AC * 1.4 = DC ???

[sliberty]

OK, I know it is not that simple. Recto tubes lose 20-50 V, depending upon the type. Yada Yada Yada. But I've always understood that when using dioides, you could approximate the rectified voltage by multiplying the unrecitified voltage by 1.4. And for many amps that I have built, this has been reasonably close.....but not always.....

I built an Express some time ago (my second), and used Moose's PT. 300-0-300V should have given me 420 or maybe 400 on the plates, but i was 50-60V lower than that. I never did figure that one out. That Express still isn't running at the correct voltages.

Now, I have a KT66 Rocktal build, (solid state rectification), and again, I am seeing much lower B+ than expected. I get 353V on the plates. 375V with the power tubes pulled. This build uses a Hammond 272JX, which is again 300-0-300V, so the plate voltage should be 400-420V I would think.

I replaced the diodes, but, just like with my Express, this made no difference.

What part of all of this am I missing?

Thanks,
Steve

[Colossal]

Steve did you actually measure the PT unloaded on the bench to see if it was really putting out 300-0-300VAC? Also, are you using a bleeder resistor on the reservoir cap? If so, could it's value be off by an order of magnitude, so, 22k, instead of say 220k?

[sliberty]

Thanks for looking.

My bleeders are a pair of 100K's in series, but just for the hell of it, I measured them to be sure that they are OK - 197K total.

The unloaded transformer is giving me about 585 (approx 292-0-293). Hammond configured it for 115V and 125V primaries, and I am using the 125V primary (either way, I'll be ~5V off given my typical 120VAC supply). Today, I measured my input voltage at about 119VAC, so it seems about right that the total secondary would be a little lower than 600V. But this doesn't explain the significantly low B+.

[LeftyStrat]

Actually that sounds about right to me. The 1.414 figure is for unloaded. Under load 1.2 is usually ballpark. So 293*1.2 = 352.

[roberto]

Yeah, it depends also on bias.
I've seen some PT builders indicate voltages at two or three biases.
It's useful to better understand how the PT reacts.

[sliberty]

The unloaded voltage can't depend on bias. And it is 375 vs 410 (calculated value).

[roberto]

My answer was referred to LeftyStrat's post.

[sliberty]

Ok, so going back to that...

1.2 * 293 = 351.6 (loaded)

But

1.4 * 293 = 410.2 (unloaded)

So, my loaded voltages are in line with this, but my unloaded voltages are at 375. Still a bit baffled.

BTW:This 1.2 multiple is new to me. I've never seen that before.

[roberto]

This is what I was talking about.
1.2 is pure empirical, it depends on bias and on PT HV current.

Have you ever checked PT AC voltages unloaded,
or you simply look at trafo specs?

[David Root]

Try connecting it to the 115V primary tap. That will give you a bit more juice.

[Cliff Schecht]

Try connecting it to the 115V primary tap. That will give you a bit more juice.

He's gonna cook his heaters!

[LeftyStrat]

I didn't mean it as absolute. Merlin's PS book gives a similar rule of thumb.

Obviously the capacity of the PT and the load it is subjected to can vary this quite a bit.

But if the loaded B+ is around 1.2*AC It's not something I would worry about.

Also, you really need to pull all the tubes to get the true unloaded B+.

[Cliff Schecht]

Ahh... This friggin rule. David Root and myself went back and forth a few times in email trying to figure out if a bias tap he had would provide enough negative voltage for the amp he was working on. After making a few dumb and/or wrong guesses, I said screw it and actually sat down to run some quick simulations. It was only then that I was able to give him a definite answer and I felt bad for putting him through my confused thought process. Even when running a single diode (half-bridge) rectifier the voltage is multiplied by 1.414 but for some reason I kept thinking because the rectifier is skipping every other cycle it would only be 0.707 multiplier.

It's the simple rules that are the hardest to not only remember but apply correctly. I'll be the first to admit that Ohm's law is deceptively simple as well as things like AC->DC rectification. Any EE can tell you the 1.414 is derived from the square root of 2 but when you ask them to sit down and calculate what the expected voltage from a rectifier/cap feeding a loaded transformer will be, I'll bet a good 80% of those EE's will get their calculation wrong the first time, unless they know how to predict what the loaded vs. unloaded voltages will be (I'd even be nice and assume the input voltage from the wall matches what the primary is rated at ).

Magnetics is as much of an art as it is a science and aside from the fields classes people take (more focused on EM-waves at RF frequencies), there is very little exposure to real magnetics design/application in most EE curriculums. My knowledge in this field has put me ahead of my piers in the past when applying for jobs and such, especially when other students/applicants don't even know that one can specialize in just magnetics designing (not that I do..yet).

[ToneMerc]

Now, I have a KT66 Rocktal build, (solid state rectification), and again, I am seeing much lower B+ than expected. I get 353V on the plates. 375V with the power tubes pulled. This build uses a Hammond 272JX, which is again 300-0-300V, so the plate voltage should be 400-420V I would think.

I replaced the diodes, but, just like with my Express, this made no difference.

What part of all of this am I missing?

Thanks,
Steve

Steve, are you by chance aren't running GT HP KT66's?

TM

[sliberty]

No, Shuguang