MRF286 output power ? UT Source defective parts ??

[KT6UK]

EDIT

Since starting this thread I've discovered that the major issue is the MRF286's supplied by UT Source (and likely most other vendors in Asia) - they are defective and or counterfeit. With this in mind, please read on !
+++++++++++++++++

Hi,

I'm attempting to build a VE1ALQ dual MRF286 power amp driven by a single MRF286 and am seeking a bit of advice as I've never used these devices before.

So far the results are not too good

I'm concentrating on the driver stage and with an bias current of 500mA and 1 watt of drive at 1296.2MHz I can drive the device to a drain current of 5 amperes at 28V but only get 4 watts out.

To do this I've had to tweak the input match and now have a 20dB return loss - originally the best match was down at 1152MHz of there about.

I've also had to remove some copper from the output - but seem to have run out of things to tweak.

I have the MRF's soldered directly to the heat spreader in a 0.02" trench such that the tabs align nicely with the PCB.

I have the PCB fastened down firmly and all seems good from a mechanical aspect.

The batch numbers on the MRF286's are QQ0380006 - purchased from UT source.

Any suggestions otherwise it seems I've just built myself an expensive shack heater !

Thanks Tim

[KT6UK]

OK, so the output match is way off, after a bit of tweaking for +30dBm in I'm seeing 41.4dBm out, better but still not there.

I'm thinking I'll align / tune each amplifier individually before assembling the whole system.
Stay tuned

[ZL4PLM]

Hi Ive built 4 x W6PQL's all gave 200W + out using MRF286's

so def something wrong somewhere!

they can def make the power out

do you have any pics of the board ... it may give some clues

thanks

Simon ZL4PLM

[KT6UK]

Hi Simon,

thanks for the reply, I was getting a little bored talking to myself

Anyway - made a little bit of progress with the single device driver - now getting just on 10dB of gain so with 5 watts of drive can just about squeeze 50 watts out, but the efficiency is poor as it's drawing nearly 7 A at 28V - so only about 25% efficient.

If you look on Darrels website VE1ALQ.com you'll see picture of his improved design - but I suspect the MRF286 have quite a spread on their in/out impedances and don't work with the Darrells pcb's as they come.

I've had to shorten the input line about 5mm to get a decent match and by about 8mm on the output line.

The drive stage is just about good enough now as it doesn't need too much gain or output power - I figure 20-30 watts of drive into the pair should be plenty.

However I expect more of the same problems with the output pair so I'm wary.....

I have everything soldered and bonded down to a heat spreader so I don't think the problem is mechanical !

Any ideas apart from stating over with the W6PQL design ! ?

I'll take some photo's shortly....

Tim

[ZL4PLM]

HI Tim

where are UTsource located???

Anyway - I dont think the MRF286's are that bad - I didnt need to tune anything other than input tuning on each amp

A single device with 5w in should be about 90w out

http://www.ve1alq.com/MRF286-RO4003C%20 ... oards.html

I see he mentions an improved design which makes me suspicious

have a look here too

http://w6pql.com/xrf-286_amplifiers_for_23cm.htm

also see

http://www.ve1alq.com/MRF286-RO4003C%20 ... ed_amp.JPG

I think his tuning / matching isnt up to speed

his board looks somewhat different too ... the matching sections look short.

my W6PQL boards need about 12-15w drive for 200w+ out

I only need about 12A at 28v for 200w

So have you seen these mods?

73 Simon ZL4PLM

[VK3QI]

Hi Tim,

I have 4 of the W6PQL dual MRF286 boards configured for high poer with a single MRF286 stage to drive the combination.

I have it throttled back to run 400 watts out and that takes about 40 amps, with all biases set individually for proper balance, but on average 500 mA per unit.

Initially I had some issues with getting maximum power out and that was caused by there being too much space between the edges of the boards and the ceramic body of the MRF286 chips. -

By reworking the position of the boards, I was able to get to a position where the MRF286s virtually snapped into position with the boards hard up against the ceramic body. That made an enormous difference. It seems that the spacing is extremely critical in reproducing the results that Jim W6PQL obtained.

If you look at the reference sheets supplied for the MRF286 including the test bed circuit that they use for evaluation, you will see that the tolerances at the board.ceramic body interface are critical.

see: http://www.nr6ca.org/pdf/xrf286rev3.pdf

The key measurement is the ceramic width which is specified at 9.65 to 9.91 mm - check your distance between the input and output board edges - it should be less than 10.0 mm width, preferably closer to 9.9 mm.

The W6PQL method of input tuning is to use a very small ceramic trimmer - it works well and takes the guess work out of snowflaking. I note the VE1ALQ does the same with a high Q trimmer.

I note that on the VE1ALQ page:

http://www.nr6ca.org/pdf/xrf286rev3.pdf he shows the 21mil copper spacer hard up against the ceramic body - is that the case in your setup?

Good luck with the project - it is great to be able to sit back on Field Day contests with a conservative 400 watts out feeding into the 2.4 metre dish!

Cheers

Peter VK3QI

[KT6UK]

Hi Simon, Peter

thanks for helping me out with my project - I'm hoping I can rescue it without starting over !

Just to set the stage, my amplifier looks very much like the 3rd picture in this link http://w6pql.com/xrf-286_amplifiers_for_23cm.htm and is built on the same 4003 material. The output match looks very similar but the input is different.

In any event I can get a good input match and drive the MRF286 up to adequate DC input - the issue seems to be with the output match.

Darrell recommends a 0.022" spacer between the heat spreader and PCB, but I've actually got the fet's soldered into a 0.022" trench on the heat spreader (0.125" thick copper plate) thus the PCB is directly on the copper heat spreader and aligns correctly in the Z dimension with the FET tabs. The trench is about 11mm wide but the PCB's are right up against the ceramic of the package, but as Peter notes below maybe my issue is that the copper output matching pad is marginally too far away as the pads don't quite come to the edge of the 4003 material.

Simon,

yes I've seen the W6PQL design and mods and will try to replicate on my 4003 boards with sticky copper tape - far from Ideal but worth a try. Ultimately I may have to replace output PCB and maybe even the input. If I do I'm hoping the FET spacing is similar if not the same.

UT Source seems like a great company for this type of part and is based in Hong Kong - very good prices and fast shipping. Might order some more MRF286's for a full legal limit project - assuming I can sort this one out first !

I can only dream of 90 watts out for 5 in

Peter,

Yes I have the input and output PCB' s butted up against the ceramic of the cases, however I suspect the actual PCB copper pad might be a fraction of a millimeter in from the material edge.

I'll check tonight and if needs be filed it down such that the electrical pad is right up against the ceramic of the FET.

At the same time I think I'll also make sure the heat spreader is nice and polished and that the underside of the output PCB is in good electrical contact.

As for input tuning - I have a 0.35 to 3.5pF Airtronics Hi Q multi turn piston trimmer which works great - I can get a 25dB return loss match which is dead stable, but I did have to reduce the input stub length by about 5 mm which is just the gap needed to accommodate the trimmer in any event !

The lessons I'm learning from this single device driver will hopefully make the pair alignment much easier and maybe help others should they stumble across this thread!

Thanks for all the help / tips guys - I'll report back on progress.

BTW this project is part of a mast head mounted pre-amp / PA I'm putting together to use with my IC-9100 once the 23cm module arrives.

73 Tim KT6UK

[ZL4PLM]

this project is part of a mast head mounted pre-amp / PA I'm putting together to use with my IC-9100 once the 23cm module arrives.

btw

have alook here

http://www.youtube.com/user/britindus



73 Simon

[KT6UK]

Dang it Simon!

The UX-9100 is not available in the USA till end of June so I'm waiting and using my home brew 23cm multimode in the meantime - check out my qrz.com page.

I'll use your video as a guide for the install once the UX-9100 arrives !

BTW just borrowed a thermostatically controlled heat gun from work to help remove the output PCB from under the MRF286's in anticipation of them needing to be moved closer - will attempt this feat later today !

Fingers crossed

73 Tim KT6Uk

[KT6UK]

So there was about a 0.25 mm gap between the MRF286 ceramic case and the PCB output pad - closed it up, no dice

Replicated the same physical match as close as I could to the W6PQL using copper tape carefully cut, installed and soldered in place - same result, no dice.

Only about 30 odd watts out for an Id of 6Amps in both cases. Seems like I'm missing 60 watts of RF.

Beginning to think I should give up on the board and try a single device PCB from W6PQL - if that goes well then I could try a dual.

I have little confidence the existing dual VE1ALQ will work with the MRF286's I have to hand, but will give it a go just for kicks.

The output match is significantly larger / longer than on the PQL solution and I don't have the tools to analyse what is actually going on.

This is turning out to be one very expensive amplifier project, but I will complete it some how !

73 Tim KT6UK

[KT6UK]

Ok - so reluctantly I've come to the conclusion after extensive testing that this amp can never work correctly - the input and output matches are way off and probably beyond recovery without some serious hacking of the PCB which would likely then compromise the mechanical integrity.

Plan B

Where to buy some W6PQL PCB's ?

Seems Jim W6PQL no longer makes them - any of you guys know of a source ?

Any ideas / suggestions ?

73 Tim KT6UK

[VK2GOM]

Have you looked at Ron VK4DD's amp designs and PCB's? www.vk4dd.com

73 - Rob VK2GOM / G0MOH

[VK4CDI]

Hi Tim
Haven't had time to do anymore with mine, other than close up the PCB to the MRF286...no change, as per your results.
It is looking like an expensive shack warmer.......

73

Phil VK4CDI

[ZL4PLM]

seems there is still a market

perhaps Jim can be persuaded!

??

Ok - so reluctantly I've come to the conclusion after extensive testing that this amp can never work correctly - the input and output matches are way off and probably beyond recovery without some serious hacking of the PCB which would likely then compromise the mechanical integrity.

Plan B

Where to buy some W6PQL PCB's ?

Seems Jim W6PQL no longer makes them - any of you guys know of a source ?

Any ideas / suggestions ?

73 Tim KT6UK

[ZL4PLM]

Sorry Tim

hope it works with the heat gun

Dang it Simon!

The UX-9100 is not available in the USA till end of June so I'm waiting and using my home brew 23cm multimode in the meantime - check out my qrz.com page.

I'll use your video as a guide for the install once the UX-9100 arrives !

BTW just borrowed a thermostatically controlled heat gun from work to help remove the output PCB from under the MRF286's in anticipation of them needing to be moved closer - will attempt this feat later today !

Fingers crossed

73 Tim KT6Uk

[ZL4PLM]

Rob

last time I looked Ron didnt do anything for 23 - he also built a W6PQL

cheers

Simon ZL4PLM

Have you looked at Ron VK4DD's amp designs and PCB's? http://www.vk4dd.com

73 - Rob VK2GOM / G0MOH

[ZL4PLM]

Tim & CO.

FYI from my ukmicrowaves yahoo reflector

On Wed, 2011-05-18 at 07:46 +0000, Simon L wrote:
>
> Has anyone got one of these running?
>
> A few of the VK's are having real issues.
>
> Any feedback from any of the ukmicrowave gang would be helpful
>
> rgds
>
> Simon ZL4PLM
>

Hi Simon
These amps need to be built EXACTLY as specified by Darrell, this
includes the depth of the shims, mounting of the MRF286s etc.
Here is a quote from some of the mail between VE1ALq and guys that have
had the amps:-

> Now on a more serious note:
> I have received 8 to 10 different Images of the stallions of the
> Amplifier Boards, and NO two (2) are the same, or as designed for
> installing.
>
There is extensive documentation and videos on the Darrells website.
When using those devices and getting the amount of power out of them
requires care when building. The drawings are on VE1ALq's web site, and
those that have followed them exactly have working units.
Mine isn't finished yet due to the long queue of things to be done.
I've cc'd my reply to Darrell
HTH
--
Best wishes / 73
Richard Bown

E-mail: richard@g8jvm.com

nil carborundum a illegitemis
########################################################################
Sent by Evolution 2.32.1 on Fedora FC14 x86_64 Dual core AMD Athlon 4400
Ham Call:G8JVM
Maidenhead QRA: IO82SP38:LAT. 52 39.720':N LONG. 2 28.171 W (degs,mins )
QRV HF + VHF Microwave 23 cms:140W,13 cms:100W,6 cms:10W & 3 cms:5W
########################################################################

[VK3QI]

Tim,

Have you tried snowflaking the output circuit to see what changes you can make to the output tuning?

A note of caution - I hope you are taking precautions when you are playing around with the board. Even 30 watts at 1296Mhz can be dangerous in certain circumstances, especially if you move your face nearby.

I NEVER operate my amp with the lid off at full power and if any tuning is required, I make certain that only one of the 8 bias circuits is active if the lid is off.

Cheers

Peter VK3QI

[KT6UK]

Hi Guys,

So I made a mistake - when I applied the W6PQL dimensions to the VE1ALQ board, I failed to account for the fact that VE1ALQ uses 32mil board while W6PQL uses 20mil material - so the impedances would be off.

Going to have to correct the line widths for the different board thickness, but will try a different mounting set up first.

Rob,

I've had a very good e mail chat with Jim W6PQL and he has no plans to make anymore boards, likewise Ron VK4DD has no stock - so it seems I either hack these VE1ALQ boards up (and hope they work) or I have to make or buy my own W6PQL boards.

Sad thing is I've put a lot of work into the metal casing / machining etc to accommodate the VE1ALQ boards - much of which will need to be re-done

Simon,

Interesting comment about having to use the shims with these boards - I had presumed the shims were just for Z alignment and I have designed them out by having a 0.022" trench in the heat spreader.

Turns out I have a spare set of VE1ALQ dual device boards and FET's which I can mount as he recommends with the shims and solder the FET's down on a flat spreader. So I'll give that a shot and see what happens !

That said I'd be very surprised if using the 22mil shim rather than a 22mil trench makes that much difference.

I did receive a strip of thin sticky copper tape with the boards - but have no idea what its for, snow flaking maybe ? Will check the VE1ALQ site again for info.

Peter,
any snow flaking seems to reduce what little power I have. I always have to remove copper and I note the input match seems to be down at 1160MHz out of the box, but I can (and have) fix that pretty easily. It's just the darn output match.

Noted about the hazards of 23cm RF and yes I'm careful not to get exposed to it

Phil,

Did you build yours exactly as Darrell specifies, ie using 22mil spacers and soldering the FET's down etc ? Did you also have the PCB's right up against the ceramic of the MRF286's ?

I'll report back with the results of my latest attempts to generate copious amount of 23cm RF energy - rather than just heat !

73 Tim KT6UK

[VK4CDI]

Hi Tim
I built mine a la W6PQL, 50mm x 110mm x 6mm block of copper with achannel milled in so the PCB just fits between device tabs and the copper block. Whole lot is bolted to the heatsink, using 6mm x 100mm wide Al bar as spacers. Devices are bolted in, I was planning to solder them in once it was working.......
Initially I had maybe 1mm gap , but have since fixed that so the ceramic butts up against the edge of the PCB....
This project is going on the backburner for a while, get some of the more urgent stuff done....
I got a strip of copper with the single device PCB also, but I am not going to even start on it....

73
Phil VK4CDI