Ringing on shunt capacitor?

Welcome and feel free to ask questions about class E, mosfet's, and AM.
You must be registered to post.

Ringing on shunt capacitor?

Postby KM5TZ/6 » Tue Mar 15, 2005 1:13 pm

I'm seeing a high frequency ringing on my shunt capacitor (500pF doorknob) at about 50MHz. I see some of this ringing in the gate drive though its hard to say for sure if this is real as scope probe grounding is critical.

With no shunt capacitor there is no noticable ringing (on drain or gate). Probing the circuit shows the ringing to be worst right at the capacitor and decays towards the FET drains. I tried a different type of doorknob and the ringing changed frequency (lower).

I susspect this is normal as I have seen ringing in some of the scope shots posted on the web but I wondered if anyone has a more a more detailed explanation? I susspect it is related to resonance within the doorknob caps. Comments?

David.
KM5TZ/6
Member
 
Posts: 26
Joined: Mon Dec 13, 2004 1:47 pm

Postby Tom WA3KLR » Tue Mar 15, 2005 3:33 pm

Hi Dave,

I did a lot of simulations of the Class E PA operation recently (using LTC SwitcherCad III) and after adding (by guess) realistic series inductances to the FET (voltage controlled switch with defined resistance) and shunt capacitors I saw the ringing you are now referring to. As you say, a recent photo showed the ringing on the drain line about the same as I got in my simulations.

What I came up with after some analysis of the situation is this:

The ring frequency during the period the FET is ON was the result of the FET package inductance divided by the number of FETs in parallel + shunt cap. series inductance, and the shunt capacitor value. In my simulations this came out to 5.333 nH and 1178 pf = 63.5 MHz.

The ring frequency during the period when the FET is off is again the FET package inductance divided by the number of FETs in parallel + shunt cap. series inductance, and the shunt capacitor in series with the net total capacitance of the FETs in parallel. This came out to 5.333 nH and 500.4 pf = 97.4 MHz.

The results were approximate to the ring frequency observed in the simulations. This is the closest I could come to defining the ring frequency observed.

Also, the ringing seen on the drain could become cancelled out when the product of one FET's capacitance x one FET package inductance = shunt capacitor x shunt capacitor's series inductance. This is not to say there is no ringing at the FET die.

In the simulations, the shunt capacitor and the FET capacitance is a fixed capacitance. In real life, the FET's capacitance is voltage dependent.

This change in the FET capacitance while the FET is off (ring frequency changing over that half cycle) could be seen in one of the recent photos. It's unescapable as there will be series inductance on the FET package leads and the shunt capacitor.

It is best to minimize the inductances to lower the peak ring voltages at the FET. Looking across the bare FET switch in a simulation allows you to see the higher voltage that is present across the die, as opposed to only being able to see the voltage out on the FET leads in real life; one good use for a simulation.

I was able to measure the series resonant frequency of a capacitor fairly easily by having a pigtail coax from an r.f. signal generator across the capacitor and also putting a scope probe across the capacitor and watching for a minimum in the r.f. voltage. Note the frequency and solve for series L. (You know the capacitor value.)

I hope this is a help.
Tom WA3KLR AMI #77
Tom WA3KLR
Member
 
Posts: 47
Joined: Sun Mar 21, 2004 12:30 pm
Location: Quakertown, Pennsylvania

Postby KM5TZ/6 » Wed Mar 16, 2005 10:46 am

Thanks Tom!

What is the predicted current in the shunt cap?

I'm using 250VAC 30A relays to switch in additional caps. I have 500pf hard wired and then switch in one or two additional 500pf caps. Plan is 500pf on 40m (hardwired), 1000pf on 80m and 1500pf on 160m. The relays have about 18pf capacitance between the contacts and the actuating coil (ground). I'm just wondering if this might be a problem?

Regards,

David.
KM5TZ/6
Member
 
Posts: 26
Joined: Mon Dec 13, 2004 1:47 pm

Postby frank carcia » Wed Mar 16, 2005 12:06 pm

Long leads=leakage inductance=ringing
Take apart your rice box final and look at the lead lengths around the final. The Lower your operating Z the worse it gets.
The FET package is pretty low L compared to the lead lengths of the circuit. That is why I used a pc board to lay out my final.
My layout would be a pain to service so I'm hoping someone finds a better
way that is easier to service. The minimum width of a conductor should be greater than 20% of the length.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby frank carcia » Wed Mar 16, 2005 12:07 pm

Long leads=leakage inductance=ringing
Take apart your rice box final and look at the lead lengths around the final. The Lower your operating Z the worse it gets.
The FET package is pretty low L compared to the lead lengths of the circuit. That is why I used a pc board to lay out my final.
My layout would be a pain to service so I'm hoping someone finds a better
way that is easier to service. The minimum width of a conductor should be greater than 20% of the length.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby Tom WA3KLR » Thu Mar 17, 2005 9:39 am

Dave,

One comment you made in your first posting I just noticed - when you had no shunt capacitor you saw no ringing. My guess on that is the ringing frequency was around a Gigahertz and so was lossy and your scope probe and scope are probably 50 - 100 MHz bandwidth so you would not see the ringing.

As far as the current in the shunt cap., one simulation I did says the RMS current would be about 9 Amps. The peak current was 28 Amps. This is modelling 1500 Watts CW output. So idle AM out is half of this. With the relay switching you will have extra series inductance added and added series resistance. With this low impedance circuit, all of this is not good. Right at the relay, the way they are constructed with the flat blades parallel in close proximity non-inductive style. So I would think that just the relay itself may be relatively low inductance. But the over-all wiring is not and you have the relay contact resistance which may be very signficant. I don't think that stray capacitance is a problem.

The convenience of a bandswitching PA is good however. It will be interesting to see how much heating and ringing you get in that bandswitching shunt capacitor circuit.

One idea I had for a 2 band bandswitching PA (160 - 80) was to keep the shunt cap fixed and use 2 different operating voltage ranges - the R value is higher on the lower band so that the shunt cap value is close to ideal on both bands. This requires the modulator dc voltage out and modulation level to be altered for the 2 bands. Even keep the tank inductor fixed? Variable tank capacitor of course. Higher Q at the higher band. Loading cap would be the one to change - bandswitching at the 50 Ohm point not a problem. I have not re-explored this idea since I created my PA design spreadsheet. Now it would be very convenient to re-examine this approach. I tend to still favor the individual band deck approach though.

I still have to finish my federal income tax and then on to my 1936 screen grid-modulated transmitter project. Then I can dig in to the Class E hardware for real.
Tom WA3KLR AMI #77
Tom WA3KLR
Member
 
Posts: 47
Joined: Sun Mar 21, 2004 12:30 pm
Location: Quakertown, Pennsylvania

Postby frank carcia » Thu Mar 17, 2005 9:48 am

The slower your scope the better it will look.
I learned a lot when I landed a TEK 7904 that can see a 1 GHz signal.
frank carcia
Senior Contributer
 
Posts: 765
Joined: Fri Jan 31, 2003 10:48 am
Location: Enfield Ct. / Niantic Ct.

Postby KM5TZ/6 » Thu Mar 17, 2005 10:38 am

Here's a close up of one PA section. It would be hard to get the leads any shorter.

I did something very similar in my class E exciter for 160/80 and 40m and it works great!

Image

David.
KM5TZ/6
Member
 
Posts: 26
Joined: Mon Dec 13, 2004 1:47 pm

Postby ke1gf » Thu Mar 17, 2005 12:18 pm

Dave if you get your rig running and stable multibanded at warp 1 that'll be really cool.
User avatar
ke1gf
Young Brainiac
 
Posts: 229
Joined: Mon Feb 03, 2003 11:38 am

Postby Tom WA3KLR » Fri Mar 18, 2005 8:03 am

Dave,

The PA construction looks terrific. I guess I was pre-conceiving a bigger relay! As you say I don't think you can do better than in the photo.

Good luck.
Tom WA3KLR AMI #77
Tom WA3KLR
Member
 
Posts: 47
Joined: Sun Mar 21, 2004 12:30 pm
Location: Quakertown, Pennsylvania

Postby Tom WA3KLR » Fri Mar 18, 2005 8:08 am

Dave,

The PA construction looks terrific. I guess I was pre-conceiving a bigger relay! As you say I don't think you can do better than in the photo.

Good luck.
Tom WA3KLR AMI #77
Tom WA3KLR
Member
 
Posts: 47
Joined: Sun Mar 21, 2004 12:30 pm
Location: Quakertown, Pennsylvania


Return to Class E Discussion

Who is online

Users browsing this forum: No registered users and 1 guest

cron