(Sticky Me?) DIY Laser Turntable

I posted a while ago about the ELP Laser Turntable, visit the link if you've not seen this before.

Retail is >$15k for the cheapo variant. ELPJ bought all the patent rights from Finial in the US in '89, so it's unlikely there'll be a huge source of supply for more economic versions.

According to ELPJ it cost them something like 21 million dollars in R&D to design. They also say that the entire signal path is analog.

Since it's so new, there isn't much information on it. I have found two reviews of it by home users who've said that it's better (Clearer, quieter, faster...) than any other turntable they've own. If they can afford to spend >$15k on something that's virtually unreviewed, I'm guessing they probably have an okay system to compare it with.

It can play bent records because the read beams have a move in / out servo on them (Just like a CD drive).

One of the owners made a few complaints, pushed the company and eventually managed to find there were solid state opamps and other bits in read beams' output path.

Also, as I understand it, each of the beams sweeps side to side across it's groove wall and onto the lands such that the user can reposition the beam on the higher edges of the groove if they want to avoid the muck, gunk and wear at the bottom (Same way laser Galvos work for laser light shows or beam scanning mirrors that sweep the beam over barcodes at a superstore checkout).

There are five beams in all. One beam reads the height of the head from the record, the other two read where the edges of the groove are for positioning the head over the groove and the last two read the variations in wall position that represent the music. The beams are normal red lasers from what I can tell.

It got me thinking, since a lot of audionuts like to try making their own gear, would it be possible to also try making our own version of this? We benefit from not having to do this to profit margins or employ people to do bits of it for us, we'd just be doing it for fun. There are guys here great at amplifier design, others who'll be good with microprocessors, so perhaps we have enough skill freely at hand to do it as a group effort.

One suggestion I've had so far (from Karma) is that an old Laser Disc (the old 12" CD used for movies and super audio) player could be retrofitted for the design since it already contains the mechanisms for 12" disc playing; the disc spindle (Platter) controls, housing for the player, read head trasport that can track the much tighter grooves on a CD, a bay for loading the disc... It's especially suitable because these things are rapidly becoming unsupported junk, so a prototyping base would be tens of dollars or even free if you know someone about to throw one for a DVD player.

The head height servo mechanism is already present in a CD drive and could be augmented to deal with really bent records if needed.

Groove tracking is also already present, although I'm not sure if it would be happy following a record groove due to the variations in wall position. Perhaps some damping could be added to the mechanism to smooth over the variation frequency of the walls so that the head doesn't try to follow them directly, only the averaging of them. The Laser Player uses two beams and scans them across the edges of the groove to find the groove position. If the beam hits the lands (Flat top surface) of the record the beam reflection is different to that when it hits the top bit of the wall. The feedback from the two beams can then be used to create a differential (The groove width) that can be damped and used to guide the head across.

The only major difference we'd need to deal with is the read beams. A CD drive reads pits and with one beam. We'd need two and to read the walls of the groove rather than the pits. Saying that, Wobble groove DVD's store data in the sides of the groove.

The wobbly groove walls on a DVD are used as a frequency reference during recording. The groove wall frequency goes into a PLL like arrangement for synchronisation with an external oscillator. The output is a correct voltage for the spindle speed. So the whole arrangement is a closed loop spindle speed controlled (The wobble is acting like a strobe disc on a servo's spindle encoder).



Click here for some more pictures about DVD storage methods

Also, try here...

That means they almost certainly have some means of reading those wobbling groove that is different to a normal CD drive. Perhaps we could pull that mechanism from a partially dead drive and use it in a Laser Disc player to acomplish groove wall reading? The record's groove wall frequency get's referenced to an external oscillator, the (spindle) correction voltage then becomes the audio output signal. That would save time, effort and money trying to build the groove wall read head by hand; which would probably be the hardest part. Be interesting to know if the TT guys and DVD guys are using the same patent, one for each medium or are claiming they're different forms all together.

The home owner reviews I was reading mentioned the beam wavelength. Red wavelength lasers are cheap, established technology. Even the focused red wavelength, compared to a stylus, is tiny and can fit into groove wall vibrations shorter (Think high frequency) than those a mechanical stylus is capable of.

But that could be further improved by using blue lasers as they start getting more popular over the next year or two. Blue = shorter wavelength so an even higher read fidelity.

Other things to consider changing... replacing the opamps in the internal beam readout gain stages with valves.

Also, I'd like to further look at the scanning methods and what is happening with them since ELPJ aren't so clear on precisely what is scanning. I -think- the read out beams may be scanning. If that is true, it suggests that audio quality might be improved again by checking if that is absolutely necessary.

The idea is that the beams can be moved away from the stylus damage and gunk at the bottom of the groove, but a normal stylus will ride through that anyway. I'd question if the scanning mechanism needed to avoid that is helping audio quality as much as is claimed. If the beam is scanning on and off the groove, the read out method is intrinsically digital; the beam would have to sweep onto and off the groove wall to collect a discrete pulse of information about where it is, that would mean it's not a continuous output.

At first I didn't like the idea of using a laser as my auto "it's digital and SS" spider senses began tingling. But you should be able able to get a purely analog signal from this. And since you don't have any mechnical contact, resonances aren't such a problem, the stylus won't wear down over time and you can play the record as many times as you like without having to worry about rounding the groove off; I really hate the idea of damaging records during listening because I have a tendancy to play things over and over a few times each time I sit down.

Patents and law wise... provided you are only making such a TT for home use and discussing the ideas involved, it's fine. If you went on to start selling books on how to do it or kits for retrofitting LD players you'd be approaching legalville. If you were to start retrofitting TT or LD players to do this, paid or otherwise, you'd definitly be somewhere bad. Selling your own version of it without a license = sued. My brother is a Patent lawyer can could give some advice on what is suitable and what isn't.

In terms of ripping ELPJ off and stealing potential customers, question how many people who can afford >$15K on a TT are going to bother messing around with hot glue guns and smashed up drives trying to make their own. One or two at most? The only people who'd bother trying to make their own are the people who can't afford one in the first place, like me.

I thought maybe this thread might benefit from a temporary sticky whilst the idea is discussed to keep everything together.

Anyone else got any ideas or thoughts about it?

SY said:

Patents are long expired. It's public domain.

Click to expand...

I know they bought the patents in 1989. By UK law, that means the maximum time they can hold them for is until 2009, assuming they keep paying the holding charge.

But since the patent will have been filed prior to that, yes, they've probably expired, or are close to.

Even if it is public domain, we still need to figure out how to put our own together. I'm trying to find some information on the read / write heads used in CD / DVD record and rewritables. Since they have to scan the wobbling groove walls I'm hoping they will also have a read / write head at something other than perpendicular to the surface. If it's perpendicular, it must be using some funky methology to read the wobble.

"You ain't got the phunk! You're all rigid, like a breadstick!"

Here's what a Laser Disc and player looks like for those who've not seen one before...

SY said:

OK, a quick search showed one US patent issued in '89, so it will expire in a few months. The other issued in '90, so there's a year and a half left. Unless you're going to sell a lot of them before them (hah!), there's probably no significant damages if you build yourself or discuss building.

Click to expand...

Yep.

My main aim wasn't really ever to sell them, just to have one for under $15k.

I thought about the interferometer as well. I've been reading about the idea of building laser listening probes, the kind you point at windows, and an interferometer is something that's discussed alongside that.

I did a couple of quick messy pictures in paint of the idea. The thing is, as the groove moves in and out sideways, if your laser beam is aimed at a fixed angle, the point it strikes the groove wall also moves up and down (Relative to the changing height of the wall). That means when the goove moves out a long way sidways, the beam could end up striking the wall right down near the bottom of the groove near the muck and stylus wear. Could do with checking the numbers on track width to predict precisely what the beam would be doing if you to just point one at 45 degrees into the surface.

If the beam comes too close to the bottom of the groove, a glavometer like scanner could correct for this and move it further back up by strobing the beam across the wall. I get the feeling the ELPJ TT does this since they imply it can actively avoid hitting the bottom of the groove and move the beam position up and down to find the best part of the groove. If the record is scratched for example, you can reposition the beam further down, towards the middle, to miss the scratches on the top and the gunk in the bottom. I can't immediately think of any way they could be doing that without scanning the beam, since I think the groove wall variations would be too big to just position it further up or down once and let averaging do the rest.

But like I say, it depends if that is totally necessary. The smaller grooves will be deformed by the stylus and rounded over by muck faster anyway. The beam might hit (relatively) closer to the bottom of the bigger grooves, but they also have more volume to deform and round over before the distortion reaches as (relatively) far up the groove wall. A scanning galvo would make things more complicated and make it too hard for 50%+ of the people who'd like to build one I think.

Do you have the patent numbers or an address for those? The patent servers in the UK are absolutely terrible. Paying to view patents? Bah.... Defeats the whole point of them!

SY said:

IIRC (I was a Stereophile subscriber when this was introduced, and Gordon Holt wrote about it extensively), the dust and muck were indeed the biggest problems. Records had to be ultra-clean since there was no stylus to push dirt out of the way. At the same time, click-removing signal processors like the Burwen were being introduced. Could that sort of thing have been used in the Finial's production versions?

Click to expand...

Good idea. ELPJ do sell them as separate units, so there's a possibility of it.

Although, I wouldn't mind keeping my records extra clean provided it meant no wear.

I'm pretty sure I read something by ELPJ mentioning that the beam was moved away from the muck though. Also, they have scanning mirrors in one of their pictures.



Not a particularly helpful picture.

"Modulation on the individual grooves is reflected to scanner mirrors and onto left and right photo optical sensors."

Either I'm missing the beam in that picture that comes off the record, up to the mirrors, back down to the record, back off it and then into the detectors shown, they haven't drawn in the music detecting cells and the ones shown are for edge tracking... or, they mean the read beam is scanned across the groove walls by those mirrors and then the reflection goes into the detectors.

If the read beam is being repositioned on the groove wall by those mirrors, there's three ways it can be working.

1.) The mirrors reposition once and then sit still. Need to check the dimensions to see if that would work.
2.) The mirrors actively try keep the read beam on the same position of the wall by moving it up or down dynamically as the wall moved. Doing so would need a closed loop monitoring the wall positions to track where they began (A reference point) so's that the read beam could track it's position on them. That's possible since they DO have two beams tracking the edges of the grooves.
3.) And what I think is probably happening. The beams are just strobing up and down the groove wall very quickly. You take samples at the same speed as that scanning is taking place but you change the phase to change where on the wall the sample is taken from. For example, if you take a scan during a negative phase, it's near the bottom, at 0 phase, you hit the centre and positive phase you hit the top.

If number 3 is true, the scanner would need to be scanning at the same speed the surface of the wall modulated to maintain precisely the same point on it. If the scanner was moving slightly slower it would drift in and out of phase with that point. But it wouldn't need to measure precisely each time. Provided it got roughly the same bit you'd get the same effect (Off the muck at the bottom and away from the scratches at the top).

I think I'll try to find out some track dimensions and then draw a scale drawing so I can actually see where the beam would be going as the wall moved if you just held it still at 45 degrees to the surface.

I'd love it if we could get this working!

SY said:

edit: Wrong! They used an analog position sensitive detector to measure beam displacement.

Click to expand...

This is how I first invisaged laser listening devices working, relying on the beam sweeping over a photodiode. The more beam on the diode, the greater the signal and vice versa.

But from what I can tell reading about the laser listening idea, the modulation is actually due to phase interactions in the beam.

I tried the window idea out a while ago for fun in my kitchen (Not actually hooked up to any listening gear). Our kitchen is quite long. I set up a laser pen to shine on the window at the other end, bounce back and then hit the wall next to me. If I walked up to the window and pressed on it I could see the beam sweep sideways maybe a few mm. The same happened if someone shut a door or was walking around on the floor of the room above the kitchen. But not much else.

With laser listening probes, the beam leaves the laser 100% (in a perfect world) in phase. It then hits the window and reflects off at an angle into a photodiode.

As soon as it reflects off you start getting phase interactions in the beam. When someone talks in the room, the vibrations are enough to alter the phase interaction patterns.

The phase interaction patterns either sum or subtract from the total beam amplitude (phase cancellation and reinforcement). The photodiode is actually listening to the amplitude modulation cause by these phase interactions rather than the beam moving side to side.

You then need to amp up the modulation a lot to get it to a usable level. If the ELPJ turntable is also using amplitude modulation as a result of phase interactions then it's probably using high quality photon detectors to lower the noise of the amplification. That may be why they're using opamps instead of valves as well, so they can pack lots of gain into it without needing a load of stages.

I've just found the patent and am having a read through it.

Lots of information on the tracking and general idea so far but not much about the actual read beam mechanics and operation.

I see the patent claims for height adjustment (focusing), lateral adjustment (groove tracking) and tangenital adjustment (alignment with groove spiral). Height / lateral easy with a normal LD player. Tangential isn't present intrinsically but could be implemented.

I've also been bothering some microprocessor / DVD guys about 'groove wobble' clock frequencies and encoding on optical disks.

Should start turning out some possibilities sooner or later!

I found a mention of the table here...
http://66.249.93.104/search?q=cache...ly/2001/2001-06.mbox+elpj+finial+patent&hl=en[/url]

I've also found some information from one of their US distributors who wasn't too happy with ELPJ stealing their pictures and videos of the table. They complained about ELPJ's policies on trying the gear, but they didn't seem too extreme to me. More interestingly was their pricing, which seems to change depending on the day and who you are.

The distributor mentioned the table having an in built declicker. Other people suggested buying record cleaning machines.

Simple way to remove the clicking and noise problems without killing the audio quality... fit an aquarium pump (with a big reservoir) somewhere remote of the table. Have the air stream blow out thought a tiny, microscopic, pin sized tangenital hole just in front of the read head. If you don't want the airborne dust going all over the inside of the player (e.g. inside the read head), connect the air in line for the compressor at the reflection angle to suck up the airstream after it hits the grooves. Put a coarse filter in the suction line to stop the dust getting into the compressor and a fine filter / trap the out line to stop oil or moisture from being blown out in front of the read head. Sounds extreme but we're talking about very small quantities of air volume and pressure here, so a cheap eBay solution could be implemented. $50 - $100 perhaps? Maybe add a soft brush to agitate the stiffer dust before sucking it up?

SY said:

Interferometers are optomechanical A-to-D converters.

Click to expand...

For some strange reason I was reading PSD as positron or photon sensitive, which didn't help my understanding of what was going on.

Position sensitive, as you said in your earlier edit, it must be simply measuring the beam displacement, nothing to do with phase or interference.

PSD manufacturer

"Hamamatsu provides a variety of one-dimensional PSD (Position Sensitive Detector) designed for precision distance measurement. They are used in displacement meters, proximity sensing, as part of a camera's autofocus mechanism and in optical micrometers for high resolution, non-contact measurements."

Method for using PSDs as optical positioning probes

jcx said:

Further you also seemed to sneer at “digital” anything - but digital processing power is in fact the most leveraged technology that could possibly cut the cost

Click to expand...

Guess this has nothing to do with why you decided to reply right?

Please acquaint yourself with wobble groove optical discs.

Same theory but with a groove that's >100 times thinner than a record's, modulating >40 times faster than you can hear, at an amplitude of <20 - 30nm and being read by a drive that costs tens of times less, literally. Seems worthwhile thinking about it to me.

jcx said:

add up all of those (laughably optimistic) factors and its still cheaper to by the laser turntable from ELPJ

Click to expand...

1.) Audio companies love to massively inflate their prices since price means everything in audio. 21 million dollars of R&D sounds great. Just like cables with no statistical evidence behind them are worth the $15k. After all that time, money and effort, kind of amazes me I've only heard of two or three people who've actually bought or listened to one. They're also leaving it a little late to hit the market what with their patent about to expire in a few months. Makes me wonder if perhaps, A.) ELPJ are bending the truth a little or B.) The people at ELPJ are stupid enough to spend 21 million on something they'll sell so little of. Lets assume they make 100% profit on each player. They'd have to sell over a thousand of them before they even came close to clearing their R&D costs alone. Realistic if you've just made a CD drive. A record player that costs $15k? Doable, not easy. That'd be at least three a day to do it in a year. Then factor that they probably had to take a loan to get the supposed 21 million, and loans have interest. And then everything else they need to pay for to sell it, like staff.

2.) They were developing this back in the 80's, when CD's were new and incredible. I'm sure they burnt up lots of money working on it while the technology they need for it to work was so expensive. Their technology is useless to the everyday guy now. A CD drive isn't, and even still they're considered as boarderline junk (drives for $5 - 10). CDs won and rocketed off since they had the market to do so and space to develop into. The vinyl format has remained constant with no big leaps in density. Of coarse ELPJ are going to want to keep you thinking their player is something special. If they told you it was nothing more than a low resolution CD drive, why would you spend $15K on it?

3.) ELPJ own the patent on the idea and haven't licensed it out to anyone. Which explains why other companies haven't started making mass produced versions of it.

4.) Even after the patent expires, there simply isn't the market for guys like Toshiba or the other drive manufacturers to consider such a drive anywhere near as seriously as they would something like HD DVD or Blue Ray. Only audiophiles and DJ's bother using records anymore, and most DJ's want to scratch their's.

5.) Despite my sneering at digital, I actually seem to be one of the few audiophiles who's willing to even consider using a laser to read a record. Most of what I've read so far dismisses the idea purely on the basis that the laser beam can't knock dust out of the track. Pretty narrow minded considering how simply that can be corrected for the gain of zero wear and better read fidelity.

I've heard similar R&D costs being mentioned by InPhase, the people developing holographic optical storage. The two just don't compare in terms of data density or complexity of the medium / drive.

I plan to make this massive leap of costs by looking at the possibility of using technology that has already been developed and distributed to a mass market, like the LD transport and DVD read technology.

It's like asking how anyone could consider making their own tonearm read TT given the cost and specialised skills involved in originally developing the technology. Developing a DIY laser TT would be just like the first guys making their own tonearms at home.

If I was suggesting you start from scratch, build this entire thing by hand using your own homemade technology and with no reference to how optical drives work or pulling them apart, I could definitly see your point. But I'm not.

I've heard a record being played through someone's flatbed scanner.