[NLRS] MGA-86576 - eBay Replacement ??

[Dr. Gerald N. Johnson]

Good point on the stability though the internal RF feedback inherent in 
most MMICs tends to improve their stability if not terminated in too 
much narrow band reactance, like a bandpass filter. Common lead 
inductance in most microwave devices is important, the high frequency 
(as in several GHz and up) demands the least practical common lead 
reactance. FETs are definitely a different character and some source 
inductance is often applied to reduce the input impedance when used at a 
few GHz and at VHF but ALWAYS presents the possibility of oscillation at 
10 to 20 GHz. Most GaAsFet circuits include a small resistor or a 
ferrite bead close to the drain to help calm the devices. Some of the 
hotter parts used well below their highest operating frequency might 
well benefit from an output circuit using a terminated high pass filter 
to load the output in the frequency range where instability is expected. 
That filter might simply be a 1/4 pf capacitor and a 20 ohm resistor in 
tiny chip parts, preferably ATC for the capacitor so it doesn't go 
parallel resonant at at lower frequency than the gain stage's maximum 
frequency.

Indeed the choice of coupling capacitor may be crucial to 15+ GHz 
stability. Its easy for the board and strays to resonate the inductance 
of the capacitor operating above its series resonant frequency to make a 
notch in the coupling. Franco Roto illustrated that well in his article 
(on his catalog page for noise diodes) on noise source coupling 
capacitors. Tough he falsely claims its the series resonant frequency 
that makes the notch in the coupling, its actually parallel resonance of 
the inductive capacitor above its series resonant frequency. The ATC 
that DEMI sells for 1 pf for 10 GHz transverter and preamp is series 
resonant at 9 or 10 GHz depending on its mounting, which is perfect for 
the operating band but its going to not couple at some higher frequency.

As well as the length of the trace to the pipe cap resonator probe from 
the device output, or input.

I have built one PGA103+ kit from WA3INC and picked up another from 
WA5VJB at MUD who sells the G4DDK board and parts. I have a Mirage 
B25160G from the early days with a U310 grounded gate preamp and it 
doesn't seem to do much. I think one of those boards will drop in and 
bring it to life. Might help the D1010N that doesn't have a preamp also. 
None of those will do as good as raising the antennas 60 feet higher 
with low loss coax though.

The horribly tiny VMK-1225 fet from Avago is rated to 25 GHz, but 
strangly not for NF above 17 GHz and its data sheet or applications note 
mentions source lead inductance for LF but not for HF and suggests the 
vias on the source lead pad be very close to the part and filled before 
trying to mount the part. Three connections on a 0402 part chase me away 
for now but the part is available for the brave. I'm thinking there 
might be better ways to do the experimental matching with tuning screws 
and phosphor bronze strips than the classic snow flaking which is cheap 
but takes hand steadiness at microscopic scale for 24 GHz and up that I 
haven't developed. And at 71 I'm unlikely to develop.

73, Jerry, K0CQ

On 12/9/2013 12:57 AM, Ronald Bergantzel wrote:
>
>
> Hello All !! Before I would jump to throwing in any replacement MMIC or Gaasfet device, I would really give Steve at DEMI a call / email and see what he would recommend for any device replacements (dont need to "reinvent the wheel" if you dont have to  ).....This is because I have noticed that some of the newer MMIC 's that have decent gain up in the microwaves MAY only be 'conditionally' stable in specific circuits (my experience shows that the 86576 falls into this category )...... I personally have been using the PGA-103 as 'second stages' up through 1296 , but the gain starts to fall off there. I like the PGA103's because they are very stable , predictable and have very low Noise Figures up through 1296 ..... Considering Jon's application , I would do some online researching IF Steve at DEMI did not have any thing to go with in the transverter .....   If this were my 2304 xvrtr and I needed a new preamp stage and could NOT find a suitable STABLE
>   replacement MMIC device , I would put a bypass in the xvrtr and use an external ATF54143 Gasfet preamp ( I can get a complete preamp down to the size of a postage stamp for 1296 that has less than 0.70db Noise Figure and over 16db Gain ) on the outside or adapt the xvrtr board for that device ..... The reason I wanted to throw in my 2cents worth is to mention the issues that could arise with CONDITIONAL or UNCONDITIONAL stability of the specific devices ... I have been using some of 'old - school' preamp designs but they only work to a certain extent with a limited amount of devices out there today - ( i.e - the ATF 54143 is only CONDITIONALLY stable from around 200 mhz and down , above that the '54143 pretty much is an UNCONDITIONALLY stable device )......   Thanks for reading my rant - just dont want anyone to think that just any MMIC could work in place of a MGA-86576 - this instability issue could really be a problem if a different MMIC that had
>   somewhat different S11 / S22 specs was installed . 73s !!! Ron KA0RYT