Mercury Power Supply Design - Checking for Efficacy

P

Paully

New member
Excuse the cross-post from Tube DIY, and hopefully Doc won't ban me for mentioning a tube rectified power supply, but here it goes.   Grainger and I are going to build an 866jr full-wave (two tubes) rectified stereo amplifier. Why?  Because they glow!  45 outputs with a 76 driver. For the power supply according to the tube manual (link below) if our max input voltage is 500V we should choose a choke input to follow the rectifiers with a minimum value of 2.1 Henries. That is followed by a capacitor with a maximum rating of 4.2uf. The manual says we can increase the inductance and then increase the capacitor in proportion. We are planning on using an LCLC setup of a 5H 400V choke / 10uf cap / 5H choke / 10uf cap being fed by a 375-0-375 power transformer. It is going to be a stereo amp, not monoblocks, with four of these rails, one for each driver and output. Sound good?

Last question, I obviously have the filament chokes in my Paramounts.  They seem like a very good option.  Any reason not to use them with 45's in the new amp?

http://www.tubebbs.com/tubedata/sheets/084/8/866Jr.pdf
 
Are you really planning to use mercury vipor in your living room?
Wikipedia:
Mercury compounds are toxic, highly persistent in the environment, and present a danger to humans and the environment. The use of large quantities of mercury in fragile glass envelopes presents a hazard of potential release of mercury to the environment should the glass bulb be broken. Some HVDC static inverter stations have required expensive clean-up to eliminate traces of mercury emitted from the station over its service life. Steel tank rectifiers frequently required vacuum pumps which continually emitted small amounts of mercury vapor.

Also, just curiosity, what are the objectives of this project ?  Are you after the best SET 45 sound or there is something else ?
Thank you,
Alex
 
From what I have read, and its worth what its worth, there is less Mercury in one of these vacuum tubes than in a typical CFL light bulb.  Maybe, maybe not but I have plenty of CFLs in the house with plenty of mercury and don't worry about accidentally breaking one when I have them throughout the house and I am more likely to break one of those than one of my precious vacuum tubes.  Not worried about continual release during operation, its not like the vacuum tube is going to work if the vacuum seal is broken.

The point is to build a vacuum tube rectified SET amp and not much more, whether it will sound better than solid state is open to question.  I wouldn't want to state that it would when I have no idea and the guys around here who design the amps think SS rectification works much better in the bass.  I want to play around and have some fun with it.  And the MV rectifiers look interesting.
 
Personally, I'm using tube rectifier with my 300b set. It's Mullard cv378 (version or 5u4g). I tried number of SS rectifiers (soft, fast, whatever) and ended with tube rectifier with CLC (20mkf/30H/660mkf) filter.
I did not calculate any values for the design you've described, but looks like with the values you have, there will be 120hz buzz and LF roll off , due to very low value of capacitors, and these values are direct result of rectifier tube selection ( lets not to worry about mercury for a moment).
By looking tube data, seems like 5y3, 5u4 etc. are better candidates. Also you can easy compare tube with SS.
Also there are a lot of other things to consider and compare - driver tube/topology, autobios vs. fixed one, direct coupling vs. cap/transformer, output transformer/topology.  All this will probably change sound more than type of the rectifier. Just built oversized power supply and worry about other things.  That is my 2c ...
Google please, Flesh and Blood by Reichert - he has pretty good advice on PS.
Good luck with project - 45 is fun!
 
Paul owns the 866JR tubes already.  The last C in the LCLC filter should be increased to give a larger storage capacity as well as better filtering.  The choke input for the filter is required for the 866JR.  

After the tubes for rectifiers there will be separate LCLC filter circuits for each 76 and each 45.

Both the driver and output tubes will be cathode biased, much like my Paramours.  

Paul has also run his Paramounts with 45 outputs and this is why he wants to scratch build this amp.
 
That is why I asked what are the objectives.  Ok, so the objectives are to utilize what is avalability and built paramount with the glow. Correct?
 
The circuit will resemble the Paramour and Paramount.  A driver capacitor coupled (possibly not) to the output.  Parafeed output, the iron is already bought.  It is the power supply that is still up in the air.

The post above is modified with a little more information.
 
Again, it's just an idea, nothing more: why not to built on the sheet of plywood outboard ps and try different tubes, capacitors etc. with the paramount. That will give you very good reference point - I guess you know sound of you amp. well enough.
 
The inductance from the data sheet only ensures that the peak rectifier current is not exceeded. To get good regulation, the first inductor should be at least equal to the kOhms of the load. For a 375v transformer you will see about 335v, and if the whole amp draws 80mA, the equivalent resistance is 4.2K, so you need at least 4.2 henries. If on the other hand there is a separate LCLC feeding one driver at 5mA the the equivalent resistance is 67K and the choke must be 67 henries. The voltage will still rise to over 500v at startup, until the tubes draw full current. Traditional practice was to use a bleeder resistance to set a minimum current (usually 10% to 20% of the total) and size the inductor for that current, so that the voltage won't rise that high even if a tube goes out.
 
Sounds like a mess.  So by having four filtering circuits, one for each major tube, I drop the current for that section from the total for the amplifier and therefore have to raise the inductance of the inductors in that section.  I am just starting to look up equivalent restance, but say for a single 45 drawing around 30 ma, we are talking about a choke somewhere around 30H, give or take, if I am starting to get the gist of this plus it needs to be around 600V And that doesn't even begin to show what size caps we need which will be large.  Ugh.  Anyway, I will look up the equations you used to figure out the Henries and decide if this is still worth the playing around given the cost.
 
Remember, for small-current stages (drivers) you can use small-current chokes (small and cheap).

Power supply filter theory is fairly simple and widely documented; google "choke input filter" for instance.
 
With just 2, 45's and 2, 76's. your not drawing a lot of current. Maybe try 2 power supply's, one for each side, or one for drivers, one for outputs...John 
 
I read on a DIY site that 83 mercury vapor tubes which are similar to 866's glow in proportion to their current draw. Someone who had an amp that drew 30ma was complaining that the 83 rectifier didn't light up much. He was referred to a youtube video (Bach Toccta D-moll on a 1000w audio tube), which showes an impressive array of 83's with a bright blue modulating glow. More current more glow less current less glow. 
 
Gave up the mercury rectifiers, just too much trouble.  Modelling on PSUD a Pi filter with the following parameters to feed a 45 amp with 76 drivers and a voltage regulator with a GZ33.  Planning on a rail for each driver tube and output tube, which is hard to model on PSUD but essentially the first cap totals to 50u for all rails combined.  Going to use a 475-0-475 transformer (need about 425V out on each rail for what we are planning.  The CLC filter for each 45 is 15u/20H(181 ohms)/20u.  For the 76 it is 10u/10H(270 ohms)/20u.  PSUD doesn't seem to have any issues with what we have done for warnings.  Anyone see a problem such as too taxing on the rectifier, etc...?  Any thoughts are welcome.
 
Anyone trying to check this with PSUD II will need to use a 50uF cap for the first cap in the string.  Without that the instantaneous current will be too high.  I figured that was only a little cheat.

To clarify there are 2 rails/filters for the output tubes and two rails/filters for the driver tubes

We are hoping that the 76 will have enough voltage swing to get the 45s to a comfortable volume (note not to clipping) with Paully's speakers, 100dB+ sensitivity.  He doesn't listen loud at all.  We do not ever want the 45s to produce maximum power.
 
According to GZ33 datasheet C max = 60uF. Also with 475+475  and capacitor  input  it will be around 465v DC on the first cap.   Looks like HT for 45 is OK.  10H choke probably is not enough for 76 to filter properly. You may need to enlarge capacitors and add load to get more current through the choke.  
Here is the datasheet http://www.r-type.org/pdfs/gz33.pdf
BTW,  why 4 HTs ?
HTH
Alex
 
The plan is a choke for each rail so we have 4 chokes (2 for each 45 and 2 for each 76).  Does that explain the 4 HT (I am drawing a blank on HT, sorry)?  We will do a lead off 15u cap for each 45 and a 10u for each 76 (total of 50u leaving us well under the 60u total).  Do you think bumping it up to 20H on each 76 would be enough?  We can certainly bump the end capacitors to get more smoothing on each rail as well, I don't think that would hurt anything.  I will go try that on PSUD.  Thanks!
 
Ok. Let's talk about one amplifier. I presume we are talking about 2x mono,right ?
So, you have one full wave rectifier (correct?),capacitor loaded ( that is the cap. With max value 60uF). What's next ? L-C-L-C where first half of the filter loaded by 45 and 2nd by 76 ? Or parallel lL-C for 45 and L-C for 76 ?
 
You must log in or register to reply here.

Similar threads

T
Stereomour OG Power Transformer (and my mods)
2
Replies
22
Views
66K
Paul Birkeland
Paul Birkeland
S
Snow Creek Shuffle interrupted: SR-45s, TFA-2004 Paramounts and @#$&Nakamichi
Replies
0
Views
11K
ssssly
S
D
drive an 833A SET with either an headphone amplifier or a 45 or 2a3 SET.
Replies
4
Views
9K
Paul Birkeland
Paul Birkeland
Ivan303
BeePree2 Done...
Replies
12
Views
8K
Ivan303
Ivan303
Ivan303
Not your father's Paramour
Replies
2
Views
6K
Ivan303
Ivan303
Back
Top