Advice needed: AMS unit (aka rotary encoder) for MDS-W1

[imkidd57]

I am repairing an MDS-W1 deck, which has quite a few problems including both AMS knobs which will not work properly; there is no spring to the 'push' action and no response from the W1 deck itself to any push/pull movement. One of the knobs won't show a change on the display to rotating it, but you can feel the clicks. Thanks to minidisc.org there is a service manual, and I took off the front panel of the W1 and located the left side panel board with the AMS unit on. The potentiometer-like unit is described in the manual as a 'rotary encoder', and was relatively easy to desolder the pins and lugs and remove it from the board.

 

With the expected lack of a direct Sony spare part, I managed to disassemble the encoder to investigate and found it irrepairable, with some of the internal parts reduced to fragments which explains why it wasn't working! Looking on Farnell's website, I think I've found a possible replacement in terms of the various physical measurements, the right pin positions and is a push-on switch, but being completely new to rotary encoders I would like some advice on the electronic digital aspects of pulses per rotation (PPR) and detente (number of clicks felt when turning the spindle).

 

When disassembled, the original unit had a disc with radial copper stripes, of which there were 20. Can I assume that these were responsible for the pulses, as very small copper pickups - one of which was still in place in the wreckage - would have contacted the stripes and could have generated the pulses. If so, then would this be a 20 PPR rotary encoder? The number of clicks felt on another working MDS-W1 deck that I have is definitely 20 for a whole rotation, so this is should be the detente.

 

The possible replacement from Farnell is 24 PPR (and detente of 24), so would this make any difference to the performance and suitability? I can't see that an extra four clicks would screw things up as it doesn't seem to be doing anything other than change an observable number or menu option on the display.

 

Any advice and opinions gratefully received!

 

 

 

[kgallen]

Usually an Alps or Bourns and 18 or 24 pulses per rotation are usual. 
 

24 should do it. The main issue I’ve found is finding one with the correct shaft length.

 

I replaced one on my Tascam MD-CD1 but I’ve not needed to do a Sony.

 

Think I got mine from RS.

[imkidd57]

Fantastic - thanks very much!

Yes the one I located was an Alps; I’ll check RS as well.

 

[EDIT] Well Farnell didn't have the push switch version, but RS have a Bournes unit that seems to fit the spec. I'm picking a couple up this morning and will feed back success or failure ;-).

[kgallen]

On 2/15/2023 at 9:49 PM, imkidd57 said:

[EDIT] Well Farnell didn't have the push switch version, but RS have a Bournes unit that seems to fit the spec. I'm picking a couple up this morning and will feed back success or failure ;-).

Great! Please let us know how you get on and ping back if there are any other queries we might help with!

[imkidd57]

I can report that the Bournes RE unit is mostly physically compatible; the five (2 + 3) electrical pins are exactly right for distance and perfectly fit the holes in the panel control board. The two larger support lugs/clips were slightly too wide to fit in their rectangular holes in the board, but it was easy to file a bit off each side of the lug; I'd say about 0.5 - 1mm in total). Then it was a good snug fit onto the board and easy to solder everything on.

 

Very slight potential problems I can see are:

• the base of the replacement RE unit is thicker than the Sony original, so the shaft may protrude further out in the recess of the W1 front control/display panel. However it is only 1-2mm at most, and there may be some leeway in the AMS knob recess that can accomodate that. Otherwise the shaft length could be reduced with a bit of coarse filing.
• The diameter of the shaft is very slightly larger than the original, so it is a very tight fit for the AMS knob. I was reluctant to push it on fully as the plastic forming the D-socket of the AMS knob is thin and likely to be brittle. Even though there is a metal grip ring, I would rather try and reduce it with some sandpapering - it wont require much reduction though.
• The pressure needed to actuate the momentary push-on switch is more than the original. I am hoping that it's not enough to move the whole deck when being used - the WI is quite a light deck! Somewhere in the specification sheets it was possible to choose a different pressure for the push-on function, but I can't remember whether it was for the Bournes units.

Positive aspects are that the rotary clicks feel just like the original, and shaft is aluminium; consequently will be much more robust than the plastic of the original that was prone to snapping.

 

I've not had time to replace the board and test electrically but will do over the weekend.

 

New rotary encoder soldered in place:

 

 

 

Encoder unit size comparison: original Sony part is the black one.

 

[imkidd57]

Update...

 

Well it turns out that the thickness of the encoder base is quite important, as the recess in the front panel had exactly 4.8mm depth for the original to fit. When mounted on the PCB, the newer Bournes unit thickness of 6.6mm won't allow the PCB to lie flat where it belongs on the front panel, and fixing it with the screws would mean the PCB would bow out and put at risk the PCB traces, amongst other hazards! Also the push switches present on other parts of the PCB won't engage properly with the corresponding buttons on the front panel.

 

I thought of filing some of the front panel plastic structure away to accomodate the thicker encoder base, but that puts at risk the separate springy bits of plastic which connect the surrounding three panel buttons and provide the back-spring support to them when pressed.

 

Further searching around what is available in the UK reveals a low profile version of the Bournes encoder (model no prefix PEC11L) with a 4.5mm base, and there is a 20 PPR / 20 detente; 20mm shaft; momentary push switch.

 

Overall the code is: PEC11L-4120F-S0020. Farnells have it in stock so I will order a couple of those which should probably fit better.

 

 

 

 

 

 

[BearBoy]

Ah, how frustrating! Good luck with the lower profile encoder. Let us know whether that component does the job as I am sure the information will be very useful for other people facing a similar issue.

[imkidd57]

Thanks - well there has been some good progress. The low profile encoder fits great and allows the PCB to be replaced properly in the front panel, without any stress on it or the other switches. Also, the plastic AMS knob fits better on the shaft and no filing is needed to get it on (or off).

 

Reconnected the ribbons and wires, powered on the deck and it works! Rotating the dial moves through tracks on MDs, and the push-on switch plays the track. In the menu the rotation similarly moves back and forth through the options, and the push-on selects the one arrowed. However, going forward through the numbered tracks on an MD, or going through the menu options is via an anticlockwise rotation, whereas it should be clockwise (checked on a working MDS-W1 deck).

 

I don't know enough electronics to understand if this can be corrected, say by swapping the pins of the A and B contacts where they are soldered to the board (see 'quadrature output table' in previous post, lower left hand diagram, where I presume CW and CCW are the rotation actions). Would anyone be able to suggest a way of reversing the current action?

[kgallen]

Yes it would be A and B contacts. Strange the rotation should be reversed. This will be down to which of A or B leads. Usually those two signals are 90 degrees out of phase with, I presume, A leading. How strange, but otherwise great news. 
 

Is there anything in the data sheet for the part you used that might suggest this, or that it’s something that exists as a part variant?

[imkidd57]

Yes it's weird! I've looked through the PEC11L datasheet but to my untrained eyes I can't see anything about variants with different rotation directions. I have attached it in case there is something I missed.

 

It may come down to something I've done in the soldering; e.g. accidentally bridged to another pin or something! I will go back and check under magnification.

 

 

pec11l-777793.pdf

[kgallen]

Yea agree, don’t see anything in the data sheet. A leads B as I expected. 
 

I’ll aim to do some more research when I get on a ‘proper’ computer (rather than a phone!). I’ll dig out the part I used for my Tascam (I realise it will be different to what you likely need but it will give us a comparison).

[kgallen]

The two encoders I bought for my Tascam MD-CD1 are:

 

MD Pitch control: Bourns PEC12R-4230F-S0024 (=Alps EC12E24244A4), RS 781-6809

 

MD Jog: Bourns PEC12R-4225F-S0024, RS 781-6811

 

Both are 24 step with switch, D shaft (first being 30mm, second 25mm).

[kgallen]

In pictures:

[kgallen]

Interesting yours is ‘PEC11L’ and mine are ‘PEC12R’.

[kgallen]

Think the difference is the PEC11 has a bushing thread and the PEC12 is non-threaded. 

 

[sfbp]

L and R - are those the direction of "increase" when turned?

[kgallen]

Just now, sfbp said:

L and R - are those the direction of "increase" when turned?

Well I was musing that but I don’t see that in the data sheets. For both, CW rotation suggests A pulses before B. But that’s looking on a phone. I really want to get the two data sheets up side by side to compare…

[imkidd57]

Thanks very much for all the info, kgallen!

 

Second time around I chose the PEC11L as it was a low-profile version, and couldn't see a PEC12 equivalent and I never thought there would be a difference in rotation. All the PEC12Rs had a base which was too thick (~6mm); even the PEC11R I got first time. It's true though that both PEC11s are threaded. Here the three encoders to show the difference:

 

 

[imkidd57]

As a practical update, with your advice kgallen I reversed the connections A and B of PEC11L onto the PCB with some thin insulated wire, and the encoder rotation is now as it should be: turning clockwise increases the track number count, and also progession through the menus. Push switch still works as expected and the whole left-hand side PCB fits back onto the main front panel.

 

Because of the lack of support by the two pins now not soldered onto the board, as a precaution, I added a collar to the threaded bit of the shaft so it would fit exactly in the hole of the main front panel (you can see that the Sony original part is a lot thicker (9mm) and a snug fit in the hole). This should help counteract any inadvertent sideways stress on the remaining PCB connections during rotation of the AMS knob.

[kgallen]

2 minutes ago, imkidd57 said:

As a practical update, with your advice kgallen I reversed the connections A and B of PEC11L onto the PCB with some thin insulated wire, and the encoder rotation is now as it should be: turning clockwise increases the track number count, and also progession through the menus. Push switch still works as expected and the whole left-hand side PCB fits back onto the main front panel.

 

Because of the lack of support by the two pins now not soldered onto the board, as a precaution, I added a collar to the threaded bit of the shaft so it would fit exactly in the hole of the main front panel (you can see that the Sony original part is a lot thicker (9mm) and a snug fit in the hole). This should help counteract any inadvertent sideways stress on the remaining PCB connections during rotation of the AMS knob.

Thanks for the update and info. That is weird. I wonder if Sony had a custom part made for some reason. I’ve suspected that before - the headphone volume control on one of my MDS-E10 or E12 didn’t seem to be a standard part. That mounts to the PCB to the side so the flat on the shaft needs to be at a different position to get the knob marker to line up between the min and max position. Also I couldn’t find the correct resistance (I can’t remember now if it was 1k or 5k).

 

For @sfbp I think the ‘L’ is maybe low profile and the ‘R’ is probably ‘regular’.

[imkidd57]

21 minutes ago, kgallen said:

Thanks for the update and info. That is weird. I wonder if Sony had a custom part made for some reason. I’ve suspected that before - the headphone volume control on one of my MDS-E10 or E12 didn’t seem to be a standard part. That mounts to the PCB to the side so the flat on the shaft needs to be at a different position to get the knob marker to line up between the min and max position. Also I couldn’t find the correct resistance (I can’t remember now if it was 1k or 5k).

 

Not at all - thanks so much for all your help 😀! I'll add some more pictures during drive B modification, and hopefully will be a useful resource in case a fully compatible encoder can't be found.

 

I noticed that the same Sony part number for the encoder is listed for the MDS-JE520 and JB930.

[kgallen]

Fab!

If you write something maybe you’d post it on the wiki too?

 

Add a repair section on this page:

https://www.minidisc.wiki/equipment/sony/deck/mds-w1

[imkidd57]

Well repairing drive B side turned into a flipping nightmare! The solder ring pads surrounding the holes for A and B of the encoder became detached - probably too much messing about with the soldering iron, sucker, flux, desoldering ribbon, and IPA. Then the A and B legs fell off the second encoder after trying to carefully bend them out of the way to reverse the connections. Track repair with fine wire was successful but was then very hard to connect the wires to the stumps of A and B; and now the encoder itself may be knackered 🤣. To highlight the nightmare, I ordered some more encoders from Farnell but all I received today was a bag of hexagonal nuts 🤪. I do wonder at their warehouse QC...

 

On review this whole affair may be better solved by reversing the A & B connections at the end of the circuit board where the tracks from the pins lead to a ribbon connector that joins with the central display board, rather than messing about with the fragile pins of the encoder. The PCB appears to be one-sided; so now that the ring pads for A and B have gone, and the tracks have become non-continuous, after looking at the diagrams in the service manual I am proposing to take jumper wires from each A & B leg and solder to the reverse points on the pads of the distant ribbon connector (pins 8 & 9). It's not very far (~2 cm) and should be far easier to manipulate than bending the encoder legs again.

 

Any thoughts?

 

Here is the schematic from the SM:

 

 

 

 

... and PCB mask with actual positions:

 

[kgallen]

Yea just cut the tracks and solder to the through hole pins of CN671 - possible since those traces don’t go anywhere else as you’ve spotted from the PCB layout. Pretty easy if you’ve got some solid core kynar mod wire like this: https://www.ebay.co.uk/itm/262293229036?mkcid=16&mkevt=1&mkrid=711-127632-2357-0&ssspo=cPqkJZIKSwG&sssrc=2349624&ssuid=nsjPidUXSvu&var=&widget_ver=artemis&media=COPY

 

I’m still a bit bemused that the rotation is reversed! 

[imkidd57]

Many thanks again @kgallen. Sorry for late reply - Farnell stock control in a bit of a pickle but I've fianlly got more encoders now. Will report ASAP.

[imkidd57]

OK well finally back with a report of success with the Drive B encoder. Did the connection reversal for pins A and B of the encoder (1 and 3 on the PCB diagram from the service manual). No need to cut the tracks as both solder ring pads had been lost but I still had to undo the track repair. Also, in the process of de-soldering the encoder for the nth time, the solder pad for pin 4 also came off, so had to make a new bridge connection for that.

 

Another figure of the rather messy repair: hairs on the area were from cotton buds during cleanup. Pink stuff is nail varnish remaining from the earlier track repair.

 

 

Anyway after covering the area with Kapton tape and reassembling the rewired PCB back into the front panel, all functions of the encoder were exactly as anticipated (clockwise + counter clockwise) and the push-button switch works fine. Earlier worries about the amount of force needed to push the switch were allayed and it is not so forceful as to move the whole deck.

 

With the above success I revisited the left-sided encoder (Drive A) of the W1 and rewired it according to the principle above, which is a much more solid repair as all the legs can be soldered to the board and the bridge wires added. It's a different wiring layout, which I can include in a repair section of the MD Wiki as requested.

 

Thanks everyone, especially of course @kgallen for all the advice.

[BearBoy]

Good to hear you've got it all sorted 🙂 And I'm sure what you've discovered along the way will help other people faced with a similar issue in the future.