Replacing M5218A in Servo Circuitry

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi,
I would like to tweak the laser servo circuitry in my Nak CD-Player.
The unit (OMS-5II/7II) is now 19 years old and error correction is not good, even though I swapped the laser 3 years ago and aligned it perfectly. 0.3mm is the barrier (scratch width) where one hears clicks and the laser starts to skip.
Any modern unit will easily go up to 1mm.
There are three opamps used in the servo circuitry, M5218A, which were also often used for audio signals.
What could be a replacement? Would a faster slew rate help the circuit to react to scratches much faster?
All the best,
Salar
 
Administrator
Joined 2004
Paid Member
Salar said:
Hi,
I would like to tweak the laser servo circuitry in my Nak CD-Player.
The unit (OMS-5II/7II) is now 19 years old and error correction is not good, even though I swapped the laser 3 years ago and aligned it perfectly. 0.3mm is the barrier (scratch width) where one hears clicks and the laser starts to skip.
Any modern unit will easily go up to 1mm.
There are three opamps used in the servo circuitry, M5218A, which were also often used for audio signals.
What could be a replacement? Would a faster slew rate help the circuit to react to scratches much faster?
All the best,
Salar


Unlikely, concealment and error correction are a function of the complexity (cost point) of the player's digital signal processing chipset, not the analog servo circuitry..

Take better care of your disks if possible or find a player that handles the errors so generated..

My Shigaclone is not great in this respect either, but I like its performance otherwise and live with the occasional skip which is actually rather rare - only happening with a few of my wife's disks which have not been well treated.
 
Hello Kevin,
no, I am talkung about the analog circuitry which keeps the laser on track and reacts to scratches. Not interpolation. I have attached the circuitry.
All the best,
Salar
 

Attachments

  • oms5eii_39-40bw.jpg
    oms5eii_39-40bw.jpg
    93.5 KB · Views: 223
Administrator
Joined 2007
Paid Member
Hi Salar,
I'm afraid it wont help at all, swapping the I/C's. This is one area of performance that has greatly improved since the first generation players.
If I mention Nyquist stability, we think of amplifiers and feedback. The same problems exist here but with the additional complexity of a "mechanical/inductive" component in the loop-- the tracking and focus coils. The characteristics of the "analogue" stage are carefully designed and very critical. Careful setup can help, particularly EF balance if the player has an adjustment for it and focus bias.
You would have to follow the manual, but I think you have done this already.
 
Administrator
Joined 2004
Paid Member
Salar said:
Hello Kevin,
no, I am talkung about the analog circuitry which keeps the laser on track and reacts to scratches. Not interpolation. I have attached the circuitry.
All the best,
Salar


Dealing with read errors is done in dsp. Only in the case where the EFM waveform is incorrect can you make any improvement in performance. Improving servo performance** is not going to make the player magically read the missing data, some players just have much more robust error correction algorithms that can cope with longer data dropouts.. Your player apparently isn't one of them.. (Nor mine if that helps..)


Note:
** It is unlikely that you need to or even can based on the comments of the previous post. The mechanical subsystem is far slower than the existing op-amp..
 
If you are having trouble tracking actual cd's, I can add my two cents.
In my experience, it is completely inadequate to align an OMS-5/7 with test discs & the 'by the book' procedures. Works ok for the actual optical 3-axes alignments, but not the servos. Best to do the servo adjustments with the widest variety of cd's(and cd-r's if the laser is healthy enough) you can, and align for the best compromise between worst & best discs. Otherwise, you'll perpetually have tracking or noise problems with some discs.
 
OMS trackability

Hi Salar

I have repaired many OMS5/7 II and they have two main problems: the disc motor and the head.

After 19 years, maybe your disc motor is worn: put a disc in play and when the disc is at his max RPM, put the on/off switch at off.
Put your ear near the spinning disc and listen.
If you hear a low frequency rattle, your disc motor is worn: it is loose in his bearings. Replace.

The head: your "new" head is maybe 19 years old. It has been stocked a long time.
Check the head suspension and check if you have side and bottom play: i mean if the head can move on his axe left and right side. Most important. check the down play. Be carefull!!
With a cottonswab, press gently on the lens and see if it can go down a little. If not, it will be unable to read properly some CD's. The lens rubber support is crushed.



If your head has no down play, you can play with the focus offset ( check the eye pattern and adjust for best signal ) to "correct" this situation but keep in mind that the servo circuitry will work harder. Adjust the EF Balance with a 1mm scotch tape on a test disc. Listen to a tone and adjust for the least error play.

Replacing the M5218A will do nothing for your problem.

Nakamichi OMS-5/7 are completely differents units: they have a 1 beam laser and a completely different electronics circuitry.

Good luck
 
Hello to all,
thanks for your advices!
The laser is three years in use old, I bought it ten years ago, when it became clear, that Nakamichi would vanish from the european market. The lens moves freely.
I also swapped the disc motor (RF-310TA-11400). Because the shaft of the present motors is too long, i had an aluminium spacer milled which holds a brass bearing. It was little tricky to fix it on the motor without adding any additional load to the axis. I also added a brass tube in the disc clamper to make it fit tighter on the motors axis to prevent imbalance. (which was there before, the clamper had a little play horizontally when it presses a CD on the disc table)

So I guess the Nak is mechanically better than it was before my mods.
But the weak error correction has remained.
I also have a second-generation Toshiba XR-Z70, the Tosh is a bit better than the Nak in error correction.
The disc with simulated defects is Digital Recordings "CD-Check". The error size ranges from 0.375 to 1.5 mm. As I wrote, the Nak fails at 0.375 already, the Tosh passes with maybe two dropouts in 10 seconds. The defects are not one black line or dot, but four lines in a 90 degree angle. So with one turn, 4X0.375 have to be compensated.
Another player of mine, a CDP-X5000 from the mid nineties can easily compensate up to 1.2mm.

Highspeed, do you still remember setting the E-F balance of the OMS?
I am using the YEDS-18 Test disc for this. But the E-F Signal does not look as clean in real live as it is shown in the service manual.
In test mode with the spinning YEDS-18, it E-F resembles a sine wave, which has to be adjusted to swing equally around 0VDC.
But no matter how correct I adjust it, (or how much it is out of adjustment) it changes frequency and amplitude randomly very fast (but can somehow be adjusted around 0V)
Thus effect was also before any mods and with the old laser, so this is normal behaviour?

The eyepattern does not look very clean, also no matter what laser. But I do not know wether I should expect a clean eyepattern because it is not documented in the service manual. Strange, because this is the first thing one should look at.
The lines are not sharp, the outer edges are blurred and noisy. Checked it on two scopes.
To make the nak safely play CDRs, focus offset has to be raised a bit, this also raises the amplitude of the eyepattern.
I spent many days with adjustment, also with tracking/servo gain and Dropout detection circuit, but 0.3mm is the barrier...
I attached a photo of E-F balance from the service manual. Maybe they used an external trigger to make the waveform look stable?
All the best,
Salar
 

Attachments

  • oms5eii_38.jpg
    oms5eii_38.jpg
    76.9 KB · Views: 173
Administrator
Joined 2007
Paid Member
Hi,
When using scope make sure you use the correct grounds as this can make the also RF look noisy. The eye pattern should ( not all players achieve this ) be very sharp and well defined. The diamond shape should be "steady" , any slow vertical bounce in the waveform indicates that the turntable is not true and that there is an up/down motion as the disc rotates. Check by looking at the outer edge of disc and turn slowly by hand.
The EF balance will be correct if it is symetrical about zero. Can be useful to adjust/check while repeatedly using track search and slow search I use to find on some players. Try it ? Can you feel any play in the spindle motor bearing.
Ultimately I think you will find it's just the way it is, my first CDP, a Sony CDP101 was not good on error correction and playing less than perfect discs (from the library etc ) but a cheap and cheerful Philips CD150 found it no problem.
If it's any help Philips Test disc SBC 444 A states that any player should be able to play without audible errors a 500 micrometer interuption in the recorded layer, also a 300 micrometer black dot as a single one off defect. Now this isn't very severe but that's the "standard" minimum as it were.
 
Hi Mooly,
the test point for GND is not defined. I used common ground of PCB/Chassis for this.
What did you mean with The EF balance will be correct if it is symetrical about zero?
That it swings around zero (equally positive/negative), or that the the bottom swing of the signal "sits" on the zero line?.
I have two disc tables for the machine, both will result in a vertical swing. So nothing to do about that.
All the best,
Sal
 
Administrator
Joined 2007
Paid Member
You need the ground that the RF amp uses ( as close to the rf amp ic as possible ) and to measure the RF as close as possible to the output of that amplifier. This does make a huge difference to how clean the signal appears in a lot of cases. Focus bias will affect the clarity of the eye pattern. If you connect probe tip and ground lead together on the scope and then touch both these to the chassis does the trace on the scope thicken due to noise ? If it does try a 10 ohm in the ground lead to the scope probe.
EF balance. Going from memory now, but try AC coupling the scope, and doing track searches while looking at the EF error waveform. As you keep doing this switch the scope back and forth between AC and DC coupling and try and get it so there is no shift as you do this. Long time is I have done one, but it should be fairly clear where the correct setting is.

Edit, ground = pin 7 of RF amp and use the designated EFM testpoint. The CX20108/9 chipset is nearly first generation and is the same as the original Sony D50 Discman (The first ever).
Dont think you are going to get this much better.
 
Hi Mooly!
Thanks again!
So E-F should look like in the attachment of this post?
Never tried to measure close to the RF- amp, I am curious to do so, but it will take some weeks until I have the time to do it.
When I measured E-F balance , i first short circuited probe to ground (DC coupled) centered the flat line as reference on thze scope and then started measuring.

The D-50 is also part of my small collection of early players, never knew they had the same chipset, as well as the CDP-502ES/552ESD. I bought the service manual of the 552, because the setup in the Nak SM is described very badly. But even though the OMS and 552 have CX20108 and CX23035 in common, the setup routines are completely different...
Did you ever try tracking/servo gain and D.O.D in the nak?
All the best,
Sal
 

Attachments

  • oms5eii_38swing.jpg
    oms5eii_38swing.jpg
    46.3 KB · Views: 117
Administrator
Joined 2007
Paid Member
That looks fine. As you turn the pot you will see the waveform shift and become non symetrical. It shows up more if I remember in track search. Oh by the way, Iv'e never worked on a Nak CPD. Worked on loads of Sonys, the D50 used the CX20108 etc. We used to sell a Fidelity player that was based on this chipset and these were always failing (the 20108). Changed probably around 6 to 8 altogether, never once on any other make of player. Must have been some PSU problem or spike seeing them off.
As for the other adjustments, as I say I don't think you will improve it a lot. For tracking gain connect scope to the tracking error waveform testpoint and as you turn up the gain a low frequency component will begin to appear in the waveform -- try it -- you want to back off slightly from this point so the fundamental just dissappears. Thats about it really :)
 
Hi Mooly,
so. many coincidences! Well, the 20108 works for 19 years now, another could be used from a defective Nak (PSU) in if something should ever happen. I have a Fidelity B1 from that area, which I fixed after a transistor blew, but never got the quiminiscient current right. It will make way for a Shanling amp end of Summer.

So I should try to adjust tracking gain with a defective disc (like CD-Check) and watch for the low frequency component caused by the dropout?

BTW,could decoupling the M5218A give a small benefit? Decoupling was not used in the Nak at all, except some ferrite beads for the CX20108 and CX23035...
Thanks again,
Salar
 
Administrator
Joined 2007
Paid Member
Hi,
Use any good disc. As the tracking gain is increased the tracking error waveform will slowly start to oscillate up and down at a very low frequency. Adjust the tracking gain nearly up to this point, but so this fundamental is absent. The low frequency component isn't caused by drop out, it's to do with the stability of the whole tracking servo loop ( nyquist again :) )
Why not try starting another thread seeing if there are any NAK owners on the forum and ask how good there machines are. I really and trully think this is as good as it get's on this. Technology, in particular the software used in servo's etc has improved dramatically, as have the actual optical units themselves.
 
Instead of starting a new thread, I dig this out.
I found 2 service manuals for cd - players using the CX20108 / CX 20109 / CX 23035 chipsets:
Denon DCD-1700 and Sony CDP-552 ESD. Very different designs for the tracking servo.
But one intersting detail: Sony and Denon installed decoupling capacitors (is this the right term?) between the supply voltage and GND of the IC´s mentioned above.
Nakamichi did only use a ferrite bead + decoupling cap for CX23035, nothing else.
I will try decoupling caps on the CX20109/CX2018, maybe I get a slight improvement...
All the best,
Sal
 
Going back to your earlier question- The 'eye pattern' output of the laser on the OMS-5/7 should look UNUSUALLY clean, cleaner than you'll see on any other player. To acheive this, first & foremost you have to do as I mentioned in your other thread, measure BOTH the optical power output AND the laser current draw. If you can't get 0.24mW without exceeding the rated current by more than maybe 15-20%, or can't get above 0.20mW without exceeding the current spec by much, you WON'T be able to get solid performance, and won't be able to use that laser. Also, as I said in a previous post, assuming you do the above successfully, you need to do the servo alignments on MANY different cd's, NOT just a test cd, nor just a few cd's. You HAVE TO align for best compromise settings among at LEAST a dozen different cd's, if not 2 or 3 dozen. TRUST ME; I've service these players since they first came out, as a Nak warranty tech for many years. I know them like the back of my hand.
If you get the laser output properly adjusted & it's reasonably close to output & current specs, but still can't get clean eye pattern on the scope, go back to the mechanical optical alignments & make damn sure you have it idealised.
 
Hello Stephen,
I did already perfect alignment. Laser Power is not in question.
The eyepattern looks as good, (or as worse) as from another player with KSS-240A.
It is only a question of going beyond proper setup.
Do you have any idea, why Sony and Denon used decoupling capacitors between VDD and GND for the CX-20108/20109 chips and Nakamichi did not? Did Nakamichi miss something, or ist there only a theoretical benefit from decoupling in thi relatively slow layout ?

Oh, btw, the sled motor is not produced any more, manufacturer is Copal.
Any idea for replacement?
Spindle motor is still produced, Mabuchi RF-310T-11400. It has a longer shaft nowadays, . Instead of cutting the shaft, I had a brass spacer milled with an additional bronce bearing. Works perfect.
All the best,
Salar
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.