"True" Peak Reading Meter

[n3lrx]

Anyone know where I can find a skizmatic for a true peak reading meter?
I found a few that are pseudo (MFJ type) peak reading, using a capacitor.
But I can't find any real peak readers. A few kits, which don't provide the schematic (otherwise you wouldn't buy it) Are there any available on the web? I tried several keyword searched on Google but only came up with the fore mentioned designs.

I know that a true peak reading meter uses either transistors or an op-amp (or both) and a handful of parts. Fake ones just rely on a capacitor and a diode or two. I'd hate to have to do it the MFJ way.

[k4kyv]

Why not just use a good monitoring scope to see what your peaks really look like?

Avoid the vertical amplifier; feed a sample of your rf directly to the CRT deflection plates for the most authentic indication.

[Pete, WA2CWA]

Peak Reading Bargraph Meter (Construction), QST, May 1981, page 31
Peak Power Reading Meters, CQ, Sep. 1983, page 54
or
Try this circuit; probably more than you need but the basic stuff is there:
http://www.edn.com/articles/pdfs/EDN/20010621/14di31.pdf

[KL7OF]

[]
Avoid the vertical amplifier; feed a sample of your rf directly to the CRT deflection plates for the most authentic indication.


DON..Where do you access the deflection plates?  and how is the TX coupled?   Thanks... Steve

[n3lrx]

I forgot to mention it's going to go on an analog meter. You know, the monkey swing? LOL

Don, that would be nice but I don't have one! Care to donate?

[KF1Z]

I am no expert.... by far.... but,

As far as I can tell, one might be better off building a meter such as the LED bargraph one with op amp instrument amps.... I've been looking around for a schematic to do just that..... then noticed your post..

LEDs & op amps respond very quickly to voltage changes much faster than a mechanical meter movement....and I like lots of blinky lights!


But, barring having a scope.. and really wanting to add "almost true" peak reading to an analog meter...
The MFJ approach may be reasonable... Look at the MFJ page, and download one of the manuals, they have schematics in them....
I believe you'll find the add-on kits for the mfj meters are transistors and caps....
Only the "stock" meters use just a capacitor to hold the meter up alittle so it can respond to the next peak a little quicker.

There is also a schematic on the TIS page (arrl) for addiing peak reading circuitry to a heathkit meter. may be of some help.

http://p1k.arrl.org/~ehare/TISarticles.html

[WA1GFZ]

Somewhere on the web there is a circuit for converting a bird 43 to peak reading.
You need a sample and hold circuit. It can be built with a switch and op amps.
Look in a national op amp applications book I'm sure you can find one. Then you need to build a trigger / delay circuit.
It charges a cap then shuts the switch off so it doesn't discharge. The hi Z op amp doesn't bleed off the charge. Better the parts the longer it will hold peak value.

[AB2EZ]

I agree with several of the previous comments. I have a peak-reading adaptor that I use with my Bird wattmeter... that works pretty well. But if you really want to catch the peaks, and to display them on an analog meter or an LED bargraph, you need something more sophisticated.

My peak-reading adaptor is the one designed by KE9UW and sold by Hi-Res Communications, Inc.

http://home.att.net/~hi-res/html/pdc-1.html

The peak-reading adapter I have is very much like the circuit in the TIS article that KF1Z referenced, except it places the diode-capacitor peak detector inside a feedback loop to create a "precision" peak detector.

Basically, to read peak power (rather than average power), you break the connection between the power-sensing circuit and the meter. Then you terminate the power-sensing circuit with a resistive load that is approximately equal to the resistance of the meter coil. [For a Bird meter, this is ~1400 ohms). That way, the power-sensing circuit behaves as it would if the meter were still connected. The voltage across this terminating resistor is used as the positive (+) input of an op-amp (the first of two op-amps). This first op-amp's output drives a diode-capacitor peak detector, with an adjustable bleeder resistor to set the peak reading decay time constant.  The output of the peak detector feeds the positive (+) input of a second op-amp... which is set for unity gain. I.e., the second op-amp has a 10kohm resistor between its output and its negative (-) input. The output of this second op-amp feeds back to the negative (-) input of the first op-amp via a 30:1 voltage divider. Thus, the closed loop voltage gain from the (+) input of the first op-amp to the output of the second op-amp is ~30. The output of the second op-amp also feeds a series resistor, which, in turn, drives the meter. The meter resistance and the series resistor feeding the meter form an (approximately) 30:1 voltage divider. (For a Bird meter, the series resistor is ~ 40k). Thus, the net voltage gain, input-to-output, of the peak detector circuit is adjusted to be unity by selecting the output series resistor and tweaking the input resistor of the first op-amp (the one that terminates the r.f. power sensor). An adjustable d.c. offset voltage is used to null the meter when no r.f. is applied to the power-sensing circuit.

All of the above is, as I said, pretty good... but the problem is that the true peak of the rf envelope may not last long enough, and, given the limited output voltage of the op-amp, may not produce enough current for a long enough time to fully charge the capacitor in the peak detector (even with the benefit of putting the peak detector inside the feedback loop). Furthermore, because of the effect of the bleeder resistor, the analog meter (and/or the person watching a faster display) may not register the absolute peak.

For example... the circuit I am using employs a 10uF peak detector capacitor, and a bleeder resistor which is adjustable from 20k ohms to 1 meg ohm. Thus, the decay time constant can be set between 200 ms and 10 seconds. However, it takes a lot of charge to change the voltage across the 10 uF capacitor. If you wanted to catch a peak that was .1 volt in size and .1 ms long, you would need to inject 10 ma into the capacitor for the duration of the peak. If you were trying to catch a peak that was 1 volt in size and .05 ms long (0.25 of the period duration of a 5 kHz sine wave), you would need to inject 200 ma into the capacitor for the duration of the peak.

Using a relatively small peak detector capacitor (.1 uF to better ensure that it charges to the peak signal), looking into a very large load resistance... and using a sample and hold circuit to measure the voltage across that small capacitor would be better (as WA1GFZ points out). An FET-based op-amp presents a large load resistance (looking into the "+" input). To hold the charge for one second, one would need a load resistance of about 100 megohms (time constant = 10 x hold time). One could implement the "sample-and-hold" functionality by turning on a switch to discharge the peak detector capacitor once per second. In that case, the peak-detecting capacitor would tend to hold the peak voltage during each 1-second interval, until discharged. This switch must have a high enough impedance in the "open" state so that it doesn't significantly load the peak detector capacitor. As an example, one could use a 555 timer to turn a suitable switch on and off. The switch could even be a relay.

The combination of a precision peak detector (i.e., diode and capacitor inside a feedback loop) and a sample-and-hold circuit would do a great job... and although it sounds a bit complicated, it is just a few garden variety op-amps, a 555 timer, and a few miscellaneous discrete components (resistors, capacitors, diode, and a suitable switch that can properly discharge the capacitor)

Stu

P.S. After I wrote this, I went back and changed the peak detector capacitor in my Hi-Res board from 10 uF to 1 uF... and I changed the value of the bleeder resistor from ~100kohms to ~ 1 megohm. There was a significant improvement in the ability of my Bird / Hi-Res combination to show the true peak power levels... without any problems in any of its other functionality.

[k4kyv]

From my experiences, the popular "ham radio quality" wattmeters are pieces of crap.  I did some checking of my transmitter a few weeks ago, using a Mirage wattmeter that I had calibrated against a Bird 43 years ago, along with a thermocouple rf ammeter and resistive dummy load.  From the looks of things, none of my transmitters make the nominal 75% efficiency of class-C operation.  Both my homebrew rigs check closer to 50%.  The broadcast transmitter checked out at over 90%!  I need to check the calibration of the DC voltmeters and ammeters in the Gates, plus I suspect the Mirage gives a high rf output reading.  Even with a 100% modulated signal per the oscilloscope, it shows peak output of less than three times the resting carrier power.  The thing makes a good SWR indicator, and relative power output meter, but right now I don't have a clue what my actual power output really is to anything closer than +/-  2 or 3 dB..

I use the scope to determine that the waveform is undistorted and that the rig is not overmodulated while still making clean positive peaks.  I can use the trapezoid pattern to determine modulation linearity. 

So I roughly guess that my output efficiency is somewhere between 50% and 70%, while the signal is clean and not flat-topping.

My monitor scope is a highly modified Heapshit HO-10 (making the thing actually useful).  It has provisions for direct coupling to the deflection plates.  Other workbench type scopes that I have owned have also had provisions for direct coupling.

If you have a spare scope around to dedicate as a monitor scope, it isn't difficult to go into it and wire the vertical input to go directly to the deflection plate, bypassig the built-in vertical amp.

Further evidence of the dumbing down of amateur radio is that when I ran the service shop at the two-way radio company in the early 80's, all the ham radio appliance manufacturers produced a modulation monitor scope to match each one of their equipment lines.  Yaesu had  models to match the FT-101, 301, 901, 101ZD, etc.  Kenwood also sold monitor scopes to match their lines.  I don't recall about Icom and Ten-tec did, but I would think it likely that they did.

A few years ago, I was looking through some of the advertising brochures from the major appliance manufacturers at their booths at Dayton, and  not a single one offered any kind of monitor scope, period.  One of the reps at one of the booths told me they had been discontinued because "no-one wanted to buy them anymore."

No wonder you hear so many crappy,  splattering SSB (an AM too) signals on the bands these days.

I still say that operating a phone transmitter (AM or SSB) on the air without a monitor scope is a lot like driving at night with your headlights turned off.  And I trust my thermocouple rf ammeter the most for determining my real rf output power.

[AB2EZ]

Last night I made some measurements of the waveforms of the signals along the path from my Bird power sensor => through my Hi-Res peak-reading adaptor => to the Bird 30 uA meter. Here are some interesting things I found. I was using a 0.33 uF peak detector capactior and a 3 Megohm bleeder resistor (decay time constant ~ 1 second). I was measuring voltage waveforms with a scope with a 1 Megohm input impedance. I was using a 2500 watt Bird sensor (2500H) in conjunction with my KW-1... feeding a dummy load.

1. The output of the Bird power sensor (terminated in the 1400 ohm input impedance of my Hi-Res peak reading adaptor) produced a voltage waveform which perfectly tracked the time-varying waveform of the precision peak detector I have in my off-air monitor. As expected, the output of the Bird power sensor is ~ 30 uA x 1400 ohms (peak at full scale output) ~ 42 mV (peak at full scale output). So I concluded that the sensor is perfectly capable of tracking the "true peaks"

2. The output of my Hi-Res peak adaptor, when terminated in a 1400 ohm load resistor (to emulate the Bird wattmeter) did an excellent job of detecting the true peaks. I.e., with 100% modulation, the peak voltage at the slowly varying  output of the Hi-Res peak adaptor was twice as large as when there was no modulation.

3. (Interesting) When I substituted the Bird meter for the 1400 ohm load resistor at the output of the Hi-Res peak adaptor... I noticed some surprising behavior. The slowly varying voltage at the output of the Hi-Res peak adaptor, as measured on my scope, seemed to vary somewhat strangely with time, and when modulation stopped, it actually went negative for about a second until everything settled down to the carrier level readings. I verified that this is not an RFI phenomenon. I suspect that this behavior is due to the dynamic physical characteristics of the meter movement (not the inductance of the meter coil... which would have to be ~ 500 Henries to cause this phenomenon!).  The meter must be acting like a generator as the needle moves backwards under the force of the spring. I measured the meter movement with my multi-meter, and determined that it has a static resistance of ~1400 ohms, in both directions of current flow, as expected.

4. In any event, the Bird meter did not reach the true peak reading... and I have concluded that the 1 second decay time constant of the peak detector in the Hi-Res module is not slow enough to be compatible with the dynamics of the Bird meter. I plan to try a longer time constant.

Stu

[WA1GFZ]

Stu,
A good sample and hold circuit would do it. Then build a trigger circuit with the time dely you want.  Also how about a buffer amp with ahigh z input so the charge doesn't bleed off the cap?

[AB2EZ]

WA1GFZ asked: "Also how about a buffer amp with a high z input so the charge doesn't bleed off the cap?"

The Hi-Res Communications Inc. unit has the peak detector feeding the "+" input of an op-amp which is specified as having a minimum of 100 megohms of iinput resistance... so the decay time constant is governed by the bleeder resistor value.

With the Bird meter, it looks like 0.33 uF in parallel with 5 megohms (my latest selection) is just a tad too fast (1.66 seconds decay time) to catch the peaks of a 1000 kHz modulated AM signal... but close. 

Rather than mess around with a large number of smaller resistors in series, I ordered some 10 megohm 1/4 watt resistors from Mouser. Those should be an excellent compromise between getting a good peak-reading, and having an acceptable decay time constant (~3.3 seconds).

Summary:

Using the Hi-Res circuit in conjunction with a Bird meter

a. 1400 input load resistance (to match the Bird meter's resistance)
b. 0.33uF peak detector capacitor and 10 megohm bleeder resistor (3.3 second decay time)
c. Feedback loop voltage divider: 20 kohms / (20kohms + 560kohms)
d. Output series resistor: 1400 ohms x 560,000 / 20,000 =  39,200 ohms

Stu

[AB2EZ]

Hi!

As a follow-up... I ended up using a 1uF peak detector capacitor and a 4.7 Mohm bleeder resistor with my Hi-Res peak adaptor. The decay time constant is, of course, 4.7 seconds. My Bird wattmeter tracks perfectly, over its full range... when alternating between the "straight thru" position (i.e., no peak adapter module between the sensor and the meter) and the "peak reading" position (i.e., with the peak adaptor module between the sensor and the meter)... with a slowly varying rf signal. However, when 100% modulating with a 1 kHz sine wave, the "peak" reading is still lower than the true peak (even with the 4.7 second decay time). I.e., with 100% modulation (1 kHz sine wave) of a 360 watt carrier, the peak reading is only about 1200 watts (instead of the true peak of 1440 watts).

It looks like it's time for me to give up on this approach; and use a readout that doesn't have the dynamic limitations of a Bird meter (which was designed for average power measurments). If I'm going to sample the power (with a sample and hold circuit), I feel that I might as well go to a fully digital approach: sample the signal at the output of the sensor at ~48 kilo-samples per secord, convert the samples to binary, store them in a RAM, use a software-defined algorithm to select and display the peak sample value. In fact, I've put in an order for an Array Solutions PowerMaster that does just that.

[n2bc]

You may want to take a look at N8LP's kit here:

http://www.telepostinc.com/

I've just finished building my second one.  Larry (N8LP) is continuously developing the firmware and adding functions as the user base grows and new ideas pop up.  In addition to the stock SWR & Power, peak and average, it measures Z, Phase angle, R, X, dBm, Return Loss. 

There are about 300 of them out in users hands now, all reports are vy positive. 

Just a happy user here.   

Merry Christmas all,   Bill  N2BC

[AB2EZ]

Bill

"You may want to take a look at N8LP's kit here:

http://www.telepostinc.com/       "

Yes... I was looking at these, and I agree that, based on what I read on line, these look like excellent units for measuring true peak power and a whole bunch of other things.

Happy holidays

Stu

[WA1GFZ]

Not a bad price for a real tool calibrated to a real standard.
I was impressed with Larry's product ever since his QEX article
Glad he turned it into a real product.