Mercury vapour 1/2 wave rectifiers.

[Diabolical Artificer]

Digging through a box of valve's I came across several of the above, namely an 866A, an 872A, and a CV1029 or 10E and thought I'd see if I could fire one up.

I started with the 872A first as it looks a bit more like a normal valve and it has a reasonable datasheet available. First finding some transformers wasn't easy, the filament for this is 5v at 6.75A and it really needs to be centre tapped too. The only thing I had was a big Tek scope tfmr 6.3v @ 8A run a variac/iso tfmr combo. From what I could glean it's recommended to power the filament for 20mins/half hour before you apply HT.

Next finding an HT tfmr proved difficult, all I could find was another old scope tfmr with 0 720v 1070v @ not very much milliamps. I also needed to find a suitable choke, HV cap and HV/beefy resistors.

After the requisite warming period I applied HT and the valve lit up with a lot of arching going on from the anode to the big tube thing which is I presume the cathode, but the filaments the cathode right? I had to make a virtual centre tap for the filament tfmr with two 390r W21 WW resistors to take the HT off.

So, the test set up was 1070v tap to anode, HT off the centre tap of the filament tfmr, into a 300mH choke, into a 30n 3000v cap, tother end to ground/0v tap of HT tfmr with a big ceramic 2k R as load to keep the current in check. After the arching I increased the load to 2k + 820r + 560r in series, the valve now runs with a blue glow all over the big tune/cathode at first, the settles to an internal blue glow after a few minutes.

HT as measured at the load was around 450v, which surprised me at first, but considering the weakness of the HT tfmr, and only one "diode" (IE the MVR), it's not surprising.

Is there supposed to be arching in normal running condition's? Secondly, would it be ok to use two different rectifiers, an 866 say and an 872A? Or would a silicon diode and resistor do for the other side of a full wave rectifier? I'll wind a better HT tfmr when I get a mo, as to be better able to test these beasties. More on the other rectifier's later.

Andy.

[Al (astral highway)]

Hi Andy,

It shouldn’t be arcing at all. The Peak Inverse Voltage is 10kV or something, max AC in, 3535V. Peak current 1A, ave current 250mA. It was designed for use with a choke of 8H and capacitor around 1.5uF.

The ‘bit tube thing’ is the anode. The cathode is the heater.

There’s no need for a centre tapped ‘virtual’ heater. This was designed as a polyphase (1 leg of 3 phase ) valve.

It’s not viable to mix silicon and valve half wave rectifiers to get full wave rectification. I think the design forward voltage drop of this valve is 15V or something under typical load conditions.

Pretty glow from these things , right ?

[Diabolical Artificer]

Hi Al, maybe it isn't arching as such, IE abnormal operation, maybe it is just the elements settling in after not being powered up for so long, see pic. the video's I've seen online show a steady glow, unless the PSU is under load, IE it pulses with music peaks for instance. When first powered up there were bright blue dots periodically appearing on the disc (see description below) which "popped" accompanied by a "arc".

The valve has a disc just under the top cap, which I presumed was the anode, then a round tube under that, with a gap of about 1/4" between the two. The "arching" was flashes of blue light dancing off the disc, onto the tube. The filament is inside this tube, you can see an orange glow under it after it's been powered up for a while, but the filament doesn't glow like a normal valve.

I got "HT should be taken off filament centre tap" from the first schematic on the datasheet, and it also recommends this in words on another similar valve datasheet, though I did try connecting HT as per usual off one tap, couldn't see a difference.

One thing I was curious about was leaving the filament powered with HT being applied, this is a big no no for a normal valve, as it leads to cathode poisoning, why is this not the case here? Different filament/cathode?

Lastly I did notice a difference in whether a choke (and choke value) or capacitor is first after the valve. The HT transformer starts to buzz and the glow is less "settled" with more inductance.

I'll upload a video to Utube at some point so you have a better idea of what's what.

Andy.

[GMB]

One thing I was curious about was leaving the filament powered with HT being applied, this is a big no no for a normal valve, as it leads to cathode poisoning, why is this not the case here? Different filament/cathode?

It is a different kind of valve altogether. More like a thyratron.

In my experience these rectifiers are rather temperamental and take a long time to settle if they have been out of use for a while. The warm up time is needed to vapourise the mercury that makes the plasma that conducts the power, so unlike a normal valve the current flow involves the ionisation of a gas rather than just electrons.

I have only seen them used in situations where their very constant forward voltage drop was the feature, i.e. where the load varies but you don't want too much variation of the rectified voltage.

You have to be careful with them because they can carry a huge lot of current, just before they die.

[G4XWDJim]

The recommendation for these valves as far as I remember if they’ve been moved about is to leave them with the heaters on for half an hour before applying HT. Old utimers using them in transmitters found that occasionally they’d take a fuse out and then be ok for days before doing it again.

I used them for a short while in my ET4336 transmiitter but then swapped them for 3B28s which had no mercury and worked well.

Jim

[Bill]

I have a Philips 70watt PA amp with two AX50 MV rectifiers. HT is 800v . If I remember right in a voltage doubler circuit fed to PP EL34s.
There is a manual switch to power heaters before switching HT on, but no recommended time delay before doing so. There is no arcing in the rectifiers, just a gentle blue glow like gassy 807s.

[GMB]

Then you are lucky. I suspect it depends on how much mercury liquid condenses out when the rectifiers are off and where it condenses.

I have a particularly troublesome one which needs a seriously long warm-up to avoid lots of arcs and sparks.

[Diabolical Artificer]

Thanks gents, that gives me some idea of what's normal, IE the popping and sparking is contamination of the electrodes with mercury droplets. The area on the inside of the glass, around the centre is free of mercury, but there are still droplets higher up and lower down.

I think what it needs is a good few hours running powered up, which means winding a decent trmr capable of a few KV at a decent current.

To stop any over current condition's, I've put a 500mA fuse in the HT lead to anode, a long one. BTW, how do these things fail; do they im/explode or anything crazy like that?

I've been away from the bench for a few days, will have more of a play today and will fire up some of the other MVR's I have.

Andy.

[kalee20]

Originally Posted by Diabolical Artificer

I think what it needs is a good few hours running powered up, which means winding a decent trmr capable of a few KV at a decent current.

Why do you need that? Mercury vapour rectifiers only drop 15V or so when conducting. So an off-the-shelf 100V 1A transformer will do, with a 60W or 100W light bulb as resistive load. You don't even need a filter inductor! (I seem to recall that mercury vapour rectifiers shouldn't be operated with a capacitor input filter anyway).

[G6Tanuki]

MV rectifiers are indeed prone to be a bit temperamental; a piece of gear I used to support had a pair of 866A: it had a separate transformer for the filaments, along with an Octal-based thermal delay tube (bimetal-strip relay) to provide the necessary minute or so's delay before applying power to the HT transformer. it also had a 'no volts' relay interlock which shut the whole thing down if there was a brief loss of power. I was always distinctly nervous about it!

Running MV rectifiers for a while with just heater-power applied, after they have been moved, or if they've been out of service for a period, was always considered good practice.

[turretslug]

Yes, I recall that choke-input filtering is required to even-out and lengthen the current pulses into the first capacitor as otherwise the low forward impedance makes it vulnerable to cathode-stripping. (As mentioned, a resistive test load wouldn't need this though).

[Diabolical Artificer]

"Why do you need that?" I suspected the 872 wasn't running right, confirmed by scope the HT - no rectification as such, scan looks like an amp with X over distortion. That and as these are supposed to be for EHT rectification seems better to test them at their intended working voltages. Lastly it's good practice.

"(I seem to recall that mercury vapour rectifiers shouldn't be operated with a capacitor input filter anyway)." The combined 866 and 872 datasheet gives figures for choke and condensor filters, with condensor operation being derated, as per normal rectifiers.

Lastly as I'd never used these things before and didn't have much of a scooby doo, I followed the "instruction's" to the letter. These things are a bit scary for someone who's never used one, now I've had a play and had a chinwag with you chap's, I'm more relaxed and less cowed.

Fired up the CV1029 and 866 yesterday, both work ok and both seem happier with a straight condensor filter, by that I mean the HT transformer buzzes less, it is seriously underpowered for this job. I'll take on board your recommendation's and use a different tfmr today.

Thanks, Andy.

[G4XWDJim]

Morning Andy,

Would it help if I send the PSU circuit of the RCA ET4336 transmitter which uses 4 x 866s. It would be interesting to see how they do it maybe. Later, it’s still dark here and it’s in the garage.

Jim

[G4XWDJim]

Here's the PSU circuit.

R109 and R120 are 25 ohms 10 W each.
C122 and C123 are 10 uf each at 2kv
The chokes aren't given a value but have an Air Ministry ref 110C/2675.

Hope this helps.

Jim

[Diabolical Artificer]

Thanks Jim, that's a beefy power supply - 1A? I see they are taking the HT of the filament centre taps as I mentioned. Why would they do this and recommend it in the datasheet? Balance?

I now have another 866 that you kindly sent me Jim, so will knock up a proper PSU.

Cheers, Andy.

[turretslug]

A centre-tapped rectifier filament winding was often seen on better-quality transformers intended for higher HT current PSUs- it halves the current accumulation along a directly-heated cathode as the charge stream builds up an output current along it. Otherwise with, say, a filament rated at and drawing 2.5A from its winding and 300mA HT current, we end up with 2.8A total flow through one end of the filament and a possible hot-spot and weakness. Centre-tapping shares this build-up between both ends of the filament.

It's complicated slightly in that circuit snippet by the two filament windings and the limiting/sharing resistors into the choke, but the basic principle applies,

Colin

[G4XWDJim]

If the PA in the transmitter is not in tune the HT current can rise to more than double that 300 m/A and then there’s the modulator current so an amp wouldn’t be far out.

Jim

[hannahs radios]

The arcing you get might be due to the heater being under run its a guess but im thinking that in fluorescent lamps if you apply the starting pulse too quickly you also get arcing and sputtering from the filaments. Now I know,were talking about recipients and not lights but they are similar in the way they work you can even have a flourishing tube work as a crude rectifier if one cathode is a bit weak