Topics

Harmonics and Relay Replacement

 

One of them must be on this group!

Raj

At 13/10/2018, you wrote:
Interesting...... the price for V23105A5403A201 relays from Farnell in VK was AU$2.28 a few days ago. Its now $3.97

glenn vk3pe

MVS Sarma
 

After all seller groups would make use of any sudden spurt in demand
of a product.
After ham community identified the use and benefits of axicom relays,
the sellers would obviously raise the costs.

On Sat, Oct 13, 2018 at 2:03 PM Raj vu2zap <@Raj> wrote:

One of them must be on this group!

Raj

At 13/10/2018, you wrote:
Interesting...... the price for V23105A5403A201 relays from Farnell in VK was AU$2.28 a few days ago. Its now $3.97

glenn vk3pe


iz oos
 

Farewell Farnell!


Il 13/ott/2018 10:19, "Glenn" <glennp@...> ha scritto:
Interesting...... the price for V23105A5403A201 relays from Farnell in VK was AU$2.28 a few days ago. Its now $3.97

glenn vk3pe

Lawrence Galea
 

Did anyone check whether there is any difference in the attenuation with the different Axicom relays?
What I mean is that according to some posts it appears that the relays have different sensitivities and different inductance values.
So did anyone check or could, or intend to check whether there is any difference in using one or the other?
Thanks

On Sat, Oct 13, 2018 at 2:49 AM Kees T <windy10605@...> wrote:
Jim,

Yes, I've seen that posted by you before......but what did they do specifically for RF use ? some spacings ? shielding ? coil core material ? base dielectric ? housing dielectric ?

I also see Axicom has 4 coil power options (like most of the other relay vendors). Which one is correct and which one did you order ?

73 Kees K5BCQ

Kees T
 

Mike,

Now THAT spec from Farnell was what I was looking for.  The 151 pages have a lot of good RF information not found elsewhere.....but will take a little time to digest.

73 Kees K5BCQ  

Jim Sheldon
 

Looking at the differences between the original Hongfa relays -S (for sensitive) version in my uBITX boards and the Axicom relays Mike Doty and I installed in place of them, the Axicom had lower coil resistance in the 400 mW coils vs the 200 mW coils fot the Hongfa relays.  That means fewer turns of wire in the Axicoms and subsequently less inductance.  That may be a huge factor in the isolation between the two types.  The coil resistance for the Axicom that matches the Hongfa in DC spec aldo has the exact same listed coil resistance.  

I'm going to order 5 each of the 150mW, 200 mW and 500 mW versions, put sockets on my 3rd V4 board and see how much difference there is between the 3 other types relative to the already installed 400 mw version.  Hopefully that will give us some definitive answers.  Be Monday before I can order them so it will be a few days before any testing can be done.

Jim Sheldon, W0EB

Kees T
 

I've looked over that 181 page AXICOM spec sheet and there are 2 things that set them apart from the herd.

Common with the herd: they are "Signal Relays", 2pF between contacts, 4pF to the coil, have gold plated contacts, etc ....all good but not unique to Axicom).

1. The solenoid relay coil is located over the top of the relay armature which will reduce the coupling between the two, especially since the armature is shorter and not routed over the end of the solenoid relay coil.

2. The solenoid relay coil on the particular Axicom relay ordered .......was    V23105A5403A201  the "403" means the option for a 400mW coil.  Now that is a difference because the "typical" relays have the 200mW coil option. More power means more current and probably fewer turns in the coil and potentially less coupling between the solenoid relay coil and the armature. 

Both #1 and #2 are not possible to fix outside the relay, may reduce coupling some between relays by adding series inductors. Capacitors across the relay coils have no effect. Capacitors to ground ? TBD

73 Kees K5BCQ

Kees T
 

Since it has been proven that there is coupling between the relay solenoid coil and the long relay armature (apparently a lot more on the non Axicom relays) .....what can be done with the original relays since internal surgery is out. I have been focused on adding series inductance to the relay solenoid coil. Results in previous posts. Well, there is also less coupling on the original relays if you reduce the RF on the solenoid coil by bypassing both coil ends with short leads to ground at the relay. I'm using 47nF capacitors and see a 3dB reduction.    and that's at Rigol power levels.

73 Kees K5BCQ

 

Jim,

My take on this relay business is that both contacts are being inductively coupled.

Imagine the NC contacts on one side - it is a loop and will be inductive. The same goes
for the other pair on the other side..

By adding a socket we may not get the best rejection from Axicom relays. I have also
socketed KT1/2/3 .. the relays have arrived and I am at the farm.. so a week more
of suspense.

If someone with a SA like 815 can feed in the tracking gen through a 47 ohm resistor
to one NC contacts in/out and take the Analyser input from the other NC contact set
and then sweep say up to 100MHz, we get some enlightenment. I am eagerly waiting
to try for myself.

Raj

At 13/10/2018, you wrote:
Looking at the differences between the original Hongfa relays -S (for sensitive) version in my uBITX boards and the Axicom relays Mike Doty and I installed in place of them, the Axicom had lower coil resistance in the 400 mW coils vs the 200 mW coils fot the Hongfa relays. That means fewer turns of wire in the Axicoms and subsequently less inductance. That may be a huge factor in the isolation between the two types. The coil resistance for the Axicom that matches the Hongfa in DC spec aldo has the exact same listed coil resistance.

I'm going to order 5 each of the 150mW, 200 mW and 500 mW versions, put sockets on my 3rd V4 board and see how much difference there is between the 3 other types relative to the already installed 400 mw version. Hopefully that will give us some definitive answers. Be Monday before I can order them so it will be a few days before any testing can be done.

Jim Sheldon, W0EB

Jim Sheldon
 

Raj,
I thought the same thing, but maybe, even using sockets, just for the experiment, I could get some relative comparisons between the types to see if any were better than others. I suspect the 500mw version, Arrow part# V23105A5503A201 will be the best because they have the lowest coil resistance, fewest turns on the coil. The only thing that worries me using that relay is it's current draw - I'm hoping it won't be too much for the switching transistors in the uBITX. The 400mW relays are working and switching OK but the additional current from the lower coil resistance may become a problem and I'm not sure yet.

I don't want to risk ruining the board by soldering and unsoldering the relays several times so planned on using sockets but ONLY on the one board I'm planning on setting up as a test station. I'll build it up on a board but use double shielded Teflon insulated COAX from the uBITX boards 2 pin connector to the BNC output and make all digital and audio wires twisted pair to the front panel jacks, encoder and switch. The construction will be open, no case but the uBITX board will be mounted on a 16 gauge aluminum plate to give a good underside ground plane and good low Z grounds for all connections.

I have some good, gold plated machined pin socket strips that are very low profile and will use those to create the sockets. Again, for these tests I will only look for which of the 4 relay types gives the best harmonic attenuation between types and then I will solder that set into a uBITX for the final results. Only way I know to test all 4 types without possibly ruining a good board from too much soldering/de-soldering of relays. A repaired trace would negate the original test results as it's characteristics would be different.

Hope people can understand what I'm trying to say here.

Jim

------ Original Message ------
From: "Raj vu2zap" <@Raj>
To: BITX20@groups.io
Sent: 10/14/2018 1:06:47 AM
Subject: Re: [BITX20] Harmonics and Relay Replacement

Jim,

My take on this relay business is that both contacts are being inductively coupled.

Imagine the NC contacts on one side - it is a loop and will be inductive. The same goes
for the other pair on the other side..

By adding a socket we may not get the best rejection from Axicom relays. I have also
socketed KT1/2/3 .. the relays have arrived and I am at the farm.. so a week more
of suspense.

If someone with a SA like 815 can feed in the tracking gen through a 47 ohm resistor
to one NC contacts in/out and take the Analyser input from the other NC contact set
and then sweep say up to 100MHz, we get some enlightenment. I am eagerly waiting
to try for myself.

Raj

At 13/10/2018, you wrote:
Looking at the differences between the original Hongfa relays -S (for sensitive) version in my uBITX boards and the Axicom relays Mike Doty and I installed in place of them, the Axicom had lower coil resistance in the 400 mW coils vs the 200 mW coils fot the Hongfa relays. That means fewer turns of wire in the Axicoms and subsequently less inductance. That may be a huge factor in the isolation between the two types. The coil resistance for the Axicom that matches the Hongfa in DC spec aldo has the exact same listed coil resistance.

I'm going to order 5 each of the 150mW, 200 mW and 500 mW versions, put sockets on my 3rd V4 board and see how much difference there is between the 3 other types relative to the already installed 400 mw version. Hopefully that will give us some definitive answers. Be Monday before I can order them so it will be a few days before any testing can be done.

Jim Sheldon, W0EB


 

GM Jim,

At 14/10/2018, you wrote:
Raj,
I thought the same thing, but maybe, even using sockets, just for the experiment, I could get some relative comparisons between the types to see if any were better than others. I suspect the 500mw version, Arrow part# V23105A5503A201 will be the best because they have the lowest coil resistance, fewest turns on the coil. The only thing that worries me using that relay is it's current draw - I'm hoping it won't be too much for the switching transistors in the uBITX. The 400mW relays are working and switching OK but the additional current from the lower coil resistance may become a problem and I'm not sure yet.
400mW is 33mA that should be OK for the transistors. I will try lower current one when I find a cheaper source or a vendor who will send
me by USPS.


I don't want to risk ruining the board by soldering and unsoldering the relays several times so planned on using sockets but ONLY on the one board I'm planning on setting up as a test station. I'll build it up on a board but use double shielded Teflon insulated COAX from the uBITX boards 2 pin connector to the BNC output and make all digital and audio wires twisted pair to the front panel jacks, encoder and switch. The construction will be open, no case but the uBITX board will be mounted on a 16 gauge aluminum plate to give a good underside ground plane and good low Z grounds for all connections.
Thats sensible, I put in sockets for the same reason. That board with L5/7 changed had spurs below -50dbm and its put to rest.
I was thrilled with Mike Doty's find (Remember Roy Doty ? - PopSci). I ordered the same 400mW relays as I thought that lower
power ones MAY be a little different in construction. Just KT1,2,3 change should do the trick. The other two don't play much
of a part in harmonics leaking through.

I tried a lot of mods, most gave a few dbs of improvement while others made things worse. The fix that makes dramatic improvement is
the root of the problem. For spurs the TX signal leaking into the first mixer was cured with TX losses with a 45MHz filter.. further thinking
led me to L5 and its position made a difference but not a fix.. the SMD for L5 made a big difference. Sort of a eureka moment!

High Q inductor in this board layout is not in best interest of impedance matching.. we are only stepping up 50 to 650 Ohms and back.
I need to experiment with a tapped toroid 2:7 without the caps and see - that might also fix the spurs issue. I wish someone would try
that and report - I am at my farm.

I have some good, gold plated machined pin socket strips that are very low profile and will use those to create the sockets. Again, for these tests I will only look for which of the 4 relay types gives the best harmonic attenuation between types and then I will solder that set into a uBITX for the final results. Only way I know to test all 4 types without possibly ruining a good board from too much soldering/de-soldering of relays. A repaired trace would negate the original test results as it's characteristics would be different.

Hope people can understand what I'm trying to say here.
I read you! I added the same pins. With a collaboration like this group we will make a great and simple exciter / QRP rig!


Jim

Raj



------ Original Message ------
From: "Raj vu2zap" <@Raj>
To: BITX20@groups.io
Sent: 10/14/2018 1:06:47 AM
Subject: Re: [BITX20] Harmonics and Relay Replacement

Jim,

My take on this relay business is that both contacts are being inductively coupled.

Imagine the NC contacts on one side - it is a loop and will be inductive. The same goes
for the other pair on the other side..

By adding a socket we may not get the best rejection from Axicom relays. I have also
socketed KT1/2/3 .. the relays have arrived and I am at the farm.. so a week more
of suspense.

If someone with a SA like 815 can feed in the tracking gen through a 47 ohm resistor
to one NC contacts in/out and take the Analyser input from the other NC contact set
and then sweep say up to 100MHz, we get some enlightenment. I am eagerly waiting
to try for myself.

Raj

At 13/10/2018, you wrote:
Looking at the differences between the original Hongfa relays -S (for sensitive) version in my uBITX boards and the Axicom relays Mike Doty and I installed in place of them, the Axicom had lower coil resistance in the 400 mW coils vs the 200 mW coils fot the Hongfa relays. That means fewer turns of wire in the Axicoms and subsequently less inductance. That may be a huge factor in the isolation between the two types. The coil resistance for the Axicom that matches the Hongfa in DC spec aldo has the exact same listed coil resistance.

I'm going to order 5 each of the 150mW, 200 mW and 500 mW versions, put sockets on my 3rd V4 board and see how much difference there is between the 3 other types relative to the already installed 400 mw version. Hopefully that will give us some definitive answers. Be Monday before I can order them so it will be a few days before any testing can be done.

Jim Sheldon, W0EB

Lawrence Galea
 

Jim
I agree with you regarding the lowest number of turns.
My theory is that the relay contacts are acting as the primary of a current transformer inducing the RF current flowing in the contacts into the great number of turns of the relay coil which would explain the RF voltage found and posted by someone on the relay supply lines.
It's like a normal current transformer but with the primary on the outside of the coils with the strong RF field inducing the voltage in the relay coli.
Apart from the possible blowby from the contacts to the other contacts due to the strong RF field, the RF voltage on the relay coil would also induce a current in the other contacts reducing the isolation.
It is true that one can say that the current in the relay contacts would cancel each other as one is going in one direction and the other in the opposite direction, but the intense RF field and the small spacing could lead to unexpected results.
It could also be considered that the contacts are acting like a small loop inducing a current in the relay coil.
Remember that a transmitting loop has a very strong field in its vicinity.
This is why I asked earlier on whether someone had or intend to check whether the results are the same with the different relay coil inductance.
The lower inductance ones would have a lower voltage induced in the relay coil due to the smaller number of turns which can be checked by the RF voltage on the relay supply lines and also leading to better isolation.
This is just my theory.
Regards
Lawrence

On Sun, Oct 14, 2018 at 1:39 PM Jim Sheldon <w0eb@...> wrote:
Raj,
I thought the same thing, but maybe, even using sockets, just for the
experiment, I could get some relative comparisons between the types to
see if any were better than others.  I suspect the 500mw version, Arrow
part# V23105A5503A201 will be the best because they have the lowest coil
resistance, fewest turns on the coil.  The only thing that worries me
using that relay is it's current draw - I'm hoping it won't be too much
for the switching transistors in the uBITX.   The 400mW relays are
working and switching OK but the additional current from the lower coil
resistance may become a problem and I'm not sure yet.

I don't want to risk ruining the board by soldering and unsoldering the
relays several times so planned on using sockets but ONLY on the one
board I'm planning on setting up as a test station.  I'll build it up on
a board but use double shielded Teflon insulated COAX from the uBITX
boards 2 pin connector to the BNC output and make all digital and audio
wires twisted pair to the front panel jacks, encoder and switch.  The
construction will be open, no case but the uBITX board will be mounted
on a 16 gauge aluminum plate to give a good underside ground plane and
good low Z grounds for all connections.

I have some good, gold plated machined pin socket strips that are very
low profile and will use those to create the sockets.  Again, for these
tests I will only look for which of the 4 relay types gives the best
harmonic attenuation between types and then I will solder that set into
a uBITX for the final results.  Only way I know to test all 4 types
without possibly ruining a good board from too much
soldering/de-soldering of relays.  A repaired trace would negate the
original test results as it's characteristics would be different.

Hope people can understand what I'm trying to say here.

Jim


------ Original Message ------
From: "Raj vu2zap" <rajendrakumargg@...>
To: BITX20@groups.io
Sent: 10/14/2018 1:06:47 AM
Subject: Re: [BITX20] Harmonics and Relay Replacement

>Jim,
>
>My take on this relay business is that both contacts are being
>inductively coupled.
>
>Imagine the NC contacts on one side - it is a loop and will be
>inductive. The same goes
>for the other pair on the other side..
>
>By adding a socket we may not get the best rejection from Axicom
>relays. I have also
>socketed KT1/2/3 .. the relays have arrived and I am at the farm.. so a
>week more
>of suspense.
>
>If someone with a SA like 815 can feed in the tracking gen through a 47
>ohm resistor
>to one NC contacts in/out and take the Analyser input from the other NC
>contact set
>and then sweep say up to 100MHz, we get some enlightenment. I am
>eagerly waiting
>to try for myself.
>
>Raj
>
>At 13/10/2018, you wrote:
>>Looking at the differences between the original Hongfa relays -S (for
>>sensitive) version in my uBITX boards and the Axicom relays Mike Doty
>>and I installed in place of them, the Axicom had lower coil resistance
>>in the 400 mW coils vs the 200 mW coils fot the Hongfa relays.  That
>>means fewer turns of wire in the Axicoms and subsequently less
>>inductance.  That may be a huge factor in the isolation between the
>>two types.  The coil resistance for the Axicom that matches the Hongfa
>>in DC spec aldo has the exact same listed coil resistance.
>>
>>I'm going to order 5 each of the 150mW, 200 mW and 500 mW versions,
>>put sockets on my 3rd V4 board and see how much difference there is
>>between the 3 other types relative to the already installed 400 mw
>>version.  Hopefully that will give us some definitive answers.  Be
>>Monday before I can order them so it will be a few days before any
>>testing can be done.
>>
>>Jim Sheldon, W0EB
>
>
>
>




Jim Sheldon
 

Raj,

Rather than just using double shielded coax from the 2 pin RF plug on the uBITX to the BNC, I carefully installed gold plated female SMA connector and that should help some but I'm just using the SMA on my test rig. The other ones are faring well enough and legal so I'm leaving them as-is. I changed my mind on the machined pin sockets and modified some old Mica filled DIP sockets I had on hand from many years ago. I pulled the unused contacts out so have Mica dielectric between pins and these sit up about 2mm off the board. Should be interesting to see the results. Be Wednesday before all the parts get here so I'll report the results in a few days.

I will add the SMA inductors in place of L5 and L7 though and found the Murata 1206 ones at Mouser. 10 on order and that will cover the 5 rigs I have.

Jim - W0EB

------ Original Message ------
From: "Raj vu2zap" <@Raj>
To: BITX20@groups.io
Sent: 10/14/2018 7:57:58 AM
Subject: Re: [BITX20] Harmonics and Relay Replacement

GM Jim,

At 14/10/2018, you wrote:
Raj,
I thought the same thing, but maybe, even using sockets, just for the experiment, I could get some relative comparisons between the types to see if any were better than others. I suspect the 500mw version, Arrow part# V23105A5503A201 will be the best because they have the lowest coil resistance, fewest turns on the coil. The only thing that worries me using that relay is it's current draw - I'm hoping it won't be too much for the switching transistors in the uBITX. The 400mW relays are working and switching OK but the additional current from the lower coil resistance may become a problem and I'm not sure yet.
400mW is 33mA that should be OK for the transistors. I will try lower current one when I find a cheaper source or a vendor who will send
me by USPS.


I don't want to risk ruining the board by soldering and unsoldering the relays several times so planned on using sockets but ONLY on the one board I'm planning on setting up as a test station. I'll build it up on a board but use double shielded Teflon insulated COAX from the uBITX boards 2 pin connector to the BNC output and make all digital and audio wires twisted pair to the front panel jacks, encoder and switch. The construction will be open, no case but the uBITX board will be mounted on a 16 gauge aluminum plate to give a good underside ground plane and good low Z grounds for all connections.
Thats sensible, I put in sockets for the same reason. That board with L5/7 changed had spurs below -50dbm and its put to rest.
I was thrilled with Mike Doty's find (Remember Roy Doty ? - PopSci). I ordered the same 400mW relays as I thought that lower
power ones MAY be a little different in construction. Just KT1,2,3 change should do the trick. The other two don't play much
of a part in harmonics leaking through.

I tried a lot of mods, most gave a few dbs of improvement while others made things worse. The fix that makes dramatic improvement is
the root of the problem. For spurs the TX signal leaking into the first mixer was cured with TX losses with a 45MHz filter.. further thinking
led me to L5 and its position made a difference but not a fix.. the SMD for L5 made a big difference. Sort of a eureka moment!

High Q inductor in this board layout is not in best interest of impedance matching.. we are only stepping up 50 to 650 Ohms and back.
I need to experiment with a tapped toroid 2:7 without the caps and see - that might also fix the spurs issue. I wish someone would try
that and report - I am at my farm.

I have some good, gold plated machined pin socket strips that are very low profile and will use those to create the sockets. Again, for these tests I will only look for which of the 4 relay types gives the best harmonic attenuation between types and then I will solder that set into a uBITX for the final results. Only way I know to test all 4 types without possibly ruining a good board from too much soldering/de-soldering of relays. A repaired trace would negate the original test results as it's characteristics would be different.

Hope people can understand what I'm trying to say here.
I read you! I added the same pins. With a collaboration like this group we will make a great and simple exciter / QRP rig!


Jim

Raj



------ Original Message ------
From: "Raj vu2zap" <@Raj>
To: BITX20@groups.io
Sent: 10/14/2018 1:06:47 AM
Subject: Re: [BITX20] Harmonics and Relay Replacement

Jim,

My take on this relay business is that both contacts are being inductively coupled.

Imagine the NC contacts on one side - it is a loop and will be inductive. The same goes
for the other pair on the other side..

By adding a socket we may not get the best rejection from Axicom relays. I have also
socketed KT1/2/3 .. the relays have arrived and I am at the farm.. so a week more
of suspense.

If someone with a SA like 815 can feed in the tracking gen through a 47 ohm resistor
to one NC contacts in/out and take the Analyser input from the other NC contact set
and then sweep say up to 100MHz, we get some enlightenment. I am eagerly waiting
to try for myself.

Raj

At 13/10/2018, you wrote:
Looking at the differences between the original Hongfa relays -S (for sensitive) version in my uBITX boards and the Axicom relays Mike Doty and I installed in place of them, the Axicom had lower coil resistance in the 400 mW coils vs the 200 mW coils fot the Hongfa relays. That means fewer turns of wire in the Axicoms and subsequently less inductance. That may be a huge factor in the isolation between the two types. The coil resistance for the Axicom that matches the Hongfa in DC spec aldo has the exact same listed coil resistance.

I'm going to order 5 each of the 150mW, 200 mW and 500 mW versions, put sockets on my 3rd V4 board and see how much difference there is between the 3 other types relative to the already installed 400 mw version. Hopefully that will give us some definitive answers. Be Monday before I can order them so it will be a few days before any testing can be done.

Jim Sheldon, W0EB


Kees T
 

i agree with Raj and that an excellent "standard" test configuration, for all to use, would be to look at the coupling between the N/C contacts on both sides which would include the relay armature length and compare the 2. Axicom relays are "in the mail".

73 Kees K5BCQ

Kees T
 

Plots of the coupling between N/C contacts on either side of the relay (which will include the armature) and what they look like with the addition of a 47nF capacitor to ground at each coil terminal.  You can see the dB of isolation improvement.

#1 HK16F-SHG "16 pin" relay (like the original uBITX relays)
#2  same as #1 but with the 2 capacitors added (see less coupling ....a good thing).
#3 NAIS TQ2-H "10 pin" smaller relay which does not have the long armature
#4 same as #3 but with the 2 capacitors added (less coupling .....a good thing)
#5 Axicom .....TBD waiting for the relays.
#6 same as #5 but with the 2 capacitors added (less coupling .....a good thing)

73 Kees K5BCQ




Mike Doty
 

For those that are following this topic, last night I "dissected" one of the Axicom 403 relays and one of the HFD27 relays.  Here are the pictures of "what's under the hood".  One shot of the HFD27 and two of the Axicom to make the plates between the contacts and the coil a little clearer.

Mike Doty
 

Here is the second shot inside the Axicom 403 that should have been in the previous post...

iz oos
 

Looking at pic #2 seems to me that adding a 47nF capacitor to ground at each coil terminal as Arv hypothesized would make the uBitx a leap close to be legal compliant with respect to the harmonics.

Il 14/ott/2018 20:31, "Kees T" <windy10605@...> ha scritto:
>
> Plots of the coupling between N/C contacts on either side of the relay (which will include the armature) and what they look like with the addition of a 47nF capacitor to ground at each coil terminal.  You can see the dB of isolation improvement.
>
> #1 HK16F-SHG "16 pin" relay (like the original uBITX relays)
> #2  same as #1 but with the 2 capacitors added (see less coupling ....a good thing).
> #3 NAIS TQ2-H "10 pin" smaller relay which does not have the long armature
> #4 same as #3 but with the 2 capacitors added (less coupling .....a good thing)
> #5 Axicom .....TBD waiting for the relays.
> #6 same as #5 but with the 2 capacitors added (less coupling .....a good thing)
>
> 73 Kees K5BCQ
>
>
>
>

Joe Puma
 

Yea but changing the relays looks like the real winner here. 


Joe



On Oct 14, 2018, at 5:31 PM, iz oos <and2oosiz2@...> wrote:

Looking at pic #2 seems to me that adding a 47nF capacitor to ground at each coil terminal as Arv hypothesized would make the uBitx a leap close to be legal compliant with respect to the harmonics.

Il 14/ott/2018 20:31, "Kees T" <windy10605@...> ha scritto:
>
> Plots of the coupling between N/C contacts on either side of the relay (which will include the armature) and what they look like with the addition of a 47nF capacitor to ground at each coil terminal.  You can see the dB of isolation improvement.
>
> #1 HK16F-SHG "16 pin" relay (like the original uBITX relays)
> #2  same as #1 but with the 2 capacitors added (see less coupling ....a good thing).
> #3 NAIS TQ2-H "10 pin" smaller relay which does not have the long armature
> #4 same as #3 but with the 2 capacitors added (less coupling .....a good thing)
> #5 Axicom .....TBD waiting for the relays.
> #6 same as #5 but with the 2 capacitors added (less coupling .....a good thing)
>
> 73 Kees K5BCQ
>
>
>
>

Ripley
 

Hello All,

 

I installed the Axicom relays in my uBITX (ver 4) test bed this afternoon and ran tests 3.8, 7.2, and 14.2 Mhz. Here are the before and after snapshots. Yellow is SSB and magenta CW.

 

Pic #1 is 7.2 MHz before relay change

Pic #2; 7.2 Mhz after relay change. Note that the CW third order harmonic was brought down by about 7 dBs but still exceeds the -43 dBs threshold for FCC compliance

Pic #3; 3.8Mhz before relay change

Pic #4; 3.8 Mhz after relay change

 

The biggest drops were in SSB mode as others have pointed out already.

 

So you know the base line of my build. The 45 Mhz crystal for spur reduction has been installed but only the crystal itself. Driver transistors Q92, Q93, Q96, Q97 were replaced with metal can 2N2222A and the emitter resistor changed to 11 ohms. AGC was added along with a manual RF gain control.

 

Ripley

KD8UYQ

 

 

Sent from Mail for Windows 10

 

From: Joe Puma
Sent: Sunday, October 14, 2018 5:57 PM
To: BITX20@groups.io
Subject: Re: [BITX20] Harmonics and Relay Replacement

 

Yea but changing the relays looks like the real winner here. 

 

 

Joe

 

 


On Oct 14, 2018, at 5:31 PM, iz oos <and2oosiz2@...> wrote:

Looking at pic #2 seems to me that adding a 47nF capacitor to ground at each coil terminal as Arv hypothesized would make the uBitx a leap close to be legal compliant with respect to the harmonics.

Il 14/ott/2018 20:31, "Kees T" <windy10605@...> ha scritto:
>
> Plots of the coupling between N/C contacts on either side of the relay (which will include the armature) and what they look like with the addition of a 47nF capacitor to ground at each coil terminal.  You can see the dB of isolation improvement.
>
> #1 HK16F-SHG "16 pin" relay (like the original uBITX relays)
> #2  same as #1 but with the 2 capacitors added (see less coupling ....a good thing).
> #3 NAIS TQ2-H "10 pin" smaller relay which does not have the long armature
> #4 same as #3 but with the 2 capacitors added (less coupling .....a good thing)
> #5 Axicom .....TBD waiting for the relays.
> #6 same as #5 but with the 2 capacitors added (less coupling .....a good thing)
>
> 73 Kees K5BCQ
>
>
>
>