A Clean Canvas - 24 Pill Class E Rig Construction

[WA1GFZ]

Brent,
I would solder the source leads a lot shorter up closer to the body of the part.

[K1JJ]

Brent,

It may be an illusion from your picture, but are the drain and gate leads clearing the grounded bus bar OK? I almost made the mistake of mounting my MOSFETS too close to the bus and had to sharply bend up the leads to clear. It all worked out, but I shud have spaced them 1/8" farther away.

T

[W1IA]

Brent,

It may be an illusion from your picture, but are the drain and gate leads clearing the grounded bus bar OK? I almost made the mistake of mounting my MOSFETS too close to the bus and had to sharply bend up the leads to clear. It all worked out, but I shud have spaced them 1/8" farther away.

T

Its an illusion..plenty of clearance. Frank I will redo the solder on source leads. I got the atc caps from Bruce yesterday. I am going to try and finish today. I'm still waiting on the .47pf bypass caps.

[WA1GFZ]

A long source lead adds inductance which shows up as negative feedback against the drive signal.

[W1IA]

Today's Update:

Installed the feeds for the IXDD414's, transorbs, and shunt caps.

[K6IC]

Been an interesting thread.  Thanks for all the photos,  Tom,  Steve,  Brent (and others ?)   Am just now going back and re-reading all.
 Someday will get together my act  and start my CLass E project.

72   Thanks     Vic

[W1IA]

Finished Deck!
Minus a barrier strip for 12 volts (VCC rail for driver chips); OK..the relays for 160

[W1IA]

Been an interesting thread.  Thanks for all the photos,  Tom,  Steve,  Brent (and others ?)   Am just now going back and re-reading all.
 Someday will get together my act  and start my CLass E project.

72   Thanks     Vic

Join the fun Vic....lots of the same rigs being built!
Everyone does a slightly different layout

Brent W1IA

[K1JJ]

Thanks, Vic -   There will be more posts on the subject.   


Looks FB, Tina.

Brent -  It's hard to tell by the photos, but do you have bypass caps right at each 12V lead of the 414's?  It's important to do this. Caps at the end of the bus are good, but add them to the leads too.

T

[W1IA]

Thanks, Vic -   There will be more posts on the subject.   


Looks FB, Tina.

Brent -  It's hard to tell by the photos, but do you have bypass caps right at each 12V lead of the 414's?  It's important to do this. Caps at the end of the bus are good, but add them to the leads too.

T

Yup...I found some .47's electrolytics  at 35 volts; very small and effective..I installed them next to the leads on the 414's. Also put some small 10 uf electrolytics (orange) at the end of the busses. I may put some .001 and .1 's to round out the values. Very compact design grasshopper

[K1JJ]

Brent,

Yes, that's the idea. Though, electrolytics are poor at RF. Take a look at my latest RF deck pic above. I stacked .001, 0.1 and .47 disc and film caps together - all connected directly to each VCC pin of each chip.   Frank hammered me to being anal about this.

The other thang is to use quality ceramic caps at the output transformer bypasses. (modulated B+) Get rid of those orange drops. Even using those expensive chip caps there is not overdoing it. I first used orange drops and saw parasitics. After I went with zero-lead length chip caps from Jay, it was as stable as a rock. Look at my latest pic closeup of the PC board I made to hold the 3/8" square chip caps that are better designed for RF.

Also, consider bypassing the incoming 414 VCC main lead from the power supply on the heatsink. And, bypass the modulated B+ at the heatsink for RF. I even slid some ferrites over these leads in addition to the RF bypassing.  (Remember to consider the capaciatnce added to the PDM filter output when adding to the modulated B+ lead if the cap is of significant value)

T

[W1IA]

Brent,

Yes, that's the idea. Though, electrolytics are poor at RF. Take a look at my latest RF deck pic above. I stacked .001, 0.1 and .47 disc and film caps together - all connected directly to each VCC pin of each chip.   Frank hammered me to being anal about this.

The other thang is to use quality ceramic caps at the output transformer bypasses. (modulated B+) Get rid of those orange drops. Even using those expensive chip caps there is not overdoing it. I first used orange drops and saw parasitics. After I went with zero-lead length chip caps from Jay, it was as stable as a rock. Look at my latest pic closeup of the PC board I made to hold the 3/8" square chip caps that are better designed for RF.

Also, consider bypassing the incoming 414 VCC main lead from the power supply on the heatsink. And, bypass the modulated B+ at the heatsink for RF. I even slid some ferrites over these leads in addition to the RF bypassing.  (Remember to consider the capaciatnce added to the PDM filter output when adding to the modulated B+ lead if the cap is of significant value)

T

The .15's are mylar film...they ran in the old rig..cool as a cucumber. Ok on the .47's... Looks like I will have to replace them

[K1JJ]

My orange drops ran cool as a cucumber too.  But I had parasitics until I changed them out.

Do a GFZ experiement.  Put one in series with a 5 ohm carbon resistor and then put an RF sig gen signal across them. Use a scope probe to find how much RF divides across the resistor and cap at 3.9 mhz, etc.  If the impedance of the rig is 2.75 ohms, then you want the cap impedance to be < 1/10th = 0.275 ohms.   Is it? You want most of the RF across the resistor in this test. If not, you will get some degree of RF getting into your PDM modulated DC cable and the transformers will not be grounded well for RF = instability... NG.    

Jay/W1VD  has a source for these 0.1 uf chip caps for only about $1 each. Mount them on a PC board like I did and get zero lead length.  I used three in parallel for 0.3 uf total per module.  RF= dead meat.

All this quality bypassing is cumulative. Do everything you can in each circuit and the rig will never let you down in bad situations.

Hey, I'm just parroting what Frank taught me...

T

[WA1GFZ]

OD caps suck at RF and most Films suck at RF because of the way the lead is terminated to the film. It causes a high ESL, effective series inductance. The fact that it runs cool means nothing. You in effect have an RF choke in series with the cap which limits the RF current so the cap stays cool. There should be ceramic caps with short leads in parallel to reduce the reactance of the bypass.
This is very important at the VCC of the 414 drivers so there is plenty of stored energy when the device turns on.
All these little issues mean the difference between clean drive and crap drive.
Also between blowing FETs and not blowing FETs.

[KF1Z]

If you're using circuit board ...
Might as well use the surface mount version of the IXDD414's...
For even less lead length.

 


Tom what are those "chip caps"  ?

Not that I plan on building any class-e rigs... but just for reference.

[WA1GFZ]

AVX .1 uf 200volt Jay bought a bunch of them when we were building BB IRB boards last year.

[W1IA]

I used three in parallel for 0.3 uf total per module.  RF= dead meat.

I assume you adjusted the last pole on the PDM filter too accommodate the total capacitance.
I need to take 2 steps backwards now....oh well. It will be worth it in the end.
 

[K1JJ]

I used three in parallel for 0.3 uf total per module.  RF= dead meat.

I assume you adjusted the last pole on the PDM filter too accommodate the total capacitance.
I need to take 2 steps backwards now....oh well. It will be worth it in the end.
 

Yep, sure did.

More info:  I'm finally getting around to the fine tuning aspect of the rig... I wondered how much additional stray capacitance is really there generated by going thru the shutdown board, leads, fets, transformers near the chassis, etc. (in addition to the 0.3 uf X 4 = 1.2uf)   I used a cap checker and found a total of 1.2125 ufd. So there's about .0125 uf additional stray. That's not bad. So adding in the bypass caps and adding in the measured stray uf is a good estimate - then subtract the result from the last PDM filter cap value...

Actually, I did it the easy way - I disconnected the last coil and simply padded the last cap until the total measured was right on.


BTW, if you're really anal about it, be sure to include the damper diode which has about 1100 pf itself.

T

[K1JJ]

Brent,

Here's a couple of new tips:

I'm in the fine-tuning phase of the rig now.  I was trying to cure the last tiny bit of hum in the system. It was hard to find. It was so slight that nobody wud hear it on the air, but still bothered me in the mod monitor. I cured it by tying a large shield strap from the PDM generator board ground to the chasis. The small wires did not do it - it needed the wide strap. I also did some more bypassing and better wiring routing within the PDM chassis box. Now the generator is cleaner and there is no instability of any kind. The tiny hum it was generating is gone and now the quiet room ambience is louder than any hum.  I've never had a rig with such a low level of hum before.

Also added an audio delay - now keyup and unkey are totally quiet with no signs of spikes on the scope. I was getting tired of that darn HV relay  'CLACK!!'

T

[K6IC]

Finally got around to listening to the audio file... Sounds teriffic,  Tom.  Well balanced audio with nice sparkle.  FB work OM !     Vic

[KF1Z]

BTW, if you're anal about it, be sure to include the damper diode which has about 1100 pf itself.

T

And the C and L in the wiring harnes to the rf deck.
\
s

[WA1GFZ]

correction. OD and film caps are fine in high Z tube circuits where a couple ohms of reactance don't matter. A different issue when you are playing in the low Z sand box.

[W1IA]

I'm back from a road trip to Texas..what a drive Phew! Was different when I was getting paid to do it. Anyway, I will be putting in an effort to finish this project over the next couple of weeks. Looks like I'm a bachelor for a couple of months

[W1DAN]

Guys:

One thing I do that makes T/R smoother is I ramp the PDM generator chip up from no output to normal output by using a capacitor on the UCC25701 disable pin. My transmitter comes smoothly up in about 1/4 second.

As I am running the old 6 FET model without a driver chip, I also avoid a T/R relay. Lastly I keep the HV on all the time. All this results in smooth T/R with about 10db loss in received RF to my FT1000...and a simple system as I am lazy.

Dan

[K1JJ]

A follow up on fine tuning Rico:

I got new reports of a spike on key up and some hum.

The hum was cured by the PDM generator feed forward circuit - it's amazing how well it worked! I was able to zero out all my hum to a totally flat carrier.  The small hum remaining from the audio will hopefully go away when I get the balanced audio cables set up.  Evidently it was HV pwr supply hum. I'm surprised others get away with their supplies using the same amount of filter C without needing the feed forward option on their E-rigs - while I had a bit of hum without it..  

I think I licked the key up pulse problem too. The VFO enable was coming on too late (#2) and amplitude goosing the finals on key up. I moved the VFO onto the #1 sequencer position and the spike went away.  I cud see it happening at the vfo driver output when the HV was on.  I still have a double carrier key up, but the first carrier is down about 30db and is simply the 414 RF drivers bleeding thru at about a watt or so.  I tried moving the ant relay delay to later position (#3) - to try to block the first carrier, but then the finals ran out of time to dump their RF during unkey. I could see the fuzziness at about 1/2 way thru unkey. With a longer ant relay hold, the scope picture looks clean on unkey.

I listened to the VFO and I hear it move barely 3-4 hz when I modulate it with a big tone. Hopefully if it still has problems we can ID it on the air with some more help from the guys.  

This fine tuning is very time consuming. We're almost there and Rico Suave will be rolled into the corner as an official rig.

Many thanks to all the guys who helped identify the various problems on the air. Sometimes this is the only way to identify the finer issues. The clues are invaluable to finding the problems.


Dan - Tnx for the suggestion. I did try the cap but it didn't help my particular spike problem. It later turned out to be the VFO keyup, so it was unrelated....

T