Dropping screens voltage

[JC1]

Screen current is proportional to plate current. Just as plate current will increase with signal (in a class AB amp), so will the screen current.
Also consider tube failure modes, eg screen may short to other elements within the tube, resulting in significant fault current; do you want the screen supply to be robust enough to cope with those currents without being damaged?
Pete

Pete[/QUOTE][/COLOR]Right, I thought about that last night after I posted this. So are you're saying that lower screen voltages to plate voltage without using an ultra linear TX to accomplish this would damage the tubes?

Or, with screen current raising with plate, thought considerably less of course, more current = more wattage = more robust zeners?

Ok, getting back to first base, a lot of people will place their zeners on the CT of the PT which of course drops the B+ porportionally to the zeners, but, why not just leave the plate voltage higher as long as the tubes are rated to except that voltage & just lower the screens to their intolerance range? The only way to do that of course with a conventional guitar amp non ultra linear PT is to lower the screens with either a large B+ voltage dropping resistor in series with the choke, thus effecting the whole preamp section too, or the zener diode trick in series from the screen B+ node to the screen resistors.

Hasn't anyone else tried this? Am I missing something?

Plan B would be the B+ dropping resistor in series with the choke & Plan C just lower the whole B+, plate included, with a zener to CT.

[R.G.]

Well, that's not the only way.

A resistor in series with the screen only drops the screen voltage because there is screen current. The resistor follows V = I*R.

You can do other stuff. A regulator tube might help, but there are cheaper, smaller, cooler answers. A power MOSFET can be set up as a source-follower to provide a voltage to screens that is a proportion of the B+, or if you work at it a bit more, a voltage that sags with plate currents. The hidden gem here is that the MOSFET can be set up to limit current when things go really bad, so you don't get things melting, overheating, or starting fires. And the voltage, following %, and current limits can be set with small adjustable trimmers.

For things not in the signal path, high voltage MOSFETs offer a great way to do advanced things without eating up more tubes and sockets.

[JC1]

RG,

I'm going with Plan "D".....just lower the bias to those KT90's a little as not to run the tranny too hot. Not getting all the bang for the buck I wanted for because of a underratted PT but it still gives me enough power to be heard on stage. The problem is the output current isn't enough to drive these properly. My PT felt a bit warm, especially on the bench without the fan I installed inside the combo blowing air over the PT & output tubes. As long as these sound good biased colder then I'm good to go. I was hoping for 55 - 60% with 46w disapation per tube but not with this PT. I have plenty of filament current to spare & a 5v tap for my switching & LEDs. It was difficult to find the right PT with all those features and enough B+ current to boot.

I did some more digging & found that my PT for my 2203 is 350-0-350, 290ma. That drives 4 EL34's rated 25w each to about 100w with no problem but 201ma doesn't drive 2 KT90's rated at 46w any where close to 92w with appox. the same voltages.

I've measured up to about 76w when I first set them up but now I'm getting less than 70, or about 67w before clipping. That was with 500v plate & 55ma biasing. 92ma is max. disapation, 55ma about 60%. Now I'm at 48ma or 52% and that's not taking inconsideration the screen currents or 5.5ma which technically would put me at 46%, 48-5.5=42.5ma.

So actually measuring 55ma at 500v and subtracting screen current had me at 53.8% and that was running my PT a little hot to the touch with no fan on it. I get 520v late at night when the AC creeps up to 125vac.

I'm using a 1 ohm resistor between the cathode & ground.

I've read were others using these tubes or KT88/6550's biased them colder as to not over tax their PT.

Anyone else experienced anything similiar?

[R.G.]

I'm going with Plan "D".....just lower the bias to those KT90's a little as not to run the tranny too hot.

No problem, just mentioning the option.

Not getting all the bang for the buck I wanted for because of a underratted PT but it still gives me enough power to be heard on stage. The problem is the output current isn't enough to drive these properly. My PT felt a bit warm, especially on the bench without the fan I installed inside the combo blowing air over the PT & output tubes. As long as these sound good biased colder then I'm good to go.

If you're willing to do some experimentation, you may get closer to the edge you want. Transformers in general can run hotter than people think. The iron and copper are good up to temperatures way beyond boiling water. It's the insulation that breaks down. There is an old rule of thumb that if you can hold your index fingertip on the transformer core when it's hot, then it's not too hot. The reasoning is that most people will involuntarily pull their fingertip off a metal surface that's hotter than about 130F. That happens to correspond to an internal hot spot temp OK for 105C/Class A insulation, and that's the lowest commercial rating in most cases. You could also measure it with a thermometer or IR temp scanner. So if it's running a bit warm, you might be able to measure it and find that it's probably OK. At least you could make an intelligent guess about how close you were. If the trannie says something like Class 155 or Class B or anything except 105 and/or Class A, it will run even hotter without damage.

If you do the test, you have to let the thing get fully warmed up. This may take a few hours - the thermal mass is quite large on bigger transformers. So one way to do this is to start cold, then take temp readings every half hour or 15 minutes. The temp will follow an exponential rise, and you can calculate the final temp from the rise measurements.

But just playing it safe works too.

[JC1]

No problem, just mentioning the option.


If you're willing to do some experimentation, you may get closer to the edge you want. Transformers in general can run hotter than people think. The iron and copper are good up to temperatures way beyond boiling water. It's the insulation that breaks down. There is an old rule of thumb that if you can hold your index fingertip on the transformer core when it's hot, then it's not too hot. The reasoning is that most people will involuntarily pull their fingertip off a metal surface that's hotter than about 130F. That happens to correspond to an internal hot spot temp OK for 105C/Class A insulation, and that's the lowest commercial rating in most cases. You could also measure it with a thermometer or IR temp scanner. So if it's running a bit warm, you might be able to measure it and find that it's probably OK. At least you could make an intelligent guess about how close you were. If the trannie says something like Class 155 or Class B or anything except 105 and/or Class A, it will run even hotter without damage.

If you do the test, you have to let the thing get fully warmed up. This may take a few hours - the thermal mass is quite large on bigger transformers. So one way to do this is to start cold, then take temp readings every half hour or 15 minutes. The temp will follow an exponential rise, and you can calculate the final temp from the rise measurements.

But just playing it safe works too.

Here's a link to the PT.....Class A insulation (105ºC).

Hammond 274BX Power Transformer
The 270 Series power transformers use the convenient 4 hole above chassis "X" type bell end mounts. Each unit features a 115 VAC, 60 Hz primary and two secondary filament windings, one 5.0V for tube rectifiers, plus a 6.3V for heater filaments. Leads are a minimum of 6" in length. Class A insulation (105ºC). These transformers are conservatively rated and CSA certified. VA: 198, Secondary RMS: 375-0-375, DC mA: 175, Fil. #1 Rectifier: 5.0V/3A CT, Fil. #2 Heater: 6.3V/6.0A CT.

http://www.parts-express.com/pe/show...number=122-235
http://www.hammondmfg.com/pdf/EDB274BX.pdf

One link says 175ma & the other 201ma. I think I have the 201ma but I'll double check when I get home
I have a IR temp scanner so I'll check that out with the fan off. Would you check it out at just idle, or full out stage level?

I wasn't on playing loud until maybe Saturday if I sit in with some friends for a gig.

What's the difference between a X15 & X19?

[R.G.]

[COLOR="#0000CD"]Here's a link to the PT.....Class A insulation (105ºC).
Hammond 274BX Power Transformer

First of all, I'm downright impressed with the info that Hammond has given you. Specifying winding resistance is a huge benefit for what you're about to do.

The way the pros do it is to use the copper windings as a thermometer. Copper has a temperature coefficient of resistance of nominally 0.003862/K. What that means is that for every degree it gets hotter, the resistance rises by a factor of 0.003862. So if you're looking at 105C as being the maximum temperature for long insulation life (that's what Class A/105 means) then you have 105C-35C = 70C rise that you can allow. (I used 35C as a guess at the internal air temp.) A 70C rise means that if the copper got that hot, its resistance would increase by a factor of 70*0.003862 = 0.2703.

The red-red winding resistance is specified at 81.35 ohms +/-2%. So if you measure it cold, you'll find out what your ohmmeter thinks it is, which varies with calibration. Then run it to heat it up. When it's fully hot, working at maximum design temperature, it will rise to 81.35*0.2703 = 103.3 ohms. Again, your ohmmeter will vary a bit, but what you're looking for is the percent increase. You have to disconnect the PT from power and from the circuit to measure this, but if you use a rectifier tube, you can just pull the rectifier tube and measure at the input pins to the rectifier.

There are sources of error in this process that the pros would work to try to minimize, but it's more accurate than measuring the surface temp and relying on the variation of the surface temp. Surface temp is good, resistance change is better.

I have a IR temp scanner so I'll check that out with the fan off. Would you check it out at just idle, or full out stage level?

You should check it at your full-load playing levels; that is, the power output you need it to run. And you need to do this for some hours. Transformers have a long thermal time constant, which means it takes them a long time to heat up to their final level. The way the pros do this is to check the temperature every 15 to 30 minutes, and watch how much it changes with time. It changes a lot at first, then as it gets close to final temperature, changes more slowly. In fact, you can take a number of measurements, enough to detect the tapering-up effect, and mathematically estimate the final temperature without actually getting up to the final temperature.

This same technique works with the IR scanner, but it tells you the estimated final surface temperature. That may well be good enough for you, and it's so much easier to do, that it's probably a good place to start. If you only get to an estimated 120F final, you (probably...) don't need to go the full banana and worry about taking change-of-resistance measurements.

And when you do this all, you'll be vastly better informed than most players about what's going on inside your PT.

[JC1]

It's getting hot, about 120F or so, but with a fan running across it shouldn't be a problem.
It's probably good I didn't install a heavier current provided PT because my OT is for a 50w amp though I sure it can handle more with no problem, it's a Drake 784-139. It's not getting hot or giving me any problems so it should be ok with the higher wattage.


Change of subject,

Right about clipping, I'm getting 162ma on one side of the OT & 154ma the other & 458vdc on the plates using the 8 ohm output into a 10 ohm resistive load (closes I had was a 10 ohm 400 resistor).

458 x .162 = 74w
458 x .154 = 70w

I have a dc bias balancing circuit to get both tubes the same at idle. I also have a 25k balance resistor between the 2 PI plate resistors to balance out the PI ac audio to the output tubes. So, with everything balanced, my OT is slightly out of balance, causing the difference in output current draws.

What I do is unbalance the PI output just enough to equal out the output currents on the OT. I read a lot about guys having issues with this and this idea appears to be the best solution to solve it.

[Rhodesplyr]

A suitably rated zener could be subbed in for the OC3, in the Leslie arrangement for the screen supply. That's probably your simplest option.

Interestingly, current-production Leslies use Zeners instead of 0C3s, and the failure rate for the Zener diodes is much higher than for the regulator tubes, which hardly ever fail. They first used one 100V, 5W Zener, with major problems, then switched to two 50V, 5W Zeners in series, which are more reliable, but still not as reliable as the 0C3.

[Steve Conner]

I have a hi-fi amp that was designed for EL34s, and I run KT88s in it. I had to reduce the idle current below what the tubes are capable of, because the PT was getting dangerously hot. Even so I think it sounds better than it did with the EL34s.

[Bruce / Mission Amps]

I have a hi-fi amp that was designed for EL34s, and I run KT88s in it. I had to reduce the idle current below what the tubes are capable of, because the PT was getting dangerously hot. Even so I think it sounds better than it did with the EL34s.

Steve, did you let it ride or change the OT impedance around for the KT88s?

[Steve Conner]

Nope, didn't change anything. The OTs are 30 watt ultralinear with 16, 8 and 4 ohm taps, but the NFB comes off the 8 ohm tap and I found that no matter what the nominal speaker impedance, they sound best hooked up to that tap. The extra power of the KT88s comes in handy driving 4 ohm speakers off the 8 ohm tap.

[JC1]

Interestingly, current-production Leslies use Zeners instead of 0C3s, and the failure rate for the Zener diodes is much higher than for the regulator tubes, which hardly ever fail. They first used one 100V, 5W Zener, with major problems, then switched to two 50V, 5W Zeners in series, which are more reliable, but still not as reliable as the 0C3.

What is the 0C3s you are refering to & do you have a schematic for the Leslies using zeners?

[pdf64]

Think of them as tube zeners, see http://www.mif.pg.gda.pl/homepages/f...41/g/GL874.pdf and http://www.captain-foldback.com/Lesl...matics/122.GIF
Pete

[JC1]

I have a hi-fi amp that was designed for EL34s, and I run KT88s in it. I had to reduce the idle current below what the tubes are capable of, because the PT was getting dangerously hot. Even so I think it sounds better than it did with the EL34s.

Did you notice any crossover distortion biasing those KT88's colder? Have to calculated what their idling at?

[JC1]

Nope, didn't change anything. The OTs are 30 watt ultralinear with 16, 8 and 4 ohm taps, but the NFB comes off the 8 ohm tap and I found that no matter what the nominal speaker impedance, they sound best hooked up to that tap. The extra power of the KT88s comes in handy driving 4 ohm speakers off the 8 ohm tap.

Are you matching the Z of the speakers to the Z of the OT or mismatching using the KT88's?