Excessive Wow & Flutter in Sony WM-2 (0.78% RMS with 8Hz Spike)

Hi Norm,

Thanks for the note! It’s reassuring to know I’m not the only one who couldn't track down official W&F specs. I’ve restored a few 1983 Panasonic boomboxes in the past that sit around 0.2% WRMS (against a 0.3% spec) and sound perfectly fine. Given how superior the Sony WM-2's mechanical build quality is by comparison, this performance hiccup really caught me completely off guard—a unit engineered this well shouldn't be performing like this.

Like you, I always give my audio gear the full "spa treatment"—sparing no effort just for my own peace of mind (and because I genuinely enjoy the process).

Before starting this project, I watched several videos and read through @Valentin's tutorials, so I was well aware of those door mechanism bearings jumping into the abyss! However, this "mystery ball bearing" was much smaller than the two door detent bearings. Naturally, the donor unit’s shaft was clean, so who knows where it came from.

The belt is certainly still a suspect as well. While Marian’s products are consistently top-notch, strange defects or batch variations can always happen. I usually buy belts from Neil at Russell Industries—great quality and prices—but he doesn't have the right size in stock at the moment. Would you mind sharing the US eBay vendor’s name so I can check them out?

I'm currently waiting on the replacement parts to arrive so I can start swapping components and see where that leads. I hope to take a victory lap (or maybe two) at the end of this project.

Best,

Alex

 

Sure, Alex, I bought from https://ebay.io/m/DXPppR Speaking of Marian, I received my order for the head block lifter and a few other parts today. I always get a little anxious when ordering from overseas with no reliable tracking, unless one wants to pay the $50 for DHL. Cheers! Norm

 

I would be happy with 0.3% WRMS on this particular unit. It's not a great figure, but it's ok for most genres and will only be audible with music with very long note decays.
A good figure on this unit would be ~0.2% or a bit lower (0.17-0.18% for example), but anything below that is extremely unlikely to be achievable.

0.1% WRMS is already DD territory, hence why when belt driven mechs obtain such figures it should be treated as the exception, not the rule.
There are some belt driven mechanisms that can obtain such figures but they're few and far between.
Contrary to what some believe, 0.3% is a figure many walkmans (even high-end ones) won't be able to outperform no matter what's done to them.

Regarding W&F specs, most belt driven walkmans don't have such specs and there's a reason for that.
From the ones that do, Pioneer has a <0.28% WRMS on their mech, AIWA gives a <0.45% WRMS even on their high-end devices, Hitachi gives a 0.2% WRMS, etc.
Interestingly enough the belt-driven devices some people take as reference (SONY WM-FX35 - 0.15% on a consistent basis) do not have a spec.

 

Thanks, Valentin, appreciate the insight. Cheers! Norm

 

Thanks, @Valentin—really appreciate the context and insight here.

It's interesting comparing this to my Sony WM-GX670 (which you helped me fix a while back). That unit consistently hits now 0.13% WRMS in both directions. It’s obviously a much newer design with a lot more plastic in the mechanism, but it performs. Still running without that one missing capstan fiber washer, too.

@worldwide137 Norm, thanks for the link! I will order a belt for WM-2 from him as a test.

Overseas orders are tricky and expensive: I try to avoid them if I can help it. Fixyouraudio has proven to be 100% reliable.

 

The GX670 mechanism and variations of it that exist in the entire WM-EX/GX range can obtain low figures, however this is highly dependent on motor and capstan bushing wear.
Seen units in this range (particularly ones with big motor like EX7, EX9, EX20, etc.) obtain figures close to 0.1%.
Ones with smaller motor (like GX670) can obtain low figures, but nowhere near as consistent. On those anything between 0.15% and 0.30% can be expected depending on wear.

Interestingly enough units like FX35 which have fully plastic mechanism (including the capstan bushings) tend to obtain good figures on a consistent basis.
Similarly some AIWA mechs do obtain low figures on a consistent basis, would even dare day your average AIWA is more likely to score better than your average SONY in the W&F department.
Also the brass flywheels I noticed matter less than many think. Seen this exact WM-EX mech to obtain better figures with plastic flywheels (lower wear) than higher wear ones with brass.

 

An update on this bewildering project.

I have good news and “the same” news to report. The good news: after replacing the supply reel gear with a new one, remediating some corrosion, and resoldering all the wire connections, the donor unit works. I gave it the usual "spa" treatment. The "same" news is that neither player’s W & F is good. The original Walkman W & F dropped to 0.45% from 0.8, and the donor now sits at 0.6% WRMS.

I swapped some parts between the two, trying to nail the swine: belts (both sourced from fixyouradudio.com) - no change; flywheel capstans - things got worse (?!); driving gear assembly with its own native washer, and hysteresis plate- no difference. During the swap, I found that the washers used between the central gear metal plate are different.



The red one (Teflon, I think) is larger and also thicker (0.21 mm); the clear one is 0.11 mm. Is that normal? The Serial numbers in these players are close.

I also changed hysteresis plates (B) and tried them upside down, etc. It didn’t seem to make any difference.

It appears there are only two things left to try to solve this W&F mystery: try a new belt from a different supplier, and replace electrolytic capacitors.

Any other ideas? I'd be grateful for your input. Thanks!

Alex

 

Hi Alex! On my WM-2 repairs, the clear washer is what I have experienced. That red washer looks rather odd to me, but I would defer definite answers from the pro's on this forum. I do believe that the WM-2 is capable of doing better than 0.4/0.6% wrms. For the WM-2's that I've restored, it has always had full replacement of electrolytic capacitors and full cleaning of all the gears. Are you able to check if the capstan is even slightly bent? Have you lightly sanded the magnetic clutch gear and metal washer on a perfectly flat surface with something like 5000 grit paper? Is there any corrosion still left on the motor pulley? Even if there is some residue from a sticky melted belt on one of the gears can introduce issues. Do you know what the frequency peak is from the wow and flutter from something like the NAK T-100 analyzer? Let's get this thing sorted! Norm

 

The transparent washer is what I've seen also, although haven't seen that many of them to say for sure that's the only washer that was used on this model.
However don't think this is the problem if the torque is right. Rather the different thickness may influence FF/REW main gear engagement.

I would try a different belt for sure, FixYourAudio ones despite generally good can be non-uniform sometimes having high W&F.
Seen situations where belts from the same batch (bought from FYA multiple pcs) were very different, one obtaining 0.4% one 0.2% on same walkman.
However worth trying Russell Industries one instead, by dimension. Belt dimension is something ~45 x 1mm (inner diameter x thickness).

What brand are the caps ? Nichicon and Matsushita are the usual suspects, but even these are not always bad in all walkmans that use them.

I agree with @worldwide137 regarding pulley grooves. These in my opinion should be sanded with 800 to 2000 grit sandpaper while rotating them to eliminate any micro imperfections that may exist.
On motors I usually run them on a bench power supply (1.6-1.8V) while kepping a small folded sandpaper inside the groove to clean it.

 

@worldwide137 , @Valentin,
Thanks for your insights, gents!

There was no corrosion on the motor pulley—just on the battery terminals, and some of the leaked acid ate a small portion of the paper/copper shield on one unit. The original belt was still intact, but stiff and dried up like a piece of wire.

I cleaned the motor pulley groove on both motors using a fiberglass pen and sanded them with 2000-grit paper per Valentin's method. I also cleaned the flywheel groove, though without assisted rotation. Watching the capstans spin, there is no visible wobble or eccentricity, though I haven't checked if either shaft is slightly bent relative to the base plate. Also, I haven't tried the ferric chloride capstan etching treatment that some techs use (like L.T. Tape Services on YouTube).

I haven't pulled the caps yet, so I can't verify the brand just yet. They are small black cans with a blue stripe on the negative side—likely Nichicon.

By the way, the red (Teflon?) washer that sits between the metal plate of the central gear and the magnetic gear is made from the same material as the thrust washer (sits inside the shaft of the capstan flywheel) and the spacer washer that sits under the circlip that holds the capstan in place; same color and thickness, 0.21 mm.

This is the "original problem" WM-2. The speed is unstable on this player, and seems to vary more than I'd like to see.


This is the "donor" WM-2



I’ll try a fresh set of belts from a different source once they arrive, and I already have replacement capacitors on order. I'll update the thread with the results as soon as I make progress!

Thanks again,

Alex

 

Alex - It sounds like you are doing all the right things, this bugger is just being difficult. As a matter of practice, I too use ferric chloride with extreme caution. No more than a 2-3 second dip and I ensure by all means to not etch the shaft that comes in contact with the capstan sleeve.

I'm no genius, so I often refer cautiously to Gemini. This is what it said about your 8hz spike:

Primary Causes of an 8 Hz Spike

• Capstan Shaft Contamination: Hardened tape oxide, shed binder, or adhesive residue stuck to the capstan surface, creating a micro-bump against the tape once per turn.
• Bent or Eccentric Capstan Shaft: Runout, burrs, or a slightly bent shaft introducing physical eccentricity into tape travel.
• Bushing / Bearing Wear: Excessive radial play, scoring, or dried lubricant in the bronze capstan sleeve bearing, causing the shaft to wobble or chatter under load.
• Flywheel / Gear Drive Defect: A nicked tooth, flash, or debris on a drive gear or flywheel surface coupled directly to the capstan shaft on a 1:1 drive ratio.

 

I should also mention, some have commented in other forums that ferric chloride is a bad idea. Why is that? If not properly handled, it can etch unevenly, thus causing wow and flutter issues. I suppose that does make some sense. However, I do feel that if you clean all tape residue/grime/corrosion first from the capstan and quickly dip it for only a couple of seconds, it should be fine. At least, I've had good success with this technique. But, something to be mindful of. Would be interesting to get Valentin's expert opinion. Cheers. Norm

 

Norm,

Thanks for the advice on the ferric chloride treatment. I’ve had reservations about chemical etching myself, so I'm holding off until every non-destructive option is exhausted. I don't want to risk "cutting out the kidney when I need to push out the kidney stone". If I get them to a listenable condition, I don't care if it is 0.35 % WRMS. I am not going into the rabbit hole of chasing perfection.

I also ran the W&F spectrum through Gemini and got similar diagnostics analysis; the spectral breakdown helps narrow down rotational frequencies. That's why I decided to swap various parts to nail the culprit.

What puzzles me most is that both units share nearly the same performance bottleneck. Two identical mechanisms failing at the same baseline makes me wonder if there's an issue with my re-assembly, belt batch consistency, or lubrication tolerance.

This WM-2 is prove-it-to-me stubborn. Looking at its beautiful layout and over-engineered build quality, it’s almost ironic that my humble Panasonic RX-F5, which is full of plastic parts, easily kicks its butt at 0.2% WRMS.

Best,

Alex

 

There's no reason to have reservations about the chemical etching, in a worst case scenario it will simply do nothing. Outcome which I would say is likely, but nonetheless worth trying just to eliminate it.
Only important part is to etch just the part that touches the tape (with cotton swab) and be careful no ferric chloride gets into the part that sits in the bushing.

Fact 2 identical mechanisms fail in the same way can suggest one of two things: either it's a common failure pattern (like worn capstan bearing) or more likely something "external" to the mechanism (would put my bets on the belt) is causing the failure.

 

@Valentin,

Mechanically—aside from the capstan etching option—I feel I’ve taken it as far as I can. I’ve gone over every potential friction point multiple times.

While waiting for the new belts and capacitors, I measured the original belts. They sit around 0.89–0.90 mm thick, whereas the FixYourAudio belts measure 0.98–0.99 mm. A ~10% increase in belt thickness seems quite significant for a mechanism engineered with such tight mechanical tolerances. Perhaps that extra mass, which adds rigidity, is driving up the W&F? Will see. At this stage, swapping to a proper 0.9 mm belt and recapping the servo circuit are my best bets to resolve the issue.

 

Very interesting matter.

I have two WM-2, one of each generation . I bought second one ( 2 gen) as a donor to fix the 1 gen unit volumen issues. When I open the second one I saw the volume circuitry was no the same, than the first gen. After installing the pot volumen on 1 gen I meant only one channel works well. I have to open the pot and fix It by glueing the metal traces to the plastic piece which carry them, heating with the solder the two spots which hold It all. Now the problem is the apreciable W&F. Usually is worse at the begginig, after several minutes the performance improves but still is high. Don't have the hardware to measure It but I bet higher than 0.6% . The central gear is the thick one. Not torque problems and the motor runs quietly and smooth. I like the sound of this walkman, maybe a little lack of high headroom but mediums,
bass and dinamic are very good in my opinión, Indeed deserves the efforts and keep on watching this thread for any clues.

Now I'm trying to fix the 2 gen unit, I got a pot volumen from an old Aiwa and place an order to Marian for the belt. But I have some wires that I don't kmow where exactly go and appreciate much a SM of the gen 2. @euroflash, could you please share It. If I am understand well you have find It.

Many thanks and good lack with the prohect!

 

@JUAN ANDRES MONTIEL DIEZ, send me a PM.

It appears there were several modifications after Supplement 2, as both of my WM-2 units vary slightly from the schematics in capacitor count and values.

Quick update on the repair: I received a set of replacement belts from a local supplier that matched the original 0.9 mm wall thickness of the factory Sony belts. As @Valentin predicted, switching to the thinner belt vastly improved Wow & Flutter.

I also recapped the donor unit, though it didn't make a noticeable difference in W&F. The original caps were still well within operating specs—the 220 µF caps matched the new replacements almost exactly, while the smaller 10 µF and 22 µF caps showed roughly double the ESR of the new ones, but still acceptable. By the way, the original caps had no brand markings, but they were obviously of good quality to hold up that well for 45 years.

I will post an update once I recap the primary WM-2 and sort out the "left channel tick" that recently developed on the donor unit (before recap). I read about a similar issue on this forum, here-> Sony WM-2. The difference is that "the tick" in the left channel in my unit only happens when playing tape. Hopefully this will help me to narrow down the problem.