Topics

uBitX V6 no transmit power Round 2 #ubitx-help #ubitxv6 #tx


Aaron K5ATG
 

Earlier I had a uBitX that had no transmit power. It was determined by this group to be Q90 which is believed to have been blown by a nearby transmitter. I sent it to HFSignals and they tested it and verified that it was Q90. 
Now I have a replacement uBitX V6 and it is doing the EXACT same thing. Here are the measurements of transmit power by each band:

80 meters- 0.13 watts

40 meters- 0.13 watts

30 meters- 0.18 watts

20 meters- 0.12 watts

17 meters- 0.16 watts

15 meters- 0.18 watts

12 meters- 0.07 watts

10 meters- 0.09 watts
Here is my complete shack set up:
Astron RS-12A Power Supply
Any Tone AU-778uv 2m/70cm FM transceiver hooked to Diamond F718A Base Station Antenna 40 feet RG8 feed point at 20 feet elevation
HFSignals uBITX V6 hooked to MFJ-1982MP 153' long endfed antenna fed with RG8 coax feed point at 9 feet elevation. On my uBitX I installed back to back 1N4148 Diodes between K1 relay pins 12 and 16 (This was supposed to protect Q90)
The separation between the feed points of both antennas is 25 feet horizontally and 10 feet vertical. 
Both antennas are grounded with a ground wire coming down from feed point to 8' ground rods driven 7' 10" into the ground.
I have an AM Broadcast transmitter about 1/2 mile from my house where I can hear it on my uBitX.

I know that there are a lot of really smart people in this group. So I would like to know what do you think that could be blowing the Q90 transistor?  
I would like to find and resolve the problem so I do not have to keep replacing the transistor. 

--
'72
Aaron Scott
QRPARCI# 16443
GQRP# 16389
4SQRP # 1080


Curt
 

Aaron

you did not describe how you are matching your 153 feet long antenna?  if you are not matching it with an antenna tuner - the rig may not output much power.  possibly nothing is damaged.  a random wire cannot be directly mated to a rig expecting a 50 ohm load impedance. 

wait I am puzzled - isn't that antenna you list 132 feet tall, an end-fed half wave antenna?  its SWR might be reasonable on several bands (I have a different make - mine is bad on 40m but okay on several of the bands). 

do you have stuff to make a simple 50 ohm dummy load?  or purchase a dummy load kit from a domestic supplier?  or borrow one from someone in your radio club - you just need something that handles 4 watts or more CW - and not key it very long, just long enough to get a reading on your wattmeter.  I assume you are keying your ubitx with a straight key or providing a short circuit at the key jack?  talking into a microphone may not read much on your wattmeter unless you can do a very steady strength whistle?  SSB behaves different than FM. 

we are amateurs and we learn by doing.  great to have you in our worldwide community. 

73 curt wb8yyy

PS - sounds like you need to construct a homebrew AM broadcast reject filter - that topic has been here recently. 


Aaron K5ATG
 

Curt,
Thanks for the reply,
On my end fed I have a 49:1 unun  or balun, whatever MFJ stuck in there, that came with the antenna. I used the antenna with a borrowed Kenwood TS-440S with no ATU and I did not have problems with it. I have an N7DDC ATU that I need to finish. I'm hung up between building a case for it to be stand-alone or installing it in the uBitX case and use it as a built in ATU with the uBitX
When I test it for power, I have tested it with it connected to a 500-watt dummy load and an MFJ-864 wattmeter, and after that, I test it with a QRPometer from NM0S that is a power meter with a 50-ohm 16-watt dummy load. Both devices had measurements that were real close. 
I have a Broadcast Interference Filter kit from Kit-Projects, I am waiting on the instructions for that. 
--
'72
Aaron Scott
QRPARCI# 16443
GQRP# 16389
4SQRP # 1080


Jerry Gaffke
 

153' of wire up in the air 1/2 mile from an AM broadcast station could be trouble.
Perhaps even bad enough to blow out the diodes in addition to Q90.

Possible the VHF rig did it, I'd pull the antenna from the uBitx 
when operating VHF till proven safe.

An electrical storm could also do this.

A $1k+ rig like the Kenwood might have better front end protection than the uBitx.

I'd build a diode RF probe, unplug the antenna from the uBitx and
check the RF voltage across the coax coming in from the MFJ antenna system.
Could do this with a scope instead if you have access to one, that AM station is
way down there around 1mhz.
If you measure more than half a volt RMS, that's bad.

I doubt this is an issue of an antenna tuner.
First off, if MFJ has built this thing correctly (not a given), should be able to 
transmit on all but 30m with a reasonable SWR
    https://mfjenterprises.com/products/mfj-1982mp
The antenna appears to be a clone of the myantennas.com EFHW-8010,
which I have success with.
And you were apparently measuring power into a 50 ohm dummy load,
so the question of maybe needing an antenna tuner is moot anyway.
But not a bad idea to obtain some way of measuring antenna SWR.

Shipping stuff round trip to Mumbai is a waste of time and money.
You might need to find somebody in the club who can show you how
to swap out Q90.

You can determine if Q90 is healthy by checking the quiescient operating point.
The base is biased from 12vdc through a series combination of 1000 ohms and 470 ohms.
So assuming zero base current (a high beta in Q90), the base voltage should be
  12 * 470/(1000+470) = 3.84 volts dc   (hopefully within a half volt of this figure)
when you press push-to-talk but do not give the rig audio.
The emitter of Q90 should be about 0.7 volts below that, or 3.14 volts dc.
These voltages can be checked with any DVM, the other probe on system ground.

Jerry, KE7ER 


Curt
 

Aaron

sounds like you are wonderfully equipped. 

now how do you key the rig to transmit?  if you only push the microphone PTT button - note on a SSB transmitter this does NOT output anything.  if you speak into the microphone, the meter should move a little but it will not indicate much power - it may not be fast enough to keep up with your speech.

I image you don't have a CW key.  a little piece of wire can connect across the key connector - and when connected across it should key the transmitter into outputting its full signal.  you should see a few watts.  try a few bands and do not transmit for long.  if you have CW output, then its a matter of having enough drive from the supplied microphone.  lets test CW output first.  not likely you will have 2 ubitx with the same issue. 

if CW is putting out a few watts - then you can try to speak into microphone slow and loud to see if you get a reading on your power meter - I mean slurring your speech dreadfully slow - I might need practice to do it myself!  actually easier to have a nearby ham listen for you. 

I will caution that many ubitx arrive with the BFO calibration way off on receive - meaning SSB may not sound ideal?  it takes some patience to calibrate this. 

let us know if you have CW output?  of course with dummy load or antenna attached. 

Curt


Evan Hand
 

Aaron,

I believe that I may have given you bad information on the correct diode to use.  Instead of the 1n4148, a pair of Schottky diodes (1n5819 or BAT41-TR) may be a better choice.  Here are 2 links on protecting receiver inputs.  That is what we are trying to do here because Q90 is in the receive path all of the time.

https://www.kb6nu.com/protect-receiver-inputs/
or
http://www.kk5jy.net/rf-clipper/

Adding a lamp to dissipate some of the power is also suggested in what I have found.  I have not installed this type of protection so do not know.
73
Evan
AC9TU


Jerry Gaffke
 

Regarding back to back diodes, it gets messy.

The diodes were found to be needed on the Bitx40, a number of reports 
on that rig having a blown transistor at Q13 (in the position Q90 is on the uBitx).
If Q90 really is getting blown away, probably best to add the diodes.

We were using 1n4148's for protection on the Bitx40, not Schottky diodes.
Seemed to work.

The "bulb" must be a very tiny incandescent, as stated in this old chestnut from Raj:
https://groups.io/g/BITX20/message/19105

Why an incandescent bulb instead of a resistor?

 The first line of the document in the link that follows reads: "Tungsten-filament incandescent lamps exhibit a very-high positive temperature coefficient of resistance with the cold filament resistance being approximately 10% of the hot filament resistance."
 http://www.allegromicro.com/~/media/Files/Technical-Documents/an295012-A-Primer-On-Driving-Incandescent-Lamps.ashx
Also, when the bulb glows, you know you're on thin ice.   ;-)
But a 50 or 100 ohm series resistor help if you don't have a proper bulb.


On the uBitx, Farhan has warned that adding back to back diodes can 
add spurs to the receiver, since the uBitx only has a 30mhz LPF at L1,2,3,4
where the old Bitx40 had a narrow band 7mhz filter.
Farhan was of the opinion that most of the RF that was trashing Bitx40's
would go right on through the 30mhz LPF and T2 to be snuffed out by the diodes
at the first mixer.  So I would not add the diodes to a uBitx if they aren't needed.
   https://groups.io/g/BITX20/message/35656
    
Perhaps put a socket in where Q90 is, and carry some spares. ;-)


Some more old links from the dim past:
 https://groups.io/g/BITX20/topic/7715656
 https://groups.io/g/BITX20/topic/4353441


This hasn't been discussed before, but might be best in a high RF environment 
to move K3 pins 9,11,13  such that receiver input is taken after the
transmit LPF's at TP7 instead of directly from the antenna port.
This way VHF energy will be knocked down before it gets to Q90, and anything
below 30mhz gets knocked down by the first mixer as previously discussed.
Opinions?

Jerry, KE7ER


Evan Hand
 

Jerry,

According to Aaron, he did install the diodes, 1n4148s, per the picture on the original post.  That was before he used the second rig, so it appears that it did not protect Q90.  I was the one who had suggested the 1n4148.

I had done more research in the past and again just yesterday.  In the material in the links that I had provided, I found the 12volt lamp to act as a power sink on high RF input.   The reason for the lamp, as opposed to a resistor, is that the lamp has a very low resistance at room temperature that climes quickly when heated by the input RF.  I know that I had read multiple people say that is common protection for some commercial rigs in the distant past in this group.

At this point I am waiting for Aaron to get back to us to verify that:
1 - Q90 is truly blown through voltage measurements
2 - That the power was measured on CW vs SSB

73
Evan
AC9TU


Evan Hand
 

Jerry,

After reading your email again, I realized that you already stated the reason for a lamp as opposed to a resistor.  Q90 failure seems to be a somewhat frequent issue.  I have responded to more than 2 people with the diode solution.  I may have been in error on that, however, I do believe that Q90 is being destroyed by high RF or static discharge.

I am going to do some more research on any commercial rigs that I can find the schematic to see how those are protected.

Thank up for the information, and I apologize for any errors on my part including not reading your email correctly.
73
Evan
AC9TU


Jerry Gaffke
 

Evan,

Been busy with other stuff and not following the group much.
Didn't realize there were so many Q90's getting knocked out.
This should be addressed.

In the dim past on the Bitx40, 1n4148 diodes were used.
Anything called a 1n914 these days is likely the same part, an older part that is not spec'd so tight as the 1n4148.
These, as you know, are standard silicon diodes, start to conduct milliamps around 0.6 volts of forward voltage.
Your Schottky diodes start conducting at half that and thus offer more protection, but may cause more spurs (birdies) in the receiver.
It is important that these be low capacitance signal diodes, something like a 1n4007 power diode may suck away too much signal.

References I'm seeing put the bulb first in line coming in from the antenna, then the diodes.
This keeps the diodes from getting burned out from a strong signal.
A case could be made for adding series resistance (50 ohms?  a bulb?) after the diodes, to ensure that the diodes take more current than Q90 does.
The 1n4148 diodes are rated at 300ma continuous current, no spec on max continuous base current for the 2n3904 but I suspect it is more like 50ma.
The emitter resistors at Q90 do offer some series resistance, so we already have a little bit of protection there.
The correct incandescent bulb may not be easily at hand, I suspect two series 50 ohm resistors, one before and one after the diodes, would work well without
just too much attenuation of the received signals.

If Q90 burned out when the VHF transmitter was turned on, better move the VHF antenna.
And stick with the 5 watt setting on the VHF transmitter when the uBitx has its antenna plugged in.

Let's assume Arron takes voltage measurements on his antenna with a diode RF probe and determines
that the nearby AM transmitter is the issue.
Arron described his MFJ-1982HP antenna as 153 feet long.
I suspect it is more like 132 feet, resonant on 80 meters.
One option would be to cut the wire down to around 66 feet, should tune the length with an antuino, nanovna or SWR meter.
That makes it a MFJ-1984HP, you lose 80 meters and some of the WARC bands, but the shorter length would
likely absorb less energy from that AM transmitter.  The shorter length would also allow the antenna to be
oriented in a way less likely to catch that AM transmission.  My guess is that broadside to the transmitter
would be worst case, things like nearby metal roofs would seriously mess with this.

2n3904 resistors are cheap.  If Q90 is blowing out so often, a jig could be made with a socket for a 2n3904
to try out various protection schemes.  Drive it with a signal generator at several volts rms  (a transmitter)
from 500khz through VHF.  

As I suggested in my previous email, the diodes in that first mixer do not provide any protection for Q90
in the case of a VHF signal due to the 30mhz LPF at L1,2,3,4. 
Perhaps add a second 30mhz LPF in the receive path immediately after K3 pin 11 near the antenna port.
Likely an easier solution than messing with how K3 pins 9,11,13 get wired as previously suggested.

Jerry, KE7ER


Evan Hand
 

Jerry,

Thank you again. 

I have done some more research and I get mixed hits on the 1n4148 vs a Schottky like 1n5819.  More important is that I see a more complex solution like this one:
https://www.hamradiosecrets.com/ham-radio-projects.html#FEP
Ham radio projects - a front-end protection circuit for receivers.
 Not sure the reasons for the capacitors in the circuit other to block any DC.  It might reduce broadcast input.

In the presentation from Bill Leonard (N0CU), he also has the capacitors in the Back to Back Diodes with Loss (Light Bulb) slide.  He put the red question mark on the slide, not me.



I looked up the datasheet on the 1n3600 and they are fast switching diodes similar to the 1n4148, not Schottky.

As far as I can tell, the original Back to Back diodes should work.

Aaron,
Can you check the resistance on the diodes when you remove them?  I wonder if they opened up because of overload from the nearby transmitter.  That seems to be unlikely, though I do not know the strength of the nearby station.  If it were that strong I would imagen that you have a lot of difficulty in receiving anything on the uBITX

73
Evan
AC9TU


Jerry Gaffke
 

Evan,

You can have large static voltages build up on some antenna systems.
Could be several thousand volts.
An RF choke to ground at the antenna port should be sufficient to deal with these.

Though in Arron's case, the EFHW matchbox should be shorting that static to ground.
Here's a group on facebook actively discussing these antennas:  
    https://www.facebook.com/groups/EndFedHalfWaveAntennas/

And on the uBitx, L1,2,3,4 and T2 provide a DC path to ground during receive.
When transmitting, T11 of the uBitx should be a DC short to ground
through the transmit LPF inductors.


So the choke is probably not needed in this case, but the blocking caps that NC0U shows
are not a bad idea.  I'd be worried that voltages could far exceed his 25 volt cap.
And I agree with his red question mark regarding the extra cap in series with the diodes.

While we may be protected from DC, we still need to deal with 
large RF fields, such as that AM transmitter, or Arron's VHF rig.
That's where the back-to-back diodes and the incandescent bulb become important.
 
A lot of the protection schemes that work for others may not work on a uBitx.
The base-emitter junction of the 2n3904 at Q90 is likely more sensitive
to excessive voltage than the antenna port of a typical ham rig.

Jerry, KE7ER


Bill Cromwell
 

Hi,

Omce upon a time I was told that those diodes' mostly fail short, not open. I had a two different pairs fail over the years and yes..they were short circuit. The radio front ends (both Ten Tec) survived:)

I am replacing the Q90 in my V3 and diodes are going to be installed.

bark less - wag more

On 10/27/20 11:06 AM, Evan Hand wrote:
Jerry,
Thank you again.
I have done some more research and I get mixed hits on the 1n4148 vs a Schottky like 1n5819.  More important is that I see a more complex solution like this one:
https://www.hamradiosecrets.com/ham-radio-projects.html#FEP
Ham radio projects - a front-end protection circuit for receivers.
 Not sure the reasons for the capacitors in the circuit other to block any DC.  It might reduce broadcast input.
In the presentation from Bill Leonard (N0CU), he also has the capacitors in the Back to Back Diodes with Loss (Light Bulb) slide.  He put the red question mark on the slide, not me.
I looked up the datasheet on the 1n3600 and they are fast switching diodes similar to the 1n4148, not Schottky.
As far as I can tell, the original Back to Back diodes should work.
Aaron,
Can you check the resistance on the diodes when you remove them?  I wonder if they opened up because of overload from the nearby transmitter.  That seems to be unlikely, though I do not know the strength of the nearby station.  If it were that strong I would imagen that you have a lot of difficulty in receiving anything on the uBITX
73
Evan
AC9TU


Aaron K5ATG
 

Sorry, this is off topic/ on topic, well I don't know. 
I'm going to try to pay attention to this but I will not be as active for a bit. We are experiencing an ice storm and it has so far done some damage to my property. The end-fed antenna is now a thing of the past. 

In this pic you can see that my power meter and stuff has been ripped from the house from a neighbors tree. Somehow we have not lost power. 

This is pretty much what the rest of my yard looks like. 
No one at my house has been hurt, everything that has been damaged will be fixed by insurance. So I will be busy for awhile. 

--
'72
Aaron Scott
QRPARCI# 16443
GQRP# 16389
4SQRP # 1080


Jerry Gaffke
 

Ouch!
On the bright side, it's an opportunity to get those antennas configured better.  ;-)

Jerry, KE7ER


On Tue, Oct 27, 2020 at 10:36 AM, Aaron K5ATG wrote:
We are experiencing an ice storm and it has so far done some damage to my property. The end-fed antenna is now a thing of the past. 


Viktors Miske
 

Ouch. Power box is one thing...  But the antenna, RIP EFHW.


Evan Hand
 

Ouch and double ouch!

Just glad to hear that everyone is OK.

let us know when you are back on radio stuff.
good luck
73
Evan
AC9TU


Bob Lunsford
 

The classic rational for those who say that a lamp being used as a dummy load is that it is not only not a 50 Ohm load, it is also varying resistance depending on the temperature of the lamp's filament. Also keep in mind that it was a common thing used with relative low power transmitters, the transmitter was built around tube and not a solid state amp. While I have also used a pilot lamp on a QRP transmitter I built for 7040 KHz, it was thrilling to see the low power transmitter light up even though not quite to full brilliance.

More nostalgia. Today, using a lamp for a load or a power indicator is definitely old school and for me, I'd now rather wait for a proper device than tempt fate and see smoke instead of merely a bright spot in the corner.

Bob — KK5R

On Tuesday, October 27, 2020, 8:29:44 AM EDT, Evan Hand <elhandjr@...> wrote:


Jerry,

According to Aaron, he did install the diodes, 1n4148s, per the picture on the original post.  That was before he used the second rig, so it appears that it did not protect Q90.  I was the one who had suggested the 1n4148.

I had done more research in the past and again just yesterday.  In the material in the links that I had provided, I found the 12volt lamp to act as a power sink on high RF input.   The reason for the lamp, as opposed to a resistor, is that the lamp has a very low resistance at room temperature that climes quickly when heated by the input RF.  I know that I had read multiple people say that is common protection for some commercial rigs in the distant past in this group.

At this point I am waiting for Aaron to get back to us to verify that:
1 - Q90 is truly blown through voltage measurements
2 - That the power was measured on CW vs SSB

73
Evan
AC9TU


Bob Lunsford
 

Have you considered a Zener diode with calculated breakdown if it is indeed a voltage thing? I remember Heath using a 36V Zener on the power output transistor on the HW-7/-8 to avoid killing off the final output transistor in case of high SWR with an improper or nonexistent load.

Bob — KK5R

On Tuesday, October 27, 2020, 8:39:00 AM EDT, Evan Hand <elhandjr@...> wrote:


Jerry,

After reading your email again, I realized that you already stated the reason for a lamp as opposed to a resistor.  Q90 failure seems to be a somewhat frequent issue.  I have responded to more than 2 people with the diode solution.  I may have been in error on that, however, I do believe that Q90 is being destroyed by high RF or static discharge.

I am going to do some more research on any commercial rigs that I can find the schematic to see how those are protected.

Thank up for the information, and I apologize for any errors on my part including not reading your email correctly.
73
Evan
AC9TU


Jerry Gaffke
 

I built the 35W 6146 single tube crystal oscillator out of the 1969 ARRL handbook as a novice.
Very exciting to watch the 40W bulb light up, smell the RF burns when I put my finger
near the inductor wound on a bit of broomstick for the pi network output.
Them FT243 crystals could apparently handle a fair bit of current, none of them smoked.

For a quick but proper dummy load to be used with the uBitx, put a couple Mouser 660-MOS3CT631R101J in parallel

For an improper device, consider Joe's (WB9SBD) dummy load, using electrodes in a water bath:
https://ae5x.blogspot.com/2020/10/dummy-load-with-chemical-impedance.html
Joe's powerpoint presentation was awfully long, here's my Cliff Notes, dimensions judged from his photos.
Electrodes are two 12 awg copper wires (2mm diameter) 6 inches long (16cm) and about six diameters apart.  Add baking soda till it is 50 ohms, about one teaspoon (or 5 ml).  The water will boil after 50 minutes at 100 watts.  SWR of around 1.0 through 60mhz, 1.5 at 150mhz but pulling the wires somewhat out of the water can readjust for an SWR of 1.0.  Joe reports good results up to 180mhz.  Has been working well for years.

If I was somewhat older, Joe's dummy load might make me nostalgic for an old electrical lighting system:
    https://www.powerstream.com/1922/battery_1922_WITTE/batteryfiles/chapter17.htm

Jerry,  KE7ER 



On Tue, Oct 27, 2020 at 08:49 PM, Bob Lunsford wrote:
The classic rational for those who say that a lamp being used as a dummy load is that it is not only not a 50 Ohm load, it is also varying resistance depending on the temperature of the lamp's filament. Also keep in mind that it was a common thing used with relative low power transmitters, the transmitter was built around tube and not a solid state amp. While I have also used a pilot lamp on a QRP transmitter I built for 7040 KHz, it was thrilling to see the low power transmitter light up even though not quite to full brilliance.
 
More nostalgia. Today, using a lamp for a load or a power indicator is definitely old school and for me, I'd now rather wait for a proper device than tempt fate and see smoke instead of merely a bright spot in the corner.
 
Bob — KK5R