unbypassed cathode and frequency response

[nickb]

There is nothing in between grid stop was bypassed and signal applied straight to the grid. It is just one triode connected we don.t talk how this triode works in some context, just by a measuring error generate by who knows reasons. The main suspect are the cheap 1× probes delivered with the scope. I ordered some decent 10x probes expected to come tomorrow. Thanks for Your answers.

In the meantime put a probe on the grid and see what the frequency response is.

[catalin gramada]

Is there a reason why you don't show a schematic?

There is nothing to show, but if it helps, this is what I measured 5 min ago with a cheap Fluke meter, just for comparison the results in "RMS" average values are more close to reality than my scope did. Measured at 400cps and 10kcps.

[bea]

Which means that You must have a massive capacitive load in Your scope/probe setup You should check for.

[J M Fahey]

Well, originally you had -5.5dB loss ; but now, measuring with a Multimeter which is not really an Audio measuring instrument but more of an Electrician´s tool (meaning it won´t exactly shine measuring 10kHz, if at all), you have only 1.1dB loss?

Considering this is being measured by an inadequate instrument (which might justify the 1.1dB loss displayed) , it maks me think there was some error in the first determination.

But please repeat it using the test instruments you have today , you might have a surprise.

[catalin gramada]

I work in a truck all week, next Saturday will be. Till then new 10x probes will come, maybe not as crappy as ones allready have.Thanks for support.

[J M Fahey]

Ok.

I suggest you build an RF Probe.

The 40´s to 60´s way to have a DC meter measure *FLAT* up to several Mega Hertz , and work perfectly well with modern meters, including dollar store $10 ones , since all measure DC very well and this turns AC , any frequency, into DC.





A couple points:

* original one uses a germanium diode to measure feeble values, maybe down to 10mV or so; you can use a silicon diode, preferably a "signal" type such as 1N4148 or 1N914 but here we don´t need exact value, just a comparison, and this by itself will measure same at 400Hz than at 10kHz (or 100kHz for that matter).

So we only compare "whatever it puts out" at 400Hz to whatever it puts out at 10 kHz and we are done.

* using silicon will make it somewhat deaf at low levels, since silicon has a threshold of almost half a volt , no big deal if we are measuring some 4 or 5V RMS so about 7V peak.

* since we are Audio guys, I would increase input capacitor to .047 or .1uF

No , we are not grossly loading the tube plate with it, diode will charge it to signal peak value in a few milliseconds and then will stop pulling current from source, so it will be effectively an open circuit

Or to be more precise, tube plate will "see the "separation" resistor shown in the schematic.

Again, old timers used 4M7 to "convert peak reading to RMS" into their 10M input impedance VTVM´s ; we can use 1M because our meters are not *that* high input impedance but in any case we are getting a comparison voltage.

If 400Hz and 10kHz both show, say, 5V DC (which would be somewhere between Peak and RMS), so be it, we are satisfied that it shows flatness (or lack of it).

* RF probes are built straight at the probe itself, to minimize capacitive load; once we got AC turned into DC cable can be any length; only precaution is to use shielded cable (even guitar quality), since it *is* a high impedance DC signal, and prone to getting inwanted outside interference.

This is a good way to make them, some build it inside a plastic BIC pen body or similar:





I bet you have everything needed (if you are flexible with diode choice, that is) and have it working in less than an hour