[Amps] peak voltage rating for load cap

Manfred Mornhinweg manfred at ludens.cl
Sun Aug 20 16:22:32 EDT 2017

Hi Alan, Jim and all

> You've been quiet for a while.

Yes, I have been building some stuff, so I was busy.

> I've mentioned a UK supplier before,

Yes. I remembered the fact, but not the address...

> but for others here it is
> http://www.charcroft.com/pdf/tab_silver_mica_capacitors/SBH-Dipped-Radial-datasheet.pdf
> You can buy online here
> http://www.charcroft.com/ebuy/shopdisplayproducts.asp?id=24&cat=RADIAL+DIPPED+SILVER+MICAS
> but at around $4.50 US each plus shipping and customs it is going to
> be expensive, although perhaps only 20-25% the cost of ATC ceramics.
Yes... The price is midway. Way better than the ATCs,or the doorknobs I
have seen, or even the C0G chips at Digikey and the like. But still a
bit stiff for a hobby project, specially after factoring in the 60 pound
  international shipping charge, and the app $100 "service charge" raised
by courier shipping companies at delivery time - plus taxes, of course!

Looks like I will use homebrew caps.

> These parts are intended for transmitting use and are specified for
> both current and voltage but sadly not for ESR.

Their ESL must also be rather high, given their long and slender shape.

> I've seen these capacitors fail with wildly excessive voltage but
> they seem to hold up OK to over-current abuse so long as there is
> some airflow over them.

OK. Their voltage rating would require sticking to the lower
capacitance, higher voltage types, which means using two or even three
in parallel at each place, on the lower bands. So we would be looking at
about 40 capacitors rather than 18. Which after adding all charges would
set me back something like $500. No go!


> ##  Why would a blown capacitor in a LP  filter  take out the LDMOS
> device ??   Where is the high swr protection located, before the mess
> of bandswitched LP filters... or after the mess of LP filters ?

The SWR protection normally goes after the filters, because the SWR into
the filters is rather high, due to the presence of strong harmonics.
This will be a high efficiency amplifier running in saturation, so the
3rd harmonic will be really strong, and a diplexer filter is not an
option because then the amp would cease to have a high efficiency!

A very lax additional SWR protection could be added before the filters.
That's a good idea anyway, to protect against wrong filter, or open
relay contacts.

> Is there any other
> way to protect the LDMOS devices in the event of a cap failure in the
> LP filter ?

I was indeed thinking of one. Given that these devices are supposed to 
be resistant to overvoltage (avalanching and limiting it), it would be 
plausible to make a protection that senses supply current and the 
forward and reverse signals from a directional coupler. These three 
signals would then be used to first calculate the actual power delivered 
to the load (forward^2-reverse^2), and the result would be subtracted 
from the supply current value, which due to teh fixed supply voltage 
represents input power. That gives a result that is proportional to the 
effectively dissipated power. If this goes beyond a certain limit for a 
certain time, the amp would be shut down.

The weak spot in this scheme is that certain nasty loads can cause 
strong reactive current in the FET, that does _not_ cause a large 
increase in supply current. So the protectio would not trigger in this 
case, but the drain current may exceed the LDMOSFET's absolute max 
value, and fuse the bonding wires.

So the already complex protection system would need to be complemented 
by some sort of RF drain current sensing. Which means putting current 
sense transformers into a place of the circuit where the impedance is 
extremely low, and thus we need short, direct and broad connections. Not 
easy to do.

> ## There are a lot of caps out there that are not RF rated.  IE:
> like the pair of 6 kv rated disc ceramics that alpha and others  use
> for a ..plate block cap assy.

Yes. Manufacturers probably test such parts and then use the ones that 
work well. For a homebrewer, specially one located in technical 
wasteland, where every single part has to be imported, with a 3 month 
delay in customs, such trial and testing isn't very practical.

> ##  same deal with your garden variety .01 uf @ 1 kv disc ceramic.
> They are NOT RF rated.  However, the  2 kv  version of a .01 uf will
> definitely handle more current than the 1 kv version. 

Typically yes, but not necessarily. Take the case that the 1kV and 2kV 
caps have the same type of dielectric. In that case the 2kV one needs to 
use a thicker layer of that dielectric, which means that to maintain the 
same capacitance it needs to have a larger area. If this area is 
metallized with the same metal thickness as the 1kV capacitor uses, then 
the ESR of the 2kV capacitor will be higher, due to the longer effective 
conductor length. The larger size allows it to dissipate more heat, so 
it will probbaly be OK at the same or somewhat higher current, but 
probably not twice the current. And if the current is limited by the 
metallization right around the spot where the wire terminal is soldered 
to it, then the current rating will actually be the same for both caps.

Now take the case of both capacitors having the same dimensions, with 
the 2kV one using a different dielectric to achieve the higher rating. 
The metallization, terminals and solder joints might be identical for 
both, thus the current handling would also be identical. But the RF 
voltage rating might actually be _lower_ for the 2kV cap! This is 
because the dielectric material having a higher voltage breakdown per 
thickness very likely has some worse characteristics in other fields, 
and one of them would typically be its loss factor!

So, judging the current handling ability of capacitors by their voltage 
rating, even on a relative basis, is treading on very thin ice.

> ##  But stuff like TV doorknobs are also not RF rated for current....
> and the 20-30-40 KV rated types  handle less current than the real
> deal  15 KV rated  HT-57s.... with both cases being say a 500 pf
> value.

Yes. The important point is: Voltage ratings and current ratings are 
quite independent, and even DC voltage ratings and AC voltage ratings 
are independent. A higher DC voltage rating does not automatically imply 
a higher AC voltage rating, let alone at RF.

> ##  I suppose you could also go crazy, and use lower value rated caps
> ..in parallel, to handle more current, but then the device count goes
> way up... and so does the $$  spent.

Yes. In fact I'm doing that for bypassing and coupling. I need a bypass 
capacitor that can comfortably handle 20A of RF current. And it needs to 
have extremely low ESL. The DC voltage will be just 50V, so a 100VDC 
rated cap should be fine. But 20A??? So I ordered 100 dirt cheap chip 
capacitors rated at 100nF and 100V, in 1206 size, in X7R dielectric. I 
will try to test one or two of them to destruction, and then decide how 
many of them to use in parallel. If 15 to 20 of them in parallel can 
handle the current I need, then this will be no problem with the 
available space and layout of the PCB. If they can't, then I will need 
to look for some other solution.

> ## My HT-50 / HT-58   1000 pf @  5 kv rated caps I use for padding my
> air variables on the 160M tuned input would run warm, with 200 watts
> CCS applied.  They are only rated for 1A CCS.    The 500 pf version
> of the same  HT-50 / HT-58 is rated for 3A  CCS.   With 2 x 500 pf
> caps in parallel, the total rating of each padder assy is now 6A
> CCS... = stone cold.

That shows how important it is to do things right. Heat is the enemy of 
electronic parts.

> ##  manfred, could you make hb caps  from  say teflon, sandwiched
> between cu straps or plates ?   I have seen that done............ on
> lower valued cap assys.

Yes. That's in fact my best fallback solution. Instead of teflon I 
prefer using mica, because it has a higher dielectric constant, higher 
hardness (so a burr on the copper tape doesn't as easily damage it), and 
is easier to find. The homebuilt caps look a bit stone-age-like, but 
they do work.

Fine adjustment of such hb caps is done by compressing them in a vise 
while the capacitance meter is connected, and stopping at the right spot.

They end up pretty large, for the capacitance values needed on 160 and 
80 meters, but the the coils are large too, and the caps are still 
smaller than the coils, or doorknob caps.

On my website, at the end of this page, you can see some small ones, for 
a VHF transmitter:



Visit my hobby homepage!

More information about the Amps mailing list