EMRFD Message Archive 14465

Message Date From Subject
14465 2017-12-23 04:23:36 QRP Labs High Speed Class E Drivers for Paralleled BS170

It's a 74ACT00 not HCT. Even faster ;-)

Remember that you are losing 0.7V in the key-shaping transistor Q6 and if you have the reverse polarity protection Schottkey diode fitted then you are also losing 0.4V there. So the PA really only sees 11V of the 12V you supply the QCX.

For reference for people who don't know it: the QCX CW transceiver web page is http://qrp-labs.com/qcx and the manual contains the full schematic and also charts of the power output vs supply voltage I measured on mine. If you get less power output than you expect, into a dummy load, and you are confident of the accuracy of your measurement, then it is possible your LPF is attenuating at a lower frequency than it should. You could try removing one or two turns from each inductor L1/2/3.

I put up a new monoband 30m EFHW antenna a week back. After a lot of work tuning and trimming I got it to read 1.01:1 on the antenna analyser. I had my 30m QCX transceiver on during the days last week, listening while I dealt with the pile of QRP Labs email, and jumped in for a QSO using the onboard microswitch as key, if I heard a CQ. Had great fun with QSOs all around Europe. 

73 Hans G0UPL 

1. High Speed Class E Drivers for Paralleled  BS170
    Posted by:  kc0vsj@yahoo.com kc0vsj
    Date: Fri Dec 22, 2017 5:08 pm ((PST))

Using the 3.3v output from the Si5351 to drive three paralleled BS170's in the QRP-Labs QCX is accomplished with 74HCT00N IC's.  The HCT series parts are fast, and have a turn on level below 3.3 volts, and running  on 5volts they make a nice level converter/driver as well as making possible some logic for keying etc.

 Is there a minimalist solution for a two transistor (NPN/PNP pair) that will let me do this at speeds over 30Mhz.  Complementary 2n2222/2n2907 open collector design  rise and fall times seem to be  less than the HCT switching times and I am wondering if there is a faster part.

 My interest in doing this is to do Manhattan builds of the QRP-Labs Class E design and compare it to the Cripe design.  I have a 30m QCX with 1.5watt output at 12volts which isjust  a bit under what Spice simulation seems to show.  The Cripe design seems to have more output and I am just trying to figure it out.