More power supply ?s

[GibsonGM]

Ok, so I'm putting together a PS for a Matchless Hotbox/Tube Reactor clone.  Just for experimenting I am doing it ugly...using a 12V transformer rectified/regulated for the 6V heaters, and I have an big old tranny with several taps. The 6V side of things is fine.

   I'm using a set of taps on the other tranny that put out ~200V (low as she goes)...fed that to my rectifier (4x 1N5404's), still have about 200V...after I go thru the filters (47, 47 and 22uF caps + correct 22K/10K resistors for B1 & B2), this has bounced up to about 300V, at least 80V higher than I want to send my 12AX7's, and too close to my 350V cap rating for my liking.  I'm not that experienced w/power supplies, I know enough to work safely - I assume this is occurring because of voltage 'building up' in the caps.  Question is how to return to the 'normal' output of this transformer, and get myself back into the 200 to 220V realm??   

1 thought I had was to remove the rectifier and go to a half wave arrangement, which is more inefficient and therefore would put out less voltage?  Or decrease the value of those caps, but then I would probably introduce a lot of noise and ripple.

Thanks,

~Mike

[R.G.]

Is your big ole tranny isolated??? MAKE SURE that the windings you take the output from is electrically isolated from the windings you put the AC power line into. An ohmmeter test between them will get you started. If they are connected, you can't use that setup without putting your life in danger.

I'm not sure how you made your measurements. If when you say "fed that to my rectifier (4x 1N5404's), still have about 200V" you did this measurement with a DC meter and no filter caps on the output of the rectifier, that could be correct.

200Vac (RMS sine wave) produces a peak value of 200*1.414 = 282.8Vdc. But if there's no filter cap on it to hold the peak value, your meter might read the average DC value, which is about Vpeak times 0.637, or 282.8*0.637 = 180V. When you add a capacitor to that, the cap holds the voltage to the peak minus any loading ripple.

I have to admit that with no other data, I'm at a loss on the "bounced up to 300V" part. That's ...impossible? Nah, it's what you're reading. Strange? Certainly that.  Is there any loading on the output at all? Say a couple of 100K resistors to ground?

As a safety pointer, put a 100K 1/2W resistor across every single filter cap. That will bleed the cap down to ground within about 30 seconds to a minute after you turn AC power off and keep the caps from holding a "snake in the grass" deadly charge for your fingers to find.

[GibsonGM]

Thanks R.G., you gave me some things to think about.  I did in fact make sure primary & secondary are isolated....this is a Buss transformer, with a 120V primary to either 220 or 440 output from the tapped secondary.  I am using the 200v taps.   The 1st measurement was under no load, from the transformer alone, and then with a 4 diode full wave rectifier.   Both read about 200vac and 200vdc respectively.
I wasn't very aware of the peak-holding function of the filter caps - makes perfect sense, that's just my inexperience with power supplies showing!  So the 300vdc output of the whole thing is now logical.....I'd like to get that down to ~200 so I don't cook 12AX7's, am not too sure about just using a voltage divider after the output of my rectifier and filters?
I will add a load of 200K, using 1/2 watt resistors and see what I get then, maybe it just needs the load to 'come back down'?
BTW, this is where I am getting this from - everything is per the diagram except for the transformer (which is really too much transformer for the job): 
http://www.storm-software.co.yu/diy/index.php?project=tube_reactor

There's already a provision to bleed the caps using the 'standby switch'; thanks for reinforcing the concept - after powering it down w/meter still attached, I watched the 200v remain present for quite a while before hitting the bleed switch!!  Potential mankiller, that stuff...

___________Edit

It dropped to 270VDC with the 200k load...according to the site I got this from, the expected load will be about 20mA, 10 to 20 times higher than the current flowing during this test.  Since I'm under the 300V maximum for 12AX7s, I'll probably go ahead and throw the circuit together using this setup and hope for some more sag.  Does this sound reasonable?  For a permanent build I would purchase the Hammond transformer, but for experimental purposes this could work out....(I am using a separate transformer to deliver the heater current with no troubles).   Thx