Thoughts on Low Power and fewer turns on LP filter toroids


Phil/K3UT
 

I've completed a 20M QCX+ and have just fixed a friend's 30M QCX+. Both are working nicely, but at 13.1 volts I get about 3watts output on both, using the onboard RF tool to measure (and less on my Elecraft wattmeter). 

To achieve that output, I had to take two turns off L1 and L3 and one turn off of L2 and fiddle around a little with the spacing of the windings. I did that with both radios.

Any thoughts why the number of turns differs so greatly from that specified in the build instructions? Both kits were shipped about the same time. Could it be a quality control issue of the toroids? I measured the toroids on the 20M version before I installed them, and they were spot on before I took the turns off. I didn't measure them after the turns had been removed.

I'm satisfied with the power output of the units.  Just curious why they aren't meeting the specs as stated in the manual?

Phil, K3UT


ajparent1/kb1gmx
 

Phil,

THe AL used for winding of toroids has a variation of 10-15% so depending on
how tightly wound  plus individual batches you get variation.

The Iron powder material is really like many high volume materials to get it
as close as they tend to be takes a significant bit of effort.  I strongly point
to the spec.


Add to that the caps used are 10% maybe 5% so when you plug in the variation 
it does get to be significant.  That said I'se used those very same table and
many over a 25 filters then tested them ( 5 bands * 5 per) and then tested
them and they were remarkably consistent.  Parts used were not the same
batches but were drawn from my own stocks some quite old.

Also did you adjust L4?

All 4 of those inductors interact and due to reflected impedances and determine
power and harmonic content.

Allison
------------------------------
Please reply on list so we can share.
No private email, it goes to a bit bucket due address harvesting


Arv Evans <arvid.evans@...>
 

In addition to Allison's comments, the filters are designed 
to work with 50 ohms input impedance and 50 ohms output 
impedance.  If either of those impedances are not 50 ohms 
the filter cutoff frequency will be impacted.

Arv
_._


On Tue, Sep 8, 2020 at 11:29 AM ajparent1/KB1GMX <kb1gmx@...> wrote:
Phil,

THe AL used for winding of toroids has a variation of 10-15% so depending on
how tightly wound  plus individual batches you get variation.

The Iron powder material is really like many high volume materials to get it
as close as they tend to be takes a significant bit of effort.  I strongly point
to the spec.


Add to that the caps used are 10% maybe 5% so when you plug in the variation 
it does get to be significant.  That said I'se used those very same table and
many over a 25 filters then tested them ( 5 bands * 5 per) and then tested
them and they were remarkably consistent.  Parts used were not the same
batches but were drawn from my own stocks some quite old.

Also did you adjust L4?

All 4 of those inductors interact and due to reflected impedances and determine
power and harmonic content.

Allison
------------------------------
Please reply on list so we can share.
No private email, it goes to a bit bucket due address harvesting


Michael.2E0IHW
 

Apart from soldering issues, this must be  the most confounding disincentive  for faint-hearted wonnabe QSX+ers ...

There must surely be a concise winding-checklist for genuine radio "amateurs" ?

Michael 2E0IHW
...........................................................

On 08/09/2020 17:57, Phil/K3UT (ex W3HZZ) wrote:
I've completed a 20M QCX+ and have just fixed a friend's 30M QCX+. Both are working nicely, but at 13.1 volts I get about 3watts output on both, using the onboard RF tool to measure (and less on my Elecraft wattmeter). 

To achieve that output, I had to take two turns off L1 and L3 and one turn off of L2 and fiddle around a little with the spacing of the windings. I did that with both radios.

Any thoughts why the number of turns differs so greatly from that specified in the build instructions? Both kits were shipped about the same time. Could it be a quality control issue of the toroids? I measured the toroids on the 20M version before I installed them, and they were spot on before I took the turns off. I didn't measure them after the turns had been removed.

I'm satisfied with the power output of the units.  Just curious why they aren't meeting the specs as stated in the manual?

Phil, K3UT


Hans Summers
 

Hi all

This problem has always come up at intervals on this group. 

It used to be that we could blame suspect capacitors for being lossy since several were found on occasion. 

Not any more! All capacitors are now NP0 ceramics manufactured by Vishay and procured from Digikey. All the toroids are from Diz @ kitsandparts.com. I don't think it at all likely that there are any quality issues with these US sourced parts. 

Which leads us just back to variations in component tolerances, winding styles, soldering (check grounded pins of capacitors which are harder to get soldered well to the groundplane!). And variations and errors in power measurement. All these variations are the most likely cause all along on most previous occasions too... but for sure the capacitors and toroids in the LPFs are really always good now. 

73 Hans G0UPL 

On Tue, Sep 8, 2020, 21:44 Michael.2E0IHW via groups.io <blumu=btinternet.com@groups.io> wrote:
Apart from soldering issues, this must be  the most confounding disincentive  for faint-hearted wonnabe QSX+ers ...

There must surely be a concise winding-checklist for genuine radio "amateurs" ?

Michael 2E0IHW
...........................................................

On 08/09/2020 17:57, Phil/K3UT (ex W3HZZ) wrote:
I've completed a 20M QCX+ and have just fixed a friend's 30M QCX+. Both are working nicely, but at 13.1 volts I get about 3watts output on both, using the onboard RF tool to measure (and less on my Elecraft wattmeter). 

To achieve that output, I had to take two turns off L1 and L3 and one turn off of L2 and fiddle around a little with the spacing of the windings. I did that with both radios.

Any thoughts why the number of turns differs so greatly from that specified in the build instructions? Both kits were shipped about the same time. Could it be a quality control issue of the toroids? I measured the toroids on the 20M version before I installed them, and they were spot on before I took the turns off. I didn't measure them after the turns had been removed.

I'm satisfied with the power output of the units.  Just curious why they aren't meeting the specs as stated in the manual?

Phil, K3UT


Curt wb8yyy
 

Phil

the 5 watt outside is specified for higher voltage than 13.8 volts.  okay a few claim more, but I trust the designer's data. 

your rigs are working wonderfully 3 watts to 5 watts is a small fraction of one S-unit !  only your power meter knows the difference. 

okay to tweak a rig when building - otherwise we may as well order it already built (funny thing, that option is coming!). 

all seems well - enjoy its radiations - wow these QCX have wonderful receivers! 

73 Curt


Phil/K3UT
 

I’ve always kept a 50ohm dummy load on the output. I will take a look at L4 tomorrow, first by compressing and then spacing the windings.  Then i may remove a turn. More later. 

Phil K3UT 


Brian George
 

Just a passing thought ...if component tolerances are a key factor contributing to low output when using the 20m and 30m LPFs then why aren't about 50% of constructors having to ADD turns and why is it only on 20m and 30m? I had exactly the same issue but a VNA scan showed that the 3dB cut-off points were ideal for both 20m and 30m - about 11% above the highest operating frequency. What I did discover was that the insertion loss was a little higher on the 20m LPF and the passband on the 30m LPF sloped down towards the cut-off pojnt instead of being flat. 

73
Brian
G3ZOH  


Shirley Dulcey KE1L
 

If the inductor values are low on a low pass filter, which is why you might want to add turns, it will cause the cutoff frequency to be higher than it was designed to be. You won't detect that issue by seeing low power output. It will cause the radio's harmonic suppression to be reduced but you'll need a spectrum analyzer to find that, and the radio will probably still meet the standards for spurious emissions. It could also change the response curve a bit, but again you won't detect that without equipment to measure the filter response. (I'd probably look for that by attaching a noise generator to the input, probing the receiver just after the low pass filter, and feeding that signal into either a spectrum analyzer or a wideband SDR. That would show the filter cutoff and shape, including any passband ripple.

If loading the circuit at that point turns out to be a problem, an alternate place to test would be immediately after one of the receive preamps -- on a QCX+ that's either JP7 or JP8. That has the downside of also bringing in the shape of the bandpass filter that is one of the functions of T1 and its neighboring components.


On Wed, Sep 9, 2020 at 5:21 AM Brian George via groups.io <george.familyuk=ntlworld.com@groups.io> wrote:
Just a passing thought ...if component tolerances are a key factor contributing to low output when using the 20m and 30m LPFs then why aren't about 50% of constructors having to ADD turns and why is it only on 20m and 30m? I had exactly the same issue but a VNA scan showed that the 3dB cut-off points were ideal for both 20m and 30m - about 11% above the highest operating frequency. What I did discover was that the insertion loss was a little higher on the 20m LPF and the passband on the 30m LPF sloped down towards the cut-off pojnt instead of being flat. 

73
Brian
G3ZOH  


Hans Summers
 

Hi Brian
 
Just a passing thought ...if component tolerances are a key factor contributing to low output when using the 20m and 30m LPFs then why aren't about 50% of constructors having to ADD turns and why is it only on 20m and 30m? I had exactly the same issue but a VNA scan showed that the 3dB cut-off points were ideal for both 20m and 30m - about 11% above the highest operating frequency. What I did discover was that the insertion loss was a little higher on the 20m LPF and the passband on the 30m LPF sloped down towards the cut-off pojnt instead of being flat. 

The number of turns specified in the manual tends to be a bit on the high side. Why? Because it is the theoretically calculated number of turns. But the theory assumes an evenly spaced turn distribution around the core which in practice is not achieved, due to several factors including there is always a gap between the ends of the windings, and that perfectly evenly spacing is impossible. Furthermore there are variations in winding styles, some people wind the wire very tight against the toroid former and for others, more loose windings leave a bit of an air gap between the toroid and the wire. 

I am quite happy with the turns specification in the manual being a bit on the high side. Why? Well, have you ever tried adding a turn to a toroid? It's certainly a lot easier to remove them :-) 

Another point is that a straightforward plot of the characteristics of the LPF on a VNA or spectrum analyzer is not the whole story either. This is because the BS170's end of the LPF has to be matched to the impedance of the PA. Since capacitor C30, the three BS170s, and L4 (all these are elements of the Class-E resonant drain load circuit) are also variable due to component tolerances/winding styles, the impedance match is an additional factor. In my opinion you just have to play with it. 

As an example. Here I am testing and aligning some QCX+ kits today which have been built here at QRP Labs, for selling as Assembled radios. Most are for 40m and are generally easier. But I have one on the bench right now which is a 30m unit. Initially the Vpp on the 'scope was 24V which is just 1.4W of output power with 12.0V supply. 

Time to play! 

I removed two turns from L3 and spread the turns out - I find L3 often needs a turn or two removed, it is the inductor closest to the PA and I think getting this impedance matched better is a good first step. Then I bunched up L2 and L1 a bit. In the end I now read 4.6W with 12.0V supply, increasing to 6.2W with 13.8V supply. From 1.4W to 4.6W all just by altering the inductances of L1/L2/L3. Remember removing turns is not always the answer - in this case, two turns removed from L3 reduces the inductance; but L1 and L2 I just squeezed the turns (none removed) which increases the inductance. 

73 Hans G0UPL
http://qrp-labs.com


Brian George
 

Thank you Hans and also Shirley for your helpful replies. All understood and noted.

73
Brian
G3ZOH

ps - Hans - hope you and your family are keeping well. 


Phil/K3UT
 

Oh, I am happy with my rigs and their output. (And their ability to get out!). I was curious why they weren't the spec'd output wasn't like the build manual.

I wanted to raise this as a discussion about some of the reasons why there seems to be variation in output of the QCX and that seems to be random among builders. 

It's nice to hear why we only remove turns (hard to put them on) and learn that the toroids and caps are now FB and - as Hans says - to get the output to change it's "time to play"

Thanks all for the discussion and the education. Maybe by the time I build my fifth QCX, I will understand how it works (and be ready for a QSX!).

Phil, K3UT


Ian VA7ITM
 

(Excellent presentation at the GQRP Conference last weekend Hans.)

My QCX+ 40m with all the trimmings is in the mail and I'm looking forward to building my first HF radio soon.

Having the above paragraphs from Hans added to the manual would be very useful.

A couple of questions...

1) I have an LCR meter. Rather than wind the documented number of turns, should I instead add windings until I measure the documented inductance?
2) Given Han's comments about bunching up some turns, does it make sense to add a blob of hot glue to the toroids once I am happy with the power output? I saw earlier comments about not adding heat shrink sleeving and am wondering if similar thoughts apply?

Kind Regards
Ian VA7ITM


Phil/K3UT
 

OK gang, after reading Hans' post, I decided it was my time to play, so I took the QCX+ 30M that has two less turns on L1 and L3 and 1 less turn on L2, and I squeezed together turns on L2 and L1, leaving L3 alone. Power increased from 3w to 4.8w at 13.5 volts. 

Did the same with my QCX+20 (L1, L2 turns squeezed) and the output increased from 3W to 4.9W.

Time to get on the air!

Phil, K3UT


Ronald Taylor
 

Ian, on the 40 meter QCX/QCX+ I've had experience with several 40 meter versions and all of them came out with 4-5 watts out using the prescribed number of turns on the low pass filter toroids. It's 20 meters and higher that some folks are seeing the necessity of removing a few turns. No heat shrink or hot glue are necessary on the toroids. Good luck with your kit ... Ron


On Wed, Sep 9, 2020 at 12:30 PM Ian VA7ITM <ian.mcalpine@...> wrote:
(Excellent presentation at the GQRP Conference last weekend Hans.)

My QCX+ 40m with all the trimmings is in the mail and I'm looking forward to building my first HF radio soon.

Having the above paragraphs from Hans added to the manual would be very useful.

A couple of questions...

1) I have an LCR meter. Rather than wind the documented number of turns, should I instead add windings until I measure the documented inductance?
2) Given Han's comments about bunching up some turns, does it make sense to add a blob of hot glue to the toroids once I am happy with the power output? I saw earlier comments about not adding heat shrink sleeving and am wondering if similar thoughts apply?

Kind Regards
Ian VA7ITM


Steven Dick, K1RF
 

My own experience is the following.  I recently completed a 40 meter QCX+ and used one of the ebay LC200A LC meters to adjust inductances up or down as needed by squeezing or adding/removing turns.  Operating at 12v, it puts out 4.75 watts on an accurately calibrated meter.  That's about as good as you can get on a 12V supply. I did the same with an early QCX, also on 40 meters which has been put through a lot of rough handling and upgrades to the latest hardware revisions and software. That one put out 3.6W with a 12V power supply.  I squeezed or spread coils to get the power up to 4.15 watts.  My goal was to get above 4 watts with a 12V supply. I could have gotten more power by further playing but stopped there and made a nice QRP contact with it. Didn't have to mess with removing turns.  I was surprised to find that squeezing or spreading turns on L4 had almost no effect on power output.  That's part of the resonant class E circuit.

-Steve K1RF 

------ Original Message ------
From: "Ian VA7ITM" <ian.mcalpine@...>
Sent: 9/9/2020 3:30:52 PM
Subject: Re: [QRPLabs] Thoughts on Low Power and fewer turns on LP filter toroids

(Excellent presentation at the GQRP Conference last weekend Hans.)

My QCX+ 40m with all the trimmings is in the mail and I'm looking forward to building my first HF radio soon.

Having the above paragraphs from Hans added to the manual would be very useful.

A couple of questions...

1) I have an LCR meter. Rather than wind the documented number of turns, should I instead add windings until I measure the documented inductance?
2) Given Han's comments about bunching up some turns, does it make sense to add a blob of hot glue to the toroids once I am happy with the power output? I saw earlier comments about not adding heat shrink sleeving and am wondering if similar thoughts apply?

Kind Regards
Ian VA7ITM

Virus-free. www.avast.com


Steve in Okinawa
 

This is an ancient patent, suggesting that toroidal inductors can be tuned using a variable magnetic field. Thoughts? https://patents.google.com/patent/US2915637A/en
JS6TMW


Hans Summers
 

Hi Steve

It works. I did it. But the range of inductance variation is not huge. See http://hanssummers.com/huffpuff/fast.html - my VFO with Huff Puff method stabilization, using a relay coil to vary the magnetic field a nearby toroidal inductor is sitting in. This stabilized VFO is inside my 40/80m Receiver http://hanssummers.com/polyphase with which I had my first ever ham QSO in March 2002. 

73 Hans G0UPL
http://qrp-labs.com

83 

On Thu, Sep 10, 2020 at 3:45 AM Steve in Okinawa <sfab43@...> wrote:
This is an ancient patent, suggesting that toroidal inductors can be tuned using a variable magnetic field. Thoughts? https://patents.google.com/patent/US2915637A/en
JS6TMW


Steve in Okinawa
 

Very cool project, Hans. Now who will be the first to make a PPP (permeability power peaker) for a QCX? (And I noted the comment about your earthed soldering iron tip, for only today the mains GCFI blacked out the whole house when I attempted to solder some antenna coax braid.)     73,. Steve


Adrien F4IJA
 

Just my tests :

I've recently built a 40m QCX+ and it seems to work very well at first try. Anyway, I've measured the RF output power at about 1.8 - 2W with 12.3V, so quite far from the informations in the documentation.

I've winded the toroids as told in the documentation and they were well repartited around the toroids with almost the same gap beetween each turn.

I've made this evening some adjustments by squeezing or expanding the loops around the toroids.

Finally, I can now achieve about 4.2W at 12.3V (so, 2W+ more than before), just by squeezing L2. L1 and L3 are well repartited around their toroids.

Here is the picture :



So, I'm more than happy, and I didn't remove turns. This mod took me only about 5 minutes. So, if you have some RF output issues, it's an idea to try to improve that the simple way.