It is a little better than the average sound card, in that it can measure with the scope out a little higher in frequency before it runs out of steam, but a little worse in that the scope and signal generator use lower bit depth, 8 bit for the signal generator, and 11 bit for the scope. Offsets are built in to zero/center calibrate the voltage on the scope, and to calibrate frequency and other things. I think the average sound card has only 192k sample rate, whereas this has 2mhz sample rate, which would give better data in some regards. Seems like the designer made some wise choices to maximize usefulness.
Does low signal level give rosier data for higher frequency levels in addition to lower? one thing a person could do is build an amp, and then compare data driving the inputs and measuring the response across secondary load, and get an idea of the difference between the response of a number of different output transformers. The graphs can be saved and exported. I’m not sure if they are exported as data or as images, however.
Does low signal level give rosier data for higher frequency levels in addition to lower? one thing a person could do is build an amp, and then compare data driving the inputs and measuring the response across secondary load, and get an idea of the difference between the response of a number of different output transformers. The graphs can be saved and exported. I’m not sure if they are exported as data or as images, however.
Bit depth times sampling rate is information rate and can be transformed reversibly as long as everything fits within the Nyquist sampling rate criteria. (Sorry for the plural - somewhere's AI objects to the singular). Transformer testing stretches the bounds of even modern soundcards pretty tight. I'd actually meant that more as conceptually, so I could get a better handle on the machine.
The matter of frequency magnitude response of iron-core transformers is so often mistaken that I hesitate to even comment - without very carefully and boringly defining everything that one says, it's easy to be misleading, or worse. The most primitive description of electricity vs. magnetic flux is almost always simplified to a monotonic B/H curve that doesn't really actually occur in iron cores, and things progress from there. But I digress. The map is not the world.
At "high" audio frequencies the core is less important, so transformer level is less important. The valves work into increasingly difficult reactive loads with higher frequency (and lower frequency!), and that's measurable. Your plan of measuring an actual working machine is, of course, the Gold Standard.
All good fortune,
Chris
The matter of frequency magnitude response of iron-core transformers is so often mistaken that I hesitate to even comment - without very carefully and boringly defining everything that one says, it's easy to be misleading, or worse. The most primitive description of electricity vs. magnetic flux is almost always simplified to a monotonic B/H curve that doesn't really actually occur in iron cores, and things progress from there. But I digress. The map is not the world.
At "high" audio frequencies the core is less important, so transformer level is less important. The valves work into increasingly difficult reactive loads with higher frequency (and lower frequency!), and that's measurable. Your plan of measuring an actual working machine is, of course, the Gold Standard.
All good fortune,
Chris
Pardon me please, guys, I have to stand corrected. Obviously I've had a very heavy fellacy in thinking. My reasoning for the given example is now: Neglecting screen current, both primaries are driven by the same current, due to Kirchhoff's law. As both secondaries are in parallel, the same output voltage applies to both. The one with half the primary impedance than the other will deliver sqrt 2 times the output current than the other. Hence, the new output current calculates to 1 + sqrt2, and the new output impedance is 1/(1 + sqrt2) = 0.414 of what is written onto the xformers.
As yet said, due to the different AC voltages at the cathodes and the plates Norman Crowhurst's ideas of cross-coupling the cathodes and plates via capacitors won't be applicable. With the lower impedance tranny at the cathodes and constant screen voltages the design rather resembles ultralinear - and what Quad did in their Quad II.
Best regards!
I got a box full of them from Stan (ESRC) several years ago. The data sheet I have is Raytheon. Most are indeed Raytheon / LYFA tubes but a few are Japanese. They are interchangeable in a new design amp if the socket is wired correctly. Pin 2 of the socket must be connected to pin 4 and / or pin 10. The Raytheon versions will indeed go to 125 WPC. I never tested the Japanese tubes. The 6HJ5 is a 6DQ5 in a compactron bottle without a plate cap.