Very low output power (0.1W)

Post Reply
S53SL
 

Posts: 9
Joined: 31 Mar 2022, 21:24

Very low output power (0.1W)

Post by S53SL »

Hi again folks,

in this topic I have mentioned I had a problems with efficiency. And at that time I measured that:
S53SL wrote: 31 Mar 2022, 21:42 My measurements:
80m: 1.2W 25%
60m: 1.8W 40%
40m: 2.8W 40%
30m: 1.5W 32%
20m: 1.9W 41%
Then someone pointed out I have too many windings on T1, so I removed one and measurements and output power were the almost same. Then I noticed FET Q3 was overheating, so I replaced it with another BS170 I had by hand. When I did that, I noticed that measurements has worsened significantly. I started to measure 0.06W to 0.10W on the output. I then bought new BS170 and replaced all 3 of them. And I still get around 0.10W on the output. It's same on all bands or worse. I even rewound the T1 coil, as it has sharp inner edge and I thought I have damaged the insulation on copper wire. This haven't helped either.

I then proceed to measure what's going on on oscilloscope.
Probes
Probes
IMG_20220409_173317.jpg (228.91 KiB) Viewed 3319 times
Measurement:
Oscilloscope
Oscilloscope
IMG_20220409_173614.jpg (163.74 KiB) Viewed 3319 times
Vpp in ANT_OUT is only 8.48V
So the output is around 0.10W with around 5% efficiency. I notice now that FET Q2 is overheating when I go on transmit.
Heating
Heating
2022-04-09-17-42-24.jpg (52.83 KiB) Viewed 3319 times
What do you advise me to do? What more can I measure or change? It seems there is some problem with FETs but I don't know what. This new ones are supposed to be BS170 from Fairchild.

Thank you for your advices.
73,
S53SL Gregor
73 de S53SL
AC9YY
 

Posts: 133
Joined: 21 Mar 2022, 00:04
Location: Illinois, USA

Re: Very low output power (0.1W)

Post by AC9YY »

Since all the FETs are in parallel and one is getting very hot it must be bad.
wa2t
 

Posts: 180
Joined: 01 Jan 2022, 16:30

Re: Very low output power (0.1W)

Post by wa2t »

Hi Gregor,

Just a thought/cross check, have you populated the two SMT FET's on the back side of the RF board? If yes, those are not to be installed at the same time as the BS170's.

73,

Robert, WA2T
S53SL
 

Posts: 9
Joined: 31 Mar 2022, 21:24

Re: Very low output power (0.1W)

Post by S53SL »

So... I have removed all the BS170 FETs... just to be sure. As it is almost impossible to resolder them because of large heat dissipation of ground pads, I came by solution to just solder wires through pads of one transistor and then solder all the FET transistors on that 3 wires.
transistors on wires
transistors on wires
IMG_20220409_200510.jpg (228.82 KiB) Viewed 3293 times
If someone will be redesigning the circuit, I recommend to make contacts to the ground thinner, so there is less heat dissipation to ground plate. Also there is via nearby that unnecessarily goes nearby, that could be moved further away (marked red). If you use copper braid to remove solder it is easy to damage this via, especially with all the heat needed to clear those holes.
vias and pads
vias and pads
IMG_20220409_210931.jpg (191.37 KiB) Viewed 3293 times


Anyway back to the story. After I did that contraption with FETs, I powered it up. It seemed it was working, it showed something over 5W on 3.5MHz. I've checked other also other bands, and I'm not sure when following happened. Either by switching and transmitting on some other band or when I turn it over and SMA connector unscrewed a bit. Suddenly I saw output is over 14W! I switched it off and power it again. I noticed that on power supply that truSDX draws 0.8-1W, on receive! When pressed transmit it was again over 14W. I could even hear interference in nearby computer speaker. This time all transistors were heating up... on receive. After about 30 sec they were above 100°C. I captured this:
thermal 100c
thermal 100c
2022-04-09-19-56-57.jpg (84.32 KiB) Viewed 3293 times
So I thought, I have probably blown those transistors as they constantly drew current.
Then I replaced them with another 3 new FETs ,which was this time easier due to wire contraption. I made sure connector was properly screwed this time and I squeezed Lx1 coils a bit in an attempt to reduce power. I powered it up and figured I retuned to square one... Power was this time less than 0.10W... Hm... ok something is not right. And then again I unsoldered the transistors and soldered 3 new ones... Same result, very low output power. When I turned on thermal camera this time, no transistors were heating up on receive and on transmit one of them start to heat up. Mind that these are another 3 new transistors. In total I replaced 12 transistors so far.

So again I don't know where I stand. It seemed to be working for a while, and then it stopped. At least I can now quickly replace transistors if needed.

Any new ideas what could it be?
73 de S53SL
AC9YY
 

Posts: 133
Joined: 21 Mar 2022, 00:04
Location: Illinois, USA

Re: Very low output power (0.1W)

Post by AC9YY »

Sounds like fake BS170s.
Faraaz VK4JJ
 

Posts: 215
Joined: 30 Dec 2021, 21:56

Re: Very low output power (0.1W)

Post by Faraaz VK4JJ »

S53SL wrote: 09 Apr 2022, 19:48 So... I have removed all the BS170 FETs... just to be sure. As it is almost impossible to resolder them because of large heat dissipation of ground pads, I came by solution to just solder wires through pads of one transistor and then solder all the FET transistors on that 3 wires.

Umm, I'm not sure that is the best/safest way to layer FETs like that?!

I have not have problems removing BS170s from the board. If you cannot desolder these FETs with relative ease, this is a sign you a.) need a higher power output iron, say at least 100W b.) a solder iron tip with a larger surface area (e.g. large chisel tip) c. more flux d. good solder sucker e. good wick.

Edit: actually, I'll straight-up say what I'm thinking. That is an INSANE way to install delicate FETs. You have gone through twelve FETs - I am not surprised in the slightest. Likely every single movement of this portable radio risks shorting your semiconductors out.

Please please - remove all of this, get a nice strong hot iron, a solder sucker, desolder carefully and properly and then start from scratch.
------------------------

Image

Also, it is difficult to see from the resolution of your image but I have annotated.

1. There appears to be specks of solder all over the board. It is not clear if this is just splatter (which should be cleared up) or areas of damage. I'm especially interested in the area marked '???'

2. There appears to be quite a few dry joints - suggest once you have removed the BS170 abomination, remove and refresh all those solder joints with plenty of flux and a hot, powerful iron.
DoubleHaul
 

Posts: 7
Joined: 28 Mar 2022, 23:59

Re: Very low output power (0.1W)

Post by DoubleHaul »

I was able to clear the two non-ground fet through holes with solder wick and flux. Once that stopped being effective I added some solder to the pad (counterintuitive I know), then hit it again with flux impregnated wick. It took a while to clear both holes on all three fets. Once all but the ground pads were clear I added some solder to the ground pad, then applied heat and pushed the ground lead of the new fet through first while the solder was liquid. I think that the key here is adding some solder so that the heat transfers all the way down through the old solder inside the through hole.
S53SL
 

Posts: 9
Joined: 31 Mar 2022, 21:24

Re: Very low output power (0.1W)

Post by S53SL »

I owe you an update. I've found a solution. But to recap all the process.

Unsoldering itself isn't much of a problem with powerful enough iron. More problematic is getting solder out of the holes, especially on ground pins, as it have tendency to get left in the middle layer. I could get most of the solder out from both sides, so there are dimples, but there was still solder left in the middle that I couldn't get out. Maybe there is a problem with type of solder I'm using or something else. I've also used plenty of flux. It helped to certain degree, but for those ground pads are difficult to manage. I've then used really small drill bit to mechanically remove that solder in the middle.
Faraaz VK4JJ wrote: 10 Apr 2022, 11:17 There appears to be specks of solder all over the board. It is not clear if this is just splatter (which should be cleared up) or areas of damage. I'm especially interested in the area marked '???'
Unfortunately that is solder mask that is damaged.
So I've managed to again unsolder and get rid of the solder from holes, at least for one of the transistors:
Image7523670586595650659.jpg
Image7523670586595650659.jpg (155.18 KiB) Viewed 2982 times
And then decided to take an approach of soldering in the header and then use protoboard to solder on FETs.
Image92355894843358271.jpg
Image92355894843358271.jpg (164.29 KiB) Viewed 2982 times
And again, results were bad. Almost no output.

Then I have removed all the FETs again, and measure them on component tester. 1 of them still showed it is FET, others showed either a diode or resistor. So I assumed those are broken ones. Then I went through other FET transistors that I have previously used (those 12) and found some that still showed that are functional FETs. I've then first soldered only one of those FETs and test the performance. It was giving around 1.5W on output. Then I added another one in parallel and now I got realistic 5-6W of output power. And then I added third and it was again almost zero output. Then I unsoldered that transistor again and measured it. And it was broken. That happened to another transistor. So I couldn't get it working with 3 transistors as third one kept blowing up.

So on the end I decided to keep only 2 transistors, as they were providing sufficient power. I've realised that they were heating up quite quickly. If I kept transmitting constantly for 30s they rose up to 80°C. So then I figured out I only need a way to dissipate that heat away. So I've used copper tape and tape it on FETs and then some thick copper strips to create a bit more area to dissipate heat.
Image739246513028135774.jpg
Image739246513028135774.jpg (160.15 KiB) Viewed 2982 times
It doesn't look nice, but it is functional. So that's my solution to this problem.
It seems it is hard to find matching FETs that don't destroy each other.

Note: power efficiency is now around 79% :)
Last edited by S53SL on 22 Apr 2022, 21:10, edited 3 times in total.
73 de S53SL
S53SL
 

Posts: 9
Joined: 31 Mar 2022, 21:24

Re: Very low output power (0.1W)

Post by S53SL »

And here is thermal image of what is heating up after 1 minute of constant transmitting.
Image1560731347123156521.jpg
Image1560731347123156521.jpg (116.67 KiB) Viewed 2978 times
I also found it interesting that SWR coil is heating up, but there is practically 1:1 SWR as I used dummy load.

On some coils I had to squeeze loops together as it was giving out too much power - over 7W.

So that's that. I still have to adjust filter coils using miniVNA, but I've managed to make two >500km long QSOs during testing.
73 de S53SL
sq9gin
 

Posts: 10
Joined: 03 Apr 2022, 09:52

Re: Very low output power (0.1W)

Post by sq9gin »

Hello, should these transistors be selected so that they have similar parameters (pair)?

Krzysztof
M1KTA
 

Posts: 26
Joined: 17 Jan 2022, 18:05

Re: Very low output power (0.1W)

Post by M1KTA »

S53SL wrote: 22 Apr 2022, 20:48
Unfortunately that is solder mask that is damaged.
So I've managed to again unsolder and get rid of the solder from holes, at least for one of the transistors:
Image7523670586595650659.jpg

And then decided to take an approach of soldering in the header and then use protoboard to solder on FETs.
Image92355894843358271.jpg
I understand experimentation perhaps but why was anything in Q4/Q5 on the bottom of the RF board at all when using Q1, Q2 and Q3 on the top side?
S53SL
 

Posts: 9
Joined: 31 Mar 2022, 21:24

Re: Very low output power (0.1W)

Post by S53SL »

M1KTA wrote: 23 Apr 2022, 14:19 I understand experimentation perhaps but why was anything in Q4/Q5 on the bottom of the RF board at all when using Q1, Q2 and Q3
Probably because you could solder SMD FET alternatives there in place of Q4 or Q5 instead of using Q1-Q3.
73 de S53SL
sq9gin
 

Posts: 10
Joined: 03 Apr 2022, 09:52

Re: Very low output power (0.1W)

Post by sq9gin »

Hello, self-assembly kits bought on aliexpress are a lottery, or you will hit a working or damaged :)

Krzysztof
Ohwenzelph
 

Posts: 207
Joined: 01 Jan 2022, 03:47

Re: Very low output power (0.1W)

Post by Ohwenzelph »

Matching of BS170 FETs
Yet another method of matching the BS170s.
This matching configuration allows the comparison between a Reference FET (TR1) and the group of test transistors TR2 etc etc.
The circuit is capable of comparing the difference in Pinch Off voltages as well as the difference in Transconductance over a part of the operating dynamic range of the FETs. Operation:
1) VR1 is originally set with its wiper at Zero volts wrt ground.
2) The first FET (TR1) is the Reference transistor, all others placed in the TR2 position
shall be compared against this reference transistor.
3) To test for a good match, the potentiometer VR1 is adjusted from its Zero volts position
up to the supply voltage. During this adjustment, the operator observes the variation
of the voltage between the Drain of TR1 and TR2.
4) Each transistor tested in the TR2 position which exhibits only small variations on the
voltmeter shall be selected as a reasonable match against the reference transistor. 5) Typical variations are 200 mV Drain voltage differential in the Pinch Off region. This
equates to approximately a differential Drain current of 3 mA.
6) As VR1 is advanced toward the supply voltage, the Drain differential voltages shall
typically indicate lower than 20mV, which equates to a difference of Drain currents of
0.3 mA.
7) Continue to insert several transistors in the TR2 position and make a selection of the
best matches.
8) If you find that there is a very poor match on nearly every one then it is likely to be the
Reference transistor chosen that is the atypical device and it should be replaced and
the complete test sequence repeated and a new selection of “best matches” made. 9) If you find that there is NO variation in voltage then check your circuit !
Measure the voltage from the Drain of TR2 w.r.t earth while adjusting VR1 over its total range. The Drain voltage shall vary from the supply voltage down to approximately 0.4 Volts.
10) Each measurement taken should last no greater than 10 seconds to remain within the maximum ratings of the transistors under these operating conditions.
9FC25E23-A903-45EE-80E2-3605F8089C25.jpeg
9FC25E23-A903-45EE-80E2-3605F8089C25.jpeg (223.85 KiB) Viewed 2873 times
B03DFF57-2813-496E-98BE-2519DB96A74A.jpeg
B03DFF57-2813-496E-98BE-2519DB96A74A.jpeg (98.44 KiB) Viewed 2873 times
BE9EC05E-0C0F-4292-A428-9067EE467FFD.jpeg
BE9EC05E-0C0F-4292-A428-9067EE467FFD.jpeg (30.84 KiB) Viewed 2873 times
I don't know who the author is, making the jig was pretty easy.
M1KTA
 

Posts: 26
Joined: 17 Jan 2022, 18:05

Re: Very low output power (0.1W)

Post by M1KTA »

S53SL wrote: 23 Apr 2022, 16:17
M1KTA wrote: 23 Apr 2022, 14:19 I understand experimentation perhaps but why was anything in Q4/Q5 on the bottom of the RF board at all when using Q1, Q2 and Q3
...SMD FET alternatives ....
That is what confused me, you cannot have both.
73
Dom
M1KTA
va3rr
 

Posts: 54
Joined: 30 Dec 2021, 21:39
Location: FN25eg
Contact:

Re: Very low output power (0.1W)

Post by va3rr »

S53SL wrote: 22 Apr 2022, 20:54
I also found it interesting that SWR coil is heating up, but there is practically 1:1 SWR as I used dummy load.
T2 will heat up, because it's 7 turns are in shunt with the load. Fewer turns provides better sensitivity for QRP levels. Here's a good explanation from Owen Duffy:

https://owenduffy.net/blog/?p=9138
Slava Ukraini!
Post Reply

Users browsing this forum: No registered users and 0 guests