[PaulM]

Some time ago, I flagged a rather unusual 26" Decca set with Teletext which appears to have been a prototype as it has various 'issues'. There's what seems to be design department 'handiwork' in various places, the CRT is labelled only as a Toshiba 'engineering sample' (1978 date) and the set came from the family of a late ex-Decca engineer in Bradford.

I had put this on one side for a rainy day (!) and finally its turn came round again. It has always suffered from random shut-down and re-starts. Re-soldering the SMPSU as per the recommendation in the September 1983 article in 'Television' (mentioned in the original thread) made a vast improvement. However, it still has a 'mess about' now and again, just when it fancies! Thus, I set to once again, and after a swapping out lots of 'low' SMPSU electrolytics, the last one to be changed (connected to IC601, pin4) seems to have fixed the problem. The capacitor goes to ground and is the decoupling for the 'counter' shut-down system which the TDA2581 IC features.

All well and good, but during my deliberations, I applied the DVM probe to the raw HT (C624 positive, rectified mains) in with no DVM ground connected and the whole SMPSU cycled through its shut-down and recovery mode. This could be repeated and was not a fluke. It is categorically not mechanical - this is either residual radiated rubbish and/or capacitive loading (the meter is on the bench). Since changing every electrolytic (apart from the main smoother which was checked and good), it's still incredibly 'twitchy' if you stab it with just a floating DVM probe. The bottom end of the smoothing capacitor is fine - you can connect at will - but the 'hot' end is super sensitive. Adding extra low ESR decoupling makes no difference.

My question is, is this normal behaviour for this series of Decca? It's now stable and providing I don't annoy it in this way, it's good. I've never known a SMPS do this (and I've also designed them, so I know the wrinkles . . .), but this one is a 'floater' tied to chassis ground only by a small cap and a 3M3 resistor. Apart form the venerable CTV19, Deccas have not been something that I've tussled with and I just wondered whether this was known 'normal' behaviour? Extra decoupling/bypassing does not affect things. Any Decca experts out there who might have an insight on this, please?

Many thanks for reading this rather long post!

Best regards,

Paul M

The circuit (or an approximate version) is attached form the 'Television' article.