The copper clamp applies direct mechanical pressure to the flange ends, and little to no pressure at the center. Reinhardt's clamp does not touch the flange at all, so it's better able to apply uniform flange pressure - so long as the ceramic is not compromised. Frankly, I'm not a fan of mechanical clamping. While I haven't looked into it for a while, I remember reading about some difficulty finding a thermal interface material that is thermally and electrically conductive - an most importantly - remains stable over time - in spite of exposure to high levels of current. Ampleon correctly believes that soldering the flange to the heat spreader / ground yields superior thermal and electrical results, but the price you pay is the nightmare you must endure in the event of a device failure. Given the substantial number of LDMOS failures in this amp, it seems to me that direct soldering would be worth a try, but I suspect the RFPA circuit board's mechanical clearances are too tight to allow it.
Hi Steven, great comments and I agree. I have been tempted to do the direct soldering. And for that matter as you likely know, Ampleon sells versions of the devices that have the bolt down tabs removed which implies direct pressure on the ceramic or a solder approach. Right now the clamp approach I'm using with the iridium pad/material is working really well. If I ever see another failure, I may go the direct solder approach. Pull the copper heat spreader and put it on the hot plate. :) Max NG7M
@@ng7m-ham-shack Yesterday, I decided to replace the blown BLF-188XR FET on my W6PQL based water cooled amp project (dead due to a compromised coax shield). In the interest of expediency, I cut the leads with a razor knife, dismounted the input and output boards, and separated the copper heat spreader from the water cooled "cold plate". Once the arctic silver thermal compound was removed, I set up a hot plate (outside) and placed the copper heat spreader on it. It took about 7 - 10 minutes, and I was able to pluck the body of the old device off with hemostats. I added some soldering paste and a bit of 63/37 solder, then set the replacement FET in place. I worked it back and forth a few times - to make sure the flange had a good coat of solder and to work out any voids or air bubbles. When I had it correctly positioned, I placed a small weight on the device, then turned off the hot plate. Overall, it was a remarkably easy and quick process. Given the relative fragility of the ceramic cover, I doubt Ampleon intended for this device to be clamped in place. They recommend soldering because it gives the best thermal and electrical connection, but that is problematic with the RF Kit amp, since it's a 1 board design. The tight PC board clearances require the board to be mounted first, THEN the FETs are installed - and that means you CAN'T solder the FET flanges to the heat spreader. Unless you really controlled the process, the heat would damage the board or at least compromise the solder of all of the parts!. IMHO, Reinhardt should consider redesigning the board - breaking it into two sections. Solder the FETs to the heat spreader then attach and connect the boards. No thermal compound to worry about, and no clamp or screw issues to contend with, and a vastly superior thermal and electrical connection. FET change out is arguably "harder", but given the improved thermal and electrical conditions, the need to do so is dramatically decreased. 73 - Steve KG4LJB.
I know it's been a few years since you made this video, but I'm curious if you remember what thickness of Indium you used? I just had my home brew linear shoot sparks at me because I used liquid metal, and even though I applied the thinnest film I could, it still managed to creep up the side of my MRF1K50 and shorted the drains to ground. The 600MHz MCU I use in my controller was fast enough to detect the over-current and cut power to the LDMOS so quickly, the MRF1K50 seems to have survived. I removed it to clean all the liquid metal off and reinstalled with a 100 micron thick sheet of Indium foil. I'm beginning to wonder if I should have gone thinner to aid in thermal transfer...
I don't remember what the thickness was on the indium sheets I used. Glad to hear your MRF1K50's survived. The liquid metal was an interesting experiment, but I too didn't have a good experience with it. I think liquid did have a better lower resistance to ground, but the indium pads have really worked well since the video. I haven't had any issues all these years now. 73, Max NG7M
Hi Neal, do you mean how has the setup worked with the fan I put on top? It's worked great. Very happy with it for almost a year now. Proper torqued clamps on the 189's with the indium pads and my fan setup has been great. I also punched holes in the sides of the case on the top front area to increase air flow coming into the front portion nof the amp. All these modifications have worked really well. 73, Max NG7M
@@ng7m-ham-shack Understand and I am in awe! But, my setup is a stock one as delivered 1year ago. Since Reinhard has not given a source for the clamps he uses, and this looks like a ton of work, how would it work if I kept the original implementation and buckled on the external fan to the board? I honestly suspect the fan is the single most effective cooling agent of all of the items you implemented (do you disagree?). Just wondered if the fan was the low hanging fruit of this!
@@NealK3NC I agree with you, the fan is a nice easy low hanging solution to add some extra cooling. Forcing air across the top of the devices and through the transformer is a bonus. Eventually, I hope to see Reinhard make the clamps available. If your amp is working well, I would hesitate making any big changes. The fan idea is pretty benign. The other thing I did which added to the over all air flow was to use a 1 1/2 inch greenlee / harborfriegh knock off punch in the upper front part of the case on each side. It really opened the airflow into the front of the amp along with the standard slits in the bottom. I don't think I ever posted anything about that minor modification. I think those two modifiations would be well worth the effort and I'm glad I made them. Max NG7M
The copper clamp applies direct mechanical pressure to the flange ends, and little to no pressure at the center. Reinhardt's clamp does not touch the flange at all, so it's better able to apply uniform flange pressure - so long as the ceramic is not compromised. Frankly, I'm not a fan of mechanical clamping. While I haven't looked into it for a while, I remember reading about some difficulty finding a thermal interface material that is thermally and electrically conductive - an most importantly - remains stable over time - in spite of exposure to high levels of current. Ampleon correctly believes that soldering the flange to the heat spreader / ground yields superior thermal and electrical results, but the price you pay is the nightmare you must endure in the event of a device failure. Given the substantial number of LDMOS failures in this amp, it seems to me that direct soldering would be worth a try, but I suspect the RFPA circuit board's mechanical clearances are too tight to allow it.
Hi Steven, great comments and I agree. I have been tempted to do the direct soldering. And for that matter as you likely know, Ampleon sells versions of the devices that have the bolt down tabs removed which implies direct pressure on the ceramic or a solder approach. Right now the clamp approach I'm using with the iridium pad/material is working really well. If I ever see another failure, I may go the direct solder approach. Pull the copper heat spreader and put it on the hot plate. :) Max NG7M
@@ng7m-ham-shack Yesterday, I decided to replace the blown BLF-188XR FET on my W6PQL based water cooled amp project (dead due to a compromised coax shield). In the interest of expediency, I cut the leads with a razor knife, dismounted the input and output boards, and separated the copper heat spreader from the water cooled "cold plate". Once the arctic silver thermal compound was removed, I set up a hot plate (outside) and placed the copper heat spreader on it. It took about 7 - 10 minutes, and I was able to pluck the body of the old device off with hemostats. I added some soldering paste and a bit of 63/37 solder, then set the replacement FET in place. I worked it back and forth a few times - to make sure the flange had a good coat of solder and to work out any voids or air bubbles. When I had it correctly positioned, I placed a small weight on the device, then turned off the hot plate. Overall, it was a remarkably easy and quick process. Given the relative fragility of the ceramic cover, I doubt Ampleon intended for this device to be clamped in place. They recommend soldering because it gives the best thermal and electrical connection, but that is problematic with the RF Kit amp, since it's a 1 board design. The tight PC board clearances require the board to be mounted first, THEN the FETs are installed - and that means you CAN'T solder the FET flanges to the heat spreader. Unless you really controlled the process, the heat would damage the board or at least compromise the solder of all of the parts!. IMHO, Reinhardt should consider redesigning the board - breaking it into two sections. Solder the FETs to the heat spreader then attach and connect the boards. No thermal compound to worry about, and no clamp or screw issues to contend with, and a vastly superior thermal and electrical connection. FET change out is arguably "harder", but given the improved thermal and electrical conditions, the need to do so is dramatically decreased. 73 - Steve KG4LJB.
I know it's been a few years since you made this video, but I'm curious if you remember what thickness of Indium you used? I just had my home brew linear shoot sparks at me because I used liquid metal, and even though I applied the thinnest film I could, it still managed to creep up the side of my MRF1K50 and shorted the drains to ground. The 600MHz MCU I use in my controller was fast enough to detect the over-current and cut power to the LDMOS so quickly, the MRF1K50 seems to have survived. I removed it to clean all the liquid metal off and reinstalled with a 100 micron thick sheet of Indium foil. I'm beginning to wonder if I should have gone thinner to aid in thermal transfer...
I don't remember what the thickness was on the indium sheets I used. Glad to hear your MRF1K50's survived. The liquid metal was an interesting experiment, but I too didn't have a good experience with it. I think liquid did have a better lower resistance to ground, but the indium pads have really worked well since the video. I haven't had any issues all these years now. 73, Max NG7M
Max, how would leaving the original board implementation (no clamps) but using the 'AMD' fan on top of the 189s have worked?
Hi Neal, do you mean how has the setup worked with the fan I put on top? It's worked great. Very happy with it for almost a year now. Proper torqued clamps on the 189's with the indium pads and my fan setup has been great. I also punched holes in the sides of the case on the top front area to increase air flow coming into the front portion nof the amp. All these modifications have worked really well. 73, Max NG7M
@@ng7m-ham-shack Understand and I am in awe! But, my setup is a stock one as delivered 1year ago. Since Reinhard has not given a source for the clamps he uses, and this looks like a ton of work, how would it work if I kept the original implementation and buckled on the external fan to the board? I honestly suspect the fan is the single most effective cooling agent of all of the items you implemented (do you disagree?). Just wondered if the fan was the low hanging fruit of this!
@@NealK3NC I agree with you, the fan is a nice easy low hanging solution to add some extra cooling. Forcing air across the top of the devices and through the transformer is a bonus. Eventually, I hope to see Reinhard make the clamps available. If your amp is working well, I would hesitate making any big changes. The fan idea is pretty benign. The other thing I did which added to the over all air flow was to use a 1 1/2 inch greenlee / harborfriegh knock off punch in the upper front part of the case on each side. It really opened the airflow into the front of the amp along with the standard slits in the bottom. I don't think I ever posted anything about that minor modification. I think those two modifiations would be well worth the effort and I'm glad I made them. Max NG7M