Off Topic -- Warning re surplus Nokia Tetra base station Power amps on 70cms


Mark GM4ISM
 

Hi All
a little off topic, but may be of benefit to someone.
 
Some of you may have purchased ex Dolphin Tetra base station power amps for use on 70cms
I ran one recently on 70cm EME at about 225W op for a weekend and came close to destroying it. A fact i didn't notice until i looked inside recently.
 
I already spotted the fact the the DC feed for the PA devices was through small plated through holes, and not adequately rated for high power operation, I bypassed this on my amplifier and thought I was OK
Wrong
The output transmission line in the vicinity of the onboard directional coupler had heated up enough to start delaminating the board and the soldered joint to the output coupler was now a classic dry joint.
Much more operating time on this amplifier would have resulted in failure
 
I am starting to write up my findings, and those of Russ G4PBP on my web page with notes etc on using the amplifiers on 70cm, but this is a heads up to hopefully prevent anyone destroying one of these lovely amplifiers.
 
Mark GM4ISM
 
 


Rob, M0DTS
 

Thanks for pointing that out Mark...

Unfortunately when out portable today testing on 70cm Digital ATV i found another way of generating the same effect, very bad SWR....

Oh well.. will have to replace the o/p cct with some coax now!
I've had two years of good service at ~180W.

Maybe you should not have mentioned it...hi

73

Rob


Mark GM4ISM wrote:


Hi All
a little off topic, but may be of benefit to someone.
Some of you may have purchased ex Dolphin Tetra base station power amps for use on 70cms
I ran one recently on 70cm EME at about 225W op for a weekend and came close to destroying it. A fact i didn't notice until i looked inside recently.
I already spotted the fact the the DC feed for the PA devices was through small plated through holes, and not adequately rated for high power operation, I bypassed this on my amplifier and thought I was OK
Wrong
The output transmission line in the vicinity of the onboard directional coupler had heated up enough to start delaminating the board and the soldered joint to the output coupler was now a classic dry joint.
Much more operating time on this amplifier would have resulted in failure
I am starting to write up my findings, and those of Russ G4PBP on my web page with notes etc on using the amplifiers on 70cm, but this is a heads up to hopefully prevent anyone destroying one of these lovely amplifiers.
Mark GM4ISM
www.dc2light.co.uk <http://www.dc2light.co.uk>


Christopher Bartram <cbartram@...>
 

Having designed TETRA PAs, I assume that these surplus amplifiers are designed
to deliver 50 - 60W pep, and will produce >200W simply because they were run
'backed-off' in order to get good large signal linearity with a relatively
poorly-performing lineariser.

I should stress that I don't know these amplifiers, but the problems which Mark
and Rob describe aren't black magic!

The issues which will arise from running the amplifiers at higher powers will
(probably) be due to the choice of substrate for the microstripline output
matching elements. If, as I suspect, the amplifiers are built on 0.62 or 0.31mm
FR4 to save money, the major issue will be resistive heating due to dielectric
losses, not dielectric breakdown. While you can't do a lot about the latter,
the heating due to resistive losses in the substrate can be dealt with in the
same way as you'd deal with losses in a resistor: heatsinking!

The power dissipated in a microstripline can be calculated, but as a rule of
thumb, those in a 200W PA will probably only dissipate a few watts.

The characteristic impedance of a microstripline is relatively uneffected by
heatsinking structures placed on top of the conductor. A U-shaped heatsink of,
say, 0.25mm copper sheet soldered to the conductor will radiate the heat
dissipated.

One tool which anyone designing, building or modifying a solid state power
amplifier should possess is an infra-red thermometer. These are cheap and being
non-contact, can be used to look at a the temperatures of PA components while
it is running. Thermal design is often neglected in amateur designs.

Vy 73

Chris
GW4DGU


Russ Stewart <g4pbp@...>
 

Christopher Bartram wrote:
Having designed TETRA PAs, I assume that these surplus amplifiers are designed to deliver 50 - 60W pep, and will produce >200W simply because they were run 'backed-off' in order to get good large signal linearity with a relatively poorly-performing lineariser. I should stress that I don't know these amplifiers, but the problems which Mark and Rob describe aren't black magic!

The issues which will arise from running the amplifiers at higher powers will (probably) be due to the choice of substrate for the microstripline output matching elements. If, as I suspect, the amplifiers are built on 0.62 or 0.31mm FR4 to save money, the major issue will be resistive heating due to dielectric losses, not dielectric breakdown. While you can't do a lot about the latter, the heating due to resistive losses in the substrate can be dealt with in the same way as you'd deal with losses in a resistor: heatsinking!
The power dissipated in a microstripline can be calculated, but as a rule of thumb, those in a 200W PA will probably only dissipate a few watts.
The characteristic impedance of a microstripline is relatively uneffected by heatsinking structures placed on top of the conductor. A U-shaped heatsink of, say, 0.25mm copper sheet soldered to the conductor will radiate the heat dissipated.

One tool which anyone designing, building or modifying a solid state power amplifier should possess is an infra-red thermometer. These are cheap and being non-contact, can be used to look at a the temperatures of PA components while it is running. Thermal design is often neglected in amateur designs.

Vy 73

Chris
GW4DGU


------------------------------------

Yahoo! Groups Links




Hi Chris and I guess you are right on all counts.

I am actually using one these amps but envoked radical surgery to stop it burning up. It currently gives 280W o/p cold and drops to 240W when hot. It easily does 250W PEP !

The surgery involved firstly, cutting out the PA section from the main board and remounting it on the original heatsink. I left the bit of board which was the o/p line of the driver stage, (for matching purposes). I connected a piece of UT141-50 to this line and put a 10pf ceramic trimmer across the the coax / stripline transition to the gnd plane; this gave a 1:1 match when the o/p peaked (1st time ever :-)

At the output end I cut off the SWR monitor line part and stripped off the 50ohm track feeding it, right back to the Sageline combiner, just leaving a small pad which the Sageline is connected to. The used a piece of UT141-50 to bring out the RF. Without the line removal I could only get 230W cold.

It was interesting to note that when I stripped off the stripline, the bit of board under where the o/p N-type spigot was soldered, was already black + burned even under the amps normal operating condition !

The o/p Sageline combiner is at or just over it's power limit at 280W and seems to be the primary heat source; you can't hold your finger on it after 5mins of carrier !!! So I blow the combiner and am currently testing use of a 20mm fin (copper strip) soldered to it's length in the air flow. I will know how well this works shortly .... need an IR camera for Christmas for all this PA work Hi. It would be easy to replace the Sageline with an external Wilkinson in UT141-75 as an alternative.

Finally and most importantly, as Mark found in his original mods, you must 'beef-up' the DC feed to each device and not rely on the via's and pc tracks than feed the DC pwr; this will result in a 'burn-up' and general mess.

If anyone needs it, I have photographs of the finished PA, these may already be on Mark's website.

All I need now is someone to work on 70cms :-(


regards de Russ G4PBP


Russ Stewart <g4pbp@...>
 

Russ Stewart wrote:
Christopher Bartram wrote:

Having designed TETRA PAs, I assume that these surplus amplifiers are designed to deliver 50 - 60W pep, and will produce >200W simply because they were run 'backed-off' in order to get good large signal linearity with a relatively poorly-performing lineariser. I should stress that I don't know these amplifiers, but the problems which Mark and Rob describe aren't black magic!

The issues which will arise from running the amplifiers at higher powers will (probably) be due to the choice of substrate for the microstripline output matching elements. If, as I suspect, the amplifiers are built on 0.62 or 0.31mm FR4 to save money, the major issue will be resistive heating due to dielectric losses, not dielectric breakdown. While you can't do a lot about the latter, the heating due to resistive losses in the substrate can be dealt with in the same way as you'd deal with losses in a resistor: heatsinking!
The power dissipated in a microstripline can be calculated, but as a rule of thumb, those in a 200W PA will probably only dissipate a few watts.
The characteristic impedance of a microstripline is relatively uneffected by heatsinking structures placed on top of the conductor. A U-shaped heatsink of, say, 0.25mm copper sheet soldered to the conductor will radiate the heat dissipated.

One tool which anyone designing, building or modifying a solid state power amplifier should possess is an infra-red thermometer. These are cheap and being non-contact, can be used to look at a the temperatures of PA components while it is running. Thermal design is often neglected in amateur designs.

Vy 73

Chris
GW4DGU


------------------------------------

Yahoo! Groups Links




Hi Chris and I guess you are right on all counts.

I am actually using one these amps but envoked radical surgery to stop it burning up. It currently gives 280W o/p cold and drops to 240W when hot. It easily does 250W PEP !

The surgery involved firstly, cutting out the PA section from the main board and remounting it on the original heatsink. I left the bit of board which was the o/p line of the driver stage, (for matching purposes). I connected a piece of UT141-50 to this line and put a 10pf ceramic trimmer across the the coax / stripline transition to the gnd plane; this gave a 1:1 match when the o/p peaked (1st time ever :-)

At the output end I cut off the SWR monitor line part and stripped off the 50ohm track feeding it, right back to the Sageline combiner, just leaving a small pad which the Sageline is connected to. The used a piece of UT141-50 to bring out the RF. Without the line removal I could only get 230W cold.

It was interesting to note that when I stripped off the stripline, the bit of board under where the o/p N-type spigot was soldered, was already black + burned even under the amps normal operating condition !

The o/p Sageline combiner is at or just over it's power limit at 280W and seems to be the primary heat source; you can't hold your finger on it after 5mins of carrier !!! So I blow the combiner and am currently testing use of a 20mm fin (copper strip) soldered to it's length in the air flow. I will know how well this works shortly .... need an IR camera for Christmas for all this PA work Hi. It would be easy to replace the Sageline with an external Wilkinson in UT141-75 as an alternative.

Finally and most importantly, as Mark found in his original mods, you must 'beef-up' the DC feed to each device and not rely on the via's and pc tracks than feed the DC pwr; this will result in a 'burn-up' and general mess.

If anyone needs it, I have photographs of the finished PA, these may already be on Mark's website.

All I need now is someone to work on 70cms :-(


regards de Russ G4PBP



------------------------------------

Yahoo! Groups Links




THE LATEST !!!! ..... it's dead.

The newly modified Tetra 70cm amp passed away peacfully last night whilst testing with DF9IC :-)

It passed my 1hr / 5min thermal cycling test. I did not do my 80% carrier on test for > 1hr test as I reckoned the Tetra would set on fire, as it is being run too far above it's design spec. for comfort. I opened it it up expecting to see 'burned things' but there were none !

Everything looked new and shiney except the device gates had 26V on them :-(

It looks like an overdrive or maybe a sequencing issue, the SWR was fine but I thought I had the sequencing sorted. Mark suggested it could be the common problem with Jap rigs of a high peak switch on RF pulse. I never thought to check this; the IC810H has only ever driven valves histortically. Very remiss really missing such an important test, as I was trying to push the BLF 548's to near their maximum rating.

So a salutary warning again to those trying to push these amps above 200W, yes they will do nearly 300w but for how long I will now never know.

Anyway it died honourably in the persuit of DX so it will recieve a hero's burial later today ..... RIP TETRA


regards de Russ ..... G4PBP