how do i tame a my pt???

[nasajonathan]

here is my situation:
1) i don't want to make an exact clone of a marshall 18 with all the real knobs, handles, etc etc b/c
2) i have no real money, so instead of a heyboer...
3) i purchased a generic pt from allied electronics. i bought one with what i thought would be a little head room (375-0-375, 150ma, 5v@3a, 6.3v@5a) for around $40.

here is my problem(s)
1) i have a 5 volt filiment supply instead of a 6.3 that the ez81 takes. i figured that i can just replace it with a gz34 and not loose too much, or is there some large difference between the two?
2) without a load on the pt, i could get as much as ~530 volts on the caps ideal!! which is about 200 more than i need. so i thought i would make a simple voltage divider down to the 325 and off i go. however, it isn't that easy. i can get the voltage pretty stable with large power resistors, but if i load down the pt with the amp, it will pull the voltage divider way out of spec. then again, if i use a smaller value resistors for the voltage dividers, then i am not passing hardly any current through the amp itself. it doesn't help that i don't really know what kind of load the amp is on the power supply because i haven't built that part yet, nor do i want to if i can't nail down the voltage. any suggestions??? i am caught in a very chicken and egg situation and my brains are getting scambled...

if anyone has any insight into my probelm, i would be greatly appreciative....

jk

[Shea]

When I use PTs that only have one 6.3v winding, I use it for the rectifer as well as all the other tubes. You can do that with the EZ81 rectifier because it is indirectly heated, i.e., the cathode and heater are two separate elements, and they are not connected to each other.

That tranny is pretty overspecced - high voltage AND current. But anyway, let's think about what we can do to make it work. This sounds like a good candidate for using a zener diode in the HT center tap. Get on with a zener voltage that equals the amount of voltage reduction you need. If you want to get close to the stock voltages, make it 85 volts. Connect that zener between the HT center tap and ground, reverse-biased. The power disspation rating should be at least 150ma x the zener voltage.

If you can't find a zener with a high-enough rating, there's a way to use a MOSFET in combination with the zener that will be able to handle plenty of power for your needs. Check out the paper "foolin' with fets" at http://www.geofex.com - it tells you how to do it.

As you probably know, it's almost impossible to predict what the B+ voltage will be with any precision, since it's affected by the load on the tranny and the impedance of the rectifier. For that reason, it's even harder to guess what the zener voltage will have to be. Although the Marshall schematic shows 300 volts DC coming off the rectifier, you can make a fine sounding amp with as much as 390 off the rectifier, if you use a 220-ohm cathode resistor for the power tubes.

Some other ways to handle the voltage would be to use a power resistor in series between the cathode of the rectifier and the rest of the circuit. I wouldn't bother with a voltage divider. The current passing through the tubes at idle will cause the voltage drop across the resistor.

Another way to compensate for high B+ voltage is to use a bigger cathode resistor.

The reason why I recommend the zener is because your trannie's HT voltage is so high, and it supplies more current than your amp will be using at max signal. To try to bring the plate voltage down using a series resistor in the B+ would require a resistor with a pretty high power dissipation rating. You might have to heat-sink it, and you'd have to be careful where you put it. Besides, there is such a thing as too much power supply sag.

And, the problem with using bigger and bigger cathode resistors is that you'll end up with such a big voltage drop across the cathode resistor that you won't be able to drive the grids positive WRT cathode - i.e., no power tube distortion. Unless you make up for it with lots more gain, but then you'd be designing a whole different amp.

So, you need a voltage drop that is too big to be accomplished with resistors alone, IMO, and hence I recommend the zener trick.

Shea

[glennyg]

If you are willing to use 6v6`s instead of the el84`s you could build a 6v6 version. With a 5ar4 rectifier you will see a b+ of around 450vdc. With the rest of the circuit the same , except for a few resistor changes here and there I think you would have a kick`in amp.

Glenn Gertis

[tedm]

With my 18watt Moxon amp with 272FX PT, I have gone to a 220R cathode resistor.

With previous experiments, I've also lowered all my voltages by 15% by plugging the amp into a dual socketed box with a light bulb in series with the AC cord (160W bulbs for this amp).

My AX84 single EL84 amp has the light bulb socket built into the amp, but needs only a 75W bulb to drop voltages 10%. Make sure your heater voltages are +/- 10% of rated spec. With my AC of 121VAC, most all my amps are on the + side of heater voltage specs, and I like to keep my variac out in the garage.

Also, about other tubes, with the 272FX PT, 6L6WGBs work fine, with socket adapters, and gives a nice fender blackface tone, plate voltages are on the low side for a Fender with the EZ81 rectifier (~340-370) but the clean sounds are fantastic, and the crunch is awesome.

So, with the overpowered PT, you can get creative and try out some interesting variations, in addition to lowering it to clone specs.

Ted

[cosmonaut]

what resistor changes would be in order to use 6v6's and a 5ar4? You have got me real curious

[Shea]

I would recommend giving each 6V6 its own screen resistor of at least 1k. You could do that and keep the single 100 ohm resistor too.

The cathode resistor and bypass cap might stay the same, but you would check how much power the 6V6s are disspating at idle to be sure.

If you're getting blocking distortion (i.e., flubby bass) from the 6V6s, try reducing the size of the grid bias resistors from 470k to 220k, or maybe even 100k. Each time you reduce the grid bias resistors by half, double the blocking caps from the pi in order to keep the bass response at the same level. "Huh?" you'll say, "weren't you just saying that the reason for changing the resistors in the first place would be flubby bass? Why would I be keeping the bass response at the same level if flubby bass is the problem?" To which I say, "reducing the size of the grid resistors would reduce or eliminate the flubbing by presenting a smaller impedance to groudn for the grid current." You may try making the blocking caps even bigger, since circuits designed for EL84s tend to roll off more bass than circuits meant for other power tubes.

It's kind of hard to spell out in advance everything that you might want to tweak. Just get your octal sockets wired up and use those 1k screen resistors, and take it step by step from there.

BTW, Ted, how did you plug octal-based tubes into your Moxon? Got a photo?

Shea

[tedm]

Hi Shea,
I did it on the bench with your chassis out upside down on the bench, and the EL84s pulled, and careful attaching of heavy gauge alligator clips from the NTE sockets to your amp sockets or corresponding turret lug or screw on 2nd socket.

Here's a couple of pics of the NTE $4 8 pin screw sockets:

www.tedm.com/amps/nte1.jpg
www.tedm.com/amps/nte2.jpg

Is the changing/addition of screen resistors for tube longevity or for tone? I liked the tone very much with just the plug and play. I did review the tube data sheets, and all seemed close enough, as the resulting voltages and dissipation proved out.

One thing I want to make and test with the 6L6s again is an SS EZ81 plug in (dual 1N4004s) replacement, as my Coppercap puts out less (about 3V) VDC than my glass tube EZ81s.

Ted

[nasajonathan]

some updates... i went a head and hooked up the gz34 which seems to work fine. the output voltage with no load is just a shade over 530 and my cap can is only rated to 500, so i had to shut it off quick. i don't know how it will react with the amp actually attached...

onto the pt... i think i overlooked the most simple solution of all....
i plugged my power supply into a variac and just turned it down till i had about 350-390 on the output (unloaded). hopefully what i can do i finish the amp and adjust the variac until i have the ~300 on the b+ when it is loaded, and then adjust the series resistors on the b+ unti i can turn the variac up to my wall voltage of around 119V. the only thing i can think of that might be bad is that the filiment voltage won't be high enough when the variac is turned down to 90 or so as it is now... but the filiment glows and it seems to work, so until i know better, i am just assuming that everything is okay.

jk

[tedm]

Tube life may be shortened significantly if the filaments are run less than 10% of spec. I would consider a seperate filament transformer if you are planning on running with low filament voltages.

[nasajonathan]

tube life... whatever! i can deal with a little shorter tube life if i can get the amp up and going. plus, the amp was only on for a few minutes while i probed around to get the voltages down to 325 and tested that the amp itself worked.

what i eneded up doing is just putting a series of 20w 500Ohm resistors in series right off of the rectifier (make sure to REALLY heat sink them unless you like the smoking amp effect) and a 5 pole single throw switch so i can select how much resistance i want (from 500 ohm to 3K). this ended up working very well because now i was able to take the power supply off of the variac and just select which voltage i wanted on the center tap of my ot. with the 1K series resistance, i get about 330V on the ct, which works perfectly! amp sounds REALLY great! as a pretty young guitarist, i have only played on ss amps, so i didn't really know what people were talking about when they said that tube amps sounded so good, but they really do! great bass responce, tight harmonics, good sustain and compressions.... wow! the only problem i have left to solve is that when i am plugged into my trem channel with the volume off, i can turn up the normal channel volume and get full sound out. it is actually kinda nice b/c i can turn up the trem channel to how i want it and then boost the normal channel volume for a little edge. any idea where the feedback into the normal channel is coming from?
thanks for all your help!
jk

[tedm]

Sounds like you've got the voltages in great area for dual EL84s.

I don't know how the channel interaction works on the 18watt since my Moxon One has only a single normal channel.

However 5E3 folks often take advantage of the interaction between the used and unused channel. Some go to great lengths to be able to subtley change tones with the unused channel. Check this out:

http://www.redwooddigitallab.com/whiz.html

[zaphod_phil]

The elegant solution is you use a choke, say around 3H to 5H. Your power supply should have a cap of around 2uF or 4.7uF, followed by your choke, and than another filter cap of a more usual value, say 32uF, which feeds your power stage via the center tap of the OT. That will get you very close to the voltage you want. You can adjust the exact voltage by adjusting the value of the first (lower value) cap. Higher voltages will give you a greater DC voltage, lower values will give lower voltage.

With a GZ34 for example and 2uF, 4H and 32uF, I estimate you'll get around 350V DC. Now all you need to do is to find yourself a surplus choke. You'll probably get a bit less hum and better headroom too.

here is my situation:

3) i purchased a generic pt from allied electronics. i bought one with what i thought would be a little head room (375-0-375, 150ma, 5v@3a, 6.3v@5a) for around $40.

2) without a load on the pt, i could get as much as ~530 volts on the caps ideal!! which is about 200 more than i need. so i thought i would make a simple voltage divider down to the 325 and off i go. however, it isn't that easy.

[Shea]

That describes a choke-input filter. It certainly would drop the B+, but also make the voltage more regulated. Would there be much less rectifier sag as a result of the better regulation? Most of us regard rectifier sag as a good thing. These EZ81 / 6CA4 rectifiers that we're using are the highest impedance full-wave rectifiers that I know of.

Shea

[zaphod_phil]

Yes, that describes a slightly-tweaked choke input filter, as it also has a smallish capacitor at the input. The benefit of the rectifier sag vs more headroom is itself a matter for debate, and not one that I myself would dare to take sides in! ) The choke will tend to remove any fast transient sag effects, but leave slower rectifier sag in place, eg resulting from playing loud power chords. You will also get some sag in the 18W due to the way the screen voltage is supplied via the resistor of around 2k ohms. In any case the bottom line here is the general-purpose PT that's being used, and the question is how to find the best way of getting the B+ in range for an 18watter. This means that some compromises from the standard design will have to be made somewhere along the line, whichever approach is used. The fun of it is trying these ideas out.

[Shea]

[reply]The choke will tend to remove any fast transient sag effects, but leave slower rectifier sag in place, eg resulting from playing loud power chords. [/reply]

I considered the latter possibility too. Maybe that slower sag is what we really hear anyway, so no harm done? Also, the DC resistance of the choke itself would add its own slow sag, wouldn't it?

Shea

[Shea]

Whoops, ignore those ineffective codes. I forgot which board I was on!

So, did the founders decide not to have an edit function?

Shea

[zaphod_phil]

I suspect the rectifiers' sag under load will be more significant than any due to the inductor's own internal resistance. I have to say this is my own gut feel, rather than on the basis of careful calculations...

I considered the latter possibility too. Maybe that slower sag is what we really hear anyway, so no harm done? Also, the DC resistance of the choke itself would add its own slow sag, wouldn't it?

Shea

[zaphod_phil]

Also, we mustn't forget (like I just did) that one of the main places you get sag in the 18W is from the resistive cathode biasing in the power stage. You can alter its transient response by changing the values of the bypass cap (or even the bias resistor). One posting I saw from Paul Ruby somewhere suggested you get a nice sag with 1000uF or above - ie the thousands of uF. Whether everyone would want that huge a bass response at the same time, is anither matter....