to limit bass response or as a highpass filter instead of a real x-over it can be higher but you're probably better of by adding a parallel resistor after the cap to form a real passive filter.

Originally posted by macpaulster
How is the power rating of the 2021B chip compared to the 2020?

Technically they are the same but the 2020 because of it's packaging can be made to run stable in 2 Ohms loads something which the 2021B or 2024 cannot. Generally a 2020 is better in the bass region while the 2021B has better imaging but it's minor differences at best and mostly depend on other factors like schematic design, board layout and quality, and accuracy of the components used. Both the 2020 and the 2021B is generally considered superior to the 2024 in all aspects, though.

[valleyman]

I have driven my ZD5s which are around 82dB efficient in a small room with a ta2024. I didnt push them really hard but I never got it to clip

[soundcheck]

Saturnus:
The cap builds a highpass with the series resistor of the input stage !?! What else do you need to form a highpass?




BTW: Newer "implementations" of some Tripath amps get along without coupling cap! Though this would need a complete separated powersupply (source and amp). If you run on batteries this is not at all a problem. Of course you'd loose the high-pass filter effect.

Sometimes it happens that you don't like what your hear, since the coupling caps usually hide some deficiencies of your chain.

[Saturnus]

Originally posted by soundcheck
BTW: Newer "implementations" of some Tripath amps get along without coupling cap! Though this would need a complete separated powersupply (source and amp). If you run on batteries this is not at all a problem. Of course you'd loose the high-pass filter effect.

Unless you happen to run both the source and the amp on battery, in which case you need 2 seperate batteries.

Technically you don't need anything else than the input coupling cap and the amplifiers input impedance for a highpass filter but try putting a very large resistor (10M Ohm works fine) in parallel between the input cap and input resistor to ground. Somehow bass response clears up significantly although technically having the same x-over frequency.