|
I already ordered AlN insulators from Aliexpress. They should arrive in two weeks.
?
73,
Wojtek SP5DAA
?
|
That’s great news! I would strongly suggest switching to the aluminum nitride insulators
Gordon kx4z
toggle quoted message
Show quoted text
On Mar 1, 2025, at 19:13, WZab - SP5DAA via groups.io <wzab01@...> wrote:
? Yes! Replacing the MOSFETs fixed my sBitx v3.
So I have learned two lessons:
- Running continuously in FT8 mode requires either reducing DRIVE or using a fan to improve cooling MOSFETs (especially if running at 13,8V power supply voltage).
- MOSFETs failure due to overheat do not necessarily cause shorting drain to source.
Thanks a lot to all who helped me to isolate and fix the problem.
?
73,
Wojtek
|
Yes! Replacing the MOSFETs fixed my sBitx v3.
So I have learned two lessons:
- Running continuously in FT8 mode requires either reducing DRIVE or using a fan to improve cooling MOSFETs (especially if running at 13,8V power supply voltage).
- MOSFETs failure due to overheat do not necessarily cause shorting drain to source.
Thanks a lot to all who helped me to isolate and fix the problem.
?
73,
Wojtek
|
Well, let's wait with celebrating until I replace the MOSFETs and find the device working properly.? What annoys me is that I didn't detect that parasitic "diode" between gate and source at the first measurements after disassembling the transceiver. However, maybe the damage was increasing while I tested it.
73, Wojtek? sob., 1 mar 2025, 12:40 u?ytkownik Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> napisa?:
Delighted that you found the problem. Interesting that both drains showed an AC voltage even though one device was effectively shorted.? That is surprising to me.
Wish you the best
Gordon kx4z?
? Wojtek,
Buy 10 pieces so that you can choose the two with almost the same opening voltage.
They may be the same in shipping, but there is still a difference between them.
You can install them without Aluminum nitride with the old mica, and use some kind of fan to cool the heat sink from the back.
Read another message where I wrote how I use the v2.
You can easily replace the AlN ceramic later if you look at which screw you need to unscrew to remove the heat sink.
Read and do the Bias adjustment as described here (you can also find the article on ).
?
It is important to do the adjustment with a 2A current limit!
--
Gyula HA3HZ
|
Delighted that you found the problem. Interesting that both drains showed an AC voltage even though one device was effectively shorted. ?That is surprising to me.
Wish you the best
Gordon kx4z?
toggle quoted message
Show quoted text
On Mar 1, 2025, at 06:38, HA3HZ via groups.io <gyula@...> wrote:
? Wojtek,
Buy 10 pieces so that you can choose the two with almost the same opening voltage.
They may be the same in shipping, but there is still a difference between them.
You can install them without Aluminum nitride with the old mica, and use some kind of fan to cool the heat sink from the back.
Read another message where I wrote how I use the v2.
You can easily replace the AlN ceramic later if you look at which screw you need to unscrew to remove the heat sink.
Read and do the Bias adjustment as described here (you can also find the article on sbitx.net).
?
It is important to do the adjustment with a 2A current limit!
--
Gyula HA3HZ
|
Wojtek,
Buy 10 pieces so that you can choose the two with almost the same opening voltage.
They may be the same in shipping, but there is still a difference between them.
You can install them without Aluminum nitride with the old mica, and use some kind of fan to cool the heat sink from the back.
Read another message where I wrote how I use the v2.
You can easily replace the AlN ceramic later if you look at which screw you need to unscrew to remove the heat sink.
Read and do the Bias adjustment as described here (you can also find the article on sbitx.net).
?
It is important to do the adjustment with a 2A current limit!
--
Gyula HA3HZ
|
I would wait on the insulators myself... They help tremendously to be honest.
?
David - WK4DS
|
I have finally disassembled my sBitx fully and removed final MOSFETs.? I have to cut the away to not damage the PCB. One of the was still alive (so I'll probably use it for prototyping):
?
But the second is broken (probably due to overheating):
?
So it looks like a fan is the "must have" if one wants longer continuously work in digital modes.
?
Now I have to decide if I should mount the new MOSFETs and reassemble the device as it is, or wait until the new aluminum-nitrid isolators arrive.
?
I'm also afraid about the PCB state when replacement of finals may repeate multiple times. It seems that the pads of Q16 and Q17 are quite delicate. If they are gradually damage during replacements, at ID reaching 4 A (or even more in a pulse) the PCB may gradually deteriorate and finally fail.
|
Hello Evan,
I tried that yesterday, and found that the resistance remains below 1 k. Another fact was that the DC voltage on gates with power, at minimal drive and PTT on didn't change as expected.? That's why I asked if the MOSFETs can be partially damaged, so that they don't get shorted, but their gate to source resistance is reduced. Another possibility is that TVS D8 and D9 got thermally damaged.?
One thing that I don't understand, is why the AC voltage on the output of the transformer is lower than on the input.? However, if something overload the transformer's output, the current consumption should increase, not decrease. I experience the drop of the current consumption. It was up to 6 A on 40 m, and up to 3A on 10 m. Now it increases to ca. 1A only. That supports the theory that the problem results from the damage of final MOSFETs.?
Anyway, it seems that I have to disassemble my device fully, remove the MOSFETs and test them.
73, Wojtek?
sob., 1 mar 2025, 03:07 u?ytkownik Evan Hand via <elhandjr= [email protected]> napisa?:
Wojtek,
?
I suggest verifying that adjusting the PABias1 pot does change the voltage on the MOSFETs' gates.? As Gordon stated, you need to be careful when making this adjustment.? It is very easy to drive the MOSFETs to destruction, and the current will "jump" as you get close to the nominal setpoint.
?
A safe way to test the bias control is to measure the resistance from the MOSFETs' gates to ground without power connected.? You should be able to vary from almost zero ohms (55 ohms) in the fully counterclockwise direction to nearly 10k ohms when fully clockwise.? Be sure to return the control to the fully counterclockwise position before restoring power.
?
In addition to a faulty control, one of the bypass capacitors may be shorted.
?
73
Evan
AC9TU
|
Wojtek,
?
I suggest verifying that adjusting the PABias1 pot does change the voltage on the MOSFETs' gates.? As Gordon stated, you need to be careful when making this adjustment.? It is very easy to drive the MOSFETs to destruction, and the current will "jump" as you get close to the nominal setpoint.
?
A safe way to test the bias control is to measure the resistance from the MOSFETs' gates to ground without power connected.? You should be able to vary from almost zero ohms (55 ohms) in the fully counterclockwise direction to nearly 10k ohms when fully clockwise.? Be sure to return the control to the fully counterclockwise position before restoring power.
?
In addition to a faulty control, one of the bypass capacitors may be shorted.
?
73
Evan
AC9TU
|
Those two diets you were talking about are not normal diodes they are transient voltage suppressors. They won’t conduct until the voltage gets around 10 V.
?
Be very careful with that PA bias. You can easily destroy the transistors with that. ?Turning it too far will cause the bias current to go to many amps!!
?
The strange thing is that performing the bias regulation (as described on the website) I couldn't get the increase by 200 mA in the whole range of regulation.
May the transistors be partially damaged?
?
73,
Wojtek
|
Those two diets you were talking about are not normal diodes they are transient voltage suppressors. They won’t conduct until the voltage gets around 10 V.
Be very careful with that PA bias. You can easily destroy the transistors with that. ?Turning it too far will cause the bias current to go to many amps!!
There is some odd residue near the input of your output transformer. ?
toggle quoted message
Show quoted text
On Feb 28, 2025, at 19:39, Gordon Gibby KX4Z via groups.io <docvacuumtubes@...> wrote:
? ?
?
be certain that you have correctly identified the PA bias potentiometer. It would be very odd if turning that too far didn’t destroy the MOSFETs
?
when you press push to talk the voltage on L 80 and L 81 should go up— because it is controlled by Q9. if Q9 is continuously on, then you have blown a different MOSFET, one or more in that complicated switching system around Q9. ?You might want to figure out for sure if you really can’t see switching action going on there.
?
?
?
|
?
?
be certain that you have correctly identified the PA bias potentiometer. It would be very odd if turning that too far didn’t destroy the MOSFETs
?
when you press push to talk the voltage on L 80 and L 81 should go up— because it is controlled by Q9. if Q9 is continuously on, then you have blown a different MOSFET, one or more in that complicated switching system around Q9. ?You might want to figure out for sure if you really can’t see switching action going on there.
?
?
?
|
In my board I have additional diodes D8 and D9 connected between the gates of Q16 qnd Q17 and the ground.
It seems that the voltage delivered by the bias circuit is limited by the those diodes to 270 mV.
However, when I set the DRIVE to 100, the voltage on the gates looks like in the picture below (probe x10 already considered):
?
?
So it does not seem to be limited by the diode (GND level is on the central line of the screen).
?
The picture of the board is below. The D8 and D9 diodes are above Q16 and Q17, next to L80 and L81 (which are in fact just jumpers):
|
I have one interesting finding. It seems that changing the PA_BIAS1 setting does not modify the current consumed by sBitx.
After pressing PTT It increases from 0.31A to 0.77A, and does not depend on further changes in the potentiometer setting.
The voltage on L80 and on L81 does not increase significantly when I push PTT in the USB mode (or press the key in the CW mode).
The voltage on the input of the potentiometer (node between R77 and D6) is correct.
Namely, the potentiometer is faulty or improperly soldered...
?
|
If C202 (47pf) SHORTED, you would expect potentially what is being seen. If the 80 meter filter were invoked, and your frequency is above that, the input of the filter might act like a short. If D15 shorted, this could happen, or if Q24 shorted this could happen If the teflon insulation on the orange wire melted (this is not likely) then the output of the transformer might be shorted If the antenna output connector or coax shorted, this low output might happen.
But chasing down every one of those is very time consuming.? ?Again, it might be easier to just look at what is coming out of the orange wire, disconnected from everything else.....much simpler than all the other techniques.
Anyone see something better?
Gordon KX4Z
On Fri, Feb 28, 2025 at 4:48?PM Gordon Gibby KX4Z via <docvacuumtubes= [email protected]> wrote:
Good point Evan!
But I can’t figure out is how disconnecting a Lowpass filter would result in a lower voltage. It should result in a higher voltage. ?
I tried to look for what things might be shorting out. There’s a chance that the high-voltage system is shorting out but it’s only connected by 47 pf?
?and that’s not much. ?
Gordon?
? It should be possible to test the LPF selection process with a DMM.? Use USB, set the drive to 1 and the mic to 0.? Then, measure the voltages on the banded sections of the selection diodes.? Only one pair should be close to zero; the others should be close to the input voltage (12 volts) without the RF voltage doubler.
?
73
Evan
AC9TU
|
Good point Evan!
But I can’t figure out is how disconnecting a Lowpass filter would result in a lower voltage. It should result in a higher voltage. ?
I tried to look for what things might be shorting out. There’s a chance that the high-voltage system is shorting out but it’s only connected by 47 pf?
?and that’s not much. ?
Gordon?
toggle quoted message
Show quoted text
On Feb 28, 2025, at 16:35, Evan Hand via groups.io <elhandjr@...> wrote:
? It should be possible to test the LPF selection process with a DMM.? Use USB, set the drive to 1 and the mic to 0.? Then, measure the voltages on the banded sections of the selection diodes.? Only one pair should be close to zero; the others should be close to the input voltage (12 volts) without the RF voltage doubler.
?
73
Evan
AC9TU
|
It should be possible to test the LPF selection process with a DMM.? Use USB, set the drive to 1 and the mic to 0.? Then, measure the voltages on the banded sections of the selection diodes.? Only one pair should be close to zero; the others should be close to the input voltage (12 volts) without the RF voltage doubler.
?
73
Evan
AC9TU
|
“Is it possible that the software got crazy and didn't enable any section in the LPF filter?
73,
奥辞箩迟别办”
It’s unlikely that the software is doing anything wrong. It’s more likely that some piece of hardware has a problem in order to figure that out. We have to divide this somewhere and decide whether the problem is on the left side or the right side. With your voltmeter, unable to measure DC voltage properly in the face of RF, this is going to be a little more difficult and may require the scope.
But the first problem is probably to split this into modules and figure out where the issue is : with the transformer or beyond that.
|
The below message was an answer to sa7buj.
As shown in the scope image, I have ca 4.5V peak-to-peak there when transmitting.
-- Wojciech M. Zabo?otny
My GPG/PGP keys:
standard: 8192R/7E17F00A (B53C 1F09 BEA9 ED46 51B6 ?2C2E 4344 F292 7E17 F00A)
confidential: 16384R/1FEDB03C (3412 2330 E225 D57D 70C3 ?2340 B294 9408 1FED B03C)
|
HA3HZ
WK4DS