Full wave CT grounded + bridge rect = trouble?

[jeff]

Hi guys
 
 I'm having trouble trying to understand this schematic. It looks like they are using two full wave rects for the pos and neg rails of the ICs and at the same time using a bridge rect for the bias. But because the CT is grounded they're using the two 2U2n.p. caps to isolate the bridge rectifier because the C.T. is grounded. Will this cause problems. I know when using a bridge you can not ground the C.T. but by adding the 2u2 caps you can???
 
This is the schematic for an amp knowing to have bias drift problems(not just mine, this model).
Could this be part of the problem?
Would it be wise to get an accesorry transformer just for bias? Or is it OK to run a bridge rect if the CT is grounded as long as you use those caps?
 
The part on the right(v+,v-) looks correct to me but I'm confuzed about the part on the left(bias).

I don't understand the use of both at once. Top is full schematic, bottom is partial I'm having trouble with.
 
Thanks,
Jeff

[John]

I'm only chiming in to learn something. Neither drawing makes much sense to me. The first one showing the 2 rectifiers: the first one coming off the PT makes sense if they're making a negative bias, but then from there it looks like the next one is a voltage doubler?? And I don't understand what the 2 2.2uf caps are for (I assume n.p. means non polarity) before the rectifiers.


Again, I'm just trying to learn something. I have NO clue what's going on in either one. 

[jjasilli]

Maybe this is helpful:  https://www.google.com/search?q=bridge+rectifier+with+caps&rlz=1C1RNBN_enUS433US433&espv=2&biw=1280&bih=601&tbm=isch&tbo=u&source=univ&sa=X&ei=o1afVK_uCJT-yQTf3YCoAw&ved=0CCAQsAQ#facrc=_&imgdii=_&imgrc=rB9FniqoTraMnM%253A%3B3tCs7zU2fUw7RM%3Bhttp%253A%252F%252Fwww.decdun.me.uk%252Fgc%252Ftraffo_config.gif%3Bhttp%253A%252F%252Fwww.decdun.me.uk%252Fgainclone_psu.html%3B440%3B251


And:  http://www.decdun.me.uk/gainclone_psu.html

[Merlin]

The first circuit contains an ordinary bi-polar supply to generate equal + and - rails, and then there is another cap-coupled bridge rectifier which generates a rail voltage twice as large (i.e. it is a type of voltage doubler). This is basically the same thing, drawn more clearly:
http://www.tubecad.com/january2000/page16.html

[PRR]

> I'm confuzed

That is usually a sign you have got too clever for your experience, and probably should step-back to stuff you DO understand.

I'm confuzed too. My hasty check suggested it might work, but I am NOT SURE of this. Merlin has good eye and his post encourages me to a weak "may work" opinion.

But not-understanding increases the risk of wiring errors (actual build is not what the plan said). And depending on the transformer, errors can be stinky smoke or outright fire, on the bench or inside your wall. (UK type fuse-plugs are good ideas; if you don't sport those you sure should fuse as soon as wall-power gets in your box.)

And I do NOT think this gives you voltages which are appropriate for Tube Amps. You have a + and a - main supply, which is unusual in tube audio amps. You have a "bias" which hangs very-very negative. If we pencil the winding as 100V-0V-100V, you have +140V, -140V, and -280V.... strange.

I also think these cap-coupled (not just cap-filtered) schemes can be critical of their cap values, and often very expensive for the power they deliver.

There's got to be a better way to do what you need doing.

[sluckey]

The bipolar supply is +15VDC. The oddball supplies output tube negative bias. Here's a real world application...

     http://el34world.com/charts/Schematics/files/marshall/Marshall_jcm2000_tsl60_60w.pdf

[DummyLoad]

silly sim software says it works...
--pete

[jeff]

Merlin's schematic is easier on the eye. Same idea with two excepitions: Lower voltages(pos and neg for IC supply) and the bridge rect diodes and caps are reversed(for bias)

> I'm confuzed

That is usually a sign you have got too clever for your experience, and probably should step-back to stuff you DO understand....
There's got to be a better way to do what you need doing.

This is not another one of my crazy thought experiment ideas, but from an actual amp
The v+ and v- are for the ICs and the bridge is for the bias. Seems to me like cost cutting measures to dual porpuse one xfmr???
 The reason I ask is beacuse that amp is known to have bias drift issues.

The bipolar supply is +15VDC. The oddball supplies output tube negative bias. Here's a real world application...

     http://el34world.com/charts/Schematics/files/marshall/Marshall_jcm2000_tsl60_60w.pdf

Yep, that's the amp. Well actually the TSL 100 but same idea. I got it for free because the bais couldn't be stablized and was wondering if this set up is the cause.

There's lots of info on the web about it. Mostly linked to conductive boards. The solution is to lift the bias curcitry off the board in one way or another, with various results. I'm attempting to do that but was asking because I've never seen anything like that and was wondering if that may be part of the bias drift problem. Sometimes you fix one thing and it seems to work, but the real problem is something esle. I don't doubt the board has problems but wonder if this also contributesto the problem some how.

I'm doing the suggested mods but was wondering if, while I'm in there, if it would be a bad idea to get a seperate xfmr for bias as well. 
http://www.drtube.com/modifications/jcm2000-stable-bias-mod
http://www.lynx.bc.ca/~jc/TSL122.html

[sluckey]

The circuit seems to be fine. The bias drift seems to be caused by the low quality circuit board.

[PRR]

> The bipolar supply is +15VDC.

Ah, that makes more sense: a tube/chip combo needing several supplies below full Plate Supply.

It is probably +/-30V (R-Zenered to +/-15V), plus a -60V for bias.

The Marshall drawing is really tangled, hard to decipher thus hard to re-draw correctly.

I would now assume it "works".

> conductive boards.

There's no very high impedances here. Hard to accept a board-rot problem. About 70K in the bias divider, whereas we often have 100K and 1Meg in signal circuits on the same type board.

I do think the 51V Zener is a Bad Idea. If the whole amp is over-volted (high wall voltage), we NEED the grids to go-down at least proportional to plate/screen go-up. I'd clip that Zener and re-bias "naked".

I do think the series caps C61 C62 are Bad Ideas. Especially being similar value to C69, the main DC cap for the bias add-on. Small variations in cap value could be significant variations in actual DC delivered. And a hot tube amp is a bad place for cap stability. And a cap failure (dead or just too drifted) leads to weak/no bias and power-bottle burn-up.

If I was asked to FIX this, I'd be real inclined to hang a teeny 120:12 transformer backward on the 6.3V winding (at the FIL points), to get 60 VAC which can be rectified to plenty of bias for this amp. Probably get 80VDC, which would have to be cut-down to 40V-50V DC for the bias point. The output is proportional to wall voltage, not critically affected by any part value (transformer ratio is wound-in, and R and C can drift a fair bit with no ill effect). One diode, a C, an R, another C, then two R around a trimpot to pick-off the desired bias voltage.

I do not know the amp or how ugly such a hack could get. I do think RELIABILITY is key. Good joints and a K.I.S.S. design.

Or burn the guts out and re-build it with-OUT any chips or ribbon-connectors (he-he).

[jeff]

The Marshall drawing is really tangled, hard to decipher thus hard to re-draw correctly.

I know, right?
 If you think the schematic's a tangled mess, you should have a peek inside the chassis!
There's got be at least 15 3 or 4 wire connections between the 4 boards. Wires everywhere. Total Rat's Nest!
If a well designed circuit board is a work of art, this one's a Picasso! Ha ha.

[kagliostro]

If I was asked to FIX this, I'd be real inclined to hang a teeny 120:12 transformer backward on the 6.3V winding (at the FIL points), to get 60 VAC which can be rectified to plenty of bias for this amp. Probably get 80VDC

I think PRR had a very good idea, if it was my amp I would try it


K

[jeff]

Ok thanks
Would using a 120:24v and a voltage doubler give enough bias voltage? I think the EL34s run at 470V I'm not sure what bias voltage El84s at 470V want to see.

[PRR]

> what bias voltage El84s at 470V

EL84 are NOT rated for 470 Volts!!

The original ratings were 300-330V. It had a fancy twin 7189 rated 400-440V. Since "all" EL34 guitar amps pushed/exceeded the limits, and "all" current EL84 production goes into guitar amps, I might ass*ume that the Chinese use plate-stuff more like 400V than 300V. However the good stuff costs more and price-pressure is intense. I would not expose EL84 to more than 390V, and then only if idled at very low current (20mA) and not used for all-night head-banging.

> and a voltage doubler

You are getting complicated. Bias failure is expensive. K.I.S.S.

> what bias voltage El84s at 470V

Divide Vg2 by Mu(g2); that's the most you could need. The Mu(g2) value may be obscure on the data-sheets; it is "10" for most tubes (6V6, 6L6, EL34, 6550/KT88) but 18 for EL84.

So to keep the tubes from melting in the first minutes, assuming Vg2 is same-as Vp, you need 470V/18 = 26 Volts. So 24:120 off a 6v feed is more than enough without doubling.

However I expect EL84 at 470V to fail in hours at any bias.

If you typed EL34 and the "8" was line-noise---- 470V/10 is 47 Volts, and 24:120 off 6V falls a hair short. It might work. But isn't that the problem? You may be needing something that WILL work, as near as humanly possible. Consider the relative cost of a teeny transformer against the cost of perhaps multiple full-sets of EL34 and the cost of lost gigs.

[jeff]

Thanks
 My mistake. typo I meant EL34 but typed EL84

I only asked about the 120:24 because I had a small xfmr on hand that I thought was a 12-0-12(24CT)
My original idea was to run it foward, off the 120V main and use the doubler not off the 6V.

 but when I double checked I found it's not a 12-0-12 it's a 12CT(6-0-6). That's perfect, So I'll go with your original idea

 On the original schematic the 6V heater supply does have two 100R resistors to ground for artificial center tap. So when I connect the 6-0-6 xfmr to the 6V supply should I ground the center tap or tape it off?

thank you
Jeff 

[sluckey]

Tape it off.

[jeff]

Could I remove the 100R resistors and ground the CT?
So instead of the 6V supply being refrenced by the 100R resistors it would be refrenced by the CT of the accesory tranformer?
The only reason I ask is because that would be more convinent to tie the acc. transformer to the board. Remove the 100Rs and solder the new Xfmr in those holes. Now instead of the 6V having 2 100R as an artificial CT it would be using the acc. Xfmrs CT.
Inotherwords using the new Xfmr's CT to refrence the 6V heater supply in place of the two 100Rs

[jeff]

Sometimes I can express myself better with a picture

Are these two schems are doing essencialy the same thing? It's just it would be easier for me to remove the two 100Rs and wire the Xfmr in those holes than splicing the xfmr in somewhere else. The Xfmr leads would just go where the 100Rs are now.

Would this work?

[jeff]

OK progress so far;
 
I dremmeled board around pins 5 to isolate bias voltage from board
Figured while I'm at do pins 3 as well since they go directly to OT I can keep those voltages off the board
 
Now to hook small xfmr to 6V supply and build and try to mount bias circuit.
Anything wrong with removing 100Rs and grounding Xfmr CT?(last post) Seem like it would be ok just easier to solder that way.
 
Thanks
Jeff