Possibly the most frugal high-end sounding amp?

[opus111]

I've recently diverted my attention from the Ozone DAC building onto designing a simple and frugal amp which matches the tonality and soundstage depth of my modified Xindak thus doing justice to the Ozone. Having the same level of output power isn't a requirement though, so I have been searching far and wide for chipamps which might have the timbral accuracy and 'bloom' I desire.

Having built chipamp based amps for a number of years, they produce great but not stellar sound in the main, when issues like grounding and decoupling are addressed fully. I have the hypothesis that what holds them back is the input stage, being as it is typically a LTP and a bipolar one at that. Added to that, they do tend to have inadequate PSRR by virtue of not having power supply separation between the signal and output stages. So they pollute their own patch, so to speak and filters can't be included because the necessary wires aren't brought off-chip.

I have up until fairly recently, played with the TDA7293 because its fairly unique in allowing separation between the voltage and current amplifying stages, especially so when the amplification function is split between two chips. This device is the only one I know of that permits that so the power supply issue can be fully fixed. It does however still have the ubiquitous LTP input stage which has given me so many problems in opamps in terms of excluding HF interference.

The search for a chipamp sans LTP has been running for a few weeks now and recently I've come upon some very interesting parts from Philips - the TDA856X series. The reason it has taken me so long to find these is I was looking in the 'audio amplifier' sections of manufacturers' websites, whereas these are in the 'automotive' section by virtue of being designed for car radios - thus they're not considered 'audiophile' in any sense and aren't particularly high powered. No matter for this application, its the SQ that's important, and based on a couple of days' listening only it seems I might have come across a diamond in the rough.

Details on my prototype TDA8561 amp to follow sometime, stay tuned

 

[steve williams]

Hi Richard

I am always interested in reading about your projects

 

[opus111]

Before I get to talk about the amp itself, first a bit of back story about the chip. Here's a block diagram culled from its datasheet:



At first blush you might say - "well it looks like there are opamps inside the chip, so what's making this any different from other chipamps?" 'Look closer!' I retort - those triangular shapes on the left side aren't your normal LTP input opamps. How I know this is because when I saw this block diagram I did a deja-vu with another amplifier I recently pondered over - this one :



The AD830 is a largely unsung innovation from Barrie Gilbert, one of analog design's true artists. Analog Devices really hasn't developed this architecture very much, only having about 3 chips based on it in their range. Linear Technology also has some chips using the same topology, for example the very frugal (but seemingly totally unknown in audio) LT6552.

 

[opus111]

I am always interested in reading about your projects

Thanks Steve - you might be interested in building this one then? I figured I'd start this project on here as its far less a daunting one for beginners than the Ozone DAC. Then once someone has this under their belt they might be ready for the Ozone as a next step. I was surprised how easy it was to wire up one of these chips and get superb sound. A kid could do it

 

[mep]

Having built chipamp based amps for a number of years, they produce great but not stellar sound in the main, when issues like grounding and decoupling are addressed fully.

How can the sound be "great," but not "stellar?"

 

[opus111]

I'm not sure if your question is about my listening experience or about the circuit details - care to clarify?

 

[mep]

I'm not sure if your question is about my listening experience or about the circuit details - care to clarify?

Take your pick. You made the statement that you have built chipamp based amps for a number of years and they produce great but not stellar sound. So you need to clarify what you mean by that. Whatever the case, I was just wondering how something could produce "great but not stellar sound."

 

[opus111]

I'll talk about the listening experience then - the chipamps I've made in the past have been LM3886 and TDA7294 types (and the variants thereof). These I've driven with DACs which haven't been top-flight because only recently have I designed a transparent enough digital source to reveal their shortcomings. So 'great sound' means good but not top-notch soundstage depth, good but not top notch tonality and good but not top notch dynamics.

Any clearer now?

 

[opus111]

A couple of important documents on opamps

A chipamp is really just a powerful opamp, so these two papers concerning opamps and their non-linearities have considerable bearing on the use of chipamps in high-end audio.

The first is by Barrie Gilbert who I've already cited - its quite a technical one and I'm not going to go into those arcane details he covers. Suffice to quote his second-to-last paragraph because there he's talking about the architecture I've already introduced. Here 'OPA' refers to the traditional opamp architecture (including the problematic LTP) -

Later, another fundamental building block, which I call the
Active Feedback Amplifier, or AFA, will be discussed. In my
opinion, this structure, which, unlike the OPA, has high
open-loop linearity and excellent closed-loop linearity due
to its distortion-canceling topology, and a very high degree
of versatility arising from its dual fully-differential inputs,
has the potential to eclipse the OPA in all applications
involving the manipulation of purely voltage mode signals.
Being a super-set, it can do anything that a conventional
op amp can do, plus a whole lot more.

http://www.linearaudio.nl/Documents/Are Op Amps Really Linear.pdf

The second paper is much more recent and covers real-world issues with LTPs in opamps which Barrie Gilbert doesn't address, maybe he wasn't even aware of. These issues have probably only come to light because of the greater degree of electromagnetic pollution in today's world - we live bathed in GSM and WiFi signals.

 

[opus111]

Here's a quick shot of the first prototype in action. I've decided to baptize it 'Nitro' for the dynamics. The power supply (offstage left) at present is a 19V laptop switching unit, providing around 60W max

 

[opus111]

I've built and am listening to the second prototype now :



Variations here are the input CM choke (which was off-board in the last one) and experimental PSU capacitor arrangements. The pic doesn't show it clearly but the caps are on 'blades' sticking out perpendicular to the PCB.

Here are a couple of prototype capacitor 'blades' that I've been building. The idea is to parallel many small caps (here they're 100uF/25V) to get lower inductance. The PCB is double-sided, one side +ve the other, -ve.

 

[opus111]

Here's an informative technical review (I'm ignoring the subjective part of it, for now) of a PC card amplifier by Charles Hansen - presumably the Charles Hansen of Ayre fame. Its not your typical Stereophile style technical review with traditional results from an AP, rather Charles has his own approach and its considerably more enlightening about what's going on with the product :

http://www.marchandelec.com/ftp/revpm48.pdf

What makes this relevant and interesting here is that the chip being used is a very close relative of the TDA8561 - its a TDA8566. I'll post up my observations of this article in a subsequent post....

 

[Kindhornman]

Richard,
I still need to read the articles here but have a question. Isn't there any way to add an external NFB circuit to some of the earlier mentioned opa's with the ltp outside of the internal feedback loop? Probably a stupid question, but I had to ask. I still don't understand why they would leave this section out of the loop, but somehow that seems to be the standard as you have posted. I will have to read the articles and perhaps the answers will be there.


Steven

 

[opus111]

Hi Steven

I can't quite grasp your question here. You want to continue using normal style OPAs but somehow bypassing the LTP? Its possible on the NE5534 because there are internal pins brought out which just happen to be the right ones for connecting an external LTP. But then again, why make life complicated when its possible to get a complete chip which has an overall architecture better suited?

As to why they'd leave the LTP out of the loop - I think its perhaps because they haven't realized it is out of the loop. If they have realized then probably they think that its only an issue at DC - because offset is the only thing out of the loop - and haven't understood that RF is varying the offset voltage. Or that they think RF isn't an issue at audio. Just my speculation!

 

[Mosin]

<snip>recently I've come upon some very interesting parts from Philips - the TDA856X series.<snip>

Philips makes some chips that can withstand far more abuse than chips that were made only a couple of years ago. They are extremely rugged. Three Philips surface mounted chips are used to regenerate three-phase power for the entry-level controller of the turntable I build. Those chips work very well in a power application where the demands are relatively high, so I suspect that most, if not all, of the similar chips made by Philips will be virtually bulletproof in an audio circuit.

 

[Kindhornman]

Richard,
Let me try again. Is there a way to add a GNFB loop outside of the chip that would then capture the LTP part of the circuit? An external feedback loop around the opa from output to input as you would do with something like the 5532?

 

[opus111]

I can't think of a way, no.

 

[opus111]

I'm busy drawing up the schematic and preparing the BOM for Nitro revA. Its a work in progress as I haven't filled out supplier details but here's the link to the BOM - I'll finalize it when I post up the schematic over the next couple of days.

https://docs.google.com/spreadsheet/ccc?key=0AspUyZApcGQhdFFiVW1ZZE1pb3ppX2NQN0hiMXgzdHc&usp=sharing

 

[opus111]

Here's the revA schematic. Points of note are:

a) the star grounding technique is essential to getting the optimum SQ - I experimented with arrays of electrolytics which ended up sounding worse because I had to compromise the grounding slightly because of physical constraints.

b) the CM choke needs to be of segmented construction to get best SQ. I tried with an off-the-shelf one and it flattened the soundstage considerably. This is a part you'll have to make yourselves as I've yet to find a commercial version. Or alternatively use a linear supply which generates very little common-mode noise. I'm currently using a very cheap ($7) laptop PSU 19V/3.5A.

c) the large numbers of ceramic decouplers on the PSU are necessary for getting the cleanest sound - particularly noticeable in the HF. My first prototype had only 4 ceramics on the supplies, shared between two channels, coz I figured that would be enough. But I didn't care for the sound at all - going up to 8 improved it, going to 16 made it still better. So I figure the more the merrier - soldering these directly on the pins of the IC is what I do to get the least possible inductance.

d) the bridge rectifier is to drop the 19V down to below 18V so the supply remains within the working range of the IC. Going above 18V can force the amp into a mute condition but won't blow it up. It also has the added benefit that you don't have to worry about the polarity of the external supply at all



<edit> I'd like to get this uploaded bigger but even when I try to add it as an attachment it still insists on a paltry 9kb size. Anyone got any ideas?

Until such time you can find the original larger schematic on my blog : http://www.diyaudio.com/forums/blogs/abraxalito/960-high-end-chipamp-build-project.html

 

[opus111]

Here are a couple more details about the CM choke. Its wound with around 2m of twisted pair wire, the gauge being approximately 26AWG (my wire is metric sized at 0.42mm inner diameter). I twist wires using a hand drill, with the remote end of the wire pair hung over a door handle. Here's a pic showing a few I've made - I have yet to find an international supplier of the former, this one is available on taobao though for anyone with the means to shop there.



Taobao ordering link: http://item.taobao.com/item.htm?spm=a1z10.3.w1284409897.37.lrvPrp&id=4576296793&