IC-7000 heat problem

(1/2) > >>

John Albert:
I think the main issue with the IC-7000 is the power management and component placement. Out of the 5 or 6 7000’s I repaired, I found that the highest heat concentration was near IC-1901 and R1902 (both mounted on the bottom side of the main board). R1902 is a 1.8 ohm in series with the 14v power. It feeds the 14 volts to IC1901, a 7808 3 terminal regulator. The 8 volts from IC1901 feeds voltage to just about every board including the CPU, PA, Main, DDS and DSP. After 8 hours in receive mode I measured temperatures, using a Fluke 189 and a thermal probe, that exceed 92.4 degrees “c” or 198 degrees “f” on the top side of the main board. It’s hotter on the bottom side and I can guarantee it’s hotter than hell! With a few screws removed on the main board, there’s enough room to reach under the board with my “pinky” and feel the regulator. It burnt my finger!

There’s a second 7808 (IC2001) and a 1.8 ohm resistor (R2011). It feeds 8 volts to an RF amplifier IC on the main board and the control head. Here’s where it gets strange. In previous models and that included Yaseu and Kenwood. They would feed 12 volts to the control head and regulate down to 8 or 5 volts to power logic or analog circuits. The IC-7000 regulates down to 8 volts on the main board feeds the control head. To power LED’s and the LCD backlight, there’s a switching supply that kicks the voltage up to 12 volts. A 7805, 5v regulator (IC101) also feeds off the 8 volt line and powers the logic circuits. In my opinion, the current needed to run the control head adds to the heat problem on the second 8 volt regulator and creates additional heat on the main board.

While making the heat measurement I discovered that if the top cover is removed temperatures drop about 10 degrees “c”

Peter Gant:
exceed 92.4 degrees “c” or 198 degrees “f”

feel the regulator. It burnt my finger!

All of this points to bad engineering practice on the part of Icom. There's a dead easy rule when it comes to components, if you can't touch it without burning your finger then something is badly wrong.

Tanakasan

Darrin:
92C at the component-- That could be way too high, or could be ok for that component, it depends on the context and the component. Many semiconductor companies assume that the die is at 150C or less at Tamb-max when they do their reliability assessment. I know we do for power devices where I work.

For a power device, like a regulator, with very low Tjc (thermal conductivity between die and package), I could see 92C being ok depending on what the high end ambient limit was. As a circuit designer, I wouldnt like that kind of heat, but statistically it could be within design limits an not lead to excessive failure.

The real question is, do these hot parts in the unit fail at an unreasonably high rate. You didnt mention that. Maybe you meant to? In the 6 units you repaired, was it a failure in the hot parts? If so, I think you've nailed a design problem for sure.

Still, heat kills! Not only does it kill the part getting hot, but lesser parts around it.

John Albert:
IC2602 is the AF speaker amplier that's vertically mounted on the main board and clipped the the aluminum casting. The engineers could have used a TO-220 style 7808 and mounted it same way.

Bill Burrows:
Typical semiconductors can run with junction temps up to 100 C and power devices to 150 C.  Its common to derate them in high rel applications.  Even so this means some pretty uncomfortable case temperatures could be acceptable.  The finger rule is very conservative.

It does seem like the 7000 gets warm in receive so I found a small computer style 12V fan and mounted it on to the rear heat sink with plastic inserts and sheet metal screws (like used to mount stuff in wall board).  The fan is powered off the antenna tuner connector which I don't use.  This results in cooler operation.

I took the unit in once for some warranty work and it came back with a firmware mod that runs the internal fan much more aggressively.  That seems to be a great improvement, at least on transmit.

The only mode now where the temp gauge gets to near the red zone is PSK and it is acceptable at room temperature at 30% power out (about 30W at 13.6V).

Hot spots are a fact of life in highly miniaturized electronics.  You're not going to get there if you limit yourself to 10 or 20 deg C temp rises.  Some components are going to have to run at closer to rated junction temps.

Besides, have you ever measured the temps in some of the beloved boat anchors touted on this site?  They got things inside that glow red hot :)

Navigation

[0] Message Index

[#] Next page