So, things have indeed become interesting.
The convertor card arrived. It is different to the one shown. It has no 384 (USB?) input. It has two i2s inputs, and a flappy door on the optical.
I bought a second CD player, too, as I've been enjoying the first so much. I ordered a Panasonic SL-PC364, hoping it is similar to the Technics. It arrived with a broken lid and didn't play discs. I received a refund and was told to keep/dispose of the player.
I ordered a second smaller converter card with only the optical and coax connections, which has been lost in shipping. Sigh.
I'm sitting here with the Panasonic open. Internally it is identical to the Technics SL-PC505. The circuit board is stamped SL-PC505, and the ICs are stamped with Technics decals. The only differences are the IR (705) and wired (505 & 364) remotes.
Both drive belts are perished, which will explain why it doesn't play discs. Plays fine with the case off if I give it a helping hand to rotate the carousel, then to lift the disc.
Anyway, I'm going to look for my soldering iron...
The convertor card arrived. It is different to the one shown. It has no 384 (USB?) input. It has two i2s inputs, and a flappy door on the optical.
I bought a second CD player, too, as I've been enjoying the first so much. I ordered a Panasonic SL-PC364, hoping it is similar to the Technics. It arrived with a broken lid and didn't play discs. I received a refund and was told to keep/dispose of the player.
I ordered a second smaller converter card with only the optical and coax connections, which has been lost in shipping. Sigh.
I'm sitting here with the Panasonic open. Internally it is identical to the Technics SL-PC505. The circuit board is stamped SL-PC505, and the ICs are stamped with Technics decals. The only differences are the IR (705) and wired (505 & 364) remotes.
Both drive belts are perished, which will explain why it doesn't play discs. Plays fine with the case off if I give it a helping hand to rotate the carousel, then to lift the disc.
Anyway, I'm going to look for my soldering iron...
Didn't look for my soldering iron last night. Became distracted by cleaning and de-scratching the case. Novus 2 and Novus 3 plastic polishes are excellent for scratch removal.
Picked the circuit board up again this morning to strategise whether to solder to the decoder IC, DAC IC, or a combination. Probably to the DAC because it doesn't matter if I break it, while do I need to save the decoder...
...and on the reverse of the board I saw these jumpers connecting the DAC and decoder ICs
I can't help noticing the label 384FS is uncannily like the 384 input missing from my convertor board, plus the acronym FS for Frame Select. And when I think about it, 384kHz is 24Khz x 16, which would be a nice clock frequency for a CD at 22Khz x 16, plus space for a spare bit. So I'm going to call that MCLK.
LRCLK, SCLK, SDATA, +5v, GND all speak for themselves.
Anyway, taking out my mother today, so won't get to this just yet.
And by "taking out," I don't mean popping a cap in her for insurance or inheritance purposes.
Picked the circuit board up again this morning to strategise whether to solder to the decoder IC, DAC IC, or a combination. Probably to the DAC because it doesn't matter if I break it, while do I need to save the decoder...
...and on the reverse of the board I saw these jumpers connecting the DAC and decoder ICs
I can't help noticing the label 384FS is uncannily like the 384 input missing from my convertor board, plus the acronym FS for Frame Select. And when I think about it, 384kHz is 24Khz x 16, which would be a nice clock frequency for a CD at 22Khz x 16, plus space for a spare bit. So I'm going to call that MCLK.
LRCLK, SCLK, SDATA, +5v, GND all speak for themselves.
Anyway, taking out my mother today, so won't get to this just yet.
And by "taking out," I don't mean popping a cap in her for insurance or inheritance purposes.
Last edited:
@rfbrw Well, yes and no. It's been an evening of interesting reading, spurred on by the enthusiasm of replacing the drive belts and watching the SL-PC364 burst into activity.
The original 1986 Phillips I2S is not coming out of the MN6626, which I now see outputs the EIAJ/Sony format.
Somewhere over the years, Phillips became NXP, and the I2S protocol expanded to four modes in some NXP documents, and still only the original mode in other documents.
So all I need to do is find or build an EIAJ to SPDIF converter. Or find/make an EIAJ to Phillips I2S convertor.
I can find proposed gate diagrams to convert from EIAJ to I2S. But no-one reporting success. So I'm disinclined to go that route.
On eBay I found a $25 card with DIP switches that claims to convert
I already soldered the Dupont lines to the CD player circuit board, so the next iteration will be fast...
...but it won't because now I have to wait four weeks for this new part to be delivered. Sigh. And I can't find any other alternatives.
The original 1986 Phillips I2S is not coming out of the MN6626, which I now see outputs the EIAJ/Sony format.
Somewhere over the years, Phillips became NXP, and the I2S protocol expanded to four modes in some NXP documents, and still only the original mode in other documents.
- Original "standard" I2S with offset-bit left-justified arbitrary length words
- Non-offset left-justified arbitrary length
- EIAJ/Sony 16 bit right justified
- 24 bit right justified
So all I need to do is find or build an EIAJ to SPDIF converter. Or find/make an EIAJ to Phillips I2S convertor.
I can find proposed gate diagrams to convert from EIAJ to I2S. But no-one reporting success. So I'm disinclined to go that route.
On eBay I found a $25 card with DIP switches that claims to convert
- coax -> optical & I2S
- analogue -> optical & coax & I2S
- all four I2S modes -> optical & coax
- optical -> coax & I2S
- analogue -> coax & I2S
- all four I2S modes - coax
I already soldered the Dupont lines to the CD player circuit board, so the next iteration will be fast...
...but it won't because now I have to wait four weeks for this new part to be delivered. Sigh. And I can't find any other alternatives.
Last edited:
An I2S -> SPDIF convertor card that had been lost in the mail arrived today.
Unexpectedly, it can be configured for a right-justified 16-bit data stream, so is something for me to play with this evening after work.
Configuration involves desoldering and resoldering miniscule SMD resistors, which may be beyond my skills. I'll get there somehow, even if I need to buy resistors with leads.
Unexpectedly, it can be configured for a right-justified 16-bit data stream, so is something for me to play with this evening after work.
Configuration involves desoldering and resoldering miniscule SMD resistors, which may be beyond my skills. I'll get there somehow, even if I need to buy resistors with leads.
Well that didn't work.
The soldering looks great, my multimeter confirms everything is correct, and even my daughter is impressed with the appearance.
The Toslink port lights up, but no sound from the hifi.
Looking back at the 6626 data sheet, it describes the output as the EIAJ format, which should be right justified, and LSB last.
Looking at the pin-outs, it is described as MSB first, which is a subtle difference, and might indicate left-justified.
Soldering those SMD are difficult for me - I need to stack reading glasses! - so I'm hesitant to start moving them again.
An I2S - SPDIF board with DIP switches to change formats is only days away, so easier for me to wait for that.
The soldering looks great, my multimeter confirms everything is correct, and even my daughter is impressed with the appearance.
The Toslink port lights up, but no sound from the hifi.
Looking back at the 6626 data sheet, it describes the output as the EIAJ format, which should be right justified, and LSB last.
Looking at the pin-outs, it is described as MSB first, which is a subtle difference, and might indicate left-justified.
Soldering those SMD are difficult for me - I need to stack reading glasses! - so I'm hesitant to start moving them again.
An I2S - SPDIF board with DIP switches to change formats is only days away, so easier for me to wait for that.
I picked up an oscilloscope at Amazon. I haven't used one for 30 years.
Next day delivery for $41. It's the size of a credit card, and I don't have to wait for the CRT to warm up. Amazing!
For anyone looking, it's the 154 model with 18MHz bandwidth. There are slightly cheaper ones, but with MUCH lower bandwidths.
Anyway, back to the CD player.
I left the Dupont lines trailing out of the case for quick iterations.
Notes for my own reference as well as for readers'.
Does that all seem right to me?
Next day delivery for $41. It's the size of a credit card, and I don't have to wait for the CRT to warm up. Amazing!
For anyone looking, it's the 154 model with 18MHz bandwidth. There are slightly cheaper ones, but with MUCH lower bandwidths.
Anyway, back to the CD player.
I left the Dupont lines trailing out of the case for quick iterations.
Notes for my own reference as well as for readers'.
- 384FS purple has a 17MHz 1.2v sine wave
- LRCK grey has a 44.1KHz 5.9v square wave
- SCK white has a 2.83MHz 5.9v square wave
- SDATA black has an irregular square wave, 5.9v, fractionally more than 700us, which gives me <1.43MHz. /16bits = <89KHz. Nice, let's call it 88.2KHz so that I have 44.1KHz stereo.
Does that all seem right to me?
- SDATA - Yes. 44.1KHz x stereo x 16bit = 1.41MHz irregular data.
- LRCK - Yes. 44.1KHz square. Up for left, down for right. Or vice versa.
- SCK - Uh, why is this double SDATA frequency? It needs to be the same as SDATA frequency. I may need something to halve the frequency. Or find a different line. I wonder if this is why Technics used to claim 2x & 4x oversampling? Not going to help me in 2023.
- 384FS - What? Why 17MHz? Not part of the I2S spec anyway. I'll ignore it for now. And then consider looking for a different line.
I went back to @huggygood 's comment about unpopulated digital output stage from the 6625 on other models.
I have the MN6626, and no unpopulated stage, nor anything that screams out at me.
I dug around on later models and found the Technics SL-PS740A, also with an MN6626, and with an optical out.
It piqued my interest.
I nearly weed my pants when I found the MN6626 has a single pin, unused in my SL-PC705, that connects directly to a TOSLINK transmitter socket.
I'll have breakfast while thinking about whether to connect that pin to an RCA socket, or to a TOSLINK that I can rip out of an old device, or directly onto the output stage of the converter cards I have that are now superfluous, to achieve both TOSLINK and RCA.
TOSLINK stops me electrocuting one device with another, which would be nice. Probably should insert decoupling capacitors if going the RCA route.
MN6626 to TOSLINK in SL-PS740A https://elektrotanya.com/technics_sl-ps740a.pdf/download.html
Well that couldn't be any more confusing.
In the diagram above the IC pins count clockwise from the chamfered corner & dimple.
In real life they are anti-clockwise / counter-clockwise depending on your accent.
Anyway, the SL-PS740a service manual is much more helpful than the SL-PC25 manual (remember that I have neither - I do not have a manual for my SL-PC705).
Realising that the DAC is sending a clock signal to the decoder, not the other way, I snipped that connection and connected XCK to the converter card. CD player wouldn't turn on, so I reversed that.
Then I soldered a Dupont lead to pin 23 of the MN6626 ("TX", which I infer as SPDIF for the optical transmitter), with plans to connect that and GND to an RCA socket. Oops - bridged two pins on the MN6626 with solder, and can't find my wick, so will need to get some more.
In the diagram above the IC pins count clockwise from the chamfered corner & dimple.
In real life they are anti-clockwise / counter-clockwise depending on your accent.
Anyway, the SL-PS740a service manual is much more helpful than the SL-PC25 manual (remember that I have neither - I do not have a manual for my SL-PC705).
Realising that the DAC is sending a clock signal to the decoder, not the other way, I snipped that connection and connected XCK to the converter card. CD player wouldn't turn on, so I reversed that.
Then I soldered a Dupont lead to pin 23 of the MN6626 ("TX", which I infer as SPDIF for the optical transmitter), with plans to connect that and GND to an RCA socket. Oops - bridged two pins on the MN6626 with solder, and can't find my wick, so will need to get some more.
Last edited:
Got wick
Removed the blob
Added a Dupont lead to pin 23
Connected TX & GND to an RCA
Nothing from the hifi. Hmm.
Oscilloscope shows an asymmetric square wave
Oscilloscope also shows an analogue music waveform where it should be, so I've not broken stuff. Hmm.
How is that TX line informing the optical out on the SL-PS740A in the diagram above?
I'll have a pancake while I think about it.
Removed the blob
Added a Dupont lead to pin 23
Connected TX & GND to an RCA
Nothing from the hifi. Hmm.
Oscilloscope shows an asymmetric square wave
Oscilloscope also shows an analogue music waveform where it should be, so I've not broken stuff. Hmm.
How is that TX line informing the optical out on the SL-PS740A in the diagram above?
I'll have a pancake while I think about it.
Also, PKPK should be 2.2v, not 5.1v
Something is amiss.
Compared to the SL-PS700, too, which is also 6626.
Looked at some 6625 and 6627 circuits, too. Similar circuit design, but different pin outs.
The 6625 and 6626 circuit diagrams have a MEMP-EMPH connection from the system controller to the signal processor. In the block diagrams, the MEMP feeds into the Digital Signal Interface part of the signal processor (which is where TX comes from). I can't imagine a good reason why an emphasis command is required for SPDIF, but I don't have any other ideas.
Iterations become tedious because I must remove and invert the mainboard to work on it, then reassemble to chassis to play a disc to create a signal. Tempted to cut a hole in the chassis under the ICs to speed things up.
Something is amiss.
Compared to the SL-PS700, too, which is also 6626.
Looked at some 6625 and 6627 circuits, too. Similar circuit design, but different pin outs.
The 6625 and 6626 circuit diagrams have a MEMP-EMPH connection from the system controller to the signal processor. In the block diagrams, the MEMP feeds into the Digital Signal Interface part of the signal processor (which is where TX comes from). I can't imagine a good reason why an emphasis command is required for SPDIF, but I don't have any other ideas.
Iterations become tedious because I must remove and invert the mainboard to work on it, then reassemble to chassis to play a disc to create a signal. Tempted to cut a hole in the chassis under the ICs to speed things up.