Topics

RD16HHF1 power curve flattening...some

John <passionfruit88@...>
 

Impressive, Bill.

I just tried 180pf across the primary of the output transformer, as opposed to the 33pf I had initially and as you mentioned from your simulations before it raises the power lower down to the point where I don't have the dip at 21Mhz anymore. With RV1 at 60%, 13,8V supply I get 22W at 80M, 20 at 40M, then tapering down to 12W at 15M then flat up to 10M. 

So I have wound down my RV1 to get 9 to 10W on the higher frequencies and I get 18-16W on the lower ones. As I wanted. I am now happy with the power curve, especially when compared to the initial one with the IRF510s (17W to less than 2W from memory).

73, John (VK2ETA)

Nick VK4PP
 

HI John,
Could you please summarise your mods to get these results.
Thanks. Nick

K9HZ <bill@...>
 

Oh yeah, one more thing…  Those RD16HHF1s should draw about 500ma per part (1A total) to put them nicely into the class AB2 operating band for SSB.

 

I guess we better all get our heads together and write this up and a concerted effort so others can do it too…

 

I also ordered a pair of RD70HHF1s tonight to try as an add-in amplifier board at some point.  While these parts are not exactly spirit QRP,  I do like the idea of having 100W available when I want it… maybe with the flip of a switch.

 

 

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 John
Sent: Tuesday, March 20, 2018 10:39 PM
To: BITX20@groups.io
Subject: Re: [BITX20] RD16HHF1 power curve flattening...some

 

Impressive, Bill.

I just tried 180pf across the primary of the output transformer, as opposed to the 33pf I had initially and as you mentioned from your simulations before it raises the power lower down to the point where I don't have the dip at 21Mhz anymore. With RV1 at 60%, 13,8V supply I get 22W at 80M, 20 at 40M, then tapering down to 12W at 15M then flat up to 10M. 

So I have wound down my RV1 to get 9 to 10W on the higher frequencies and I get 18-16W on the lower ones. As I wanted. I am now happy with the power curve, especially when compared to the initial one with the IRF510s (17W to less than 2W from memory).

73, John (VK2ETA)


Virus-free. www.avg.com

John <passionfruit88@...>
 

Hello Nik,

How did you go with the feedback issue?

My power amp mods in summary:  finals as RD16HHF1s, feedback resistors R261/R262 as 820 Ohms, Transformer 2T primary / 3T secondary on BN43-202, 180pF across the output transformer primary and 330pf across R87/R88. That's all...in a way.

Heavily inspired by Erhard's (DF3FY) information as displayed on uBitx.net.

Results per band at 14V supply, RV1 at %60, RD16's biased at 250mA each:
Band      output power   Total current
80M        22W                         2.9A
40M        20W                         3.2A
30M        16W                         2.3A
20M         16W                        2.7A
17M         15W                        2.4A
15M         12W                        1.6A
12M         12W                    not measured
10M         12W                        1.7A

At 22W the output transformer does get warm but not the LPFs.

Wound back RV1 to get 9-10W on the high frequency bands and 16-18W on the lower ones.

All the best,

73, John (VK2ETA)

Henning Weddig
 

even with the mods. the TX string shows a 3 dB drop in output power...

Please exchange the two single chokes L8 / L9 with a bifilar wound coil on a e.g. FT43-50 core, see the reasons for that mod in the description already cited in this forum  "ludens.cl".

BTW: how was the output power determined?

How does the power curve PIn - POut look like,

Is the output power measured at the 1 dB compresion point?

Tests on my tries to build a PA with the RD16HFF1´s using different output transformer configurations (even with a 1:4 guanella transformer with 25 ohm coaxial cable on two BN43-202 cores + Balun with 50 ohm coaxial cable on a FT43-80 torroid core + bifilar choke (FT43-50 torroid) + extra choke on the center tap  with a R-C combinaiton to terminate even harmonics) I was not able to get more than 10 W output power at P1dB.  Therefore I am really sceptical about the high power results of > 16 W.

The "hot" cores show that something must be wrong!

Did You measure the suppression of the even harmonics which should be well suppressed  (at least 40 dB) ? How good are the odd harmoncis suppressed?  If the third harmonic is only 10 dB down (1/3 of the amplitude of the fundamental wave) the PA produces square waves, which is not a performance of a linear amp.

IM performance??

Henning Weddig
DK5LV

Am 21.03.2018 um 12:41 schrieb John:

Hello Nik,

How did you go with the feedback issue?

My power amp mods in summary:  finals as RD16HHF1s, feedback resistors R261/R262 as 820 Ohms, Transformer 2T primary / 3T secondary on BN43-202, 180pF across the output transformer primary and 330pf across R87/R88. That's all...in a way.

Heavily inspired by Erhard's (DF3FY) information as displayed on uBitx.net.

Results per band at 14V supply, RV1 at %60, RD16's biased at 250mA each:
Band      output power   Total current
80M        22W                         2.9A
40M        20W                         3.2A
30M        16W                         2.3A
20M         16W                        2.7A
17M         15W                        2.4A
15M         12W                        1.6A
12M         12W                    not measured
10M         12W                        1.7A

At 22W the output transformer does get warm but not the LPFs.

Wound back RV1 to get 9-10W on the high frequency bands and 16-18W on the lower ones.

All the best,

73, John (VK2ETA)

Glenn
 

 Henning, I too am skeptical about power out above about 16W with RD16HHF1's..

In my setup here, i see flattening of the output waveform around the 14-15W mark, measured  with no LPF on the output. I amusing the G6ALU output topology. I am not yet feeding the supply to the finals via the Balun core method but intend to try that asap.

glenn


On Wed, Mar 21, 2018 at 05:18 am, Henning Weddig wrote:

even with the mods. the TX string shows a 3 dB drop in output power...

Please exchange the two single chokes L8 / L9 with a bifilar wound coil on a e.g. FT43-50 core, see the reasons for that mod in the description already cited in this forum  "ludens.cl".

BTW: how was the output power determined?

How does the power curve PIn - POut look like,

Is the output power measured at the 1 dB compresion point?

Tests on my tries to build a PA with the RD16HFF1´s using different output transformer configurations (even with a 1:4 guanella transformer with 25 ohm coaxial cable on two BN43-202 cores + Balun with 50 ohm coaxial cable on a FT43-80 torroid core + bifilar choke (FT43-50 torroid) + extra choke on the center tap  with a R-C combinaiton to terminate even harmonics) I was not able to get more than 10 W output power at P1dB.  Therefore I am really sceptical about the high power results of > 16 W.

The "hot" cores show that something must be wrong!

Did You measure the suppression of the even harmonics which should be well suppressed  (at least 40 dB) ? How good are the odd harmoncis suppressed?  If the third harmonic is only 10 dB down (1/3 of the amplitude of the fundamental wave) the PA produces square waves, which is not a performance of a linear amp.

IM performance??

Henning Weddig
DK5LV

Am 21.03.2018 um 12:41 schrieb John:
Hello Nik,

How did you go with the feedback issue?

My power amp mods in summary:  finals as RD16HHF1s, feedback resistors R261/R262 as 820 Ohms, Transformer 2T primary / 3T secondary on BN43-202, 180pF across the output transformer primary and 330pf across R87/R88. That's all...in a way.

Heavily inspired by Erhard's (DF3FY) information as displayed on uBitx.net.

Results per band at 14V supply, RV1 at %60, RD16's biased at 250mA each:
Band      output power   Total current
80M        22W                         2.9A
40M        20W                         3.2A
30M        16W                         2.3A
20M         16W                        2.7A
17M         15W                        2.4A
15M         12W                        1.6A
12M         12W                    not measured
10M         12W                        1.7A

At 22W the output transformer does get warm but not the LPFs.

Wound back RV1 to get 9-10W on the high frequency bands and 16-18W on the lower ones.

All the best,

73, John (VK2ETA)

 On Wed, Mar 21, 2018 at 05:18 am, Henning Weddig wrote:

even with the mods. the TX string shows a 3 dB drop in output power...

Please exchange the two single chokes L8 / L9 with a bifilar wound coil on a e.g. FT43-50 core, see the reasons for that mod in the description already cited in this forum  "ludens.cl".

BTW: how was the output power determined?

How does the power curve PIn - POut look like,

Is the output power measured at the 1 dB compresion point?

Tests on my tries to build a PA with the RD16HFF1´s using different output transformer configurations (even with a 1:4 guanella transformer with 25 ohm coaxial cable on two BN43-202 cores + Balun with 50 ohm coaxial cable on a FT43-80 torroid core + bifilar choke (FT43-50 torroid) + extra choke on the center tap  with a R-C combinaiton to terminate even harmonics) I was not able to get more than 10 W output power at P1dB.  Therefore I am really sceptical about the high power results of > 16 W.

The "hot" cores show that something must be wrong!

Did You measure the suppression of the even harmonics which should be well suppressed  (at least 40 dB) ? How good are the odd harmoncis suppressed?  If the third harmonic is only 10 dB down (1/3 of the amplitude of the fundamental wave) the PA produces square waves, which is not a performance of a linear amp.

IM performance??

Henning Weddig
DK5LV

Am 21.03.2018 um 12:41 schrieb John:
Hello Nik,

How did you go with the feedback issue?

My power amp mods in summary:  finals as RD16HHF1s, feedback resistors R261/R262 as 820 Ohms, Transformer 2T primary / 3T secondary on BN43-202, 180pF across the output transformer primary and 330pf across R87/R88. That's all...in a way.

Heavily inspired by Erhard's (DF3FY) information as displayed on uBitx.net.

Results per band at 14V supply, RV1 at %60, RD16's biased at 250mA each:
Band      output power   Total current
80M        22W                         2.9A
40M        20W                         3.2A
30M        16W                         2.3A
20M         16W                        2.7A
17M         15W                        2.4A
15M         12W                        1.6A
12M         12W                    not measured
10M         12W                        1.7A

At 22W the output transformer does get warm but not the LPFs.

Wound back RV1 to get 9-10W on the high frequency bands and 16-18W on the lower ones.

All the best,

73, John (VK2ETA)

John <passionfruit88@...>
 

Henning and Glen,

Thanks you for the feedback. All interesting questions.

I sincerely hope someone with the test equipment can duplicate and fill in the blanks.

The only test I can think of performing is the 1dB compression point. Since I would use an RF probe with a diode I need a high enough input voltage to make the test meaningful even after correction for the diode's threshold.

So if the gain compression is most likely to be in the final can I measure the output of the driver's stage and the output of the finals while I sweep the RV1 pot for an analysis of the 1dB compression? Also should I measure before or after the LPFs?

All the best,

73, John (VK2ETA)

K9HZ <bill@...>
 

“I was not able to get more than 10 W output power at P1dB”

 

You are working with TWO devices in push-pull that are capable of almost 20+ watts per part with properly designed support circuitry.

 

You might try a literature search on amplifiers that use RD16HHF1 parts.  The most prolific amp that comes up is the one designed by UT2FW who has sold thousands of amps of this design. 

 

http://www.ut2fw.com/node/443

 

It’s in Russian, but you can translate it in google translator.  He doesn’t seem to have a problem getting 35 watts out of his design.  I don’t either (with enough drive).  He does use a 1:3 transformer but it is also DC loaded.  And, yes I agree if the transformer is getting hot… a low of power is getting dissipated in the core rather than coupled to the antenna.

 

 

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 Henning Weddig via Groups.Io
Sent: Wednesday, March 21, 2018 7:19 AM
To: BITX20@groups.io
Subject: Re: [BITX20] RD16HHF1 power curve flattening...some

 

even with the mods. the TX string shows a 3 dB drop in output power...

Please exchange the two single chokes L8 / L9 with a bifilar wound coil on a e.g. FT43-50 core, see the reasons for that mod in the description already cited in this forum  "ludens.cl".

BTW: how was the output power determined?

How does the power curve PIn - POut look like,

Is the output power measured at the 1 dB compresion point?

Tests on my tries to build a PA with the RD16HFF1´s using different output transformer configurations (even with a 1:4 guanella transformer with 25 ohm coaxial cable on two BN43-202 cores + Balun with 50 ohm coaxial cable on a FT43-80 torroid core + bifilar choke (FT43-50 torroid) + extra choke on the center tap  with a R-C combinaiton to terminate even harmonics) I was not able to get more than 10 W output power at P1dB.  Therefore I am really sceptical about the high power results of > 16 W.

The "hot" cores show that something must be wrong!

Did You measure the suppression of the even harmonics which should be well suppressed  (at least 40 dB) ? How good are the odd harmoncis suppressed?  If the third harmonic is only 10 dB down (1/3 of the amplitude of the fundamental wave) the PA produces square waves, which is not a performance of a linear amp.

IM performance??

Henning Weddig
DK5LV

Am 21.03.2018 um 12:41 schrieb John:

Hello Nik,

How did you go with the feedback issue?

My power amp mods in summary:  finals as RD16HHF1s, feedback resistors R261/R262 as 820 Ohms, Transformer 2T primary / 3T secondary on BN43-202, 180pF across the output transformer primary and 330pf across R87/R88. That's all...in a way.

Heavily inspired by Erhard's (DF3FY) information as displayed on uBitx.net.

Results per band at 14V supply, RV1 at %60, RD16's biased at 250mA each:
Band      output power   Total current
80M        22W                         2.9A
40M        20W                         3.2A
30M        16W                         2.3A
20M         16W                        2.7A
17M         15W                        2.4A
15M         12W                        1.6A
12M         12W                    not measured
10M         12W                        1.7A

At 22W the output transformer does get warm but not the LPFs.

Wound back RV1 to get 9-10W on the high frequency bands and 16-18W on the lower ones.

All the best,

73, John (VK2ETA)

 


Virus-free. www.avg.com

Nick VK4PP
 

HI John and the group.
I did the mod with the 100uH and 1nF cap on the MIC input fo the preamp board,
It did cut the interferance/feedback out completely. However, it has also cut the MIC gain dramatically (5W vs 20W CW)
Might play with the values, smaller inductor maybe?

Thanks.
Nick.

Nick VK4PP
 

MIght try this as a 3 element:

John <passionfruit88@...>
 

Hello Nik,

I am confused. A 100uH and 1nF filter has a cutoff frequency of 503KHz, so am not sure how it can affect your audio.

You also mentioned CW, and CW works directly on the unbalance of the mixer and does not come near the audio circuit, therefore I can't understand how it affect you power out.

Are you sure there is not something else?

Also I placed the low pass filter on the input of the SSM2167 module (if you still have it in-line).

But either way it should not affect your power out.

All the best,

73, John (VK2ETA)

Nick VK4PP
 

HI John.

Sorry, I was a bit unclear.
CW power is normal, no effect there, I get the full 20+w output.

But the SSB is low now... I will check it over tonight, try it with and without again, and post up a few pics.

Cheers.

MVS Sarma
 

we can get mpsh10 in smd MMBTH10 AND  the pin out is same as present device.
 They are cheap enough on ebay.  Hope the info helps.

regards
sarma
 vu3zmv

Regards
MVS Sarma
 

On Wed, Mar 21, 2018 at 5:11 PM, John <passionfruit88@...> wrote:
Hello Nik,

How did you go with the feedback issue?

My power amp mods in summary:  finals as RD16HHF1s, feedback resistors R261/R262 as 820 Ohms, Transformer 2T primary / 3T secondary on BN43-202, 180pF across the output transformer primary and 330pf across R87/R88. That's all...in a way.

Heavily inspired by Erhard's (DF3FY) information as displayed on uBitx.net.

Results per band at 14V supply, RV1 at %60, RD16's biased at 250mA each:
Band      output power   Total current
80M        22W                         2.9A
40M        20W                         3.2A
30M        16W                         2.3A
20M         16W                        2.7A
17M         15W                        2.4A
15M         12W                        1.6A
12M         12W                    not measured
10M         12W                        1.7A

At 22W the output transformer does get warm but not the LPFs.

Wound back RV1 to get 9-10W on the high frequency bands and 16-18W on the lower ones.

All the best,

73, John (VK2ETA)


Nick VK4PP
 

So looking good so far. I am getting these output values: (RV1 adjusted(75%) for 20W on 80/40)
80m : 20W
40m : 20W
20m : 18W
15m : 17W
10m : 14W

No feedback, but I have yet to test a QSO on 20/15/10m...

I did change the pre-driver emitter capacitors (7 of) to 320pf (100+220)
I'm happy with that, not sure if I will do the mpsh10  too... maybe if it gives a flat 20w on all bands..

73, and thanks to all for the help...
Nick

John <passionfruit88@...>
 

Hello Nick,

This looks like a good result.

Just to summarise your changes: RD16 finals, BN43-3312  1T primary, 2T secondary, 320pf on all 7 emitter resistors of the pre-amp chain, 330pf across the primary of the final's winding. Anything else?

Since I am portable VK4/Brisbane area at present we could try a QSO if you want. We should be able to get 20m going as we are far enough from each other.

All the best,

73, John (VK2ETA)

Nick VK4PP
 

Hi John,
BN42-202, but a 2:4 ratio winding giving 1:4 transformation. Runs cool at 20+w with 330pf Mica cap. 220ohm feedback resistor.
Listening on 14.277 ?
73 Nick

John <passionfruit88@...>
 

Thanks Nick for the information.

Calling now on 14,277.

73, John (VK2ETA)

Ashhar Farhan
 

How much power should we put out? It is really your call. I personally prefer 10 watts. There are multiple reasons. First, between 10 watts and 20 watts, the other side will feel almost no difference. Second, at 10 watts, I can still you batteries. Third, the heatsink is managable. Fourth, I am technically in QRP land.

- f


On Sunday 01 April 2018 09:03 AM, Nik VK4PLN wrote:
Hi John,
BN42-202, but a 2:4 ratio winding giving 1:4 transformation. Runs cool at 20+w with 330pf Mica cap. 220ohm feedback resistor.
Listening on 14.277 ?
73 Nick

Gordon Gibby
 

I'm VERY impressed and interested in the ability to have up to 20W for these reasons:


1)  I see this as a way to entice many local ARES club members into moving into HF from their exile in VHF-only land....20 watts  possibility is quite useful.

2)  I do mostly digital, so I set the power level by adjusting the TX GAIN on the Signalink (or more likelly, on the homebrew interface I soldered together)​ -- so I can easily set it at a modest 10W if the rig itself can produce up to 20....with the higher duty cycle, this is prudent.

3)  I'd like to get folks locally into an 80 meter training net, and having a rig that can handle up to 20 watts SSB peaks (with its low duty cycle) at this price level is very attractive.   I can honestly tell them they will be doing 8-10 W out from the beginninng on the lower bands, and that they can double that if they wish with a bit of soldering work.


Many of these people learned how to solder only within the last year (from me) and building gas discharge tube lightning arrestors last weekend is the proper level of challenge for them!!!


Cheers,


gordon




From: BITX20@groups.io <BITX20@groups.io> on behalf of Ashhar Farhan <farhanbox@...>
Sent: Saturday, March 31, 2018 11:41 PM
To: BITX20@groups.io
Subject: Re: [BITX20] RD16HHF1 power curve flattening...some
 

How much power should we put out? It is really your call. I personally prefer 10 watts. There are multiple reasons. First, between 10 watts and 20 watts, the other side will feel almost no difference. Second, at 10 watts, I can still you batteries. Third, the heatsink is managable. Fourth, I am technically in QRP land.

- f


On Sunday 01 April 2018 09:03 AM, Nik VK4PLN wrote:
Hi John,
BN42-202, but a 2:4 ratio winding giving 1:4 transformation. Runs cool at 20+w with 330pf Mica cap. 220ohm feedback resistor.
Listening on 14.277 ?
73 Nick

John <passionfruit88@...>
 

Dear Mr Farhan, 

Thank you for your post. I also personally think that 10 watts is is a good number and I have set my rig at just under that on 15 to 10M.

Unfortunately to get to this  power level, the lower bands have to be above.

In my case that value is 16 to 17W on 80M. sloping to 9W on 15m and flat after that despite the mods that have been described in this thread (its original objective) and in other threads. The original variation as-built was, from memory, 10 to 1W variation across 40-10M, or 17 to 2W at higher PA voltage.

I am working on an 1st If shift based ALC/power limiter using the technique described in my other post and this will hopefully allow for a completely flat power curve over the bands.

But unfortunately this is expected to only work in SSB modes.

So the issue remains that the power curve as designed is quite steep.

My wish for a future version of the uBitx is, if there is one, and I sincerely hope there is, for a somewhat flatter curve. A maximum variation of 1 to 2 for example would be in my opinion quite acceptable.

While we are on the wishes, it would have made my modifications much simpler if the Si5351 was on the main board as it would allow to extend the I2C bus to whatever micro-controller we wished to use.

Finally, a big thank you for such a wonderful rig. It is such a pleasure to work with, and its customisation makes it an unparalleled match for our own needs.

All the best,

73, John (VK2ETA)