|
Apologies, it's been a long time since I worked with RF circuits like the sBitx. What ratings (power, tolerance) on the resistors for the v3 upgrade would be warranted?
|
All the through holes are 1/4 watt resistors and smd are 1206. - f
On Wed, Nov 22, 2023, 1:35 AM KO6AGA Mike < mkpapp@...> wrote: Apologies, it's been a long time since I worked with RF circuits like the sBitx. What ratings (power, tolerance) on the resistors for the v3 upgrade would be warranted?
|
On Mon, Nov 20, 2023 at 06:45 PM, ajparent1/kb1gmx wrote:
Gordon,
That is true...? save for there is a transformer between the transmission line and the Drains.
If its 1:4 (impedance) step up going out it will work the opposite ways for the return
signal for a total return loss.? So the voltage is 1/2. (sqrt(4)).? Its greater for a
1:9 (1:3 turns) case.
Still I've used 3x to 4x as a safety margin.? The 100V irf510 is fairly resistant at 24V
and less so at 32.? For nominal 12V (13.8V DC) that means 55-60V device has
enough voltage margin.
However blowing devices is often a symptom of other things going on such
as RF instability or DC instability.
--
Allison
Puzzled still how I blew an IRFZ24N (just one -- the other appeared fine) in a highly modified V2 which has huge zeners across the gates to ground which have been shown with oscilloscope photos to literally protect the gates.? ? The drive level was 68% (and 100% had been set to move about 35 watts into 50 ohms) but the device was through an un-tuned autotuner, into the input circuitry (which has a tuned circuit) of an ancient Heathkit SB-200 amplifier.? Frequency around 3.550 MHz.? ?I had forgotten to hit the "tune" button on the autotuner between the sBitx and the amplifier.? ?After only a few dits and dashes -- shorted PA Mosfet, with current limiting by the power supply and huge current indicated on an external current meter.
?
I'm guessing I created a situation that resulted in either damaging voltage or current on the 55V IRFZ24N device.? ?Voltage might be more suspect?? ?The MOSFETS were heatsinked using Aluminum nitride pads with goo, and had ben installed a year or more previously and no problems ever since until this moment.? ?
?
Definitely a learning moment for me.? ?This is the FIRST MOSFET in a PA that I've ever had go south.? ?I don't operate hours and hours like others here; I'm more of a tinkerer who just puts things together, makes a few contacts, does some winlink and spends an hour or more just listening to relatively high speed CW in the early morning for practice.? ? But I've spent many hours and hours diagnosing problems with various units.? ?Maybe the IRF510 is indeed tougher due to its higher drain voltage rating?? Dunno.
?
Gordon KX4Z
?
|
Gordon,
I don't have a Zener or suppressor on the gates of the FETs. I think about this differently and only limit the input power for protection.
My power supply is a current limiting and has a reverse voltage when it reaches and exceeds the set current.
I had to change FETs when I wasn't using it for broadcasting, because I didn't have an antenna, but it was powered by a battery and I only used it for monitoring (there was no current limitation). I don't remember the details, maybe it was on a dummy load in a setting.
That's why I mentioned that the sbitx is a male egg that needs to be taken care of. I made another modification around the power stage, which is more complicated. I did it because I couldn't do the Bias setting correctly.
The current consumption always jumped even with a small adjustment of the potentiometer.
That's why I built Didier's circuit control of the F5NPV.
The FETs are not connected to ground directly, but through an RC element (0.2 Ohm parallel 220n).
Also, the Bias voltage is temperature dependent, controlled by 6 diodes connected in series.
Yes, this is not easy to replicate, which is why I deleted my previous comment. I am currently using the device this way.
I bought 10 of the FETs and selected them with a transistor meter for the closest opening voltage, which I installed.
It worked for me. Everyone should decide for themselves how to use the device.
My suggestion of AlN (aluminum nitride) for installation was only made after the thermal photographs appeared.
--
Gyula HA3HZ
|
ps,: reverse voltage when it reaches and exceeds the set current the reverse bias: it does not reverse the voltage, but reduces it to the extent that the current consumption exceeds it.
--
Gyula HA3HZ
|
?
More thoughts on this IRFZ24N destruction.
?
1.? ?It occurred with just 3 or 4 dits / dahs with "drive" set to 68% on 80 meters -- but with the intervening tuner in an indeterminate state into the ancient SB200 amplifier.? ?
2.? ?There was no indication of a malfunction of the SB200 amplifier (a plate arc would easily explain damage to the exciter)
3.? ?The IRFZ24N is rated at up to 17A current -- which the little switching supply that I used, cannot produce.? ? I think it is rated at 12 Amps or so at best, maybe less.? ?The first indication was a total shutdown -- so the supply went into current overlimit shutdown.? ??
?
4.? The IRFZ24N is rated for 55V drain-to-source.? ? That particular device had huge 5W zeners from gate to ground, with up to 1000 pf of capacitance (so I didn't get much power at 10 meters, (laugh) ).? ??
?
5.? The best explanation appears to be over voltage on the drain due to some oddity of the un-tuned intermediate auto-tuner state.? ?Over current seems less likely since the supply can't really generate enough.
?
6.? ?Based on this failure, Ashhar's move back to 100V-rated parts seems quite reasonable from a manufacturer point of view.? However, for MY particular interests,? I note that there are available IRF520 parts? (available on Mouser)? ?with similar input capacitance to the IRFZ24N, and I could certainly replace the 5W zeners and their 1000pf capacitance (!)? with the 3pf TVS diodes that just arrived from Mouser.? [There are other IRF520's that have more input capacitance and questionable V dss ratings.....not so good....]? ? ? ?I might then have significantly more "headroom" for power? on 80/40/30 than you can get with the IRF510, and be comfortable around 30 W power output with good linearity for digital operations.? ?I like being well within "max" power for linearity.? ? ?And maybe less risk of suddenly blowing a MOSFET due to a poor termination impedance.
?
Gordon KX4Z
?
?
?
?
?
|
There are things for which there is no reasonable explanation. The proof of the pudding is in the eating, so you have to try the idea.
--
Gyula HA3HZ
|
I have done some informal tests into a 500 watt dummy load to see how well the power output will hold up on the IRF520 and I can get 55 watts continuous?on FT8 (on 80 mtrs) as long as I am willing to transmit without failure, but after about 5 minutes the power starts to fall off, I assume from the heat, and ends up at about 50 watts. In practice this is probably a poor signal so?I dont operate here. It is too high for the other parts too as I found out the hard way when it killed one of the filter switching diodes on me. These mosfets are almost bomb proof when biased properly and operated under 40 watts. I am glad Ashar did the hardware change to use them, I have operated FT8 in POTA?parks at SWRs of 2.3:1 and had no damage, the IRF24N would never tolerate that... You and Ashar are two of the best resources this email reflector has and I for one appreciate all your efforts to share your knowledge?with us. Thank you! David WK4DS On Fri, Jan 10, 2025 at 8:56?PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote:
?
More thoughts on this IRFZ24N destruction.
?
1.? ?It occurred with just 3 or 4 dits / dahs with "drive" set to 68% on 80 meters -- but with the intervening tuner in an indeterminate state into the ancient SB200 amplifier.? ?
2.? ?There was no indication of a malfunction of the SB200 amplifier (a plate arc would easily explain damage to the exciter)
3.? ?The IRFZ24N is rated at up to 17A current -- which the little switching supply that I used, cannot produce.? ? I think it is rated at 12 Amps or so at best, maybe less.? ?The first indication was a total shutdown -- so the supply went into current overlimit shutdown.? ??
?
4.? The IRFZ24N is rated for 55V drain-to-source.? ? That particular device had huge 5W zeners from gate to ground, with up to 1000 pf of capacitance (so I didn't get much power at 10 meters, (laugh) ).? ??
?
5.? The best explanation appears to be over voltage on the drain due to some oddity of the un-tuned intermediate auto-tuner state.? ?Over current seems less likely since the supply can't really generate enough.
?
6.? ?Based on this failure, Ashhar's move back to 100V-rated parts seems quite reasonable from a manufacturer point of view.? However, for MY particular interests,? I note that there are available IRF520 parts? (available on Mouser)? ?with similar input capacitance to the IRFZ24N, and I could certainly replace the 5W zeners and their 1000pf capacitance (!)? with the 3pf TVS diodes that just arrived from Mouser.? [There are other IRF520's that have more input capacitance and questionable V dss ratings.....not so good....]? ? ? ?I might then have significantly more "headroom" for power? on 80/40/30 than you can get with the IRF510, and be comfortable around 30 W power output with good linearity for digital operations.? ?I like being well within "max" power for linearity.? ? ?And maybe less risk of suddenly blowing a MOSFET due to a poor termination impedance.
?
Gordon KX4Z
?
?
?
?
?
|
David,
?
Never say never.
You say 50W is not enough on 80 meters. On dummy load, it is difficult to make QSOs, but try it with a well-sized antenna. I still use the v2, which has the IRFZ24N that I replaced and has more than 7700 QSOs on 3 bands: 7, 14, 21 MHz, which I use with a single antenna and without an ATU. Currently, the dxcc/lotw cfm is 163/147, which I achieved with a maximum output power of 20 watts. Of course, I modified a few things on the original v2, but the most important thing is that I use it with a current-limiting power supply. This power supply regulates the output voltage downwards above the set current. Thanks to this, the FETs of the final stage were not damaged. This required a little experience (I have 74 years behind me) and technical solutions.
So don't say never, because the only constant is change and we live in a changing world.
--
Gyula HA3HZ
|
On Sat, Jan 11, 2025 at 04:08 AM, WK4DS wrote:
I have done some informal tests into a 500 watt dummy load to see how well the power output will hold up on the IRF520 and I can get 55 watts continuous?on FT8 (on 80 mtrs) as long as I am willing to transmit without failure, but after about 5 minutes the power starts to fall off, I assume from the heat, and ends up at about 50 watts. In practice this is probably a poor signal so?I dont operate here. It is too high for the other parts too as I found out the hard way when it killed one of the filter switching diodes on me. These mosfets are almost bomb proof when biased properly and operated under 40 watts. I am glad Ashar did the hardware change to use them, I have operated FT8 in POTA?parks at SWRs of 2.3:1 and had no damage, the IRF24N would never tolerate that...
You've got a lot more experience than I have!? ?Ashhar has probably poured hundreds of hours into the reliability issue and he knows a lot lot more than I ever will.? ?All I want is a bit of "headroom" and a modest amount of power for digital operations or CW.? ?I've made WINLINK connections with uBitx's with only 4 watts in good conditions, so amazing things can often be accomplished.
?
What DID surprise me was the difference between different manufacturers' versions of IRF520.? ? Perhaps stupidly, I ordered some generics from Amazon.? ?Then I noticed that one datasheet I was examining for IRF520 cited a breakdown drain-source voltage of only SIXTY volts.? ? I think it has to do with the reverse-biased "body diode"? ? I looked at it multiple times, and sure enough, that is what it cited as "minimum."? ?The same manufacturer cited an input (gate) capacitance of nearly 500 pf -- much more than the Vishay device, which is stated to have 100V drain-source breakdown and 360pf input capacitance.? ? (I havent thought hard enough to consider the miller capacitance, sorry Allison!!)? ?
? ? So I think I'm going to give that device a try next.? ? I ordered some from Mouser.? ?Have to wait for them to arrive.? ? ?In the process I did some reading and better understood the use of the "linear" portion of the device characterization for a transmitter.? ? It is very understandable why Ashhar went back to the tried-and-true IRF510 for the commercially available kit.? ?
?
With all the trials and experiments and learning, this transceiver has been a huge benefit to me.? ? It might not be the best choice for an "appliance operator" -- but it has been excellent for me, and this group has taught me a TON.? ? What fun learning!
?
Gordon KX4Z
?
|
The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
I am curious to see what you get too. I am curious because you are using Vishay devices and I am using Bojack devices. Mine are also not matched at all, I just grabbed two out of the bag and soldered them into place. I have a 4 amp current limiting power supply I built back in trade school that I used to setup?my radio. That was how I stopped the avalanche current problem?on mine.?
Gyula, I think?there was a miscommunication between us earlier. I didn't mean for that to sound like I needed 50 watts. I meant that was the test output into the dummy load to stress test the mosfets. I am sorry I misled you to think otherwise.
David WK4DS On Sun, Jan 19, 2025 at 7:42?AM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote:
The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
I have concluded that the irfz24n is blowing up due to over voltage at the gate.
The high voltage was due to self oscillations when the rx bypass of pa was in effect providing a feedback path from the antenna to the PA input.
We also had too much gain at the lower frequencies that allowed very high RF levels. This was due to the unfortunate use of 2N2219s in the driver chain that runs out of steam at higher frequencies and had to be run at high current to get better gain.
On Sun, Jan 19, 2025, 6:12 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
Thanks, Ashhar
I guess that’s possible. ?It would mean the Zener’s that I had on the gate at that point failed. I’ve previously demonstrated (with oscilloscope) that they clip and at very high frequencies their enormous capacitance should reduce gain. ?
The failure I had was with the band set to 80 m and Unknown load because I hadn’t properly tuned the auto tuner. ?So I don’t know what load impedance the system saw.
I’ll have to go back and look and see if I had reduced the delay times before the failure— if so, that would be powerful data to support your thesis!!! (and I would add back in some of the delay times!)
Gordon?
toggle quoted message
Show quoted text
On Jan 19, 2025, at 08:28, Ashhar Farhan via groups.io <farhanbox@...> wrote:
? I have concluded that the irfz24n is blowing up due to over voltage at the gate.
The high voltage was due to self oscillations when the rx bypass of pa was in effect providing a feedback path from the antenna to the PA input.
We also had too much gain at the lower frequencies that allowed very high RF levels. This was due to the unfortunate use of 2N2219s in the driver chain that runs out of steam at higher frequencies and had to be run at high current to get better gain.
On Sun, Jan 19, 2025, 6:12 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
I went back through my posts and the failure of the Z 24N occurred before I had implemented any changes in the CW code.
The first change that I reported was only increasing the sampling time by 20 times. Apparently I did that after the failure of the MOSFET. ? By itself, I don’t think that would change the TR delays. ?
only subsequent to that change, did I start removing some of the delay times; so I haven’t been able to test those with a working power amplifier
So maybe I’ll put in a few of those delays with these new MOSFETs??
Gordon kx4z?
toggle quoted message
Show quoted text
On Jan 19, 2025, at 08:54, Gordon Gibby KX4Z via groups.io <docvacuumtubes@...> wrote:
? Thanks, Ashhar
I guess that’s possible. ?It would mean the Zener’s that I had on the gate at that point failed. I’ve previously demonstrated (with oscilloscope) that they clip and at very high frequencies their enormous capacitance should reduce gain. ?
The failure I had was with the band set to 80 m and Unknown load because I hadn’t properly tuned the auto tuner. ?So I don’t know what load impedance the system saw.
I’ll have to go back and look and see if I had reduced the delay times before the failure— if so, that would be powerful data to support your thesis!!! (and I would add back in some of the delay times!)
Gordon?
On Jan 19, 2025, at 08:28, Ashhar Farhan via groups.io <farhanbox@...> wrote:
? I have concluded that the irfz24n is blowing up due to over voltage at the gate.
The high voltage was due to self oscillations when the rx bypass of pa was in effect providing a feedback path from the antenna to the PA input.
We also had too much gain at the lower frequencies that allowed very high RF levels. This was due to the unfortunate use of 2N2219s in the driver chain that runs out of steam at higher frequencies and had to be run at high current to get better gain.
On Sun, Jan 19, 2025, 6:12 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
Gordon,
The zeners dont work. I measured this, the zeners cant switch that fast. You need 1n4148s two in series to the ground and two in series to the 12v. A classic clampiny circuit.
-? f
On Sun, Jan 19, 2025, 7:24 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: Thanks, Ashhar
I guess that’s possible.? It would mean the Zener’s that I had on the gate at that point failed. I’ve previously demonstrated (with oscilloscope) that they clip and at very high frequencies their enormous capacitance should reduce gain. ?
The failure I had was with the band set to 80 m and Unknown load because I hadn’t properly tuned the auto tuner.? So I don’t know what load impedance the system saw.
I’ll have to go back and look and see if I had reduced the delay times before the failure— if so, that would be powerful data to support your thesis!!! (and I would add back in some of the delay times!)
Gordon?
? I have concluded that the irfz24n is blowing up due to over voltage at the gate.
The high voltage was due to self oscillations when the rx bypass of pa was in effect providing a feedback path from the antenna to the PA input.
We also had too much gain at the lower frequencies that allowed very high RF levels. This was due to the unfortunate use of 2N2219s in the driver chain that runs out of steam at higher frequencies and had to be run at high current to get better gain.
On Sun, Jan 19, 2025, 6:12 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
toggle quoted message
Show quoted text
On Jan 19, 2025, at 10:27, Ashhar Farhan via groups.io <farhanbox@...> wrote:
? Gordon,
The zeners dont work. I measured this, the zeners cant switch that fast. You need 1n4148s two in series to the ground and two in series to the 12v. A classic clampiny circuit.
-? f
On Sun, Jan 19, 2025, 7:24 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: Thanks, Ashhar
I guess that’s possible.? It would mean the Zener’s that I had on the gate at that point failed. I’ve previously demonstrated (with oscilloscope) that they clip and at very high frequencies their enormous capacitance should reduce gain. ?
The failure I had was with the band set to 80 m and Unknown load because I hadn’t properly tuned the auto tuner.? So I don’t know what load impedance the system saw.
I’ll have to go back and look and see if I had reduced the delay times before the failure— if so, that would be powerful data to support your thesis!!! (and I would add back in some of the delay times!)
Gordon?
? I have concluded that the irfz24n is blowing up due to over voltage at the gate.
The high voltage was due to self oscillations when the rx bypass of pa was in effect providing a feedback path from the antenna to the PA input.
We also had too much gain at the lower frequencies that allowed very high RF levels. This was due to the unfortunate use of 2N2219s in the driver chain that runs out of steam at higher frequencies and had to be run at high current to get better gain.
On Sun, Jan 19, 2025, 6:12 PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote: The MOUSER order for?
?
942-IRF520NPBF 12 12 0 0.592 7.10
?MFG Part No: IRF520NPBF
?Infineon Technologies MOSFT 100V 9.7A 2
?
IRF520's from "infineon" arrived a few days ago.? ?They are marked "IR"? (International Rectifier) on the body.? ?I managed to get them installed over Aluminum nitride insulators and got the idling current set for about 0.25-0.30 A earlier today.
?
1.? Getting the tab to easily go FLAT on the heatsink proved to be more difficult than I had expected.? ?I think I didn't allow enough "slack" in the device leads when I soldered it and that caused it to be pushed into an odd angle that meant a lot of force had to be supplied by the "bar" to hold them down.? ?I didn't like that and slept on it to think it through better.
?
2.? I ended up re-heating the solder connections while the devices were held down by the "bar" to allow some movement of the leads? ?I think that helped.? ?I should have used a tiny bit more lead length.? ?
?
3.? (I'm still using the staked-in female screw jacks -- I haven't drilled through to add a back nut.? ?If they were to fail, I think I woud do that next time)
?
4.? Starting from FULL COUNTER CLOCKWISE, and using a small "wall wart" 12-VDC supply, I was able to set the idling current.? ?I try to always do this with a WALL WART with limited current capacity since I don't have a laboratory current limited power supply.? These are only few dollars and are WELL WORTH HAVING for non-destructive re-setup of one of these radios.? ? (For me, I also have a polarity protection circuit inline -- these are valuable radios and I hate the thought of destroying one with reverse polary.? ?Once I get the polar protection inserted, it never comes out, STAYS with the radio from then on.? ?)
?
5.? Based on the rotation of the potentiometer, a slightly higher (more positive) bias gate voltage was needed on the IRF520 than with the previous IRFZ24N.? ? That's just differences in the devices.? ??
?
6.? I have Ashhars recommended bi-directional TVS diodes (with miniscule 3pf capacitance) in place.? ?Soldering those with my tools was quite a feat!!? ? They have astonishingly low junction capacitance compared to what I see for typical axial leaded devices.? ?
Later today I should be able to start making some power measurements and resetting the hardware coefficients for the various bands.? ? I bet I'll see more power on the higher frequencies with the significantly lower junction capacitances.? ??
?
Gordon KX4Z
?
|
V2 circuitry with gate inductors and resistors/? ?blown IRVZ24n, both replaced with IRF520
?
OK,? I don't have the patience of Ashhar, but I have managed to set up my original V2 radio now with IRF520 MOSFETS from "Infineon" (marked International Rectifier) so that it basically achieves a linear near-maximum output power at 100% drive.? I have an old radio shack swr meter that has a voltage-derived power meter that is pretty "close" after calibration except apparenlty at 40 meters where it is a bit optimistic.? Using that meter I developed coefficients that appeared workable (though perhaps not absolute maximum) and then checked with a siglent digital scope and 10X probe with on-screen measurement of RMS voltage on a Heathkit ancient 50 ohm cantenna and converted to power.? ?The sBitx internal reading (which I have separately calibrated long ago) is similar
?
IRF520? Infineon? ?biased for about 0.3Amp total idling current
3pf TVS diodes on gates instead of the previous huge zeners (which Ashhar says don't stop the problems)
MODE = CW? ?Power measured by Siglent scope and mathematically voltage -> 50 ohm power
Frequency? ? 100% drive output power? ? hw_settings.ini? ? Comment
3.5010? ? ? ? ? ? ? ?32W? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? .0037? ? ? ? ? ? ?Gets a little less power than on 40, similar setting to IRFZ24N
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? my earlier notes suggest I got a little more power than this, with the IRFZ24N's
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? and gate zeners, so the IRF520 may be a little inferior here.
?
7.010? ? ? ? ? ? ? ? ? 33.3? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?.0028? ? ?
10.1? ? ? ? ? ? ? ? ? ? 22.6W? ? ? ? ? ? ? ? ? ? ? ? ? ? ? .0028? ? my notes indicate I was in the range of 21W with IRVZ24N & gate-zeners
14.01? ? ? ? ? ? ? ? ? 16W? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?.0032? ? my notes here on IRFZ24N don't make sense
18.1? ? ? ? ? ? ? ? ? ? ?9.33? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?.0061? ? ? ? ? ? ?? the gate circuit reactance issue here?
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?notes indicate I got 11-12W with IRVZ24N's
21.01? ? ? ? ? ? ? ? ? 16.6? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? .0054? ?notes indicate I could only get 15W with IRFZ24N & gate-zeners
24.9? ? ? ? ? ? ? ? ? ? ?12.3? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?.0063? ?notes indicate I could only get 10W with IRFZ24N & gate-zeners
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ?likely due to their huge capacitance
28.0? ? ? ? ? ? ? ? ? ? ?11.52W? ? ? ? ? ? ? ? ? ? ? ? ? ? .0063? ?notes indicate I could only get 8W with IRFZ24N & gate-zeners
? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? ? likely due to their huge capacitance
?
?
From Ashhar's earlier data, these are no particularly huge improvement or decline from what he is able to get with IRF510's, so his choices make excellent sense from a manufacturers' reliability standpoint.
?
Hope this information is useful to tinkerers who consider the higher Drain-Source IRF520 with the oriiginal IRFZ24N circuitry of the V2 version
?
Gordon Kx4z
?
?
?
From MY standpoint I like this because I do a lot of NVIS and other work on lower bands and prefer to be able to drive a linear amplifier with this rig and/or use digital waveforms on WINLINK etc.? ?With 32-odd watts I can derate a bit? and still get good power with likely good linearity.? ?Ashhar's choices are better researched and possibly even better in ways I don't even understand.
?
Because this is now so "new" since installing those MOSFETs I did not dare to do any "brick on key-down" tests.? ? The IRF520 is said to have a slightly lower junction-to-tab thermal resistance, and the aluminum nitride is a significant improvement over mica but my physical installation may not be as good as theirs, and CONTACT is the most important item other than the insulator.
?
I'm just pleased as punch to have a working V2 version again -- and several bags of good MOSFETS now available and the itty bitty TVS diodes.? ?I had bought a SECOND unit (secondhand also) and will need to tear into it to change the insulators etc.? ?In all the time I've had this sBitx this is the FIRST TIME I've ever blown a final.? ?I had a uBitx V3 and never blew those either, I think.? ?Going to 100V D-source mosfets may be a real winner and for the older V2 driver, using the TVS diodes also likely a winner.??
?
For these test's I reset the T/R switching delays to intermediate values (not right in front of me) about 2/3 of what Ashhar had.??
The CW? is set to do 20 checks/sec on the key.? ?A straight key was used to key the unit for power testing.? ??
?
?
?
?
|
While I had the repaired V2 sBitx (now with IRF520 mosfets) on the scope and dummy load, I check several different delay timings in the T/R software.? ?Ashhar has demonstrated some damaging oscillations when the T/R circuit allows for brief positive feedback.? ?I couldn't remember how all this has been remedied, but apparently it has -- I sent very fast "dits" on bands from 80M to 10M and could not see ANY signs of high power oscillations.? ? If the TVS diodes had clamped an oscillation, I should have still seen a muted version (I think) -- and I didn't see ANYTHING, even with 20mSec delays reduced to 5mSec.? ? So the suggested T/R fixes for making the first "dit" of CW more normal appear reasonable (so far).? ?Photos to follow when I get more data.
?
I did see some very low-level "crud" on the oscilloscope while the transmitter chain was active.? ?I haven't measured it further.??
?
Gordon KX4Z
?
?
|
HA3HZ
WK4DS