Mitrek as repeater - cooling?

jistabout · Post by **jistabout** » Fri Sep 17, 2004 1:20 am

Howdy. I've built a pretty nice UHF repeater using a Motorola Mitrek 100w mobile running 50 watts, with a massive fan (4.5") blowing into the PA compartment and an associated exhaust port(3"). Plenty of cool air blows onto the PA transistors. My problem is that after 30 minutes or more of continous key-down, the left rear of the Mitrek case near the heatsink still seems to me to be getting awfully hot. Doesn't seem to hurt it but I'd like to keep it cooler. I've tried several types of fans and configurations, all with similar results. Here are my questions:

Has anyone here used or is using Mitrek mobiles in repeater service, Ham or commercial?

Is cooling a problem?

Are they reliable?

Can they handle several hours of continous key-down?

Any stories, advice, comments or any other info on this would be appreciated. Thanks in advance.......

MO_TRASHER · Post by **MO_TRASHER** » Fri Sep 17, 2004 4:39 am

Our ham club recently replaced a VHF Mitrek repeater conversion that ran for over two years. As long as you're moving air across the heatsink, you should be fine.
We had ours mounted vertically with the PA up to maximize cooling.
The original cover was in place and no additional holes were put in the casting. (Other than the one for the RF connector)
We never had a transistor fail.

phrawg · Post by **phrawg** » Fri Sep 17, 2004 5:19 am

Lots of times airflow dynamics and cooling effects are never
considered when attempting to cool a device.
As you stated you were blowing air into the pa. NOT ALWAYS
A GOOD PLAN. placing a fan so that it DRAWS air through the fins
of the heatsink can be much more efficient at cooling it. Not
always is this the easiest to do physicaly, but is the best plan.
What happens is that when air is blasted at a surface to be cooled,
it forms eddy currents and ripples on the surface as it tries to flow
by and maximum exposure if the air to the surface is not realized.
The best plan is to let the fins that have already pulled the heat
from the transistors radiate the heat into the air around them
naturally, then draw that air away with a fan so more cool air can
flow in to replace it. If the transistors were meant to be blasted,
they would have much more exposed surface than they do. They
loose their heat through mechanical transfer to the sink.
Letting air come to the sink naturally allows it to flow evenly
over much more of the surface area and remain long enough
to become heated. THEN you remove the hot air and more
cool will naturally replace it. As a matter of fact if you were to
design a shroud around the fin assembly and open some entrance
holes for air, then DRAW AIR FROM, not blow air in to the shroud
the cooling would be the best possible. I also have a mitrek
based repeater and it works great. and no service in over
4 years of operation. Hope this may help. Phrawg

jistabout · Post by **jistabout** » Fri Sep 17, 2004 9:13 am

Phrawg, I've always used fans in such a way as to draw hot air away from a heated surface also, and I've tried that with this system. Results were about the same, so I turned the blower around in an attempt to guage best effectiveness. As for cooling the transistors rather than the heatsink, I got that suggestion from an article by Peter Harrison here on the 'net. I originally had two 3.5" fans set up to draw air thru the heatsink, but I had to mount them on top of the Mitrek case, which means that only 50% of each fan was actually over the heatsink and pulling air. Anyway, this repeater has an HF link on it which is why I'm concerned, because its very possible that it will be on the air for an hour or two at a time. The Mitrek is mounted vertically with the heatsink down, due to room constraints in the rack. Mabye I should change the orientation of the heatsink?? Anyway, thanks again

hwstar · Post by **hwstar** » Fri Sep 17, 2004 12:13 pm

It's probably overkill, but my design approach to does will survive the "brick on the key test". I used a duct and a 24CFM blower to direct a stream of air over the heatsink, an additional 12CFM fan blows air through the PA cavity. I also have a PIC-based fan controller and LM34 expoxied to the heatsink. With this setup and a 80 deg. F ambient temperature, the heatsink stabilizes at 107-112 deg. F. depending upon the radio ( I have 3 similar Mitreks built up this way) The controller will not turn the fan on until the heatsink reaches 100 deg. F, and PTT is present. If the heatsink reaches 140 deg. F a warning signal will be outputted. If the heatsink ever reaches 150 deg. F, the controller will interrupt the PTT line to the Mitrek and keep it interrupted until
the temperature falls below 130 deg. F.

phrawg · Post by **phrawg** » Fri Sep 17, 2004 12:16 pm

Ahhh yes. Good to see someone else understanding the
"draw don't blow theory"

I have my mitreks mounted vertically
on a sheet of rack panel aluminum with the connector to the
right end and the fins to the left as you look into the rack.
Then on the panel directly under the heat sink a large hole is
cut for a 4-5 inch fan which is mounted on the back side of the
panel so that the air is sucked around the fins from the front,
then under the heat sink then through the hole and the fan, then to
the back area of the cabinet. The hole is positioned so that just
a bit of the arc of the hole is past the end of the fins so it is not too
far under the radio. This seems to work well and is set up so it
kicks in with ptt regatdless of the temp. maybe that configuration
might be of interest to you. BTW I use 2 mitreks to build a repeater
and only 1 channel element in each and only wire the connector
to the controler with whats needed to tx or rx. Then both tx are
tuned the same and both rx likewise. If I have a failure, I just
swap mitrek positions then swap the elements between the radios
and turn it back on. Almost instant service restoration till I can
get to really repair it. Phrawg

jistabout · Post by **jistabout** » Fri Sep 17, 2004 2:20 pm

Hey Trasher, did you guys replace your Mitrek with something else? If so, what and why? Just curious.....

MO_TRASHER · Post by **MO_TRASHER** » Sat Sep 18, 2004 11:00 am

jistabout wrote:Hey Trasher, did you guys replace your Mitrek with something else? If so, what and why? Just curious.....

We replaced it with an MSR2000 base converted to repeater operation.
Since it didn't have the continuous duty PA, I added four fans to it; two under the PA blowing upthrough it, and two above the PA Blowing out the top through louvers that were added.

The whole blow/draw thing is a little convoluted to me.....I figure that you want air moving across or over the heat sink...if you can do that, you should be alright.

Cowthief · Post by **Cowthief** » Sat Sep 18, 2004 11:37 am

Hello.

I would look at what overclockers are doing with the "box".
Yes, there are fans galore in the latest modbox, but there are also heatsinks galore.
Heatsinks on the memory modules, heatsinks on the CPU, heatsinks on the graphics card, etc.
The idea is to move air, yes, but also to provide a good coupling for maximum heat transfer.
The heatsink itself needs to be kept cool(er).
You could put a huge heatsink on the finals and not need a fan at all, if room allowed for it.

bernie · Post by **bernie** » Sat Sep 18, 2004 2:07 pm

My two bits worth:
I suspect the eventual failure will be the solder on the PA transistor collector connection.

As long as the heat sink is not so hot you can't hold your hand on it, it should be OK.

One thing that I noticed on the Mitrek:
When lowering the power below a certain point side bands show up as the carrier on frequency is lowered.
This was a very early production 30W UHF unit.
Check derated amps with a spectrum analyzer to be sure this does not happen.

thebigphish · Post by **thebigphish** » Sat Sep 18, 2004 5:08 pm

phrawg, i agree with your "suck vs. blow" explanation...it is dead on. i showed it to some Industrial sciences guys who said it made perfect sense...nice jawb.

M3SGC · Post by **M3SGC** » Sun Sep 19, 2004 10:53 am

Why not try a heat pump AKA Peltier Element. Affix it to the surface that still remains hot. Just an idea anyway (DO make sure you have it working the correct way around).

jistabout · Post by **jistabout** » Fri Sep 24, 2004 11:30 pm

Okay. Here's what I ended up doing. 2 fans, mounted on the top cover of the radio, both drawing cool air through the heatsink. I also have them drawing hot air from the PA compartment via holes drilled in the lower part of the cover beneath each fan. Cool air enters the PA compartment through a series of holes in the cover between the fans. Seems to work fine. 50% of each fan draws air through the heatsink, the other 50% through the PA. This is my original configuration, and it seems best. After advice from you guys, I decided to reconsider it. Anyway, thanks so much. Hope this discussion helps someone else too

.

jistabout · Post by **jistabout** » Sat Sep 25, 2004 1:27 am

P.S. - I also re-mounted my mitrek. Mounting your Mitrek vertically with the heatsink down appears to not be a good idea.

richyradio · Post by **richyradio** » Mon Sep 27, 2004 3:58 pm

...you're asking for trouble with using anything higher than a 30 watt mitrek...when you start paralleling outputs with those wilkinson (pc board) couplers things start getting inefficient and really can't stand true 100% duty cycle...( 30 watt mitrek single output) turning the power down does not work either...they can actually get less efficient....(and possibly spur out as stated above) stick a PA on it if you really feel the need for more power....I generally don't like fans at all- all they do is suck in dirty air....give me a massive heat sink w/ the semiconductors properly spaced apart....

jistabout · Post by **jistabout** » Tue Sep 28, 2004 9:09 am

richy, I'm thinking that you're right. I've got the radio running much cooler now after hard uses, but I had to do alot to do that. I've got another 100w UHF Mitrek with a bad PA, and I might take out the PA board, run output from its exciter into an external amp such as a Micor repeater amp or some such. But how do the GE boys get away with running their converted mobiles like this? I see alot of MASTR II mobiles running as repeaters with no additional cooling!! Better PA design???

jistabout · Post by **jistabout** » Fri Dec 31, 2004 1:46 am

Just an update on this in case anyone cares

.

I still wasn't happy with the cobbled-together fan arrangement, and it still didn't work to suit me so I re-did everything and went to the system which phrawg is using. Much cleaner and works great. Pics upon request....

- DBH

jim · Post by **jim** » Mon Jan 03, 2005 4:38 pm

Heatsink position doesn't matter. Many people think that it does because "heat rises", which is true....in a gas or in the air. It doesn't matter in a solid. Putting the heatsink to the top won't make it take any more heat out of the case.
With forced air, you are not relying on natural drafting or convection currents to move the air over the heat sink.

A) Suck warm air out of an enclosed case.
B) Blow cool air over a heatsink. Look at a PC's processor chip!
Make sure to use a fan that can blow. Some are designed to be much more efficient at sucking and don't work well when their output side is put up against an object- they just "stir" the air around in a turbulance or vortex "surge" instead of forcing it through. The blade shape is what matters on the fan for suck/blow efficiency.

Oftern, modifications can be made to the internals to get more air to flow inside.

Another consideration is to have 2 speed fans. Low speed runs 100% and high speed comes on with PTT and stays on for an additional 30 seconds after dropping PTT. This can be done with an Elk or other timer. Low speed is produced with a 20 or so Ohm resistor in series with fan B+.

Instead of using many muffin fans, get a larger (1/3HP) squirrel cage blower and blow into a "manifold box." This is a simple sheetmetal or wooden box that's cube shaped. The side walls are approx. 4 times the size of the blower outlet. Use several flex ducts attached to the manifold and route them to the heatsink and wherever else you need reliable air. The advantage to a common blower is that you can essentially cool all of your equipment as they do in alot of commerical electronics and machinery. For the flexes, you can use someting as simple as small 3" dryer duct or the mega-buck silicone embedded ducting. This works well on large racks and it's QUIET! You don't have 12 muffin fans buzzing away.

jistabout · Post by **jistabout** » Tue May 22, 2007 12:56 am

Not to revive an old thread, but just wondering -

A GE Mastr II 100w mobile radio can operate (per the manual) at 100% transmit duty cycle at reduced (50% or so) power without additional cooling, and yet a 100w Motorola Mitrek mobile cannot. How come? The GE PA looks to be similar. Heatsinks are about the same. Is the GE just a better design?

George · Post by **George** » Tue May 22, 2007 4:19 am

Overall a better way to eliminate heat is to not generate it in the first place.

I really think a better way to do this is to use a 40 watt radio and, despite the smaller heat sink area, has less transistors to be making heat in the first place. In a standard Mitrek you are using four large devices. In the lower power units you are using a controlled stage and one driver.

Of course I am half awake and don't have my manual handy, but this is like the MastrII in that the amplifier essentually has a driver and controlled stage with a power amp of either the 75 or 100 watt variety.

The controlled stage/driver comes in two flavors...a fifteen or a forty watt version. Without the remainder of the power amp, there is a huge area of heat sink that has NOTHING on it. The heatsink fins are the same for 100 or fifteen watt output.

The Mitrek, on the other hand, has shorter fins for the lower power units, but I doubt it will overheat with that much aluminum under it even under repeater duty cycle.

One of my projects is using 40 and 35 watt X9000 mobiles as remote bases that could be connected to repeaters. I know the lower power units would not self destruct as opposed to the higher power units. Besides, I don't need the high power for the sites where the equipment is located.

George

Mitrek as repeater - cooling?

Mitrek as repeater - cooling?

Cooling.

Why not try a "Heat Pump".....

Mitrek cooling