Hi Alan,
I am also using a simulator, the free LTSPICE which seems to work OK,
in fact I use it with wine under linux. With this, it is easy to
predict the voltages that will develop across all components and you
can also experiment with fine-tuning of the components. For example,
in my design at 200 W output I get about 170 V peak to peak across C2.
The waveform is very close to a sine wave, so that's about 60 V RMS.
So I now have a number to use when evaluating caps against their
datasheets. The voltage rating of these polypropylene pulse rated caps
seems to depend on the capacitance for a given VDC max, so the final
result will depend on what sort of caps I can get my hands on. The
current seems to be about 33 A peak to peak (wow!), at 90 degrees
behind the voltage.
Isn't simulation a great thing for those on a budget! :-)
73, Dimitris VK1SV
2012/8/9 Alan Melia <[email protected]>:
> Hi Dimitrios, yes those should do the job. I suppose that you need a
> situation with paralleled caps where any one can carry the peak current. It
> is probably a bit marginal if you really have to rely on current sharing in
> this situation. I used up to 3 in parallel in the C2 position without
> problems....I seem to remember the peak current was about 15 amp. I did a
> simulation usig a switch mode power supply model.
> Good Luck with it.
>
> Alan
> G3NYK
>
> ----- Original Message ----- From: "Dimitrios Tsifakis"
> <[email protected]>
> To: <[email protected]>
> Sent: Thursday, August 09, 2012 1:34 AM
> Subject: Re: LF: Re: Caps for Class-E amp
>
>
>
>> Hi Alan,
>>
>> thanks for the tip, I will keep an eye open for RIFA PHE428 or PHE450
>> which seems to be the newer version of that capacitor type.
>>
>> In my PA, I used two .1 uF caps in series to make .05 uF, then three
>> of these in parallel to make .15 uF hoping that the currents will be
>> distributed, however I think, once one of them gave up the ghost, the
>> others immediatelly followed suit, so I ended up with one in every
>> two-in-series capacitor being busted.
>>
>> By the way, that self-healing effect is quite useful if you ever want
>> to create a bunch of lower value caps with non standard values :-) I
>> now have a77 nF, 45 nF, 16 nF and even a 0.3 nF out of my .1 caps! I
>> am kidding of course. :-)
>>
>> 73, Dimitris VK1SV
>>
>> 2012/8/9 Alan Melia <[email protected]>:
>>>
>>> Hi Demitris, I am not sure of the types codes but I used 1000v or greater
>>> rating, the problem often is that the caps are not rated for high pulse
>>> currents. The best kind of caps that are used are rated for pulse
>>> operation
>>> in switch-mode power supply. Normal caps will not take the current
>>> pulses.....the individual plates get disconnected from the lead wires.
>>> One of the caps I used has come to hand its a RIFA PHE428 rated 2000v. I
>>> used this on 36v Class E running about 300W. I believe WIMA make similar
>>> caps but I am not sure of their type numbers.
>>>
>>> Alan
>>> G3NYK
>>>
>>> ----- Original Message ----- From: "Dimitrios Tsifakis"
>>> <[email protected]>
>>> To: <[email protected]>
>>> Sent: Wednesday, August 08, 2012 11:03 PM
>>> Subject: LF: Caps for Class-E amp
>>>
>>>
>>>> Hello LF group,
>>>>
>>>> I am building a 200 W Class-E transmitter for LF but keep killing the
>>>> capacitors in the output circuit (both C1 and C2 according to the
>>>> schematic in Sokal's QST article). I have used a mixture WIMA
>>>> polypropylene MKP10, MKP4 and FKP1 caps and I haven't damaged any FKP
>>>> ones so far. Should I not bother with MKP and use only FKP caps or is
>>>> there any other type of capacitor what will do the job? I need a
>>>> couple of hundred of nanofarads. The voltage rating of the caps I
>>>> destroyed was 400 VDC or 250 VAC. I can see that these caps have a
>>>> decreasing AC rating as the frequency goes up, but that's what I had
>>>> in the junk box at the time...
>>>>
>>>> The mode of failure of these MKP10 0.1 400 VDC caps is interesting
>>>> too, they seem to go down in capacitance as the damage progresses.
>>>>
>>>> Any advice is appreciated.
>>>>
>>>> 73, Dimitris VK1SV/SV1DET
>>>>
>>>
>>>
>>
>
>
|
