Please help with calculations on Output transformer to get ohms l

[Justin Thomas]

What Juan said times a bazillion!
Personally, I am uncomfortable with anything more than my meter will handle, 600V...

Before positing a theory of how one <could> use a battery to check, I should ask a question... the VERY low resistance of an OT secondary would effectively be the same as shorting out a battery, correct? Which would result in a dead battery very quickly at best and a small fire &/or explosion at worst, correct?
Asked as someone who burned himself carrying a 9V in his pocket as a kid...

Justin

Edit: your 9V phone plug is likely to be DC, also - in which case, it will not work on a transformer - trannies need AC to operate.

[Gnobuddy]

Then don't work on tube amps, the power transformers put out more than my 240vAC

There are tube amps, and there are tube amps. My last DIY design uses a 48 V (RMS) Hammond power transformer. Half-wave rectifying that produces about 75 V. My power supply design includes a voltage doubler that produces approximately +150V DC for small-signal pentodes, a voltage tripler that produces about +225V for the little 6AK6 output pentodes, a voltage quadrupler that produces around +300V DC for 12AX7 triodes, and a reverse-polarity voltage doubler fed from the 24 V transformer centre-tap to generate about (-75V DC) for fixed bias experiments.

For the guitar amp output power levels I can actually use, I don't think I really ever need more than 350V B+. Less, if I take advantage of later generation sweep tubes that can drive lots of current at relatively low B+.

Music Man amps with their 700-800v B+ ought to be positively frightening.

To me, they are exactly that - frightening! I am not a certified electrician, and have had no formal training in high voltage safety. IMO it would be foolhardy and irresponsible for me to mess with 800V power - I'm simply not qualified. I hate to quote a violent Clint Eastwood character, but "A man's got to know his limitations!"

I have been inside my Fender Princeton Reverb, with 440V DC on the output anodes. I wore my Salisbury Class 0 electricians gloves and eye protection for that job. That is probably as high a voltage as I plan to deal with. I'd like the chance to grow old with my wife.

Those of you who routinely deal with those scary-high voltages have my respect, and my best wishes for your continued safety!

-Gnobuddy

[SoulFetish]

I know it's easy enough to use and 6V or another convenient voltage at an available tap and figure out what the turns ratio is with very little math...
BUT, how difficult would it be to build a 1V p-p, 50/60Hz(or whatever) oscillator feeding a source follower for driving low impedance windings. It would give you quick, simple, consistent results, and should prevent any unexpected high voltages appearing on windings of unknown transformers. Okay, so you would need a over 1V at the gate for 1V to appear at the source, but you see what I'm getting at.

[Gnobuddy]

oscillator feeding a source follower for driving low impedance windings.

Interesting idea, if it's a measurement you do often. It would probably be simplest to use a small, low-power amp IC (or module) to drive the transformer. Run it on a couple of AA cells to keep output voltages nice and low.

I'm not a big fan of the source-follower idea, because single-ended source followers are better at sourcing current than sinking it, and push-pull source followers have massive crossover distortion until you add sophisticated biasing and lots of negative feedback - at which point you basically have a small power amp!

But I think Malcolm Irving was entirely right (post #7). Even a cheap DMM can measure small AC voltages quite accurately, as long as you keep the frequency reasonably low (no more than 100 Hz, say.) A dedicated transformer ratio tester may be overkill for most of us.

-Gnobuddy

[Dave H]

I know it's easy enough to use and 6V or another convenient voltage at an available tap and figure out what the turns ratio is with very little math...
BUT, how difficult would it be to build a 1V p-p, 50/60Hz(or whatever) oscillator feeding a source follower for driving low impedance windings.

A 600 ohm output sig gen may be enough to drive it. If the OT is unloaded its impedance at 1kHz (say) won't be all that low.

[Enzo]

Ever hear a cook tell someone a dull knife is more dangerous than a sharp one? It is true.

I appreciate the desire for safety, but if I started wearing gloves, I doubt I could hold my meter probes to accurately probe my circuit. Slipping and shorting out a 400v supply will blow up a rectifier as easily as it would on a 600v circuit. Safety is safety, I use teh same skills to avaoid youching 600v that I use to avoid touching 400v, or 150v. I don't like arbitrary limits, as I might convince myself this circuit ONLY has 380v in it so I am safe, while the 450v circuit over there is dangerous. They are all dangerous, the 120v from the wall outlet can kill you. You learn to not stick your fingers on it... or you die anyway.

Frankly I am more likely to die driving to the grocery store than grabbing B+.

[SoulFetish]

Interesting idea, if it's a measurement you do often. It would probably be simplest to use a small, low-power amp IC (or module) to drive the transformer. Run it on a couple of AA cells to keep output voltages nice and low.

I'm not a big fan of the source-follower idea, because single-ended source followers are better at sourcing current than sinking it, and push-pull source followers have massive crossover distortion until you add sophisticated biasing and lots of negative feedback - at which point you basically have a small power amp!

But I think Malcolm Irving was entirely right (post #7). Even a cheap DMM can measure small AC voltages quite accurately, as long as you keep the frequency reasonably low (no more than 100 Hz, say.) A dedicated transformer ratio tester may be overkill for most of us.

-Gnobuddy

Yeah, your probably right. An opamp should give me a low enough output impedance anyway. Im not trying to drive a loaded secondary. I was just throwing out ideas. But a small DIY stand-alone module with clip leads would be quick and easy. Actually, if you could combine it with an inductance tester, it would be super handy to get a sense of bandwidth on an output tranny.

[Tom Phillips]

A 600 ohm output sig gen may be enough to drive it. If the OT is unloaded its impedance at 1kHz (say) won't be all that low.

That method has always worked fine for me. It does load down the oscillator but you just mesasure the secondary & primary voltages and do the math. No problem.

[pdf64]

So as to account for the magnetizing current, Chris Merren advises to use a big test signal, ideally measure it in the amp with a resistive load, see Vintage Amps Bulletin Board ? View topic - Mysterious Marshall JTM45 amp

[Tom Phillips]

So as to account for the magnetizing current, Chris Merren advises to use a big test signal, ideally measure it in the amp with a resistive load, see Vintage Amps Bulletin Board ? View topic - Mysterious Marshall JTM45 amp

That does not make sense to me. I can believe that the power transfer efficiency would be affected during a low level signal test. However, I do not see how the turns ratio measurement, which is what we are discussing, would be significantly affected.

[R.G.]

over

I think maybe we ought to do a sticky that amounts to a mini-course on the simplified model of transformers, and require people to complete it before offering opinions about transformers. A lot of this stuff just falls out with that as background. OK, I'm only half serious.

The simplified model of transformers has an ideal transformer in the middle. It can transform any frequency at >>exactly<< the turns ratio from primary to secondary. On the primary side, there is a nonlinear resistor, which models the core loss, and which may be ignored for most situations until you're figuring things like insertion loss and heating. There is a primary inductor across the primary, modelling - the primary inductance. This can be linear or non-linear, depending on how persnickety you are being on flux density and such. There are resistors in the primary lead and secondary lead, modeling ... yep, you guessed, wire resistance. If you're playing with high frequencies, you need to include small inductors in series with the primary and/or secondary leads to model leakage inductance. An alternative to the leakage inductor view is to model the ideal transformer with a coupling coefficient, but that's complicated for beginners. That's pretty much it for the simple tests, which, following the 80:20 rule, tell you 80% of everything about the transformer with 20% of the effort. (Note that finding the other 20% of what there is to know takes the second 80% of the effort.)

If you just want to know the turns ratio or phasing, you do and Open Circuit Test, the capitalization reflecting what the transformer guys think of it, an official test. You put an input on the primary and measure the volts on the secondary, setting this up so that as much of the complications of the nonlinear primary loss, primary inductance, wire resistance and leakage inductance is avoided as is reasonably possible, realizing that it's not perfect, but you might get 95% of your information with 1% of the effort you'd need for fancier tests.

So - how do you make the leakage inductance, resistances, primary inductance and core losses be as negligible as possible? First those pesky resistances. So don't let currents flow. Leave the secondary open. No current flows in the secondary to muck up your measurements. Some current will flow in the primary, determined by the primary inductance and core losses. That current will drop some voltage through the wire resistance and leakage inductance. To minimize the error, you use a frequency that's high enough to not let much primary current flow through the primary inductance.

You don't go mucking around with the lowest frequencies you can put through it. You'd like not to have the leakage inductance messing things up, but you can recognize that the leakage inductance is hundreds, if not thousands or tens of thousands of times smaller than the primary inductance in an OT, and ignore it. A frequency that's comfortably larger than the lowest frequency for which the primary is designed is great. A few hundred Hz will really do it up fine.

Note that if you're after easy and not exact, you can use 60Hz. It's below the nominal frequency of guitars, but most guitar OTs will do OK there. So you're a few percent off - you'll muck up the measurements by that much, so 60Hz is probably OK.

The voltage needs to be figured in. you want a voltage that's (1) large enough to move the core flux out of the small-loop doldrums, (2) small enough to stay well away from any fold-over from the beginnings of saturation, and (3) easy to measure well with what instruments you have. Working the core flux somewhere in the middle is good. You're part way there by using a frequency that's not rock-bottom. Using a voltage that is, as my grandmother used to say, "middlin" will keep you away from saturation too. Staying away from max flux also cuts the core losses a lot.

Perhaps the biggest issue is "easy to measure". You can easily muck up the measurements by having to change scale on your meter if it's not well calibrated. And let's face it - none of us ever get our meters calibrated. I don't, and I know the problem. So, pick a voltage that's about mid scale for the low/secondary side, and significantly bigger than the 200mV that's the usual internal range for the circuits inside cheap meters. I like to use a few volts. Gets the measurements up out of the mud. You want the measurement to use at least two and preferably three digits on the meter, as there is >>always<< a +/- one count error on any digital multimeter, no matter how good. Then measure the secondary voltage, change scale, and measure the primary voltage, and do the math.

Intelligent choices minimize the easiest errors, and get the remaining errors down in the noise. It takes a lot more effort to do a better job.

Putting a resistor load on the secondary makes things worse if what you're after is the turns ratio. It adds current in the secondary and primary and therefore adds error to the measurements by causing the primary voltage to be the sum of the turns ratio voltage, but dropping that by the primary current times the primary resistance. There's always some error there, but why add to it by deliberately adding secondary loading? There are other tests where secondary loading is needed - notably the Short Circuit Test - but that's not what this question is about.