Amplifier advice for 2M

[radiogareth]

I picked up a pair of lovely dual-fet PAs last weekend. I've spent some time today with a magnifying glass and ohm meter working out whats what.
Before I drive it with some RF I'd value any advice or suggestions.

2018-11-26 17.46.09.jpg (131.33 KiB) Viewed 3866 times

Its a 28 volt device with a 35dB pre-driver also running on 28 volts, preceded by a small mosfet and then an RF switch. By virtue of a lovely board layout, I've removed the small mosfet and fed the drive to the chip capacitor that feeds the RF gain block (CA2832C) which in turn connects to the dual mosfet gates. There is a LM317 for bias and when fed with Vin it regulates at 1.21 volts on the dual mosfet gate (adjustable up and down a little). There seems to be a lot of sensing circuitry (a pair of AD log detectors as part of some SWR bridge) and a couple of quad op-amps which is what the 7905 powers. I won't need any of this for what I'm doing. The output is unswitched and goes straight to a nice BNC. I hooked up my MFJ259 and on 146.5 the 'output' matching can be adjusted to 0 reactance and an R of 15 ohms. I'd assume the input is also tunable (for max output) when fired up.
So on 2M my portsdown gives out 2.0 dBm (1.6mW) according to my cheap eBay power meter. So boost that by 35dB and its going to be around 1.5 watts (30dB =x1000 so uneducated guess) which will presumably be an appropriate level to drive the final dual FET. Quite what the output will be (by extrapolating the data down to 146 MHz for the MRF275 I'd guess around 17dB) which should put me in the 50-70 watt range. Quite how linear this is at this point remains to be seen.
Assuming it all works, I can look at the spectrum shoulders and can reduce the drive via the in-line attenuator of the Portsdown. From what I've read I THINK the dual FETS will need biasing on quite hard - they seem to be running around 250mA each at the moment. How 'hard' or is it simplest to 'set on test'.
Suggestions welcome - no data is available from the makers (I tried) but it was sold as 'possibly a paging amplifier'. One did have 'Guys' written on it and its certainly VHF sized. The missing space if for a thermal switch which I'd guess would shut the amp down and sound the klaxons...
Gareth

[M5TXJ]

Gareth,

Don't forget the 50W ERP linit on 146.5MHz, unless you've got a dummy load on the mast fed with wet string, you'll be well over the limit.

73 Dave.

[radiogareth]

Noted Thanks for the reminder......

Gareth

[G4FRE]

The macom datasheet suggests 100mA per device, which is what I ran mine at on 2m before I got the i0jxx 500W amplifier

Dave
g4fre

I picked up a pair of lovely dual-fet PAs last weekend. I've spent some time today with a magnifying glass and ohm meter working out whats what.
Before I drive it with some RF I'd value any advice or suggestions.2018-11-26 17.46.09.jpg
Its a 28 volt device with a 35dB pre-driver also running on 28 volts, preceded by a small mosfet and then an RF switch. By virtue of a lovely board layout, I've removed the small mosfet and fed the drive to the chip capacitor that feeds the RF gain block (CA2832C) which in turn connects to the dual mosfet gates. There is a LM317 for bias and when fed with Vin it regulates at 1.21 volts on the dual mosfet gate (adjustable up and down a little). There seems to be a lot of sensing circuitry (a pair of AD log detectors as part of some SWR bridge) and a couple of quad op-amps which is what the 7905 powers. I won't need any of this for what I'm doing. The output is unswitched and goes straight to a nice BNC. I hooked up my MFJ259 and on 146.5 the 'output' matching can be adjusted to 0 reactance and an R of 15 ohms. I'd assume the input is also tunable (for max output) when fired up.
So on 2M my portsdown gives out 2.0 dBm (1.6mW) according to my cheap eBay power meter. So boost that by 35dB and its going to be around 1.5 watts (30dB =x1000 so uneducated guess) which will presumably be an appropriate level to drive the final dual FET. Quite what the output will be (by extrapolating the data down to 146 MHz for the MRF275 I'd guess around 17dB) which should put me in the 50-70 watt range. Quite how linear this is at this point remains to be seen.
Assuming it all works, I can look at the spectrum shoulders and can reduce the drive via the in-line attenuator of the Portsdown. From what I've read I THINK the dual FETS will need biasing on quite hard - they seem to be running around 250mA each at the moment. How 'hard' or is it simplest to 'set on test'.
Suggestions welcome - no data is available from the makers (I tried) but it was sold as 'possibly a paging amplifier'. One did have 'Guys' written on it and its certainly VHF sized. The missing space if for a thermal switch which I'd guess would shut the amp down and sound the klaxons...
Gareth

[radiogareth]

Well the good news is that the FETs are in excellent working order, it makes a great selfexciting power oscillator! If you take a peek at the earlier photo, the input is the small coax with the two ferrite beads on, this feeds into one end of the 35dB gain block (little black box below the choke) whose output goes to the FET gates. The only change I have made is to bypass the RF switch and another gain stage before the 35dB gain block.
I might try a bit of tinplate shielding or remove the 35dB block totally and try one of the small eBay amps which have shown to work well with my Pluto, subject to drive level. I can see its going to have pretty lively gain on 144, the data sheets indicates 4 watts of drive will yield over 150 watts AT 500MHz so its going to be an interesting blend of output vs linearity. The data states 150W at 400MHz, 12.5dB, 200W at 225MHz and 15dB so I'd guess 'moregain' at 144.

'Watch this space'.....

Gareth

[radiogareth]

Overnight I wondered if removing the first FET had left the front end a bit too 'up in the air' so added a 56R SM resistor from input to ground (bit like 'grounded grid!). This 56R was originally in the drain feed to the first FET so 'seems logical' to me.
Still unstable, but on 25 volts DC input and with direct drive from my Portsdown on 2M (no attenuation selected) it delivers an indicated 90 watts. The signal on my SDR looks OK for a few seconds (SR125), then jumps to a wider signal (oscillation) that continues even if drive is removed.
So my first problem is to cure the instability, then look at the drive levels and resultant linearity that is obtainable.
Gareth.

[radiogareth]

Had another go last night with a friends spectrum analyser. There's just too much gain in this gadget for stability so I moved on to trying a similar amp driven by an outboard eBay amp.
This is the eBay pre-driver, fed directly from my Portsdown with no attenuation inline. https://www.ebay.co.uk/itm/50-1200MHz-1 ... SwMmlbazre SOmeone else was using one in their Portsdown so I took a chance. Its on a re-worked board inside and device ID unknown ATM - might be a mosfet or more likely a beefy MIMIC.
Success I think, 12 watts indicated on my power meter and the spectrum look good (to me, please advise/comment!). Test card, 125KS, 146.5.
The amp is one of these, with a filter in the output chain. https://www.ebay.co.uk/itm/100W-FM-VHF- ... 1776288254 running off 28 volts. Good for 120 watts on FM on 4 and 2.
Could be a good route for 2M DATV?

2018-12-17 20.05.47.jpg (82.03 KiB) Viewed 3505 times

Gareth

[radiogareth]

By removing the 35dB gain block, various bits of coax and adding a brass shield twixt input and output I have got it stable, feeding the portsdown into a "0.5Watt 50-1.2GHz" eBay special then into the dual Mosfet gate. My RSP1 spectrum analyser shows this waveform which I count as 'just about 35dB down for the shoulders. This is 125SR test card and about 14 watts on the meter. My PSU can't deliver enough current to support 28 volt operation but I'd assume more RF at that voltage.

Capture.PNG (16.31 KiB) Viewed 3476 times

However, to get some RF out of the amp I need to bias it up quite high (2.5amps Id). If I set it at the perhaps more conventional 100mA as per the data sheet I get very little RF output. I'm puzzled by this. Whats going on? (My understanding of amplifier classes is clearly lacking but suspect its something to do with it. I have read that high bias is needed for good linearity on DATV but not to get RF?)
Gareth

[radiogareth]

Dropping the drive level by another -3dB gives this and 10 watts. Looks better IMHO.

[radiogareth]

And 28 volt PSU, -5dB attenuation, looks good at 13 watts out.