Philips N4450

[TonyDuell]

I know external photo links are somewhat frowned-on here, but there is no way I could upload all these photos. I have put some pictures of the machine here :

https://www.***********/photos/tony_d...57684388926024

[Welsh Anorak]

Hi
Always fancied one of these - they came out when I was at school and appealed to me. However a tad beyond my pocket money.
Very much of the time - reminds me of the N1700 VCR in its internals. From experience I'd change those PCB mounting electrolytics on the flip-flop PCB as they have a habit of shorting. The Lockfits and AC128s you've mentioned - I'd be tempted to uprate the latter as they never seemed up to the job in other Philips open reels of the period.
How are the reel motors? I know they were a weak point in this machine.
Good luck with the restoration - I like the replacement tapped screw receivers you've made.
Glyn

[TonyDuell]

I think it's a wonderfully complex recorder and totally 'Philips' (meaning unconventional, but it will work very well). I've wanted one ever since I heard it existed, I was very lucky to find it (and the seller, not on this forum, was close enough that I could collect it, which was a definite avantage).

I also have the simpler N4510, and in that machine 4 of the 6 AC128s in the flip-flops were leaky. So I changed the lot. I will probably change the lot in the N4450 too. Based on my experience with the N4510, I suspect silicon transistors, like the 2N3906, will work fine there. I am not sure what you mean by the PC mount electrolytics in the flip-flops. If you are looking at my photo of that module then the 4 metal cans are AC128s, not capacitors. I can't see any electrolytics on there.

I've not tested the reel motors. If all else fails I'd have a go at rewinding them. Small DC motors are not hard to do most of the time. The do spin freely though (at least after I cleaned off the Evil Goo...)

I felt that using woodscrews was a nasty cost-cutting measure, hence my replacement blocks and plates. Still better to cut costs on something like that than on the rest of the machine (which looks to be very well made).

[TonyDuell]

Thinking about it some more, I am not sure that 2N3906s will be suitable in the flip-flops of this machine. Some of these transistors switch the current to the deck solenoids (pinch rollers, etc) and I hadn't realised until I looked it up that an AC128 (the original transistor) will pass 1A.

I wonder if a BC640 would be better.

[TonyDuell]

A (small) progress report :

I finally got round to plugging it into the mains. At power-on the control system went crazy, and the left spool motor squealed like mad. To avoid problems there I powered down, pulled out the 2 flip-flop boards and powered up again.


All the switch illumation lamps came on, so did the 2 for the level meters (and the track switch controls them). The clock motor was running nicely. The capstan motor was running (with no belts obviously the capstans weren't turning), and the speed control switch does just that. Obviously I don't know if the speeds are correct but it runs.

I removed all 8 AC128s from the flip-flop boards and replaced them with 2N3906s. Maybe a bit marginal, but if they fail I can fit something with a higher maximum Ic. Put the flip-flops back in. Also removed the pulley from the left spool motor, cleaned it and put a drop of light oil on the bearing thus exposed. Powered up again. The control system seems fine, the motors and solenoids seem to run when they should. I will order a belt kit.

I've tried connecting a pair of cheap-n-nasty speakers to it. There is more background noise than I would like, probably due to the lockfit transistors. As I can easily get T092-cased versions (as I understand it, the BC547 is the same transistor as the BC147, just in a different package) I will get some. In AMP mode (by-passing the tape unit), the left channel gain seems much lower than the right. Maybe lockfit transistor problems again. I think the way to sort that out is to put a signal generator on the input and trace signals with a 'scope.

The good news is that in play mode (actually stop mode, since the electronics is set to play in that state), I get a buzz from each speaker when I bring my old demagnetiser up to the heads. Again left channel is quieter. Also works after selecting right-to-left mode and stopping it again. So the replay heads look to be doing something.

[Sideband]

Back in the day, there was a correct way of setting up playback levels using a calibrated tape recorded with a 1Khz signal. I seem to recall two levels being on the tape, one at 0dB and one at -20dB.

The record levels were set up in a similar way with a calibrated signal generator and a special 'service' tape. Firstly the bias had to be adjusted correctly then signals were fed in at -20dB and 0dB, the idea being that there was minimal difference between 'A' and 'B' (off-tape) monitoring.

[TonyDuell]

I may not need to use a calibration tape. I get low gain on one channel in AMP mode (not using the tape). If I touch the appropriate pins of the tuner input (all controls set appropriately) I get a much stronger buzz on the right channel than the left. However I have now found that a microphone plugged into each microphone socket in turn gives the same deflection on the appropriate level meter when I speak into it. This suggests to me (confirmed by the unequal buzzes when I put the demagnetiser on the heads) that the fault is in the tone control or power amplfiier areas. I will inject a steady signal and trace it with a 'scope.

[Reelman]

Glad to hear things are progressing Tony, gives me a bit of incentive to resurrect mine. When mine was working I rewired the timer switch to actually switch on and off the timer motor since it was the noisiest thing on the machine when running!
Good luck with the rest.
Peter

[TonyDuell]

Yes, my timer makes a bit of noise too. It's easy enough to disable it when testing, the motor is wired via a couple of small faston terminals to a tagstrip on the chassis rail. I do find the power supply for that motor to be typically Philips too. Not run off the primary of the mains transformer (as an autotransformer) and not run via a dropping resistor. No, it's a dropper capacitor with a VDR in parallel with the motor, I guess to allow it to run off any sane mains voltage.

The linkage in my timer that does the carry from the minutes drum to the hours drum is sticking. I will investigate that.

But I first need to look for the autoreverse fault (yet another oddity, there's a monostable with an NPN and a PNP transistor cross-coupled) and the audio fault. I want at least to know the approximate area of the latter.

[TonyDuell]

A bit more on this machine.

I thought it had a problem with the autorevese system in that if I turned that feature on, then pressed the left-to-right button the appropriate solenoids would engage. Grounding the guide would then flip it to go right-to-left, but grounding the guide again did not stop the tape.

I checked the 5 transistors and 2 electrolytics on the autoreverse module, they are all good. And then I realised that when it reverses, it waits for the tape to stop (no voltage from the reel motors) before it engages the other solenoid. During that time the tape is not moving, so the guide would remain grounded via the tape foil in normal operation. So I tried it like that, holding my ground wire on until the other pinch roller engaged. And it works fine.

So human error on my part...

As for the audio problem, it's another silly fault. I fed my test oscillator into both channels of the tuner input, set the controls appropriately selected AMP mode and pushed up the recording level slider. One meter registered, the other didn't. Tapping the slider got both to register to the same level. And my 'scope showed signals of the same amplitude at the monitor outputs and speaker sockets (after pushing up the volume slider), including when test speakers are connected.

So I suspect it's nothing more than dirty pots. Although the user interface is a slider, they're normal rotary pots with rack and skew gearing (well, it is a Philips!), so getting replacements is not going to be too hard I think. But I will try cleaning them first.

[Sideband]

I might have a few pots for this machine. Depends on value and if I can find them.....!

[TonyDuell]

When I get a suitable Tuit (circular) I will check all the pots and see if they can be cleaned or need replacing. So far the recording level pot (R721/R751. 22k log dual gang) seems to be the main problem. I would guess that things like the tone controls will have had little use and are probably fine.

[Sideband]

They can be cleaned with normal switch cleaner seeing as they are just normal pots. Getting to them can be a bit tricky and the opening isn't always were you'd like it to be....!

[TonyDuell]

Yes... Just like Philips to have slider controls and normal rotary pots. (In case anyone is wondering, I _love_ Philips designs...)

I am going to have to take the amplifier unit apart anyway, the capstan drive belt goo has coated some of the switch levers. And as I said the timer needs a bit of work. Hopefully I will be able to get to the pots when I do that.

[TonyDuell]

For those who are following this saga...

I've tried to clean up the goo and the potentiometers. The amplifier is a separate sub-chassis consisting of the front panel with all the controls, a motherboard at the top with lots of plug-in boards going downwards. The slide swithes are (mostly) on this motherboard, with operating levers on the top side.

So I took the whole lot out, took out all the plug-in boards (I did keep them in left and right channel sets), then dismantled the switch levers (which were covered in Evil Goo from the capstan belt), uscrewed the clock/timer, motherboard, scratch/rumble filter and front connector assembly from the panel. I could then move the motherboard away, and see the potentiometers.

By loosening the nut (10mm spanner) on each pot I could slide it out and remove the skew gear from the spindle. With them all removed the front panel was free. So it all got cleaned up.

I gave most of the pots a good squirt of propan-2-ol and worked them from end to end a few times. The recording level control that had given me problems I took apart (just remove the collar that couple to the skew gear, bend up the tabs and remove the bush/bearing, the front resistive track and lift out the wiper). It didn't look very worn, so I cleaned it well and bent up the wipers to make better contact.

Let's hope that does it. They may be rotary pots, but that doesn't mean they will be easy to get. For one thing they are smaller than normall (spindle looks to be about 4mm). And for another, Philips have designed it so that when the slider is at the bottom (minimum volume or level) the pot is fully clockwise. So they are all reverse log law (!). Why I do not know. They could have made the skew gear and the pins on the slider the mirror image of what they are and used normal log law potentiometers. Oh well...

It is mostly back together (and it wasn't hard). I want to clean up the timer while it is out, so I can't test things yet.

[julie_m]

Did you see this thread, which mentions some other Philips machines using the unusual NPN / PNP flip-flop arrangement, by any chance?

[TonyDuell]

I wrote several of the messages in that thread...

The N4418 is similar to the N4510 but has power amplifiers and speakers built-in. The N4510 is line level and headphone output only. Both are 3 motor, 3 head machines, only run the tape left-to-right and have 6 flip-flops (2 boards of 3 each).

The N4450 is the big one, takes 10.5" spools. It has power amplifiers (but no built-in speakers). 3 motors, 6 heads. Runs the tape both ways with autoreverse. 8 flip-flops.

There's a monostable multivibrator in the autoreverse circuit of the N4450 that uses an NPN and a PNP transistor, although both are silicon (BC147 and BC157 originally).

[julie_m]

Oops Wasn't paying any attention to the names. Just remembered it in a moment of madness .....

[TonyDuell]

No harm done

I've now got the clock/timer stripped on the bench. It needs a good clean-up. It's quite clever, the setting wheels for start and stop each carry a spring-loaded pin contact which makes contact with a flush metal stud on the time wheel in the middle, thus completing the circuit at the right time. The motor feeds a 2-stage worm-and wheel reduction system, that drives the minute time drum. Every revolution of that operates a linkage which steps the hour drum on one position.

It's easy to remove the motor and the fixed spring contacts. Not too hard to get the shafts of drums out. And a right pain to dismantle the reduction gearing.

[Reelman]

On my timer motor there is a lever, accessible from the rear of the machine that changes over from 50Hz to 60Hz operation. Is that fitted to yours Tony? Good luck with it!
Peter