Philips N4308/55 refurbishment (plenty of faults!)

[dazzlevision]

Hello,

I’m in the process of restoring a Philips N4308/55 two speed, four-track mono all-transistor tape recorder. My particular unit was made in 1972. The later production "/55" version has changes to the drive belt arrangement.

I was fortunate that none of the four belts had turned to goo, just the pause brake rubber “paddle”, which was a lot quicker and easier to remove and clean up.

I also removed all the sliding metal linkages connected to the operation keys and the metal slider that moves the two turntable drive pulleys, in order to clean up all the old grease and apply fresh.

Before I fitted some new belts and the eight turntable rubber “paddles”, I thought I’d give the electronics a quick check over. Initially, the amplifier produced a hiss on the speaker with volume set at maximum (in both play and record modes), but then it cut out – dead. Tapping the PCB could sometimes bring it back, so I first checked all the soldered joints in the audio output stage, but nothing seemed amiss.

When the fault was present, there was 25V on the collector of the AC187/01 audio driver transistor. The emitter and base biasing resistances were checked and were spot on (luckily, Philips used carbon film resistors long before UK manufacturers did). So, I then checked dc voltages in the audio output pair circuit (at all connections to the AD161/162 output pair, except for the AD162’s collector), and these were much the same at 25V (the value of the LT supply line).

I suspected the AC187/01 driver and replaced it with a used spare from “stores”, but the fault was just the same.

I then replaced the AD161/162 output pair and upon testing, one had an unusual c-e junction voltage reading when using my DMM in diode test mode (normally reads o/c in one direction and around 0.15V the other). Even so, when I switched on, the fault was still the same. I then did continuity tests on the notorious spidery Philips print tracks in the audio output stage, but these were all good.

Next, I applied freezer spray and this seemed to prompt the fault to come and go when applied to the AC187/01. I then fitted another second hand AC187/01 and that cleared the fault. I get the impression that these devices (just a low power Germanium device in an Aluminium heat sink block that can be screwed to the chassis) might suffer from tin whisker growth, in the same way as the infamous AF11x series of RF transistors. Much to my relief, fitting another used AC187/01 cleared the fault completely.

However, in record mode, there was still some intermittency (of a different sort) when the PCB was tapped; the level of background “hiss” went up and down. This proved to be due to high resistance contacts in one of the two leaf switches screwed to the AD161/162 heat sink, one of which changes the amplifier’s frequency response curve for the two different tape speeds. Gently rubbing a Servisol soaked cotton bud to the contact areas sorted this out.

Finally, the record level meter's pointer didn’t move when an input signal was applied and this WAS due to a minute crack in a section of the extremely fine print that connects T9’s (AC125) emitter to 0V.

For good measure, I removed the sliders from the three PCB mounted switches (the delicate type, with the gold plated contacts on the slider - widely used by Philips) and cleaned/lubricated them before reinsertion.

I was expecting a lot of work on the mechanical part of the recorder, but not with the electronics! I’ll let you know how I get on with fitting the new drive belts and those eight turntable “paddles”

Regards,

Dazzlevision

[dazzlevision]

Hello again,

I've just measured the resistance between the Aluminium heat sink of the defective AC187/01 transistors and the lead out wires. In both devices, there is leakage to the three leads and this does alter dramatically when the device is tapped.

Regards,

Dazzlevision

[Jimmyhaflinger]

i also had to replace the driver transistor in two of my 4308s

[Nickthedentist]

Well done.

I had one of these from a jumble sale as a teenager (late 1980s) and never got to the bottom of a very intermittent cutting-out of the audio on PB.

Nick.

[8 Tracker]

Excellent, another one saved for the nation! These intermittent faults can really try your patience. Nice that you persevered. I have 4 of these (2 4307s and 2 4308s) and no electronic faults on any of them (yet). Am I just lucky?
Mark.

[Jimmyhaflinger]

one of my 4308s also had an o/c volume pot and a dead emitter bypass capacitor causing low gain and tinny sound

[dazzlevision]

Originally Posted by 8 Tracker

Excellent, another one saved for the nation! These intermittent faults can really try your patience.

Hello Mark,

I must admit that I was getting a little fed up with tracking down this intermittent fault and initially, it didn't occur to me that an ACxxx device might suffer from the same tin whisker syndrome that the AF11x family does. I have never encountered such a "microphonic" effect in an AC1xx device (and I have plenty of equipment that uses them).

My N4308 looks to have been stored in a loft or similar unheated and probably damp environment, as the brushed Aluminium trim on the cabinet top has some corrosion in places and I suspect it hasn't done the exposed Copper track areas on the PCB much good either.

It may be that the storage environment was such as to prevent the drive belts turning to goo, as they usually do. Only the main (2mm square cross section) belt was in fact broken; the others appear to be serviceable (albeit a little stretched).

Regards,

Dazzlevision

[dazzlevision]

Hello,

Well, I've set up the quiescent current in the AD161/162 output pair and the electronics side now seems correct and stable.

So, after first cleaning all pulleys and driving surfaces, I then fitted new belts and turntable clutch "paddles", together with a new Pause brake "paddle".

The mechanism then operated correctly in all modes.

I adjusted the height of the exit tape guide post (as it had been twiddled - the red locking paint applied at the factory had been disturbed), to ensure the tape travelled over the pinch roller/capstan correctly, with no buckling or curling (and then applied fresh locking paint). All other red factory locking paint was thankfully undisturbed.

I had to fit a new wire-ended lamp to restore the illumination of the record level meter.

The next defect was a broken shaft on the volume control knob. Luckily, I had a spare from a long since dismantled N4307.

The only other issue is broken sliding catches on each side of the lid. Now, I had read on this forum that they can be swapped over, to use the undamaged side of the sliding knob latch, but when I attempted to removed them, they broke (rather brittle plastic). Perhaps I should have gently warmed them up with a hairdryer first? However, I do have a good N4307 lid in the loft and I hope they will yield some good catches (or maybe I might just swap over the Aluminium trim strip at the front, to avoid breaking these as well).

I plan to check the overall frequency response by making some test tone recordings at 3.75 ips, to give some idea of the extent of head wear.

Regards,

Dazzlevision

[Jimmyhaflinger]

the PCB tracks are tin-coated on MK1 (made in Holland) machines and painted with green solder resist on MK2 (made in Austria) machines, both are usually pretty resistant to corrosion

the catches are simple to remove once you get the hang of it, just align them with the hole and gently lift them and push them out bottom first, just don't try to flex the part that was once flexible before they got brittle

also, they still work even when broken

[paulsherwin]

Originally Posted by dazzlevision

I must admit that I was getting a little fed up with tracking down this intermittent fault and initially, it didn't occur to me that an ACxxx device might suffer from the same tin whisker syndrome that the AF11x family does. I have never encountered such a "microphonic" effect in an AC1xx device (and I have plenty of equipment that uses them).

There does seem to be increasing evidence of tin whisker growth in some AC1xx types. I wonder if it's just taken longer to become apparent because of the lack of an earthed body.

[dazzlevision]

Originally Posted by Jimmyhaflinger

the PCB tracks are tin-coated on MK1 (made in Holland) machines and painted with green solder resist on MK2 (made in Austria) machines, both are usually pretty resistant to corrosion

Yes, I also have an EL3558, which I suspect was the predecessor model to the N4308, as it has similar facilities and a very similar tape deck. The EL3558 is made in Holland and features a much larger PCB, with all Germanium transistors and solder coated PCB tracks, rather than the later green solder resist. The PCB seems rather more robust.

The main problem with the EL3558 compared to the N4308, is that it uses two more of those rubber "paddles" as turntable brakes, whereas the N4308 uses some sort of friction material as might be used in car brake pads/shoes and is long lived (no sticky black "goo"!). In addition, the large PCB covers up a lot of the underside of the tape deck, making access more difficult.

Regards,

Dazzlevision

[Jimmyhaflinger]

the 1965 EL3558 evolved in the EL3578 and the identical N4306 in 1967 before the N4308 came out in 1969

the EL3578/N4306 was similar to the EL3558 but had the same improved brakes as the N4308

anyway, solder coated PCBs were used in the N4308 too until about 1973 when the facelifted model came out and production was moved to Austria...

[ricard]

Originally Posted by paulsherwin

There does seem to be increasing evidence of tin whisker growth in some AC1xx types. I wonder if it's just taken longer to become apparent because of the lack of an earthed body.

It would definitely make sense.

I've had problems in several Tandberg 12/1200X's, where there is an AC127/128 driver pair in the output stages connected to a common heat sink (which also may be connected to some point in the circuit, I can't remember now). These apparently suffer from the tin whisker problem; well over half the machines I've been in contact with have had the problem of one or both output channels going silent, which has been aleviated by removing the driver transistors and 'zapping' away the tin whiskers and re-mounting them.

[dazzlevision]

Hello,

I have fitted a spare N4307 lid that I had and swapped over the aluminium trim strips to show the correct model number on it (luckily, the adhesive had become ineffective after 40+ years). The fixing clips on the N4307 lid were perfect.

I also fitted a perfect clip-on cover that I had (from the same scrapped N4307), for the mains lead stowage compartment on the underside of the recorder. The clip part is often broken or the cover is missing.

Cosmetically, all I need now is a speed change knob.

I have to repair my Gould OS300 'scope before I can do a frequency response check recording and am awaiting a TIP29A transistor for it's -11V LT rail shunt stabiliser circuit. When that is up and running, I can do the response test and hopefully, it will be close to the manufacturer's spec (up to 14kHz on 3.75 ips, +/- 3dB).

Regards,

Dazzlevision

[8 Tracker]

Originally Posted by dazzlevision

luckily, the adhesive had become ineffective after 40+ years

Indeed! The various trims on my few have all fallen off at various times. Even the metal deck cover comes adrift. Evo-stick seems about the best solution for a permanent repair.

Mark.

[dazzlevision]

Hello again,

It looks like the Rec/Play head is worn, as the HF response is poor.

I have completed repairs to my 'scope, so I tried (for the first time ever) to do a frequency response test by recording single sinewave tones at 100, 1k, 10k and 14kHz in succession. Recorded using a constant amplitude tone injected into the "Radio IN/OUT" socket input and playback level measured on the output pin of the same socket.

You need to set the record level using the highest frequency tone, as the record amplifier has a rising frequency characteristic, to overcome the tape’s magnetisation characteristics and the effect of the bias current.

On playback, there was no detectable output at 10 or 14k (according to the service manual, it should be +/- 3dB up to 14k). I put my scope across the record head, to see what the waveform looks like (never done that before either) and the HF bias signal is much larger than the audio tone, so it is difficult to see it.

I also did these tests on a trusty Thorn DC43 deck mono valved machine (at the same tape speed - 3.75 ips) and although there was more output at 10 and 14k on playback, the magnitude of the spot frequency tones were also minimal compared to the bias signal (across the R/P head).

Can anyone with more experience of such testing shed any light/further info on this?

Thanks,

Dazzlevision

[ricard]

Originally Posted by dazzlevision

I also did these tests on a trusty Thorn DC43 deck mono valved machine (at the same tape speed - 3.75 ips) and although there was more output at 10 and 14k on playback, the magnitude of the spot frequency tones were also minimal compared to the bias signal (across the R/P head).

It is true that the bias signal is significantly larger than the audio signal at the recording head. It's often very hard to get an idea of the audio waveform this way.

If you can disable the bias oscillator temporarily, that's one way of doing it. On valved machines it's often possible to pull the output stage/oscillator valve, but of course on a transistorized machine it's a bit more intricate.

[dazzlevision]

Hello Ricard,

That's a good idea and I will try it out. On my Thorn DC43 recorder, the erase oscillator is an ECC82 valve, which can simply be unplugged.

On the N4308, I think the erase oscillator can be disabled by unsoldering a print connection, to isolate the slider switch contact that enables LT feed to the erase o/c when both the Record and Play keys are pressed.

Regards,

Dazzlevision

[SteveCG]

The HF response depends quite critically on the recording bias level being set appropriately for the make and even sometimes the thickness of tape you are using.

Are you using LP (Long Play) or DP (Double Play) on this 1/4 track machine (you should be!)?

Also, whose make of tape are you using? - and does it have a product number/type?

[dazzlevision]

Hello StevCG,

I have checked that the bias current (equated to a voltage developed across a low value resistor in series with the R/P heads) is within the range of values specified in the Philips service manual.

As to the tape used, it is one of many secondhand reels that I have inherited with various recorders of the 1960s and early 70s. The one I have used for these tests is a "Scotch" branded tape with the old "3M" logo embossed on the spool and without a storage case. So, it's not immediately obvious what thickness of tape it is.

I realise that the electronics/electrical adjustments of tape recorders made after "low noise" tapes were introduced circa 1970 were altered to make optimum use of these. That may well cause a distortion of the record-play frequency response of earlier recorders when used with "low noise" tapes.

What I wasn't aware of, was the need to avoid use of standard play tape on four track machines. Do you know why this is? I thought the main difference was the thickness of the plastic tape, not in the magnetic oxide coating upon it...?

I shall certainly try another brand of tape and ensure that it isn't Standard Play.

Regards,

Dazzlevision