Using the 83

[Kris0603]

Hello forumers,

After reading a thread here about mercury vapor rectifiers, and their pros and cons, i felt compelled to ask this question.

I am making a stereo amp, from a schematic. I have however decided to change the tube compliment in the PSU. Going from a true dual design comprising two GZ37, LC filter (10H-470uF), my best bet right now is one shared 83, and split from the chokes on in. Will the tube be fine with this?

Some tests in PSU Designer II puts this solution spot on, giving me 260 volts from a 350-0-350, at 250 mA draw, with low ripple.

Should i go for this, or try using maybe two 5AR4s or 5U4GB's.

If i do decide to go with the 83, i reckon i will have to go gentle with it, because it has been stored for god knows how long lying on its side, and it looks to have mercury all over the lower glass detail, the one with the hole on one side. I do have some idea of the operating requirements of the 83 and other mercury rectifiers. I just want to know, is it a good idea?

rgds, Kristian

 

[analog_sa]

is it a good idea?

By all means. After all you want more magic, right?

Splitting the chokes, etc is not a problem. Just make sure you don't exceed the max current through the 83. It really doesn't like it. 250mA should be fine.

As you probably know it's a good idea to run the heaters without HT for an hour or more in the working position if the tube comes out of storage.

 

[speaker]

you probably know it's a good idea to run the heaters without HT for an hour or more in the working position if the tube comes out of storage.

I've never heard of that precaution for using an the 83. Can you explain why this is so? I would guess this reforms the mercury residue in some way?

speaker

 

[EC8010]

It's to remove possible shorts due to mercury droplets being in the wrong place and to turn the rectifier into a mercury vapour rectifier rather than a vacuum rectifier with added mercury droplets. Some manufacturers say 15 minutes, others 30, you can't go wrong be allowing an hour. You have to go through this regime each time the rectifier is disturbed...

 

[Kris0603]

Thank you all for replying.

Yes, i knew about this. RCA also recommends that the plate voltage be applied with a variac, presumably to allow the added anode dissipation to aid in vapourising all of the mercury.

As for magic, sure the mercury glow is nice, but its not that obvious on the 83. I've just had a hard time finding substitutes for the GZ37s, which i can only find at (to me) very high prices. If anyone can suggest a good substitute which will get me to where i want to be, please feel free to do so. Im really more of a function over flair kind of guy,

rgds, Kristian

 

[speaker]

It's to remove possible shorts due to mercury droplets being in the wrong place and to turn the rectifier into a mercury vapour rectifier rather than a vacuum rectifier with added mercury droplets. Some manufacturers say 15 minutes, others 30, you can't go wrong be allowing an hour. You have to go through this regime each time the rectifier is disturbed...

Makes complete sense. Thank you for the clarification.

speaker

 

[martinab2]

The 83 sounds fine for your application - they are long lasting and relatively trouble free tubes provided that you always remember to pre-heat.

Some of the 'magic' of mercury rectifiers is down to their almost constant voltage drop which is unaffected by current drawn within operating limits. Other tube rectifiers have a voltage drop which increases with current. So a mercury tube acts almost like a regulator as well.

I see you are using an LC filter - These have an excellent reputation but i haven't had much luck in making them work properly. When you run your circuit through psu designer, check it using a stepped load and watch what happens as the load changes. Chances are you will see some oscillation unless you have been very clever / lucky.

Usually I give up and turn it into a CLC adding a small cap after the rectifier (4u or so). The 83 will cope with this ok (its mentioned in the datasheet) and it sorts out the oscillations.

 

[kmtang]

Standby HT switch is needed

Please note that you need to add a standby HT switch for use with 83 mercury vapor tube.

The 83 tube needed to be heated up first before applying the HT. It's very simple by inserting a switch from the CT wire to ground.

Please read the data sheet of 83.


Johnny

 

[liberty]

Hi all.
I am trying using an 83 for lilliput amp: http://www.audiokit.it/ITAENG/KitElettr/MEGA/Lilliput/Lillman4.gif

Simulating with PSU designer II, I saw nice situation "LC" with a transformer 220V/270V-0V-270V, choke 5H, cap 470uF.
With a load of 1k-1,5kohm(150-200mA, like lilliput is) it seems to have 210V(after a little peak in the first second).

The problem is that 83 rectifier need a pre-heat for filaments, so when we give anodic Voltage, the tube run soon. Lilliput tubes(E88CC, 6080) need 10 seconds to be ready, so we have 10 seconds in which the 83 rectifier see a bigger resistance than 1k-1,5KOhm. If we make a simulation with PSU Designer II, for a load of 30kOhm, or for 200kOhm, we see that Voltage continue increasing, and reach about 320V in 7-8 seconds.
Then lilliput will starting working(filaments of 6080-E88CC) and load become 1k-1,5kohm and Voltage finally is 210V.

But...someone can tell me if it is dangerous for schematic of lilliput the initial grow of Voltage? Lilliput before starting to work(in 9-10 seconds) will see 320V!

Some books tell about LC rectifier system, and about this problem(which involvs also normal rectifier tubes). They suggest to use a resistor after LC to give a minimum load to the rectifier tube.
But this resistance must be too much low(in my situation must be about 5k o lower) and will dissipate big energy, also during normal use...

So, someone expert can help me with a better solution?

 

[agent.5]

Use a separate filament transformer for the 83. The power switch turns on the 83 filament transformer and a delay relay. Use a delay relay connected to a solid state relay to turn on the V+ transformer after say 30 seconds.

 

[Miles Prower]

The problem is that 83 rectifier need a pre-heat for filaments, so when we give anodic Voltage, the tube run soon. Lilliput tubes(E88CC, 6080) need 10 seconds to be ready, so we have 10 seconds in which the 83 rectifier see a bigger resistance than 1k-1,5KOhm. If we make a simulation with PSU Designer II, for a load of 30kOhm, or for 200kOhm, we see that Voltage continue increasing, and reach about 320V in 7-8 seconds.
Then lilliput will starting working(filaments of 6080-E88CC) and load become 1k-1,5kohm and Voltage finally is 210V.

But...someone can tell me if it is dangerous for schematic of lilliput the initial grow of Voltage? Lilliput before starting to work(in 9-10 seconds) will see 320V!

Simple solution here. Since it's already necessary to preheat the 83 cathode without HV, add two switches: one to turn on all the heaters during the preheat, and another to turn on the HV. It's the same solution one would use if building a solid state PS. That way, all the tubes are ready to go when the HV is connected and you won't get that voltage rise. It's also a good idea to avoid the possibility of cathode stripping.

Some books tell about LC rectifier system, and about this problem(which involvs also normal rectifier tubes). They suggest to use a resistor after LC to give a minimum load to the rectifier tube.
But this resistance must be too much low(in my situation must be about 5k o lower) and will dissipate big energy, also during normal use...

If you need such a large current through the bleeder resistor to stabilize the voltage, then best to consider either a choke input filter instead, or electronic voltage regulation. Again, it's not a problem if you preheat all the heaters before applying the HV.

 

[liberty]

Use a separate filament transformer for the 83. The power switch turns on the 83 filament transformer and a delay relay. Use a delay relay connected to a solid state relay to turn on the V+ transformer after say 30 seconds.

Yes,
this is for safe of 83 tube...
...but question was different!

The problem is that when(after 30 seconds) V+ starts, 83 tube immediately works. But lilliput amp tubes(6080 and E88CC) need 10 seconds to be ready to work(and, of course, to "give" optimum load to 83 tube, which is near 1k ohm). In this 10 seconds the load seen by 83 is greater then 1k ohm. this load may be greater than 10k ohm(...I think so...), and in this condition V+ go up to 320V in 5-6 seconds. So lilliput amp tubes(when will become ready to work) will find 320V! Finally the load become about 1k ohm and V+ 210V.
But what can happen in that 10 seconds, and when V+ is 320V and lilliput start to work?(lilliput amp must work with 210V).

It is a classical situation of LC filters(choke input filter). If the rectifier hasn't a sufficient load, V+ grow up to the "standard" 1,4X...

If I simulate with PSU designer II and I put 1k ohm load, I can see V+ having a little peak(240V) for a while in the first second, but become soon 210V.
The problem is(as I told) that lilliput amp need 10 seconds to work and to become a load of 1k ohm. So the 83(which have ready filaments and work instantaneously...) see different load and in 6-7 seconds charge the cap at 320V...

I am not sure this is a real problem, and if it is I have no idea how to solve. So I asked for a suggestion in the forum...

Manuals say to put a resistance, but it will be not a great solution. For example I need a 5k ohm, and it will dissipate 12W everytime(and make extra-work for 83 everytime...).

At the time I can't change type of filter because I have a transformer 270V-0V-270V and using "capacitor input filter" will give a bigger V+. And also I don't think the 83 will be glady to work with a new transformer with less Voltage. It was made for 400V-500V and I am not sure it will work well with 160V-170V(extimated voltage for cap filter input, to give about 210V).

I know, it's a mess! But I am here to learn!

 

[liberty]

Simple solution here. Since it's already necessary to preheat the 83 cathode without HV, add two switches: one to turn on all the heaters during the preheat, and another to turn on the HV. It's the same solution one would use if building a solid state PS. That way, all the tubes are ready to go when the HV is connected and you won't get that voltage rise. It's also a good idea to avoid the possibility of cathode stripping.



If you need such a large current through the bleeder resistor to stabilize the voltage, then best to consider either a choke input filter instead, or electronic voltage regulation. Again, it's not a problem if you preheat all the heaters before applying the HV.

Yes, I considered this solution.
I considered 3 switches, 1 for 83 heaters, 1 for E88CC-6080 heaters and 1 for V+.
As You know sometime 83 Mercury needs long pre-heat(30 minutes, for example, when it comes from a long travel). So it is better to separate the heaters.

But...I don't remember where and when, sometime ago I read that it is not a good idea to make work filaments of triodes without V+....

 

[kathodyne]

hi,

There seem to be 2 sides on the preheating story....

In case of a normal rectifier tube you should preheat all tubes BUT the rectifier....
Then turn on the rectifier which lets the voltage rise slowly...

(problem? the rectifier tube will be in great stress because it has to deliver current while the cathode isn't yet hot...)

OR

You preheat all tubes and switch the hv (ac from transformer)
(problem? all hot tubes get the high voltage slam)


Some say rectifiers are equipped to deal with the stress and only a little cathode-stripping occurs (replacement every 2 years)
Old radios seem to use the first method without preheating the rest of the tubes....

is worrying about this tto much or is there reaseon for concern in using one of both methodes?

 

[liberty]

hi,

There seem to be 2 sides on the preheating story....

In case of a normal rectifier tube you should preheat all tubes BUT the rectifier....
Then turn on the rectifier which lets the voltage rise slowly...

(problem? the rectifier tube will be in great stress because it has to deliver current while the cathode isn't yet hot...)

OR

You preheat all tubes and switch the hv (ac from transformer)
(problem? all hot tubes get the high voltage slam)


Some say rectifiers are equipped to deal with the stress and only a little cathode-stripping occurs (replacement every 2 years)
Old radios seem to use the first method without preheating the rest of the tubes....

is worrying about this tto much or is there reaseon for concern in using one of both methodes?

Yes,
but the problem is little different here…
I have to pre-heat the 83 rectifier, so I have only 3 possibility(if any expert have no others):

1)Pre-heating the 83 tube, and give V+(and heaters voltage) after 30 seconds to e88cc/6080, hoping that the 320V will not harm the circuit.

2) Put a resistance after the LC filter to give a minimum load to rectifier, so that V+ will stop at 220V-230V. Pre-heating the 83 tube, and give V+(and heaters voltage) after 30 seconds to e88cc/6080, hoping that the resistance not stress the 83 too much during normal work.

3)Pre-heating 83, E88CC and 6080, and, after 30 seconds, give V+, hoping that tubes will be not stressed by this, and hoping also that the amp will not have a dangerous “bump” in loudspeakers.

Second option seems to be the safest one. But it isn’t so simple to find big resistences. If I will use a 5k ohm I need something like 20W…
Lamps can be a nice idea(for example 2 x 230V/6W in series) but have 2 problems:
a) the filaments will burn, a day, leaving the 83 without a sufficient load.
b) Lamps can be “noisy” for circuit.

Any good expert has a better solution?