Golborne Vintage Radio

Hello gents, my above scope developed a fault just before Xmas.
Been sourcing components since then i.e. Caps and blown transistor Q1080.
Replaced old parts but unfortunately still not working.
I have 324Vdc at P119, Q1080, Q1030 and Q1032 collectors.
Neons flash alternately ds1024 and ds10?? Next to each other.
Any help would be appreciated.

Join the Tekscopes group on Yahoo - huge number of very good quality People on there, including a number of ex-Tektronix designers and support staff. It's the go-to place for this sort of query.

Hi Nick, thanks for the info will join Tekscopes.
Kind regards, John.

Look forward to seeing you on there! Unbelievably there are getting on for 9000 members on Tekscopes now. Very much the vast majority are lurkers.

But it is a completely well-behaved list, with zero moderation by the founder and current list manager. There is only one list too - no individual threads or topics.

Craig

If Yahoo continues to unravel as a business, I'd seriously think about getting an archive of Tekscopes together for a potential move to Google Groups.

I did this with NEONIXIE-L a few years ago. Good move, IMHO.

It is all, I believe in hand already. Yahoo are making a major mess of just about everything, and Groups is just about useless now - a sort of inexorable process that started with neo.

Craig/Nick, joined posted through Tekscopes, hardly any response only one person who felt uncomfortable in troubleshooting the Tek434 psu due to its complex arrangement but Yahoo is a complete nightmare.

The power supply for the early 434 is a primordial switched mode power supply. I've just downloaded the manual and had a shufty. We are talking a switched mode power supply dating from a 1971 instrument, and was probably designed in the late 60's.

I haven't had that detailed a look. The two flashing neons are associated with the starting the switcher going. They are a neon relaxation oscillator and should flash alternately every five seconds or so, with no precision at all. There are then two sets of discrete circuits which turn on and turn off Q1080. The turn off circuit looks at the collector current of Q1080 and when it reaches a specific value measured by T1080, turns the transistor off. How frequently this turn-on/turn off happens is controlled by U1130, a ua723 that is configured as a switching controller whose period of oscillation is controlled by the value of the +15V line. If that is the correct value it oscillates at about 25kHz. That frequency changes slightly to ensure that it stays at 15V. So the frequency of the whole switched mode supply is set by that 25kHz frequency.

The uA723 is still a current part by the way, and costs less than a quid. That is one hell of an innings for a piece of silicon.

The problem with switchers in general, and early ones specifically is that operation is somewhat arcane. First thing I'd do is unplug each transistor in turn and check it. Check one at a time, and if OK replace back in the board. Are all the diodes OK? Particularly VR1084/5 and CR1084 whose purpose is to stop the c-e voltage of Q1080 exceeding 250V. If one of those has gone open circuit that could explain a dead Q1080. Go through that lot and come back with what you find.

I'm not surprised that noone stepped up to the plate candidly. The list is littered with experienced people (me included) who have wasted whole weeks trying to resurrect a Tek switched mode supply, The top of the dresser in the spare bedroom has three 7000-series supplies that defeated me big time.

Craig

Craig makes a good point about the 723. It was a pretty early bit of silicon, designed by Bob Widlar at Fairchild c1967, the first voltage regulator chip. Unless there was an earlier and little used part. It's still capable of good performance though you wouldn't use one a new design.

Didn't realise the uA723 was a Bob Widlar design!

There is a story about Widlar. He was a mate of Barrie Gilbert, and they made a bet with each other about who could design a current mirror with the minimum number of transistors. Barrie lost because his design used three transistors, and the Widlar current mirror(as it is now known) only used two. Used in almost all op-amp chips made ever since.

For those who do not know, Barrie Gilbert was (is) a prolific British electronics inventor with over 100 patents to his name, but spent the vast majority of his life in the US. He was a leading light in the golden years of Tektronix, and inventor of the Gilbert Cell which finds its way into ultra-high bandwidth amplifiers (Tek's interest) through to four quadrant multipliers. After his illustrious career there he was poached by Analog Devices, and that was only after they agreed to build a research facility specifically for him in Beaverton. He is a fascinating character.

Craig