Adding an extension to a scope lead

[ka1tdq]

I'm looking to have a convenient way to monitor parameters on my 75 meter class E rig.  Normally, I need to pull the scope out of the closet and set it up in a cumbersome manner.  I would like to have the scope on a fixed location on my desk.  The only problem is that it's located farther away than the reach of the scope leads.

I'd like to mount two BNC barrel connectors to the front chassis of the rig.  Then run two 10' coax runs to the scope from there.  Then, whenever I wanted to check things out with the transmitter, I'd just need to plug the scope leads into the other side of the BNC barrel and monitor what I needed.

I know the extra length adds extra capacitance, but can I get away with this on 75 meters?

Jon

[W1ITT]

RG-62A is a 93 ohm coaxial cable.  It has 12.8 pf per foot, as opposed to 27.5 pf per foot for most of the common 50 ohm cables.  It may or may not fit perfectly into your connectors of choice but, at 75 meter band, it won't matter much.

[WD5JKO]

Jon,

   I'd make a resistor divider at the rigs RF output. Take a 5K, or 10K, or 15K from the RF output to a 50 ohm resistor going to ground. At the junction have a BNC connector. The attenuation will be -40, -46, -50db respectively for the values suggested. For legal limit AM (400W carrier) the 15K resistor (-50db) would dissipate over 1 watt. I'd do something like series up three 5.1K 1W to make 15K. Then run ordinary 50 ohm coax to the scope. At 400 watts carrier, you will have 141v RMS out of the transmitter, and about 1/2v RMS out of the attenuation divider

    You can do a similar thing with the high level audio. Make a divider, and run that to the scope. That way you can use the scopes X:Y option to view a Trapezoid pattern. Then turn the sweep to 1ms/div and see the modulated envelope. Use that audio channel to trigger the scope when viewing the modulated envelope.

Jim
Wd5JKO

[AB2EZ]

Jon

It would be helpful to know a bit more about what these BNC connectors on the chassis are connected to inside the transmitter (I.e an abbreviated schematic drawing.

For example, if you are observing the waveform between a drain bus and ground... you could do the following (which is what I did with my Class E transmitters):

1. Close to the drain bus, make a broadband, DC-coupled R||C voltage divider as follows:

A. Place a 15pF capacitor (rated for at least 1000V) in parallel with a  3300 ohm resistor (a 2 watt or larger rated non-inductive resistor would be okay).

B. Attach one end of this parallel combination to the drain bus, and place a 1000pF  capacitor (rated for 500V is okay) between the other end of the 15pF/3300 ohm parallel combination and ground. Keep all of the reasonably leads short.... to avoid forming a large area loop that would couple to the RF fields that are present.

C. Run a piece of 50 ohm coaxial cable (small diameter cable is okay), with the shield connected to ground at both ends, between a terminal on a terminal strip near the front panel BNC connector and the node where 15pF capacitor and the 3300 ohm resistor connect to the 1000pF capacitor.

D. At the terminal on the terminal strip, connect a 50 ohm resistor 1/2W (or larger) non-inductive resistor to ground. This resistor forms the termination of the 67:1 voltage divider between the drain bus and the terminal on the terminal strip. [i.e. 50 ohms/3350 ohms= 66.7 and (1/1000pF) / (1/15pF + 1/1000pF) = 67.7]

E. Connect a 680 ohm 1/4W resistor (or a pair of 330 ohm 1/4W resistors in series) between the terminal on the terminal strip and the center pin of the BNC connector.

Now you can connect an oscilloscope to this front panel BNC connector with a 50ohm coaxial cable that is as long as needed... provided there is a 50 ohm termination at the oscilloscope end of the cable.

The net result is the there will now be a 1000:1 voltage divider (broadband and DC coupled) between the drain bus and the input to the oscilloscope.

If you see a 0.1V peak signal on the scope, then the peak of the drain waveform is 100V.

Note that you can use a similar approach to observe the gate waveform.

Stu

[ka1tdq]

Actually Stu, that's exactly what I'm trying to do.  I drew up a schematic of what you said.

I would actually much prefer to hard wire in the test points like this rather than using scope probes.  That way, I can just swap the coaxes around on the BNC's to monitor what I want.  I'll mount 4 BNC jacks for Drain Bus 1 and 2, and Gate Bus 1 and 2. 

Sooo, what values would I use for the gate bus?  It's only about 11 volts peak, so I would need either a 1X or 10X probe.

Jon

[AB2EZ]

Jon

In the case of measuring the gate waveforms:

I suggest that you simply omit the 50 ohm termination and the 680 ohm series resistor on the right side of the coaxial cable. Connect the center conductor on the right side of the coaxial cable directly to the center pin of the BNC connector. Make sure the coaxial cable to the scope is terminated by 50 ohms at the scope end of the cable.


Having eliminated the second voltage divider [i.e. 50 ohms / (680 ohms + 50 ohms)]... you are left with a signal at the input to the scope that is (1/67) x the signal between the gate bus and ground.

You could use different voltage divider component values, but with 7V peak between the gate bus and ground, you  will still have about 0.1V peak at the scope.

Stu

[WD5JKO]

Stu,  Good stuff there....

I still repair commercial 13.56 Mhz 3KW RF generators. Using a 100X low capacity scope probe, I notice the probe tip warms considerably when used beyond about 15 seconds. During development work, something like what you propose would be more applicable.

Jon, I miss understood your first post. I thought you intened to monitor the RF output from your Class E rig. There are not too many hams that monitor the drive to the final amplifier stage and the output before the filter at the same time. Heck, most hams don't monitor anything except the meter on the rig...if there is one.

What I proposed is a way to monitor the modulated RF envelope, and the audio, such that ith a scope, you can monitor the RF envelope, or the Trapezoid pattern. Each presents information not seen in the other. That is what I do with my Central Electronics 20A station. On AM I only use the trapezoid. On SSB I use the envelope pattern.

Jim
Wd5JKO

[ka1tdq]

I've never actually done the trapezoid or modulation measurement you talked about Jim (other than read about it).  Not to start a discussion about modulation percentages, but I do limit my negative peaks and push the limit on positive.  I use Steve's modulation monitor and everything looks good there.

My concern is about gate and drain waveforms.  With everything being homebrew, I need to make sure that my gate waveform is #1 actually there, and #2 looking good.  And I need to look at the drain waveform for tuning.  I've done that all once, and can pretty much rely on the wattmeter, but I'd like to conveniently check in on it from time to time.  Granted, it is just a PW 400 watt carrier rig, but it's still good to keep everything running good.  

Jon

[AB2EZ]

Jon
Jim

Let me propose an improved version of the circuit... which is simpler, doesn't require any high power-rated resistors... and which is closer to (but not exactly the same as) the design of the business end of a x1000 scope probe.

1. Keep the 15pF and the 1000pF capacitors as originally proposed in my earlier post.

2. Replace the 3300 ohm 2W resistor with a 47k ohm 0.5W (or 1W) resistor.

3.  Eliminate the 50 ohm termination resistor on the right side of the coaxial cable.

4.  Add a 4700 ohm 0.25W (or 0.5W) resistor from the center pin of the BNC connector to ground. (This 4700 ohm resistor is a calibration resistor, and also a DC safety load to have in place if there is nothing else attached to the BNC connector)

5. Make sure that the coaxial cable between the BNC connector and the scope is terminated with 50 ohms at the scope end of the cable.

The above will produce a 1000:1 ratio between the voltage on the drain bus and the voltage across the scope's input... without requiring any high wattage resistors.

With respect to the gate bus measurement:

Keep the design that I recommended in my previous post in this thread... except the 3300 ohm resistor doesn't have to have a 2W dissipation. Since the gate voltage cannot be  greater than 15V, a 0.25W resistor is sufficient for the gate voltage probe.

Stu

[ka1tdq]

Using the smaller power resistor will work better since clearances are pretty tight on the heat sink assembly. 

By the way, I don't have a gate "bus" per se.  Each of the 8 FETs has its own driver.  I'm just going to sample one of the gates from each side (I'll monitor 2 of the 8 FET's gates).  The drains are common though, obviously. 

It's not ideal, but it's representative of what's going on with the other FETs. 

Jon

[ka1tdq]

I'm going to make a slight variation.  I'm just going to hardwire the drains and not bother with the gates.  The connections and clearances at the gates are so tight, I'm worried about the kids coming in and pushing on components.  They could short something out.  I'm using open frame construction, so it's all out there.

I plan on mounting the BNC barrels at the chassis and putting all components in one spot at the drains.  That way,  it's just one piece of coax going to the scope.  The drawing should clarify.

Jon

[AB2EZ]

Jon

The new schematic you posted is not correct. The correct schematic is attached below. You can cut the coaxial cable (shown below) anywhere you wish... to insert a BNC (or other RF-type) connector.

Stu

[w8khk]

Jon and Stu, I have been following this thread with great interest.  I am in the process of finishing up a 400 watt eight-FET Class-E rig based on the WA1QIX design, including all the boards making up the PDM modulator and power supply.

I like the idea of installing permanent monitoring ports for the gate and drain points, at least on my first Class-E RF deck.  I know the design is solid, as it has been around for many years.  I just want to be sure my implementation is correct.

With eight FETs, I would need to monitor two drain buses, and four gate buses.  So I plan to add 6 BNC connectors to the back of the rig.

Stu, rather than follow the verbal text you provided above, could you please document in schematic form both the gate and drain attenuator circuits you prescribe for monitoring?  I am sure that would be valuable to others reading this thread at a later date.

One question regarding the 50 ohm termination - I assume the termination at the scope end of the monitoring cable is necessary only when actively monitoring, ie, that port is connected to the scope.  The purpose of the 50 ohm termination is to obtain an accurate representation of the waveform sans reflections on an improperly-terminated line.  I further assume that an unused port at the load side of the attenuator circuit does not need to be terminated with 50 ohms; and will not negatively affect the signal on either the gate or drain bus due to the isolation provided by  the attenuator.  If this is not correct, please advise, and all ports, used or unused, may be terminated with 50 ohms.

[w8khk]

Stu, I realized you provided the schematic of the drain monitoring attenuator after I posted my question.  (I should have refreshed my browser before starting to post.)

I believe it is correct to assume that the gate monitor attenuator is identical to this schematic except that we must "omit the 50 ohm termination and the 680 ohm series resistor on the right side of the coaxial cable."

My apologies if I have caused any confusion.  I would still like confirmation whether the 50 ohm termination may be omitted from the monitoring port if that port is not cabled to the oscilloscope.

Thanks and regards,
Rick

[N1BCG]

See below...

[AB2EZ]

I am sorry for all of the confusion.

Attached are two schematics. The first is a 1000:1 RF voltage divider... suitable for monitoring a drain bus. The second is a 200:1 RF voltage divider... suitable for monitoring a gate bus.

In either case:

You can insert an RF coaxial connector anywhere along the coaxial cable shown
You do not need to terminate unused monitoring ports with a 50 ohm termination

Stu

[ka1tdq]

Thanks Stu for the updated schematic. I'm headed to the parts store this morning to pick up the stuff. Huh, that almost sounded like a drug deal.

Anyway, I'm hoping to get some building time in today.

Jon

[w8khk]

Stu, thanks for the clarifications.  Separate schematics totally avoid the confusion.  Adding the permanent monitoring ports will make it so much easier to troubleshoot, as well as verify proper operation after getting it on the air.

[ka1tdq]

It works pretty slick! 

In one picture, you can see the connection to the drains (the red heat shrink around the RG-58) and the BNC barrels on the front plate. 

The drain waveforms in the other picture are at 400 watts.

Jon

[AB2EZ]

Jon

The drain waveforms don't look like class E waveforms.

What is the specified bandwidth of the oscilloscope?
Do you have a 50 ohm termination across each oscilloscope input (either provided by the oscilloscope as an input option, or added externally)?

Note that you should make these measurements with each input set for DC coupling.

Stu

[ka1tdq]

I don't see an option in the menus for 50 ohm loads on the scope inputs.  I cobbled together a cantenna dummy load with the input of channel #1 and this is the result. 

At 400 watts carrier, I'm using 9.09 amps at 47.4 volts on the drains.  That equals 92.8% efficiency.

I'll have to put a T at the input to each channel and get 50 ohm terminations for it as well.

Jon 

[AB2EZ]

Jon

That looks better!

When you add the T connectors at each scope input, and the 50 ohm BNC terminations are mounted directly on the T connectors (standard 0.5W BNC terminations will be fine)... the waveforms should look exactly like the waveforms on Steve's web site. There should be no high frequency ringing on the waveforms once the inputs have proper 50 ohm terminations.

Stu

[w8khk]

Jon

That looks better!

snip...  There should be no high frequency ringing on the waveforms once the inputs have proper 50 ohm terminations.

Stu

Finally - a good application for all those leftover Thin-LAN BNC terminators and TEE connectors!