advertisement


AD815 vs Starfish

hacker

Delicious and nutritious
I finally got around to hooking it all up today, and was able to do back-to-back auditions with my 2 preamps: the AD815 and Starfish.

The setup:

Ergo IX speakers, 2x NAP-based monoblock power amplifiers, Squeezebox3 streaming audio player. I used the same PSU for the trials: a 25-0-25 Avondale EI core transformer, hexfred rectifier diodes,10,000uF smoothing. The Starfish is powered by +26V from a TeddyReg, the AD815 by local ALWSRs @ +/-15V.

The AD815 has the SRs just a couple of inches from the circuits, uses Rubycon ZLs for decoupling on-board, and has only one cap in the signal path: a 2.5uF polypropylene.

The Starfish is symmetric and uses 2x 12V Tracos, giving +/-24V rails. These are locally regulated by the 14 on-board regs to +/-12V. All coupling caps are bog-standard tants. Feedback caps are 47u MMKs. The 1K resistor in the feedback network has been swapped for a 2K, halving the gain of the Starfish to 6x. It has an ERT, not a NERT.

I approached the listening with as little bias as I could, although I must confess to having preconceptions of the AD815 being smaller, tighter, fewer components, and - most importantly - having only 1 cap per channel in the signal path. That was countered by my feelings about the regenerated 0V of the Starfish, coupled with the carefully laid out 0V routing scheme, the clever application of local regulation, and decoupling in all the right spots. My main bias was that the AD815 was already hooked up to my remote-controlled input selector and attenuator.

Playing a variety of music ranging from metal, prog rock, acoustic female and male, jazz, and electronic was a very enjoyable experience :)

The first thing I noticed was that there's more bass from the AD815. It's deeper and more enveloping. The Starfish is leaner, but tighter. After much swapping and reflection it seems that the AD815 has a slightly bloated feel to it, like the bass is a little too big and just slightly muffled compared to the Starfish. The 'fish does not want for bass, and it is more accurate, easier to follow, and more controlled than the AD815.

The presentation of the Starfish was subjectively larger than the AD815 and I felt myself more immersed in the performance when listening to it. The AD815 just squished the instruments together a little more, compressing the sound ever-so-slightly in comparison.

The midrange of the Starfish is clearly superior. Sharper, more attack, more realism, just better in all respects. Female vocals are expressed with more finesse and clarity, and there's a texture to voices that's missing from that AD815, covered as it is by the slightly thicker lower-end.

I think a lot of the difference between the two preamps stems from the thicker bass, the more rounded, bloated feel to it. The upper registers from the AD815 are amazing, probably slightly better than the Starfish - sounds like large cymbals, hi-hats, tambourines, etc are more clearly expressed and seem to have the edge on detail. The thing is... it's just not as musical as the Starfish!

All of these conclusions come from a slightly compromised Starfish, too. It doesn't have the NERT, isn't to the latest BOM, has an old Alps pot, and crappy 10p tants doing coupling duty. Still, it's definitely my favourite of the two - it's simply more pleasurable to listen to, expresses the music more accurately (I feel), and makes my toes tap more. There's a very good chance I'll be stripping my preamp out and replacing the AD815 board with the Starfish.... hah, who am I kidding? Of course I'll be stripping it out!

Carl
 
Did you have to do that - just as I was about to order some opamps to build up an amp to compare with my......STARFISH!

Mine is NERT'd BTW with Wimas for coupling and 47uF AVX caps polypropylenes for feedback.

So you got a spare opamp I might as well try it as a head phone amp - it should still blow the Headline away!
 
Lol, don't despair just yet - the AD815 is still a magnificent amp. It's just that the Starfish is better :D

It's a bit of a bugger having to go to the trouble of rejigging my entire preamp... ah well, such is DIY. It'll have to wait until next week because this weekend is beer festival weekend in Flagstaff ;)
 
I have my own private beer fest going on at Chez Case right now - yes I have no friends ;)

Thanks for your review of Fish vs AD815 Carl, something which was long overdue and very makes for very interesting reading.

I'm too am thinking the 815 would make an excellent Headamp or the more likely for me as I have the most excellent White Noise headamp a small footprint pre-amp to fit inside a hot-rodded SB3 custome case (when I finish it) to turn it into that integrated source and pre idea from a couple of session back.

Oh I just wish for more time for this sort of thing.

Nick
 
Even though my ad815 and starfish aren't optimal. I prefer the starfish.

I wonder if a buffer-stage before the ad815 would bring them closer in performance?
 
Did you try the nulling circuit ?
Also how does it sound if you lose the 2.5uF and feed the output neat (assuming your NAPs have input caps)?

No, there was no nulling circuit or input coupling caps. My NAPs have wire links instead of coupling caps, so I used the 2.5uF polyprops inside the preamp to avoid having a DC potential across the pre -> power interconnect.
 
Nice write-up, Carl
fatmarley said:
I wonder if a buffer-stage before the ad815 would bring them closer in performance?
Before that I'd try setting the AD815 up with the same input bandwidth limiting as the Starfish... to compare apples with apples IYSWIM.
 
Before that I'd try setting the AD815 up with the same input bandwidth limiting as the Starfish... to compare apples with apples IYSWIM.

I think i've just answered my own question. My old modded 62 (just the gain stage) was much more musical than the ad815 - I may have to get it up and running again.
 
I have the same experience, the 815 chip is not "musical" enough. But I have a new configuration coming up that are way better. It's a combination af a opa827 and the new buffer circuit from national 49600. The buffer is placed inside the feedback loop much like the schematic that you find in the data sheet for the 49600.
It's been up and running for a week and it is very promising. Next step is to bias the 827 in to class A and see what happens. For shure it has lots of qualities and I will publish the schematics when I feel that it is a competitor for the classic 321 circuit.
Stay tuned!
 
Conclusion:
I have been playing around with three different preamps. A Guitariste with ALWSR's regulators, a Nelson Pass B1 with Teddyregs and finaly the one with opa827 and the new buffer circuit from national, 49600, and almost standard LT317/337 regs.
Today I made the final test and I must say that the opa827+buffer amp sounds great and will stay in my system. It is at least as musical as the Naim clone but with better control and separation of different instruments etc. I have also made an experiment and biased the 827 into class A by using a 2 mA FET based current regulator. That is the best way to destroy this amp, it really kills the music.
I'm happy:D
 


advertisement


Back
Top