1977 Marshall 2203 High Plate Voltage!!

[bob p]

I always thought that the proper way to avoid voltage drops was to use low resistance wire with a gauge of adequate ampacity, rather than to overvolt at the supply end. I guess I'm fortunate that I live in an area where we've never been subjected to brownouts or rolling blackouts, and I've never seen my wall voltage sag. Maybe the fact that my distribution transformer is on a pole across the street, and that the run to my house is only about 75 feet is working in my favor. And then there's the fact that the guy who built my house used 10ga wire everywhere when he really didn't need to go that big.

My local electric utility has started a voluntary power program to avoid the problem of synchronous demand by summer air conditioners. They've deployed centrally controlled "smart meters" that are capable of providing a rolling blackout to air conditioners in peoples' homes. They only take $10 per month off of your bill if you sign up for the service, so I haven't bothered.

Going back to transmission losses -- the voltages that are used in power transmission are huge -- as high as 1.38 million volts. The reason that they use extreme voltage in power transmission is so that they don't have to transmit power by transmitting large amounts of current. The process of using high voltage to minimize the amount of current being transmitted helps to avoid resistance-induced power losses. On the transmission side, I don't think that voltage drop is such a big deal. I think that's primarily a problem and the end of the run, after the transmission voltages get reduced to distribution voltages during the "last mile."

In rural applications, the transmission system uses autotransformers that have automatic tap-switching equipment built-in, which effectively causes the transformer to act as a voltage regulator. This type of variable ratio autotransformer serves to compensate for voltage drop on the line.

Like Enzo said, most of the equipment that we're going to plug to our wall socket is going to be able to tolerate a reasonable amount of over-voltage. It doesn't care if the mains is high. In most cases it doesn't care if the mains is a bit on the low side either. Which brings us back to the question -- what was the point of changing the voltage spec at the point of service from 110 to 125?

So far everything I've read about power distribution makes it sound like there's no benefit at the transmission side. Because the use of electrical appliances in the home has proliferated so much in the past 50 years I'm wondering if the reason that the voltages have been increased by 125/110 = 14% is to allow customers to have more power for their appliances without having to draw more current. I suppose the the pessimist might say that the real reason that they do it is because the KWh meter connected to your service drop charges you for power, not voltage, and pushing the voltage up by 14% also increases your bill by 14%.

Keeping the hijack going, I stumbled upon an interesting DOE report about Large Power Transformers used for powering the grid. Some of them weigh 650,000 lb! Wow!

http://www.energy.gov/sites/prod/fil...e%202012_0.pdf

[bob p]

Who wired this amp? There are a few things in this picture that look funny to me.

[Mick Bailey]

I've never seen a 2203 with a transformer like that. But then again, I only ever see UK models. Is that correct?

[cluster]

Who wired this amp? There are a few things in this picture that look funny to me.

looks stock to me?? what do you see?

[g1]

Those transformers are used in Canadian models.
T3490 and maybe the OT is T4353 ?

[cluster]

Ok.. so I opened up the amp today and took some measurements. here is what I got (tubes in).

Wall = 120VAC
heaters = 6.3 VAC
plates = 534 VDC
screens = 520 VDC
grid bias = =51.7 VDC

Still consider the plate/screen voltage high but maybe someone can chime in??

Question: I have about bias method on this amp. Since it's vintage, it doesn't have the 1 ohm resistors to ground on the power tubes. Usually, I install these resistors to make it easy to bias but not to this amp. Second approach I use is a weber bias rite, but the one I have is not calibrated properly and is giving me inaccurate readings so I can't use that. So, I was going to try the "Shunt" method but I want to clarify how it's done. I have a really good Fluke (1000V rated) to use. Could someone here explain to me how to properly bias an amp using this method? If there is another method that I'm missing, please let me know. I don't have a scope, so that method is out. Any suggestions would be appreciated.

Thx

[Chuck H]

Over 50V bias at 500ishV is definitely going to be colder than typical (not wrong, just not typical, so, higher plate voltage reading). Why not just install a 1R resistor? g1 is cozy with the OT shunt method and I think it's a good way to go for a couple of reasons. First, the resistance of the OT primary is higher than 1R (and therefor easier to get an accurate reading) and second, it isolates plate dissipation, which is what the actual datasheets are trying to tell us. Though it's common for novice player/techs to consider the total tube dissipation and relate it to plate dissipation, that's not accurate.

You basically just set your meter for current measurement and short the probes across the OT primary CT and either plate. The considerably higher resistance of the meter should keep things in check.

[g1]

Another way to do it is measure your OT primary resistances (each half) then measure the DC across them and calculate the current.
It's safer than the shunt method.

Like Chuck said, the shunt method is a current measurement from the OT primary CT to each plate.
The danger is that in current mode, your meter probes become like a straight piece of wire. If you forget you were in current mode and try to do a voltage measurement somewhere else (like checking plate voltage) you will do serious damage. That is usually where people run in to trouble with shunt method; not while they are doing the shunt measurement, but after, when they forget to put the probes back into the correct meter ports, and change the meter out of current range.

[cluster]

Forgot to ask. In both methods, are the measurements to be taken with the tubes in or out?

[cluster]

Took some more readings.

With Tubes in:

resistance across CT/Left winding = 15.8 ohms (amp off)
voltage across CT/Left winding = 1.18Volts (amp on)
current = 1.18/15.8 = 0.0747 Amps

resistance across CT/Right winding = 14.4 ohms (amp off)
voltage across CT/Right winding = 0.97 Volts (amp on)
current = 0.97/14.4 = 0.0674 Amps

[pdf64]

Tubes in; if out, they're not drawing current and so trying to measure their current draw would be futile

I don't think we should be advocating the current shunt method, as it's needlessly hazardous, and less accurate than the OT voltage drop/resistance method, the only benefit being an instant current reading.
A competent person might choose to play fast and loose with their own safety, but they are fully aware of the hazard and what to expect.
It seems a very bad idea to suggest it to a noob

[J M Fahey]

I thought that everywhere in NA they kept the grid at the same voltage, so regions could exchange power when needed.

No, voltage is irrelevant, you can change that at will with transformers.
And anyway Country wide nets work at tens to hundreds thousand Volts, absolutely unrelated to house wall voltage.

What is important is frequency and phase.

[Chuck H]

Well that looks like a typical amount of current. Though I've never seen EL34's biased at greater than (less than?) -50V even with high-ish plate volts. Three things...

It's really important to make sure the power supply is fully discharged when you take resistance readings for the primary halves because any voltage there can skew the reading. It's been my experience that even with bleed resistors there can be a small voltage standing on the caps for a long time. I usually let the amp discharge through it's bleeders for several minutes and then connect an alligator clip lead from the main filter + lead to ground and leave it there for the measurements (but don't forget to disconnect it before firing the amp up again!!!). There will be lower resistance path to ground across the winding in the circuit this way, but it's still not significant to skew the calculation.

Be aware of your meters "null" figures. These are relatively small measures you're testing. Small enough that whatever reading your meter stands at when not connected to anything can throw off the calculation. So in each measurement you need to touch the probes together and then subtract that reading from your measurement.

And the imbalance you have is greater than what I would expect. If taking new resistance and voltage readings as above doesn't show improvement I would probably test the two tubes on the .0747A side to see if one is drawing significantly more current than the other.

[J M Fahey]

I find *BOTH* primary side current methods DANGEROUS, shunt or not.

Safe method is to "waste" 2 minutes and add 1 ohm cathode resistors, then measure voltage drop across them.

Added advantage: Math is easier !!!!

Use the mV scale and treat it as if it were a mA one ... same reading "number".

[cluster]

Well that looks like a typical amount of current. Though I've never seen EL34's biased at greater than (less than?) -50V even with high-ish plate volts. Three things...

It's really important to make sure the power supply is fully discharged when you take resistance readings for the primary halves because any voltage there can skew the reading. It's been my experience that even with bleed resistors there can be a small voltage standing on the caps for a long time. I usually let the amp discharge through it's bleeders for several minutes and then connect an alligator clip lead from the main filter + lead to ground and leave it there for the measurements (but don't forget to disconnect it before firing the amp up again!!!). There will be lower resistance path to ground across the winding in the circuit this way, but it's still not significant to skew the calculation.

Be aware of your meters "null" figures. These are relatively small measures you're testing. Small enough that whatever reading your meter stands at when not connected to anything can throw off the calculation. So in each measurement you need to touch the probes together and then subtract that reading from your measurement.

And the imbalance you have is greater than what I would expect. If taking new resistance and voltage readings as above doesn't show improvement I would probably test the two tubes on the .0747A side to see if one is drawing significantly more current than the other.

Thanks for the response. One thing I did notice on the board (bias section) is that this amp has a 65k bias range resistor instead of the typical 47k in that spot. hence this might be why I can't get lower than 41MVDC per tube. Possible that this amp's bias section wasn't properly converted to EL34? here is a photo of that section.