Hackaday Prize Entry: High End Preamps

While compact disks are seeing an uptick in popularity thanks to a convenient format that offers a lossless high-quality 44.1 KHz sample rate with 16-bit depth, some people are still riding the vinyl bandwagon of 2010. With that comes a need for the best hardware, and that means expensive cartridges and preamps designed by someone who knows what they’re doing.

For this year’s Hackaday Prize, [skrodahl] is building a really, really good preamplifier for moving coil turntable cartridges. It’s already built, it’s already tested, and the results are good: it produces between 36 and 46dB of gain, -110dB of dynamic range, and a signal to noise ratio of 79.46 relative to a 5mV input. That puts this preamplifier into the same territory as preamps sold with serial numbers, crystal lattices, and other audiophile nonsense.

The quality of this preamp comes from the design, and like any good open hardware project, [skrodahl] has made the schematic, PCB, and layout of this preamp completely open. It’s a great preamp, and a great entry for the Hackaday Prize.

37 thoughts on “Hackaday Prize Entry: High End Preamps

    1. That’s a particularly good question because this amp seem a little at odds with what I expected.

      I would have expected that there are four wires coming from the pickup to the pre-amp so that the first ground point is at the pre-amp itself. I would also have expected balanced inputs and a differential input to the op-amp or in other words current loop sensing.

      There is only one ground point for the input.

      Apparently it has good specs in any case. I would have also liked to see some decoupling resistors on +ve and -ve and also 1 meg resistors are a bit high unless things are going into a metal case.

      1. I like the specs, especially this line:
        Crosstalk: N/A – The MH-1 is dual mono
        which is a load of manure. There will be crosstalk – not a lot of it, but there will be. And this is not a dual-mono design, as it shares power supply (on the board too). True dual mono amplifiers don’t even share PCB. I also don’t understand that aversion to input and output capacitors – they don’t mess up the audio bandwidth as long as they are selected properly. For this project single NE5532 would be enough. Anything more is just audiovoodoo…

        1. I agree with most of what you said but there are reasons for keeping capacitors out of the signal loop.
          Here is some interesting reading material about distortion products introduced by caps:
          http://diyaudioprojects.com/mirror/members.aol.com/sbench102/caps.html
          Even on his very basic test setup you can see how caps can distort the signal.

          I used to read rec.audio.tubes and alt.guitar.amp newsgroups, and every now and then discussions would flare up that contrasted the different ideas people had about components, especially using “unobtanium”, for which they would pay hundreds of $.

          Some audiophiles claimed that they could hear the difference between a _silver_ **power cord** and a normal copper one. That’s of course hogwash, because to start with your house is full of copper but the last 1 meter should make the difference?
          At that point it becomes more psychological.

          1. Some years ago. O.K., some decades ago, I went to a very highend audio seller ($40K turntable, $5K cartridge, $10K moving coil pre-amp, etc.) and ended up sitting at the back while another customer had a demo to compare two sets of speaker cables. The setup? A stereo system across the end of the room, such that the turntable was physically the highest, then moving coil pre-amp, then pre-amp, then amp, then speakers sitting on the floor, such that all of the cables were tight enough to ensure that the cables only ran from higher-to-lower physically as they cascaded down the chain, this to “ensure the electrons only ran downhill”. Something wrong there… and not with the hardware.

            There’s a lot of alleged audiophile voodoo items where I’m left waffling to notice a difference, then label it, and then figure out if it’s “better” or not. But then there’s some for which I have personally experienced a meaningful, even outstanding, A vs. B difference.

            One example (1979) was a moving coil pre-amp: wired internally with copper, or silver or gold. We opened them up, and only the internal wire was different (all soldered with silver). The gold/silver versions sounded meaningfully (but not significantly) better than the copper version. Gold and silver versions sounded different from each other; none of us could say which was “better”. Best guess on why the difference, was capacitance of the wire, but no one was confident on the why.

            Another example, around six years ago, was a huge difference with DIY speaker cables. First with the typical 18/2 stranded copper lamp cord vs. a jaw-dropping detail & clarity difference going to cross-connected telfon insulated Belden antenna coax. The sound difference was ‘OMG’ with the TV still streaming through the stereo and then with a CD music source it was astounding. OMG level of detail & clarity and I heard all sorts of details in recordings that I’d never heard before.
            So I tested the cheaper version of the coax: how on earth is there an overwhelmingly significant quality/detail improvement – when it’s the same speced coax (same wire, both with teflon insulation), except the significantly better sounding one has teflon sheathing – sheathing that never touches the conductor!!! wtf? Something is going on, but I haven’t a clue what.
            And then, a much cheaper unknown speced stranded copper “audio” wire, costing little more than 18/2 lamp cord, but it sounds nearly as good as the DIY cross-connected coax for a tiny fraction of the cost.

            Based on the alleged cap mechanical/magnetic issues degrading audio sounds, for an experiment, I de-packaged the two film caps (the only caps) in an external crossover. Then the treatment.
            Teflon heat-shrink on the leads. Wrapped firmly in organic cotton, painted until soaked in shellac, wrapped (firmly, but not compressed) with hemp cord soaked in shellac. Wooden XO box, lined with cork floor tiles. Separate compartments made for each XO component, divided by cork floor tiles secured with contact cement. Wired up, no soldering: connections made with brass bolts tapped into the wood above the component compartments. Placed pieces of bees wax into the air space around the components, then poured hot bees wax into each compartment to encase each component in bees wax. Goal/result is each component in the XO is physically “heavier”, damped & vibration decoupled between components. Plugged the XOs in. Sound quality went from “I need to upgrade my amp” to “I’m never selling that amp!” Astounding difference in detail, clarity and noise.

            Then there’s a meaningful, but not significant, improvement in sound in a soldered joint vs. high-pressure crimped joint. Better detail & clarity when using crimped joints. My best guess is that the soldered joint is introducing some micro resistance and micro capacitance that have an (eddy current?) RC circuit affect that smears the detail in the sound???

            So given my first hand personal experience, I can entertain the possibility that the quality in sound is there, and would be improved by separate mono units. I’d also want to better secure and damp the caps, if not do the full de/re-packaging treatment.

          2. Canoe, did you do double blind tests? In world of audiovoodoo there is psychological bias that alters perception of sound based on materials used or price tag of equipment. Above certain price threshold differences between amplifiers, preamps, turntables, CD players, DACs for lossless audio playback or sound cards are based solely on price and brand name. And unless your cables are pieces of crap, they won’t have impact on sound at all. Human ear is not a very good instrument. Everything with the distortions below 1-0,5% should sound the same for most humans. But if you include the price tag and specs when comparing equipment by ear, more expensive and less distorting things will always sound better. Because you know they should. That’s why extreme audiophiles that spend absurd amount of money on their toys hate double blind tests so much. They don’t want to look like bunch of nitwitted suckers when their “superior” equipment turns out to be not better than setup for 2000USD or less…

          3. @ Canoe:
            wow, you put in some serious effort!
            Regarding the speaker cables, perhaps the coax speaker cables had less inductance, so that the higher frequencies were damped less? I usually associate ‘clarity’ with ‘presence’, or the highest tones within the music.
            Whereas ‘detail’ I usually associate with ‘channel separation’.
            I have seen DIY projects where they used UTP cable, but then something like 40 strands in parallel per conductor, each strand pair twisted at different turns per meter.

            I have also heard of capacitor microphonics, as you indicated, but I didn’t expect insulating everything mechanically to make such a big difference

          4. Re. I know the bias, as I’ve seen that so many times, where money is spent and therefore there has to be an improvement, even if it takes days to try to identify a difference…

            With the cross-connected Belden coax, I was primed to expect a meaningful improvement and expecting to do an A vs. B comparison in order to see if the reviews on the forums were correct (hell, I was so swayed I bought the last of a whole spool). From listening to me, my friend was expecting to hear a difference but skeptical if it was really going to be better or not. But the third person, who cared nothing about the stereo sound, was out in the kitchen and didn’t know we were doing anything, heard the difference from there and came out to see what had happened. I was being lazy and switched while the stereo was running, first the left speaker cable, then the right side; there was a big delay in doing the right side as we were so blown away by the sound coming from the left, and that’s just with the TV as source; AND, without proper terminations – just twisted the conductors together as a quick test. The difference wasn’t listening A vs. B, and trying to decide which was better, it’s an astounding night and day difference. And what I haven’t noticed mentioned, is the nasty bias created if the volume is higher, as people will conclude that the slightly louder one is “better” without noticing that it is louder. In this case, the volume was – noticeable to me anyway – a tiny bit louder, but that was entirely overwhelmingly eclipsed by the sound quality improvement. Once both speakers had the new cable, we went through song after song to hear all the things we’d never heard before. Blown away for hours. And the same the next day and the day after that.

            I’d heard various speaker cables (and other components) over the decades, most I wouldn’t take if they were free (not worth my time to hook them up), and I’d had my eye on a bi-wire set that I could hearing a meaningful improvement with but not enough to be moved to spend the money. But with the X-connected Belden Coax it was OMG. It was better everything. You don’t need a double-blind to get someone to distinguish between a clear blue sky vs. overcast by cloud.

            With the cheap speaker wire, I’d expected it to be able to play sound, but similar to 18/2 lamp cord. It got swapped in when a friend who was moving across the country talked me out of my X-connected Belden coax set as I could wire up a replacement set for myself. Only the cheap speaker wires are so good that with other priorities, in two years I haven’t been driven to get around to that yet. To assign a subjective number (which is problematic, but here it is anyways): I’d rate 18/2 lamp cord at 1, X-connected Belden regular sheath coax at 3, the stupid-cheap speaker wire at 7-8 and the X-connected Belden teflon coax at 10. Every other speaker cable set I’ve heard I’d rate at 4 or under. From the forums, the consensus is there has been better than the X-connected Belden coax available for a number of years now, finally from manufacturers (but no where near all of them; many are still crap and/or snake oil) and also in DIY designs (where there is also crap and snake oil), and providing that performance across a wide range of amps and speakers.

            With the treatment of the XO, the end result is also mind blowing. But all the XO changes were made at the same time. I have no idea which was responsible for the improvement: encasing each component separately in beeswax, decoupling by cork or the bizarre treatment of the caps. Shame I did all at the same time, as I could have not flooded the compartments and tested like that first.

            Generally, coax is the last thing you’d want for speaker cable. And while I’d highly recommend the sound quality one gets from the cross-connected Belden 89259 coax setup, I can’t recommend that design as the forums, and that design’s author (Jon Risch), were clear that there are better available (commercially and DIY) since that design first came out, and if I recall correctly, with some designs for an even lower DIY cost.

            p.s.
            On the last meter bit: that is suspected to be involved with filtering and stability/contamination in the power supply. Ultra stable rail voltages are related to reproducing the input signal as accurately as possible and to not introducing noise. Like the IEEE double-blind tests on speakers years ago, it’s hard to find an electrical measure of what is going on that corresponds with what people perceive as poor vs. good vs. better quality. With the speakers, they found that smooth lines on the graph of a speakers frequency response would result in a speaker that people perceived as better than a speaker that had a graph with lines that had sharp artifacts. The difference was attributed to cabinet vibration induced noise contaminating and/or distorting the sound. I’m sure I’m sure I’m not remembering their wording, but you get the idea.

            Like such a major difference, and in obvious sound quality, in two DIY speaker cables of identical design and implementation, where the only difference is the sheathing material, and that never touches either conductor, nor the sending nor the receiving unit grounds; like wtf? I’d read it, didn’t believe it, expected a minor difference that I’d have to hunt to describe, and then I haven’t a clue how the sound quality could be so obviously & hugely better for a minor physical difference in a component that never even touches a signal conductor.

          5. Well, I’ve found a website that seems to describe how to make what you were talking about:
            http://diyaudioprojects.com/Power/Belden-89259-Speaker-Cables/
            So perhaps in the future I’ll have a go and cobble some together, thanks for the tip.
            The sound difference that you described reminds me of my first tube amp, which was a Philips PA system, only 15W or so per channel tops, that was at least 30 years older than me. Only 3 or 4 active ‘elements’ in the signal path. (well.. not counting my computer sound card of course). Hearing instruments and nuances that on other amps weren’t perceivable. In my mind that feeling of discovery is worth putting a bit of effort in to get a good setup, even though some times the ‘improvements’ are a bit of a dead alley.

          6. That’s the basic design, but not the original author’s take on it. I use teflon heatshrink, with a half cm length of the coloured on top for ID. Just because since I did the teflon vs. other test, I’ve found a trend that teflon coated is almost a holy grail tweak for audio signal path wire, whereas PVC, foamed or not, seems to be a it’s-going-to-suck guarantee.

            That’s the dual cross-connected design. There’s also a quad cross-connected design, also by Jon Risch, apparently with improved bass response (my speakers don’t go low, so I haven’t tried a quad cross-connected). Add that name to your search. One of the things you should find, as I noted above, the consensus, including from Jon Risch, is that since the X-connected designs were released, there have been better options/designs available. I found the threads, and contact with Jon, through http://www.audioasylum.com/audio/cables/bbs.html. I haven’t been there for quite a while, so I have no idea what you’ll find. That Belden cable was also getting both rare and pricey. Do NOT use the cheaper non-teflon sheathed cable version.

          7. And this DIY cable makes everything sound better because? Magic?
            I’d call that bullshit, but there is no point. Believers will believe, and everyone else knows that already…

          8. Well.. a hobby is per my definition as something that you put time and money in and you get nothing back but satisfaction. If those people want to put in a lot of money and get lots of satisfaction because THEY can hear it, then I’m completely happy for them.

            I’m keeping my mind open about these things, and am willing to try this or that, provided that it doesn’t cost me too much money. Who knows, maybe someone is on to something? There are a lot of people who can’t hear the difference between 128k MP3 and CD quality (or don’t care). There are a lot of people who can’t hear the difference between budget lamp cord speaker cables and better 2mm thick loudspeaker cable. But I can. So perhaps the more ‘outlandish’ claims are based on something as well.
            Who knows, perhaps there ARE some differences that can be measured, but we’re measuring the wrong thing? Perhaps only between 20-20KHz, because ‘that’s all the human ear can hear’? Or perhaps by just measuring a frequency spectrum, you’ll miss things like impulse response? (Just guessing here).

          9. > And this DIY cable makes everything sound better because? Magic?
            Not interfering with the signal as much as other cables.
            It is cross-connected, so noise to the cable will be largely cancelled, so less noise gets to the speakers and less back to the amp where it would be picked up by the amp’s feedback loop, so overall less noise and less distortion of the signal.
            As stated previously, what bothers me is the difference when using the teflon SHEATHED Belden 89259 vs. their sister cable that uses a different sheathing; both cables use the same conductors and are insulated by teflon (or foamed teflon, I forget). Wish I had a scope. But that would only let me see the character of the difference, not provide me with an explanation of why there is a difference. As to the why, or how, the sheathing could make a difference, I haven’t a clue. I recall hearing proposed explanations, but none I heard stuck with me as credible.

          10. > Are you sure this isn’t better? It’s more expensive than teflon, that’s the main qualification right?

            For too many of the manufacturers, that is dead on. They need things to cling to to justify their product line and price.
            They get really pissed when someone comes out with something that can be DIY for $150 and outperforms their cables at $1000, $2000 and $4000, or more.
            Which is what the X-connected Belden 89259 cable design did back in its day. Some credit it with getting a lot of manufacturers to come out with products that performed a lot better, instead of just costing more.

  1. Most of these stats are nonsense. -110dB dynamic range? 70dB SNR? Who cares! If you’re below about 45dB you’re signal will already be limited by the source material itself. Vinyl has pretty crap stats.

      1. Not at all. Vinyl masters prepared for the medium may sound better to than the digital masters typically used. Claiming anything otherwise is just simply stating you don’t like to listen to good quality audio and prefer it to be butchered by a lossy recording medium.

    1. That depends on the quality of the record. Remember CD4? That recorded the rear channels at frequencies above 20Khz. The circuit employed a couple of bandpass filters and frequency shifters to bring them down to audible range. There was only about 4.5~5Khz bandwidth for the rear channels but still quite the technological achievement.

      Of course only using a CD4 stylus and God level cleanliness was required to prevent destroying the fine high frequency of the records. Dropping a plain old elliptical stylus on a CD4 disc was a good way to move even the most peaceful audiophile to thinking of committing acts of violence.

      Even ordinary stereo records could be pressed with frequencies over 20Khz and there are pickups that can reproduce such frequencies – for a while. Eventually even the most obsessively cleaned record will get the high frequencies worn off.

      Direct Metal Mastering made vinyl even better by removing some of the steps between cut lacquer and stamping dies. IIRC CD4 was dead and buried before DMM. DMM+CD4 likely could have expanded rear channel bandwidth considerably.

      If you ever have the chance to listen to a freshly cut lacquer, take it. There’s NO NOISE. The hot stylus cuts a glass smooth groove.

      No high frequencies was what the hullabaloo was about in the early days of the audio CD. It has a hard cutoff at 20Khz. Direct digitizations of master tapes used for cutting record masters resulted in CDs sounding ‘harsh’, ‘dull’ and ‘flat’ without any of the overtones available on vinyl. Digital processing techniques had to be developed to attempt to address such things with old recordings, while new music for the digital age just ignored anything above 20Khz. “The human ear can’t hear anything over 20Khz anyway.” Not as what we can *perceive* as sound, but it still affects how we hear.

      Then the music industry got a passion for loudness, remastering and processing old releases to equalize the volume, compressing the range to make tracks sound louder, and wrecking any nuances originally provided by variances in volume. Imagine listening to an original recording that has a quiet passage, then the same track on a reissue where the volume has been equalized and that meaningful quiet bit has become bombastically loud.

    1. Because the designer has wisely spent the time to make the signal paths as short as possible in order to reduce cross-talk and interference.

      As for ground planes – there is noting unusual about not having a ground plane in audio and it looks like the intention is to put in a metal box which is a couple of order in magnitude better than a ground plane anyway.

    1. They are socketed so you can replace op-amps with better ones. Of course it doesn’t really matter, but that’s audiovoodoo for you. for better sound quality you might try waving a dead chicken over your unit, but only clockwise…

        1. It must be freshly killed one. Black chickens are better for fidelity, those reddish-brown widen the bandwidth and white ones are best for higher frequencies. Also ducks are good if you have problems with reverberations in your temple of auditory orgasms…

    2. The sockets are there for this kit to be practical. The op amps can get overheated by the soldering iron, they might be soldered in the wrong way or the polarity of the power reversed. Being able to replace the chips is, in many cases, better than soldering them to the board.

      It’s DIY though, so you don’t have to use the sockets.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s