4.19 MHz clock driver IC

[Studio263]

I have a wall clock here that's stopped working. It is unusual because it uses an HF quartz crystal oscillator (4.194304 MHz) instead of the usual 32.7KHz sort, in the interests of accuracy apparently.

The IC has failed, there is a strong signal at the crystal but the open collector output that drives the motor coil (1 Hz?) is stuck low. The IC is marked 1150A but a search reveals nothing - are there any standard parts that do this job? The supply voltage is 1.5V.

[Nickthedentist]

Hello again!

I can't answer your query, I'm afraid, but could you post a picture of the movement as it's possible that someone might have an identical one you could cannibalise.

I presume it's a fairly early unit if it's got a "discrete" IC rather than a black blob on the PCB?

Nick

[Studio263]

Its from a Braun 4833 which I think was their first wall clock. The Big Book of Braun (or whatever it is called) says 1979 and the date code on the IC agrees. The IC is a DIL 08 package, there is plenty of room in there to modify the circuit if necessary.

Here's a picture with I found from doing a DuckDuckGo search (no Google for me!), this is the same mechanism but used in something else. They were so proud of the 4.19 MHz crystal that they even mention it on the back.

[Herald1360]

You could probably cook up a replacement easily enough using a 4521 CMOS divider chip. I wonder if the 1150A might be the same beast?

How about a 'generic' movement, you can get these locked to MSF for a few quid from eBay etc..

[Nickthedentist]

You can indeed, and that may be the only solution, but I know Studio263 will want to keep this iconic clock as original as possible.

Nick.

[Studio263]

Correct, although my wife says "either fix that thing or throw it away"!

The movement is quite interesting, it is outwardly typical of the generic type but inside as well as the HF crystal it features a "moving coil" motor where the actuator coil moves about 1/5th of a turn on each pulse. The lead-outs to the coil are tiny springs of the extension type, not hairsprings as one may expect.

I was considering transplanting the 32.7MHz crystal and driver IC from a generic clock movement onto the original PCB if I get really stuck, but it would be nice to retain the accurate (but battery-gobbling) 4.19MHz setup if that is possible.

[Nickthedentist]

Maybe compromise then (the secret of a happy marriage!). Fit a new movement for the time being and save the original one somewhere safe, as I'm sure you will be able to fix it one day!

[Leon Crampin]

I have a Philips HR 5570 clock with a similar 4.19 MHz movement. Mine runs, but the mechanical drive to the gear train from the moving coil is hopeless. It's a leaf spring which engages with a ratchet wheel and the relative tensions of the leaf spring and the wheel detent are stupidly critical. Consequently, it doesn't run reliably, although the moving coil pulses reliably at 1 Hz.

The 8 pin DIL chip is marked SAA 1114Z date code M6447Y. The date code on the HC 18U crystal can is 3-77 which would be about right.

If you can't find another chip for your clock, send me a PM. I might consider a sacrifice, as I've never liked this clock and would be happy to fit a generic movement to it.

Leon.

[Leon Crampin]

A search reveals nothing of either chip.

If you really want to restore your original movement, have a look at the Intersil ICM7213. It's in a different package and would require a higher voltage (a 3V lithium cell would power it) but it could be configured to do your job. The output curves suggest that it would drive the mc motor directly.

Leon.

[Studio263]

I’ve worked out a solution to the problem. I dismantled a few generic clock movements and found in one an 8 pin 1444E driver IC and its 32.7 KHz crystal. This IC is made by Eurosil, just like the 1150A in the original movement. The pinning was different and obviously the crystal would have to be swapped over as well, but the biggest problem was that the 1444E has a “bridge” output where the motor coil is driven bi-directionally, e.g. the current through it reverses once every cycle. This is in order to accommodate the type of motor where the armature is a small permanent magnet which turns 180 degrees each second.

The 1444E has two outputs which are normally “high” (e.g. 1.5V). They alternatively dip “low” (0V) for a short period to produce the bi-directional drive, so each one produces pulses at 0.5 Hz. This isn’t ideal for the original motor, which being of the moving coil type requires a single pulse from an open-collector sink once every second.

The answer is to add and invert, but how to do this in an appropriate micro-power fashion? The answer was to use a pair of PNP transistors (I picked BC556Bs) with their collectors and emitters in parallel to drive the load. The pulses from the 1444E were connected to these via 47k resistors; this was the highest value that still gave a “clean” waveform across the coil. I assembled the parts onto the original PCB using the original holes and some surface wiring, this way I can easily convert the circuit back to the original HF spec if I find an 1150A IC. With the cover back on everything looks correct and original, the clock is now back in service on the workshop wall.

but how to do this in an appropriate micro-power fashion?

I have driven similar movements from a square wave using a series capacitor, you end up with bi-directional pulses.

[Nickthedentist]

You're a clever man, T.J.

I'm very impressed - hope your wife is too.

Nick.