Not much, this was just easier to post a link to. I very well may use his, but I will need more gain with it, the input has 17 to 1 capacitive divider. I'll end up trying both and see how it shakes out.
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
Looks like 1.7pf.
It only matters to keep the calibrated capacitors aligned. But, how will I know if it is off by 1.5pf. I do have a couple of Q standards I can verify with, in good condition but still 50 years old.
Thanks, Mikek
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
So am I, now that i see that the output is the source. ;-)
Gosh, it's going to be hard to find a jfet that bad. Maybe add a source resistor to kill some transconductance. That would bootstrap Cgs, too.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
C302 is the key component to make the follower to rectify the input signal. It keeps the cathode voltage up when the grid goes negative. The circuit was known as cathode detector or infinite impedance detector.
To the OP: Any of the more modern detectors can be made to work with much better precision than a hand-picked tube.
The proper way with modern chips would be a log amplifier and a simple rectifier (if not included in the amplifier).
-- -TV
I think they selected for low grid current because of the 100Meg grid resistor - but that resistor is only needed for floating measurements, nearly all simple Q measurements will have the resistor shorted because the grid is connected to ground by the coil under test. This means that you can probably get by with plenty other tubes instead. Ebay lists some
2A6's at not insane prices and if you are prepared to recalibrate then an even wider choice of tubes. I think a 6AU6 pentode connected as triode (g2 and g3 to plate) was used in one Q-meter in exactly the same kind of circuit as your Boonton.Also I don't think you need worry too much about grid capacitance - any sensible tube (or fet) is going to be a few pF which is about the same magnitude as the wiring strays - and anyway it is right across that tuning capacitor and so is tuned out.
piglet
That schematic is a bit different for determining the bias voltage. The
22 kohm resistor (R301) because the other side of the meter chain appears to use R111 and R312 to set the current in the meter to zero! Using the 22 kohm resistance in the cathode leg and the curves in the link I provided give closer to 3.3 volts bias on the cathode. The 100 Megohm resistor R201 DC grounds the grid so it is 3.3 volts more negative than the cathode, much like biasing a FET.I would speculate that R310 is used to set the scale factor using a coil of known Q.
The grid is always negative with respect to the cathode. It has to be some volts more negative before it cuts off cathode current. With a 283 volt plate and a 3.3 volt cathode there is about 0.4 volts of room before the current stops changing... so yes, I guess it is rectifying a
4 Vpp signal, but it won't be terribly linear in the DC produced vs such a small AC signal. Much better than directly using a tube diode.To size the bias resistor in the source leg look at the curves for the FET and pick a drain-source operating voltage on the X-axis, then draw a load line that gives you a current you like at the 3.3 volts the rest of the Boonton is designed for. To set the drain resistor use this same current and the voltage that needs to be dropped.
I eyeballed the cathode resistor load line because the zero plate voltage current was far off the page. Are you familiar with load lines? It is a straight line connecting the points of all voltage on the plate/drain (and no current through the resistor) to all current through the resistor (and no voltage on the plate/drain).
BTW, your simplified schematics would have been fine if you had drawn the 49.9 kohm resistor being terminated to a fixed voltage rather than ground.
To verify I am doing this right, set up your Boonton and zero the meter. Then measure the voltage on the zero adjust and also on the 22 kohm resistor R301. It should be around 3.3 volts ballpark and both the same (zero current in the meter, right?)
-- Rick C
I may have complicated things by trying to make them simpler, I'll post a link to the full schematic of the Q meter section. It includes switches for Low Q, Q change, calibration and zeroing adjustment.
No, scaling for Q ranges is done by adjusting the drive voltage. You must do a math problem to get Q, unless your drive voltage is set to 20 millivolts that is 1X. If it is set to 10 millivolts it's 2 times the meter reading, 5 millivolts is 4 times the meter reading, etc.
I'll try to do that this evening. Mikek
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
4
itions.
f F r ndepth on this one. If correct, your 2 requirements are inherently contradi ctory.
It's quite different to how I understand it. The circuit uses a grid leak. The grid leak resistor causes the buildup of electrons on the grid to bias it negatively. This is done to push a tube into nonlinearity so it rectifie s as well as amplifies. But it now works nonlinearly, so the meter must be marked nonlinearly.
I'm with Bitrex in that designing a new instrument would be easier.
NT
No, the grid is DC grounded by the inductor under test.
The 100M resistor probably keeps the meter from going bonkers when the coil isn't installed.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
The grid leak resistor causes the buildup of electrons on the grid to bias it negatively.
This is done to push a tube into nonlinearity so it rectifies as well as amplifies.
But it now works nonlinearly, so the meter must be marked nonlinearly.
The meter is marker nonlinearly.
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
As John said, the grid is DC grounded through the coil under test and it would appear R202 labeled 0.02 (ohms, Mohms?)
Grid lead biasing was used with tubes that had directly heated cathodes as it was harder to bias the cathode. But this design uses indirectly heated cathodes and so biases the cathode with a resistor. But the value barely allows any current to flow so when the grid swings more negative the tube is cutoff and when more positive the current flows more, rectification. The cutoff is not sharp though, so the DC voltage produced is not linearly related to the AC signal applied.
-- Rick C
The circuit is a peak detector. C302 catches the peaks. Mike could use a diode for the peak detector, and a diode-compensated reference voltage on the other side of the meter.
The best FET I see for "What's All This FET-buffer Stuff, Anyway?" is still any one from National's Process 50. Those have the lowest input and feedback capacitance among FETs with decent forward transconductance.
The J309 is a reasonable alternative, Process 92, slightly higher-capacitance.
You can see the Process 50 family members on page 3-3 of the old handbook (pg. 49 of the linked .PDF with text)
Kleijer's BF256C is in that list of Process 50 parts, as is National's
2n4416A.Cheers, James Arthur
A diode would work if its capacitance was very low, but without gain it might kill the Q he wants to measure.
I'd use an oscilloscope.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Ok, I zero'd the meter and I have 2.66 volts on the low side. ie, the zero function is holding the low side of the meter up at 2.66 Volts.
I have ran some tests with a coil at different Drive levels (Injection Voltage) I'll refer to this as I.V. As I increase the injection voltage, the voltage on the low side of the meter increases. With 3mV I.V. the low side of the meter is at 3.36 volts.- With 5mV I.V. the low side of the meter is at 4.2 volts. With 7mV I.V. the low side of the meter is at 5.03 volts. With 9mV I.V. the low side of the meter is at 5.86 volts. With 11mV I.V. the low side of the meter is at 6.75 volts.
I'm sure this is because the output of tube cathode is driving more current through the Zero adjust circuitry and increasing the low side voltage.
Here's the voltage on PIN 5 of the tube (cathode) Measuring a coil with a Q of about 418.
With 3mV I.V. The cathode voltage is 3.44 With 5mV I.V. The cathode voltage is 4.36 With 7mV I.V. The cathode voltage is 5.31 With 9mV I.V. The cathode voltage is 6.22 With 11mV I.V. The cathode voltage is 7.23
With the Q meter at F.S. the voltage on PIN 5 of the tube (cathode) is
7.7 volts. I.V. is 11.94mV.Mikek
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
I was assuming he'd put his FET buffer in front of the diode.
Surface mount or through-hole?
Grins, James Arthur
Is the cathode also at 2.66 volts with zero signal like the "low side" of the meter? In fact, every point between the cathode and the wiper of R311 should be the same voltage, right?
Just curious, what are the DC voltages on the two sides of the plate resistor R303?
I assume that once you set the zero the operating instructions are to use a reference coil and some value of input voltage to adjust the reading with R310?
Yes, a higher voltage on the cathode will give you a higher voltage on the low side of the meter. There are still series resistances between the meter and the zero pot. Even if you measure on the wiper arm of the zero adjust pot, that voltage divider has a Thevenin equivalent of a voltage source and a series resistor. So there will always be some change in the wiper arm voltage as the circuit operates. This is factored into the series resistances used. The meter responds to the current which will scale with the cathode voltage minus the reference voltage.
You could try throwing those numbers into a spread sheet to get the difference between cathode and low side voltage (this is proportional to the meter current). Compare this voltage to the input voltage by dividing and see if it varies much. If that ratio is pretty constant the circuit is pretty linear.
Why can't you use your Kleijer amp for this? It will give you a really great no load on the circuit and it will be easy to construct an ideal diode circuit to measure the output amplitude. You might want to steal the thermocouple unit to measure the input if you don't want to use an oscope. All of this can be used to feed a simple MCU board and give easy to read measurements.
-- Rick C
Rickman, I can't get your 3:50pm on message 4-5-17 to load. Mikek
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
I think we are in different time zones, so is this the message you can't load? I'm in EDT.
-- Rick C
I don't think time zone matters.
I plan on using the high input amp developed by dagmar, although I don't need the capacitive divider on the front end. I want the high R, but want to match the 1.7pf capacitance of the original tube. I have some tube plugs that I plan to connect to the pcb I'm developing. I want a plug in pcb. I think John is probably right I can just detect with a diode and add the voltage drop back in on the other side of the meter with the zero circuit. I have a spreadsheet with measurements of Injection voltage, Tube Cathode voltage, voltage across the Meter*, Q meter reading and some calculations with that data. I'd like to send it to you if you want to look. Mikek
-- This email has been checked for viruses by Avast antivirus software. https://www.avast.com/antivirus
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.