Plexi 50W 1986 bass questions

[goldstache]

Getting ready to start my first plexi build.  Better late than never!
Got a few questions.

I have a 190-0-190 PT that won't support a 100watters filament needs, and I have chassis that's pretty perfect for a 50W.
So I'm wondering if I can go full wave bridged like the 100 watt and obtain reasonable voltages for a 50W design.  I'd like to do the bass model Marshall and perhaps add the differences between the 1987 and 1986 (cathode bias, and bypass caps, etc) on a 4pdt switch.

I noticed that the 100 watt Marshall transformers out there are 170-170, do you think my 190-190 will jive? 

Should be a neat build, some parts are on their way!

[PRR]

Might come out 60 Watts.

[goldstache]

Here is the schematic I made for a Marshall 1986 model.  I'm trying to build it with parts I have on hand.  I'm especially interested in the bias supply and the elevated heater supply. 
I'm trying to use a Heyboer Pt I have laying around (190-0-190).  It's almost spec for the 100W 1959. Using full wave bridge.  Should I use what's in the schematic I made as a starting point, or go off the 1959 bias topology? And I have a heater float in my attached schematic. 


Do you think this is possible??? 


Trying to make this jive with what I have laying around.
Thanks

[PRR]

Gonna need a return (ground) somewhere around PT and rectifier. The odd corner of the FWB seems like a plan.

No clue if the sidekick bias supply works that way or with those values. Test that before you stick power tubes in.

[goldstache]

Good eye!  Forgot that on schemo apparently.

Thanks.

I'm also entertaining power scaling via PTCT.  But that's a later nut to crack.

[sluckey]

Page 2... Move those .1µF coupling caps to the left of that junction dot. As drawn the bias voltage cannot get to the grids.

Page 3... That bias supply will not work with a bridge rectifier. Look at page 5 of this pdf file. Copy exactly the circuit at the left side of the bottom of the page. And feed it directly from the PT, not through the standby switch.

[goldstache]

Boy i really goofed that schemo.  Again, Great eye!

Thanks for help!   I'm interested, in that PDF.  You gotta link?

Time to make my turret board.

[sluckey]

Sorry, forgot. Here it is...

     http://sluckeyamps.com/misc/Amp_Scrapbook.pdf
     

[goldstache]

Sluckey, thanks!  What a concise set of documents! 

[goldstache]

Thanks to PRR and Sluckey I have modified the errors on my schematic (attached) and struck the old one above from the record. 

Time to build as written and see how she goes. 
Thanks guys.

[goldstache]

Well I got it all wired up.  I've got my variac at about 85% of 120V.  This is yielding around 430VDC at the first reservoir (50uF + 50uF @ 500V parallel).
My bias is at 32ma as of now, as I did the math .7 X 25/430 gives me 40ma or so.  I lowered the bias pot just to test play.  Sounds great! 
I have a couple concerns and questions. 

It seems I'm gonna have to go series caps for higher voltage handling.  I'm thinking another 50/50 @500v can in series with the existing one.  As for post choke voltages, I'm sure they will be close to the 500V rating of the existing cap can already installed. 
Thinking I might have to try the Zener method to drop some voltage. 

The other question/concern is that I'm using a Bridge rectifier in the amp.  The current really takes some excursions while playing and the voltage usually sags about 30V when the amp is at about 1/2 output volume.  My PT is rated for 190-0-190 @ 400ma.  I'm assuming that idle is low so some heavy travel when playing is normal.  Both the PT and the OT are barely warm, so I'm thinking it's ok. 

I was playing loud for about 10 minutes and then I noticed a little magic smoke!!!!?????  I powered down and looked around but couldn't find any burnt parts, and the smell check yielded some funky smells around the PI portion of the circuit, or around the V2a/CF tube.  I brought it up again as I figured maybe one of the dropping resistors in the HT gave it up under the higher current travels when playing loud.  The amp still works without noise and I haven't felt brave enough to repeat the "Smoke" scenario. 

Anybody got any tips or ideas? 
A note on the Zener method of reducing B+:  Since I'm using a bridge rectifier and my PTCT isn't grounded, I just slap em' between the bridge rectifier and it's grounding node right?????

Thanks.  It's a singer.  Blissful harmonic feedback in that thing. 

Thanks everyone.

[goldstache]

Here are my voltages as of now.

V1A,B
Pin 1- 176VDC
Pin 6- 178VDC
Pin 3,8- 1.19VDC

V2A
Pin 1- 152VDC
Pin 3- .92VDC
V2B
Pin 6- 266VDC
Pin 7- 152VDC
Pin 8- 153VDC

V3A,B
Pin1- 198VDC
Pin6- 210VDC
Pin3,8- 35VDC
Pin 2- 22VDC
Pin 7- 23VDC

EL34's have 435VDC on the plates and 430VDC on the screens.

Referring to my attached schematic above. 
A- 437VDC
B-436VDC
C- 307VDC
D- 266VDC
E- 251VDC

Also, my bias on the grids is -39VDC when biased cool at 32ma

[sluckey]

Nothing wrong with 500V B+. But you're gonna have to use a series set of caps on the screen node too. That's a lot of money if you use 50/50 cans like Marshall did! That's a total of six cans.

I would suggest you get four F&T 100µF@350Vdc (approx. $22 if you shop wisely) to take care of the plate and screen nodes. You'll also need two 56K/3W balancing resistors for the screen node. Then use your 50/50 cans to take care of the rest of the nodes.

Wire the plate and screen nodes like this 100W super bass...

     http://el34world.com/charts/Schematics/files/marshall/Marshall_jmp_super_bass_100w_1992.pdf

[goldstache]

That's what I was thinking.  I have some 100uF @350v F&T's.  I will rewire this week.  What do you think about the current swing?  Is it extreme?  I have to set my DMM on 2VDC range as it swings plus of 200mv when Im cranking it.  I'm monitoring current draw through standard bias test points.  So that's a swing of 200ma per tube when peaking.  And the amp is only on 5?  I'm sure the idle is acceptable, but just wondering if these wide current measurements are bad for the tubes or PT.  No way to regulate and the 100w plexi's used the bridge rect. and they are still kicking around.  I guess that's normal.  Ive only had experience with full wave and current definitely swings but not as wide as it does with the bridge.  I kinda get why, the load effects the PT much differently.  But wanted to ask if anyone thought this was extreme.

Sluckey, thanks for the bias supply tip.  Worked flawlessly!!!!!!

[goldstache]

I forgot to mention that I don't have the NFB hooked up.  It was causing a hummmmm and a whine, so I disconnected it.  I had it hooked up to the 8ohm tap.

[PRR]

A 50W amp sure is likely to swing >200mA peaks.

[goldstache]

I originally bought these transformers for an amp that called for a 190-0-190 SS full wave.  The B+ was around 250VDC for the circuit.  I was using EL34's and researching Class A operation.  The Current draw numbers wouldn't swing too much and the voltage was almost fixed at 250V under playing conditions.  Given the higher current demands having such a low HT I was surprised that it didn't peak a little more.  It is a hearty Heyboer Transformer, and perhaps the underuse of it in general led to it's "stability".  So when I went with the same iron to try and pull off this build(plexi), I was surprised that the current and voltage would sag/peak so widely!   I figured the 100W Plexi power topology with the FWB rectifier could probably get me close.  It seems to be jiving pretty close to the voltages I see floating around for a 50W. 


Pretty sure the NFB on the was squealing due to OT primary swap.  I wired the whole amp up in 1 session, so all I could stand to do, once I heard the Shhqueal, was clip it out and turn off the iron. 


With no NFB its a bit unruly and lacks the tightness, so I'm looking forward to swapping leads and giving it a listen. 


Still wondering if my thinking is correct for the Zeners to drop some B+?????? 15V 5Watt 
Without them, once the Variac is at 100% (I'm currently still at 85% of the variac 430V on the plates). I'll bet i'm gonna be a little over 500 on the plates. That seems a bit high for what I've been reading.
Is it even a big deal to be 30-40 volts off?!
Thanks for helping with my build! 

[kagliostro]

Zener reduction of B+ can be achieved in some different manner

for CT transformer here are a pair of solutions and one for non CT PT

(using a 47v 1.3W Zener + TIP147 or 15V 5W Zener, if necessary, on the 5W Zener option you can stack more Zener to drop more voltage)

Be aware

Quote from Tubeswell

(Edit: Of course, if you have a fixed-bias output stage, you'll want to take into account the power supply source for the bias voltage when considering dropping the HT voltage with this zener method. Some PTs have a separate winding for bias supply, and in that case the bias voltage will be independent of what you do to the HT winding)

Here one other

http://el34world.com/Forum/index.php?action=dlattach;topic=19277.0;attach=53605

Franco

[sluckey]

Nothing wrong with 500V B+.

[kagliostro]

Nothing wrong with 500V B+

Franco

[PRR]

> I was ....Class A operation.  The Current draw numbers wouldn't swing too much and the voltage was almost fixed at 250V under playing conditions.

Almost by definition, Class A current demand does NOT change without and with signal. With real tubes, the current may rise 5%-25% at full signal. If it rises more, or drops much, the load is non-optimum.

> I'll bet i'm gonna be a little over 500 on the plates. That seems a bit high

How high is high??

The original EL34 has a 800V plate rating. Most folks think this was insane, and probably is on present production, but it is not a low-volt tube. Screen ratings are lower, but Marshall and everybody has been pushing EL34 to and past 500V for decades (Ampeg had 580V plate in one model), and present production has to accept that.

Nothing wrong with 500V B+.

Yeah.

[goldstache]

I added series caps for the reservoir and for the Screen supply. 
Swapped the OT leads to the plates. 
Reinstalled the negative feedback (ended up using a 22K resistor with 250K pot for NFB control)
The plate voltage is around 536VDC and I'm biased at 32ma

The amp sounds good and clean.  I got used to hearing it with no negative feedback as I had it disabled due to OT primaries being swapped around upon initial fire up. 

I like it.  I also liked it at 440V as well. 

Was thinking of trying the Zener method of lowering B+. 
In the above posts there is a pic of use of zeners with no CT on the PT. 
My PT does have a CT that I have connected to the center junction of the reservoir series caps.  If I want to try the zeners, do I disconnect the FWB rectifier ground point and attach the zeners from that point to ground????
It was also noted that I should be aware that the bias rail changes voltage.  I tapped the bias supply directly from the secondary AC taps on the OT.
So, I guess im asking, does the Zener method lower the AC of the secondaries?  Or, does it shed voltage off the post rectified DC?  I was thinking the the latter, but I have to ask. 

Thanks everyone, just dialing it in now!

[goldstache]

Playing it in a bit and noticed a couple things.

When I get to about 75% up on the volume control one of the EL34's starts to flicker red, then subside shortly after I cease playing.  Verge of Red Plating perhaps.  The tubes are within 5 mils of current draw difference and this one is the hotter of the two.  I'm scared to go full volume for ears/tubes sake. 

Im seeing that Vg2 of EL34 is 500v.  Am I just seeing some over voltage on the screens?  I have 1K screen resistors on.  Should I drop a bit via RC stage to the screens or another option. 

I haven't tried the Zener B+ drop as of yet.  Still waiting to hear your input on that. 
Also, I forgot to include the specs of my OT.  It's a Heyboer 3.9k primary.  PRR, you got me thinking about the wide current excursion in while playing and that the wider the swing perhaps the wrong primary impedance might be going on.  I'm not real sure how to calculate Output Tube impedance when matching a transformer.  I'm thinking its Ra and the voltage in some relationship. 
Thanks again.

[PRR]

> not real sure how to calculate Output Tube impedance when matching a transformer.

Datasheets and known-good commercial amps.

[goldstache]

I tried the Zener method of dropping B+.  I series wired 4-15v 5w zeners cathode band towards chassis ground and anode side of diode to FWB ground node.  I got no voltage drop whatsoever.  I even changed my adjustable fixed bias to see if it would bias in at 70%.  NO DICE.

What am I missing here? 

[sluckey]

Measure the voltage drop directly across the string of 4 diodes. What do you have?

[goldstache]

Red lead to FWB Black lead to ground.

-3.3volts?
I'm I seeing normal biased diode vdrop????
Banded side does go towards ground right?

They don't show continuity.

[sluckey]

Are you saying you measure only 3.3V across that stack of zeners with the amp powered up as normal? If so, they are installed backwards, meaning they are forward biased.

The voltage drop across a forward biased zener is only about .7v just like most common silicon diodes. Zeners need to be reverse biased to see the voltage drop you are expecting.

[goldstache]

I removed the zener string and didn't get .7v drop across one.  Then I checked continuity and it's open. 

Just to clarify, Banded side toward star ground anode toward GND of FWB right?

Thanks

[goldstache]

Tried it again with some fresh zeners.  Worked this time.  Same orientation as the last string.  Kinda strange, but it seems to be legit.  I rebiased and I'm letting it burn in. 

[sluckey]

Glad you fixed it. Now that it's working measure the voltage drop across the zeners while the amp is turned on.

[goldstache]

Thanks to everyone here!  I'm seeing 39.7v. 
I'm thinking I'm good.  P=E X I
 
I've got 5-15V 5W Zeners in series. 

E X I     39.7v X 500ma to be safe
P= 19.85 Watts at 500ma

so 25W handling should be good, No????

My bias supply of course was altered too.  I'm using the one you posted Sluckey.  Do I alter the First divider or subsequent dropping resistors in that supply to give a better range to my bias pot.  I was able to dial it in, but its past 70% of the bias pot travel to do so.  Wouldn't mind tweaking it for the learning experience if nothing else. 

Again, thanks.  It sounds great where it's at. 

[sluckey]

Decrease the 47K bias range resistor to decrease the negative voltage on the grids which will increase the current thru the tube (tube runs hotter).

Increase the 47K bias range resistor to increase the negative voltage on the grids which will decrease the current thru the tube (tube runs cooler).

[goldstache]

Great! 
Again, thanks.

I finally get why Marshall sold so many of these amps!!!!!

[goldstache]

22K solved it.  For this set of Winged C's the bias pot travel is centered and it's biased at 70%.

Thanks

[goldstache]

I have a total of 5. 

I was expecting 70v or so.  Based on my B+ values before.  I think I had around 525 on the plates at around 32ma. It's biased in around 440V on the plates at 38ma right now. So I'm a bit confused as to the relationships of what's going on. 

Do you think some of the Zeners in my string shorted????

:EDITED PREVIOUS B+ VOLTAGE> oops (525V)

[sluckey]

If you feed a well filtered dc voltage into a zener, you can rely on the voltage drop across the zener to closely equal the voltage stamped on the zener.

But you are feeding 120Hz pulsed dc that ain't well filtered into the zener and the dc voltage drop you measure across the zener with a dc voltmeter will always be less. You would need a scope to see what's really going on.

[goldstache]

Makes sense. 

Now to play with some values in the preamp. 

The 1986 circuit is indeed bassy!
I like the way the brighter channel (1) sounds.  It still has a little too much thud through most of the cabs I've tried.  All at 8 ohm. 
I'm using shared cathode resistor and bypass cap (330uf). 
Should I split the bias and go with 1987 preamp?
Or......  Just lower bypass cap on the cathode from 330uf to something half that, or coupler out of the stage?

  It's really close to where I like it, just needs a little less low hz response. 

[sluckey]

The 1986 is a bass amp. It's supposed to sound bassy.

[goldstache]

True, and I seem to gravitate to bassy amps.  I usually play a blackface bassman head.  And I've owned tweed bassmans. 

But this thing is a thumper.  I've tried greenbacks in the pas and thought they were a bit bright and lacked the low end.  Perhaps that's a nice combo for this circuit. 

I guess I'll have to switch to 1987 spec pre and tweak.  I know I'll have to add the .68 to the driver before the cathode follower.

[goldstache]

Well, left it on for about 5 hours and played it periodically throughout the night as I worked on the cabinet.  Had it at about half volume with a booster pedal on.  About 15 minutes of loud playing I smoked the zeners!  I think my voltage crept back up to 525VDC on the plates and started to smoke the 1k 5w screen resistors.  Powered it down, pondered it, and called it a night.

Should I buy a proper wattage Zener with sink type?

Up my screen resistor wattage?

[sluckey]

That big amp just wants to run at 525V. NOTHING WRONG WITH THAT!

Zeners seem to work OK at voltage reduction in smaller amps. But for big amps if you really want to lower the B+, just get an appropriate power transformer.

[HotBluePlates]

The 1986 circuit is indeed bassy!
I like the way the brighter channel (1) sounds.  It still has a little too much thud through most of the cabs I've tried.  All at 8 ohm. 
I'm using shared cathode resistor and bypass cap (330uf). ...

The answer is to jumper the two channels (perhaps with a short cable from one channel's input jack to the other channel's input jack), and use the two channel volume controls to balance screech & mud.

That's the solution almost everyone arrives at with these amps, because the bright channel is almost always unusably bright, while the "normal" channel is too muddy.  You use the two volume controls like an overall EQ for the amp.

[goldstache]

Sluckey, I'm starting to feel the same way, though I do like the sound with the voltage a little lower.  And aside from that, the tubes would glow pretty bright when the volume is cranked.  Sounded OK just made me a bit nervous.  And I'm running the screens above 500V.  I plan to use it with an attenuator, out live, so I can't have it smoking on a gig.  Thanks for the help too!

I've been jumpering too, and I love that about Marshall's, but I gotta tweak it for my ears.  It's just a low end monster.

I'm gonna run some voltage tests with audio to see what things are doing around the screen supply/resistors. 

Any other input is greatly appreciated. 
I know its the classic case of wrong iron, just want to settle on a healthy design that I like the sound of.

[sluckey]

I plan to use it with an attenuator, out live, so I can't have it smoking on a gig.

If you gotta depend on it, by all means, get a proper power transformer. The only thing smoking so far is the zeners.   

[goldstache]

I'm going back to no zeners.  But I would like to add a sag resistor after the choke to get the screens down. I'm thinking the voltages on them are too high 520-25.
 Without the Zeners, biased in, its got about 530V on the plates at 33ma. 

[kagliostro]

http://el34world.com/Forum/index.php?topic=20467.msg215610#msg215610

Reply#12

But may be that is excessive if you want to drop only 30 - 40V

Franco

[goldstache]

I will comb the thread!  Thanks

I wired it up with a 1k 10w before the screens.  Choke> 1K> screen node

Brought the screens to about 498V at idle when biased.  This also brought the PI and preamp voltages a little closer to what I see floating around. They are within about 20volts or so depending on position.

Also settled on making it the 1987 50W circuit.  Gonna play it and tweak it a little more.  It's really a learning experience.  Thanks for all the guidance.