The PL81 TV line output valve.

[Jez1234]

I'd best cancel the yacht I was going to buy with the proceeds from the kit version then

[Richardgr]

Very compact! What was the rationale for the 6550? Anode current? And do you use it as a triode or as a pentode?

[Jez1234]

If you think that's compact you should see the innards.... when you make something up as you go along and have ideas for further improvements after starting on it then unfortunately you can't just stretch the casework a bit to accommodate it... The raw supply bit seen externally is just that for the main HT. There is another HT supply of -400V inside, from which a stabilised -150V is derived for a precision current source, plus of course the rectifier, smoothing and regulator for the LT supply. The chassis is used as the heatsink for the LT series pass mosfet but it dissipates only a couple of Watts or so as there is very little voltage overhead on the pass device. It is built traditionally using point to point wiring with tag strips for the valve and raw HT, -HT and LT parts but with 3 lots of circuitry on veroboard.

The 6550 is used as it was the meatiest and most suitable valve I had to hand. It is triode connected. A 6080 or 6AS7G would be considerably better. As I said I wanted to make a high performance valve bench supply but using only what I had to hand and with a zero spend. I achieved this with relatively little compromise ie the single 6550 is pretty much adequate for the task. It does however limit current to a little less then the designed 100mA at max voltage output due to it's comparatively high series resistance compared to the likes of a 6080 which should cure this.

With hindsight I'd have chosen less massive smoothing capacitors (they are 470uF each) on the main HT rail. They certainly give nice and low ripple even before the regulator etc but the issue is that after switch off then with a light load and low ish voltage, say 150V and 15mA output, they hang on their voltage for so long that the regulator gradually drops out of regulation and the output can then rise to around 300V for a few seconds.... not good if powering eg a battery valve radio or similar. If you are drawing say 50mA @250V then the caps discharge fast enough to avoid this. They were free, covered some holes nicely, and were a fit for the mounting clips I had to hand!
I plan to add a small mains powered relay at the output so it disconnects the HT when mains is removed.

[Ed_Dinning]

Hi Jez, the loss of mains relay was a standard on many commercial PSU's.
Usually configured to connect a dump resistor across rails on switch off to drop voltages below about 50v (SELV) in 5 secs or so

Ed

[Jez1234]

I considered both that method and simply a series relay. It's rather "busy" under the chassis so it would be nice not to have to fit a wirewound as well as the relay. There again with a series relay the condition of the contacts is a worry, especially when switching DC.

[Aubade65]

There is a design for an EL81 based regulated psu in Radio Constructor October 1961 p206

Regards

[G6fylneil]

Some equipment i worked on many years ago had multiple thyristor crowbars in the power supplies, which were in the kV range, so no contacts to worry about.

[Richardgr]

Originally Posted by Aubade65

There is a design for an EL81 based regulated psu in Radio Constructor October 1961 p206

Regards

Link to the magazine

Interesting, thanks! Bookmarked for later reading.

[Jez1234]

Looks a nice design. Downloaded thanks.

[DrStrangelove]

That's reminded me of the stabilised psu I built out of a 30P4 and an EF80.

No idea what I used for the reference and the thing itself is long gone, though I can still see it in my mind's eye.