Topics

New ideas for the Audio TX/RX pop and PA output stage improvements on ubitx.net

Mike Woods
 

If you look on the home page of http://www.ubitx.net you will find suggestions from Erhard Haertel DF3FY that address:

  • Audio Pops on transitions from RX to TX
  • Improved output on all bands from the PA stage (using RDF16HHF1s and a new output transformer and minor changes to the circuitry).

Erhard had difficulty posting to the BITX20 iogroups list so wrote to me directly.  I have not tested these mods, but these are two concerns about the uBITx (loud audio pops and lower power out above 40m) so I am sure that Erhard would be as interested as me in any feedback from those trying out his mods.

73 Mike ZL1AXG

--
Mike Woods
mhwoods@...

John <passionfruit88@...>
 

Excellent, thanks Mike for posting this

Great web site by the way. Thank you for maintaining it.

One point though about this mod, I can't find reference to the ferrite balun BN43-201. Would it be BN43-202 instead by any chance? 

All the best,

73, John (VK2ETA)

RCBoatGuy
 

Also, what kind of wire is recommended on the new transformer?

Thanks,

Carl, K0MWC

M Garza <mgarza896@...>
 

I am pretty sure that it is a BN43-202.  Now, where did I put them and that 24ga teflon wire....

Thank you for the great info!

Marco - KG5PRT 

On Feb 16, 2018 5:14 PM, "RCBoatGuy via Groups.Io" <ijnfan-HamRadio=yahoo.com@groups.io> wrote:
Also, what kind of wire is recommended on the new transformer?

Thanks,

Carl, K0MWC

Nick VK4PP
 

Will it matter if you wind the bn43-202 so that all the wires come out one side only? Still a 2:3 ratio. This will make placing it where the original t11 was look and fit rather well I imagine?
73 and thanks for the excellent mod!

Nick VK4PP
 

Like so

William R Maxwell
 

What counts as a turn in one of those binocular types - up one side and down the other toexit at the same end or just up one side, like a toroid?


On 17/02/2018 4:24 PM, Nik VK4PLN wrote:
Like so

John Backo
 

Bill:

One turn is up one hole and back down through the other hole,
Thus it is double what in normal on a toroid, i.e., the wire
goes twice through the ferrite to make one turn.

And a complete turn always puts the two wire ends on the same side.
E.g., you could easily build a center-tapped winding by putting 5 1/2 turns (say),
twisting a loop for a future connection, and continuing for another 5 1/2
turns to complete the winding. Then a transformer could be made with
another winding starting at the other side (say).

These binocular types make excellent transformers.

john
AD5YE

MVS Sarma
 

While implementing 1/2 turn we can bring back the wire to same side, but on the outside of the bicocular. this dcan be clearly seen in the article mixer musings.
 

Regards
MVS Sarma
 

On Sat, Feb 17, 2018 at 11:28 AM, John Backo <jabac@...> wrote:
Bill:

One turn is up one hole and back down through the other hole,
Thus it is double what in normal on a toroid, i.e., the wire
goes twice through the ferrite to make one turn.

And a complete turn always puts the two wire ends on the same side.
E.g., you could easily build a center-tapped winding by putting 5 1/2 turns (say),
twisting a loop for a future connection, and continuing for another  5 1/2
turns to complete the winding. Then a transformer could be made with
another winding starting at the other side (say).

These binocular types make excellent transformers.

john
AD5YE




William R Maxwell
 

Thanks John, I had assumed that to be the case but then realised that I had never previously thought about it. I have used them on a few occasions but always with the winding  information already specified - other persons' designs, in other words.

Bill

VK7MX


On 17/02/2018 4:58 PM, John Backo wrote:
Bill:

One turn is up one hole and back down through the other hole,
Thus it is double what in normal on a toroid, i.e., the wire
goes twice through the ferrite to make one turn.

And a complete turn always puts the two wire ends on the same side.
E.g., you could easily build a center-tapped winding by putting 5 1/2 turns (say),
twisting a loop for a future connection, and continuing for another  5 1/2
turns to complete the winding. Then a transformer could be made with
another winding starting at the other side (say). 

These binocular types make excellent transformers.

john
AD5YE





Rob Snow
 

I'm learning here.  I understand that a mica capacitor is the best type to use with RF from reading and they are the most stable.

What would the impact be of using a polystyrene or ceramic capacitor in place of the 33pf?  I'm assuming you would see distortion, but I don't know.

Mike Woods
 

On 17/02/18 10:53 AM, John wrote:
I can't find reference to the ferrite balun BN43-201. Would it be BN43-202 instead by any chance?
John - I will ask Erhard DF3FY whether this was an error.  I suspect it was.

73 Mike

--
Mike Woods
mhwoods@...

M Garza <mgarza896@...>
 

Ceramic capacitors would work fine.  Just make sure it is rated for the voltage that will be present.  I use 500 or 1KV caps, personally.  Over kill, yes, but I have not had one fail yet.

Just my $0.02.

Marco - KG5PRT

On Sat, Feb 17, 2018 at 1:49 PM, Rob Snow <rsnow@...> wrote:
I'm learning here.  I understand that a mica capacitor is the best type to use with RF from reading and they are the most stable.

What would the impact be of using a polystyrene or ceramic capacitor in place of the 33pf?  I'm assuming you would see distortion, but I don't know.


Jerry Gaffke
 

Silver mica caps work fine, were state of the art when most hams learned how to build LPF's and such.
But that was 50 years ago.
Surface mount ceramic caps of type np0 or c0g are far cheaper and just as good.
I'd recommend something rated for at least 200 volts for a uBitx transmit filter, maybe even 500v.
Probably in a 1206 or 1210 size package.
Avoid other ceramic dielectrics such as x5r and x7r, fine for bypass caps but capacitance can vary wildly with applied voltage.
Polystyrene should be fine if that's what you got.

Through-hole caps are generally more expensive than surface mount,
as handling of surface mount parts is easier and sales volumes are much higher.

Jerry, KE7ER



On Sat, Feb 17, 2018 at 11:49 am, Rob Snow wrote:
I'm learning here.  I understand that a mica capacitor is the best type to use with RF from reading and they are the most stable.
 
What would the impact be of using a polystyrene or ceramic capacitor in place of the 33pf?  I'm assuming you would see distortion, but I don't know.

K9HZ <bill@...>
 

The mistake that most people make in RF circuits is not paying attention to the current rating of the part.  Make sure your capacitor can handle the current for its service.  The voltage, size, and rating (eg NPO) etc. are the easy parts.

 

 

Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ

 

Owner - Operator

Big Signal Ranch – K9ZC

Staunton, Illinois

 

Owner – Operator

Villa Grand Piton – J68HZ

Soufriere, St. Lucia W.I.

Rent it: www.VillaGrandPiton.com

Like us on Facebook! facebook icon

 

Moderator – North American QRO Yahoo Group.

 

email:  bill@...

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of Rob Snow
Sent: Saturday, February 17, 2018 1:49 PM
To: BITX20@groups.io
Subject: Re: [BITX20] New ideas for the Audio TX/RX pop and PA output stage improvements on ubitx.net

 

I'm learning here.  I understand that a mica capacitor is the best type to use with RF from reading and they are the most stable.

 

What would the impact be of using a polystyrene or ceramic capacitor in place of the 33pf?  I'm assuming you would see distortion, but I don't know.


Virus-free. www.avg.com

Jerry Gaffke
 


On Sat, Feb 17, 2018 at 12:19 pm, K9HZ wrote:

The mistake that most people make in RF circuits is not paying attention to the current rating of the part.  Make sure your capacitor can handle the current for its service.  The voltage, size, and rating (eg NPO) etc. are the easy parts.

 

K9HZ <bill@...>
 

Yes that makes sense…

 

The current through a capacitor is given by I=C*dV/dt so the average current over time is I=C*int[dV/dt]/t… For your example: let’s use a 100pf capacitor in an 30 MHz circuit at 50 VRMS which is 140V P-P.  Iavg = 0.42 * 0.707 = 0.3 amps average.  Now the spec from that capacitor in your finding is 0.1% at 1 VRMS at 1MHZ which is 0.00028A * 0.707RMS *1000% = 0.2 Amps continuous at 30 MHz.  Yikes we are over by 50%!  Note that this DOES vary with capacitance… that’s why the larger the capacitor, the more current capacity is needed.

 

There are plenty of calculators on the internet to figure this out…. Google “capacitor Current Calculator”.

 

 

 

Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ

 

Owner - Operator

Big Signal Ranch – K9ZC

Staunton, Illinois

 

Owner – Operator

Villa Grand Piton – J68HZ

Soufriere, St. Lucia W.I.

Rent it: www.VillaGrandPiton.com

Like us on Facebook! facebook icon

 

Moderator – North American QRO Yahoo Group.

 

email:  bill@...

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of Jerry Gaffke via Groups.Io
Sent: Saturday, February 17, 2018 3:18 PM
To: BITX20@groups.io
Subject: Re: [BITX20] New ideas for the Audio TX/RX pop and PA output stage improvements on ubitx.net

 


This Vishay datasheet is fairly typical.
    

https://www.mouser.com/ds/2/427/vjcommercialseries-222796.pdf

Closest thing I see to a "current rating" for the c0g parts is the Dissipation Factor on page 1. 
Not quite enough information.
They spec a dissipation of 0.1% at 1.0 VRMS and 1 MHz,
but we might be using it at 50 VRMS and 30 MHz.

Any further advice on how to evaluate caps for use in a LPF?


Though note that this c0g spec is much better than the x7r spec,
which claims 2.5% dissipation at 1.0 VRMS and 1 KHz.

Jerry

On Sat, Feb 17, 2018 at 12:19 pm, K9HZ wrote:

The mistake that most people make in RF circuits is not paying attention to the current rating of the part.  Make sure your capacitor can handle the current for its service.  The voltage, size, and rating (eg NPO) etc. are the easy parts.

 


Virus-free. www.avg.com

Jerry Gaffke
 

I find that train of thought rather confusing.
Not sure where that 0.3amps average comes from, I'd think you could just find
the impedance from 1/(2*pi*Hz*Farads) and divide that into the 50 volts RMS to find current
rather than bother with the calculus.
Not at all clear how a 0.1% dissipation spec somehow means a max of 0.2 Amps at 30mhz.
A few more details would be appreciated.

I think most builders just try some np0's and if they get warm then replace with a bunch of smaller value caps in parallel.
I'm not aware of any Bitx40 LPF caps blowing out, and I doubt they are anything special.

Jerry


On Sat, Feb 17, 2018 at 02:11 pm, K9HZ wrote:

The current through a capacitor is given by I=C*dV/dt so the average current over time is I=C*int[dV/dt]/t… For your example: let’s use a 100pf capacitor in an 30 MHz circuit at 50 VRMS which is 140V P-P.  Iavg = 0.42 * 0.707 = 0.3 amps average.  Now the spec from that capacitor in your finding is 0.1% at 1 VRMS at 1MHZ which is 0.00028A * 0.707RMS *1000% = 0.2 Amps continuous at 30 MHz.  Yikes we are over by 50%!  Note that this DOES vary with capacitance… that’s why the larger the capacitor, the more current capacity is needed.

 

Jerry Gaffke
 

History with the Bitx40 doesn't mean much if we're building a LPF for 30mhz.
Especially if we get the power up to 20 or 30 Watts.
But after poking around some I think brand name np0/c0g's from reputable vendors should be fine, 
you don't want junk.  Smaller caps in parallel are a good idea if experiencing trouble with currents.

This guy seem to know what he's doing.  His 1.5kW 30mhz LPF uses a the "1kV type MC22 Mica Capacitor",
perhaps Mouser 598-MC22FF101J-TF.  Can get them at a reasonable price if you want to take this to extremes..  
  http://w6pql.com/a_1.5kw_lpf_for_160-6m.htm


Jerry, KE7ER


On Sun, Feb 18, 2018 at 12:06 am, Jerry Gaffke wrote:
I think most builders just try some np0's and if they get warm then replace with a bunch of smaller value caps in parallel.
I'm not aware of any Bitx40 LPF caps blowing out, and I doubt they are anything special.

K9HZ <bill@...>
 

Yes, I suppose for the amateur builder, applying power to your design in and finger testing is an option.  If it doesn’t burst into flames, I suppose its ok?  (In commercial designing we don’t often have that option so we’ll have to stick to our calculus).  It might be wise to invest in one of those IR temperature guns from Harbor Freight Tools for $20 to save your fingertips though!

 

For LP design work at the approx.. 10 watt level… this just isn’t gonna be an issue unless you are using the wrong filter for the band and the filter is dissipating all of the power.  Once you get to the kW+ region… best use your math.

 

Like I said in my original note, if you are uncomfortable with the math, there are many calculators on the internet that work.  Just google them.

 

To be clear about the specification… it says that you would be operating at 0.1% of the full/ capable dissipation properties of the capacitor if it were used in a 50 VRMS circuit at 1 MHz  Using this one data point, you can translate that to any voltage/ frequency point and ratio it to this point to determine what % of dissipation you are operating (designing) at.

 

 

Dr. William J. Schmidt - K9HZ J68HZ 8P6HK ZF2HZ PJ4/K9HZ VP5/K9HZ PJ2/K9HZ

 

Owner - Operator

Big Signal Ranch – K9ZC

Staunton, Illinois

 

Owner – Operator

Villa Grand Piton – J68HZ

Soufriere, St. Lucia W.I.

Rent it: www.VillaGrandPiton.com

Like us on Facebook! facebook icon

 

Moderator – North American QRO Yahoo Group.

 

email:  bill@...

 

 

From: BITX20@groups.io [mailto:BITX20@groups.io] On Behalf Of Jerry Gaffke via Groups.Io
Sent: Sunday, February 18, 2018 2:06 AM
To: BITX20@groups.io
Subject: Re: [BITX20] New ideas for the Audio TX/RX pop and PA output stage improvements on ubitx.net

 

I find that train of thought rather confusing.
Not sure where that 0.3amps average comes from, I'd think you could just find
the impedance from 1/(2*pi*Hz*Farads) and divide that into the 50 volts RMS to find current
rather than bother with the calculus.
Not at all clear how a 0.1% dissipation spec somehow means a max of 0.2 Amps at 30mhz.
A few more details would be appreciated.

I think most builders just try some np0's and if they get warm then replace with a bunch of smaller value caps in parallel.
I'm not aware of any Bitx40 LPF caps blowing out, and I doubt they are anything special.

Jerry


On Sat, Feb 17, 2018 at 02:11 pm, K9HZ wrote:

The current through a capacitor is given by I=C*dV/dt so the average current over time is I=C*int[dV/dt]/t… For your example: let’s use a 100pf capacitor in an 30 MHz circuit at 50 VRMS which is 140V P-P.  Iavg = 0.42 * 0.707 = 0.3 amps average.  Now the spec from that capacitor in your finding is 0.1% at 1 VRMS at 1MHZ which is 0.00028A * 0.707RMS *1000% = 0.2 Amps continuous at 30 MHz.  Yikes we are over by 50%!  Note that this DOES vary with capacitance… that’s why the larger the capacitor, the more current capacity is needed.

 


Virus-free. www.avg.com