[time-nuts] Re: 5061A HV Supply, et al. {External}

[Jim Muehlberg]

Ed,

Thanks for all the info.  Indeed, the old HP manuals are really good and 
informative.  Your comment led me to the A18 schematic - I assumed it 
would just be a block in the schematic, but alas, HP does it right and 
shows the schematic of a "Non Repairable" hermetically sealed unit!

I did rip into the supply already, but I've found no smoking gun.  All 
components are in tolerance. (though I understand it may fail under high 
voltage)Â  It opened up easily with some careful propane torch 
application.  No damage!

The 10M input to the DMM is too low, according to one poster.  I built a 
182MOhm divider and still no luck.

I don't know much about Ion pumping, but there's a few guys here at NRAO 
that have had experience.  Back in the 70's, Green Bank had some dewars 
with ion pumps, but all that now has been superseded by turbo pumps.

This little project is just a fun distraction from my real job, though 
it would be nice to have this in the lab, it's totally not necessary for 
local oscillator development.  It may help our grad student if we can 
get a mating pair of standards.  I guess the joke about have one or two 
clocks may apply here.

I've been offered an A18 for a good price, which, given the labor I've 
invested so far, is a distinct possibility. (Thanks Corby!)

Anyway, I can't stand to see this stuff tossed out.  It is a high 
precision tube and it has a 90's date code.  I wish knew more, but 
that's it.  They should have had a running time meter on these things!

(coming soon - free 5061 parts!!)

Thanks,

Jim

On 11/15/2022 5:14 PM, ed breya via time-nuts wrote:
> Jim, your A18 is likely just fine, so don't rip into it without 
> knowing more about what's going on. I knew nothing about the 5061A 
> except that it's a Cs standard. I got curious about all this 
> discussion, so found the manual and did a quick look-through. It's 
> pretty fascinating and educational.
>
> Others have already given suggestions that cover most of the pieces. 
> Here's a quick summary as I see it:
>
> 1. A Cs tube that's been inactive for a very long time needs to be 
> ion-pumped down to proper vacuum level, as indicated by the pump 
> current. The system protects the hot-wire ionizer and electron 
> multiplier by disabling their activation until the vacuum level is right.
>
> 2. The built in 3,500 V ion pump supply only provides enough current 
> to run this function for "normal" conditions like continuous or 
> "often-enough" (every 6 months) operation. When the pump current is 
> too high, the voltage drops off, but it still pumps, just at a lower 
> rate. The fact that it's drawing (excessive) current on the meter 
> shows that it must be pumping, unless the current is taking another 
> path due to some failure. If there is true pump current, it should 
> still eventually get it down far enough, if the tube is viable.
>
> 3. The procedure in the manual shows the use of an external HV supply 
> to relatively quickly get the vacuum into the right range. The 
> recommended maximum current is 5 mA! I found it kind of surprising to 
> be this high, but that's the deal. It further says that if the current 
> remains 5 mA or more after a minute, the tube is likely shot. So, to 
> get an assessment and recovery in a reasonable time, you need an 
> external HV supply.
>
> 4. I'd recommend looking for a commercial small HV PS in the 1 to 5 mA 
> range. A 1 mA one will be much easier and cheaper to find, and should 
> still get the job done, just slower. I strongly advise against rolling 
> your own with big iron or big power HV stuff like microwave oven guts, 
> unless you're thoroughly familiar with how to safely handle and work 
> with HV. This can be fatal if you screw up badly enough.
>
> A small lab type HV PS can give you a good zap, but you'll live to 
> tell about it, unless you really really really screw up. Again, 
> smaller is safer, so a 1 mA or so seems good. Small HV PSs are common, 
> often used inside equipment, for all sorts of applications, or NIM 
> racks and such. Preferably you'd want the right polarity (many - 
> especially smaller ones - are reversible either way), and variable 
> from zero to at least your needed voltage, and with V and I readouts.
>
> If you want to make one yourself, and learn about HV with reasonable 
> safety, I'd recommend starting with some small resonant-Royer 
> converter circuits - these are everywhere, in the form of CCFL tube 
> drivers for LCD backlights. You should be able to find lots of DIY 
> examples and parts to inspire.
>
> 5. Back to the measurements, as has been suggested, you simply must 
> use a high enough dropping resistance for probing the HV, without 
> overloading the output. In this case though, you don't really need to 
> worry about it, since you have a probe built in. The 200 megs of R4 
> drops the output voltage for the regulator circuit. In conjunction 
> with the resistors in the bottom of the divider, it appears the 
> feedback signal is normally +4 V. It's complicated a bit by the 
> current sensing, but close enough to be a good proxy for the output. 
> The trick is to measure the feedback with a non-loading voltmeter - or 
> at least one with very high input resistance like 10 megs or more. 10 
> megs would give a few percent error, tops, in both measurement error 
> and possible upset to the circuit operation. Many lab grade DMMs like 
> HP or Fluke have a native range over +/-10 V, so can easily look at 
> the 4V signal without loading.
>
> So, if you properly monitor the feedback in the as-is condition, it 
> will indicate the very low actual output, divided by about 875 times 
> (3500/4). Now, if you safely unhook the tube load per the manual, the 
> feedback should come up to around the proper 4V setpoint, and the 
> output voltage should be right. If not, then there's more trouble.
>
> Anyway, even if there's still a problem, I wouldn't worry about the 
> A18 yet, and first do the external HV recovery. If it seems to work 
> (low enough ultimate current), then hook it back to normal and see 
> what happens. If the current stays very high, then the tube is shot, 
> and A18's condition is moot.
>
> BTW, you may want to study up on ion pumps in general, to get some 
> idea of what's going on. I see here that there's a huge possible range 
> - well over 100 to 1 - of ion current involved, so I'm wondering what 
> ions are being pumped here, exactly. I assume it's mostly Cs ions that 
> are loose in the beam-line, where they don't belong. At first I 
> thought the whole deal was some sort of gettering function, to trap 
> out any bad contamination ions like in most vacuum tubes, but to take 
> up to 5 mA initially, I'd say it would have be a pretty crappy vacuum 
> to begin with. So, I think the ions are the good Cs ones, that just 
> need to be relocated. The manual probably explains it, but I haven't 
> seen it yet.
>
> Also speaking of ion pumps, long ago I acquired a small one for 
> developing high vacuum. These kinds of pumps need the same sort of 
> deal as the tiny one in the Cs tube, just scaled up many times. The 
> main thing is some form of current limiting (ballast), forming a 
> variable voltage range depending on the current. I forget the exact 
> numbers, but mine needed up to 5 kV at maybe 15-25 mA maximum (high 
> vacuum, medium pumping rate - the current gradually drops as particles 
> are cleared), and maybe 150 mA from 1 kV to dead short. This is where 
> big iron is ideal. I used a microwave oven transformer and voltage 
> multiplier, and a big ballast choke and resonator capacitor on the 
> primary. After much fine tuning of the basic circuits and parts, I got 
> it to nicely match the desired I-V curve of the commercial unit 
> normally used.
>
> Ed
>
>
>
>
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe send an email to time-nuts-leave at lists.febo.com
-- 

Jim Muehlberg
Senior Engineer

National Radio Astronomy Observatory
ngVLA Local Oscillator Lead

1180 Boxwood Estates Rd B-111
Charlottesville, VA 22903-4602
P 434.296.0270
F 434.296.0324
www.cv.nrao.edu/~jmuehlbe