Designing a phono stage

Status
Not open for further replies.
Hi,

I would like to say many thanks for such an important site. Very informative, but I've got a question. I would like to build a phono amp that is able to accept MM and MC.
I have some practical experience and would eventually like to build a power amplifier. Like most of the people on this forum I would like good / excellent sound, but I am reluctant to pay the massive sums on the UK high street. Are there any such circuits that I could be referred to?

I would appreciate any help that could get me started.

Thanks.
 
You might consider whether you want passive RIAA or active (generally construed as being EQ'ed via a feedback loop). You can also have a combination of both in the same circuit. Passive is better.
Also decide whether you want solid state or tube.
If you choose tube, I'd recommend staying away from circuits which split passive EQ into two or three sections. It makes the filtering easier on the initial go-round, but when you change tubes down the road, you may find that your replacement tube has a dB or two different gain, which would wreak havoc on a carefully tuned +-.1 dB EQ curve. Go for the all-in-one passive EQ, and you're guaranteed to be in good shape.
Off the top of my head:
Solid state--
http://www.passlabs.com Nelson was kind enough to post the service manual for his Aleph Ono (nice aural pun, there...)
http://www.metaxas.com They've got a passive EQ circuit as a DIY
Tube--
http://www.tubecad.com (or maybe http://www.glassware.com, I don't remember which; both sites are the same guy) has a passive RIAA circuit--two of them, actually
You might also try the Bonavolt site and others off the Links section, here.
There are schematics for various Conrad-Johnson, et. al. pieces floating about here and there, but you don't want to trust them blindly; I know for sure that there's a spurious MV-75 schematic that ain't the real thing. I believe I saw a PV-5 schematic that looked right. Caveat emptor.

Grey
 
Passive RIAA?
How much in the dark have I really been? (Don't answer that, it would be too embarrasing for me). I am more interested in Solid State at the moment, but this information will prove to be very helpful. Thanks GRollins and Geoff.
 
Passive v. Not Passive RIAA

Brownlow

If I am understanding the term "passive" as used here, passive would be a single gain stage with all the RIAA filter stuff in the feedback loop.

Not "passive", or "active" here would have a few gain stages taking care of the different stages of the RIAA curve.

Multiple filter stages come from the idea the the RIAA curve is made up of three filter sections, more or less.

To take a more official look at it than this reply, check out http://www-s.ti.com/sc/psheets/slyt...)+<OR>+(riaa)+<IN>+Page),+[10](riaa)+<IN>+A)&

Looking at that URL, I'm not betting it will work, so go to ww.ti.com and search riaa. The first technical doc you get should be the one you want. You could also right click on that address and use "save target as." That might work better.

The article talks about designing single supply RIAA filters, but that doesn't change things much.

Hope this helps,
Erik
 
An informative application note can be obtained at National Semiconductor's homepage: http://www.nsc.com
Look for AN-346.
Partly based on this, Norman Thagard has constructed a Phone stage described in the magazine Electronics World, april and may issues.
Mr. Thagard's name may be familiar to some of you, just look at the article about the A-75, from mr. Pass.
 
Hmmm...
Didn't realize that this was a tough concept. Sorry. I would have been more explicit, had I known.
RIAA equalization can be handled either 'passively' thus:

GAIN-->EQ-->GAIN

Or it can be handled via a feedback loop, thus:

GAIN-->GAIN
***|--EQ--|

Or by a combination of the two, thus:

GAIN-->EQ-->GAIN
****|-----EQ----|

(Ignore the *'s, as the forum appears to scan out white space characters in excess of one space, and the diagrams collapse without them. They're just there to space things out.)

Passive (the top one) is considered better because you don't have the time/phase delays through the feedback loop. Any of the three topologies will work. Good examples of phono stages have been constructed of each kind. Passive EQ simply reduces the number of things that can go wrong.
Important to note is that the RIAA EQ curve (the classic one, there was a later variation that didn't get widely adopted) is not one curve, but three overlapping ones; one at 50Hz, one at 500Hz, and one a little over 2kHz. Each is a simple 6dB/oct filter section. This leads to the temptation to arrange the circuit thus:

GAIN-->EQ-->GAIN-->EQ-->GAIN-->EQ-->GAIN

with each gain stage buffering and providing gain to offset the insertion loss in the intervening EQ section. (This is what led me to comment on the [non-] advisability of using this method with tubes, above.)
The total gain needed will vary somewhat, depending on whether you plan to use a moving magnet or moving coil cartridge--from 40dB (pretty much standard for average moving magnet cartridges) to 70-80dB for the really low output moving coils. An average gain for moving coils will be on the order of 65dB. This is measured at 1kHz, as (by definition) the RIAA section is a filter and does not have flat response. Insertion loss for a typical (passive) RIAA EQ will be, say, 20-25dB and this must also be added to the overall gain required from the phono stage.
Viewed as three 6dB/oct filters, it's not that bad, conceptually, but proper execution can be tricky.
Historical note, for those who may be tempted to think that passive RIAA EQ is some new tweak developed by the high-end folks...it ain't. Go dig up an ancient RCA manual and you'll find a nice little number in there using the two halves of a 7025 (the industrial version of a 12AX7) with...you guessed it...passive EQ.

Grey
 
I've just been looking at an old article in Electronics
Today from 1993 where John Linsley Hood discusses the
various possibilities in phono stage design, and then
gives a practical circuit. The latter is the same as
the one sold by Hart Electronics, and discussed here:

http://users.classicfm.net/williamshart/phono.htm

If you give me your address I can send you a photocopy
of the article.

Alex
 
Hmm,

it does seem that my understanding with Passive and Active are not quite compatible here. I would certainly say that a Passive phono is beyond me currently. Anyway, I'll say thanks again to all for sharing advice.

On a final note, Alex M does the JLH phono stage sound good? What would you compare it to?
 
JLH phono stage

I haven't put together the complete circuit, but I did
build a hybrid of the RIAA part of the JLH circuit with a
different input stage (taken from another of JLH's ETI
articles, but specifically designed for low noise).

I used this in my preamp for several years; it sounded very
good, if not quite as clear and open as the Audio Synthesis
ADEQ I use now.

I have several very positive reviews of the Hart kit, one in
HFNRR (I think) comparing it favourably with a Naim preamp
at three times the price.

There is an online review here, for instance:

http://www.digitaltheater.com/reviews/seperate12.html

Alex
 
*28* transistors!?!
Whew!
Clearly, the JLH phono stage isn't designed in the simpler-is-better mode...
I just had a second and took a quick look at the Hart electronics site. You'd think for throwing that much silicon into the thing that they could do better than a 26 dB S/N ratio.

Grey
 
Alex,
I understood that it was 14 per channel. (Hoping, anyway. Egad, the idea of 28 per channel is horrifying.) I, for one, don't regard the idea of 14 transistors per channel as encouraging. I do make an exception for circuits that run multiple gain devices in parallel to lower noise and/or output impedance, but somehow I don't think that's the case, here.
It's been a long time since I've seen anyone use the "below surface noise" phrasing. Note the vagueness--surface noise of a new record? of a ragged out party record? of a special test album? It says nothing. It's safest to assume that this thing is noisy, and that they're being vague in order to avoid being pinned down. In the past (and I mean a long time ago) people used to use semantic tricks like this to sidestep measurements that could be meaningfully compared to their competitors' numbers. (Yet another reason not to join the "Specifications Are Everything" crowd--you can't trust many of the published numbers...) Assume that -26 dB is a *best* case scenario and go from there.
Caveat emptor.

Grey
 
Status
Not open for further replies.