Return-Path: Received: from mtain-df10.r1000.mx.aol.com (mtain-df10.r1000.mx.aol.com [172.29.64.222]) by air-de01.mail.aol.com (v127_r1.1) with ESMTP id MAILINDE012-5ea34b7bec9518f; Wed, 17 Feb 2010 08:18:13 -0500 Received: from post.thorcom.com (post.thorcom.com [193.82.116.20]) by mtain-df10.r1000.mx.aol.com (Internet Inbound) with ESMTP id 565523800029F; Wed, 17 Feb 2010 08:18:11 -0500 (EST) Received: from majordom by post.thorcom.com with local (Exim 4.14) id 1Nhjla-0007oA-1b for rs_out_1@blacksheep.org; Wed, 17 Feb 2010 13:16:34 +0000 Received: from [193.82.116.32] (helo=relay1.thorcom.net) by post.thorcom.com with esmtp (Exim 4.14) id 1NhjlZ-0007o1-5E for rsgb_lf_group@blacksheep.org; Wed, 17 Feb 2010 13:16:33 +0000 Received: from mail.gmx.net ([213.165.64.20]) by relay1.thorcom.net with smtp (Exim 4.63) (envelope-from ) id 1NhjlW-000060-Es for rsgb_lf_group@blacksheep.org; Wed, 17 Feb 2010 13:16:33 +0000 Received: (qmail invoked by alias); 17 Feb 2010 13:16:22 -0000 Received: from p5DD3484B.dip.t-dialin.net (EHLO Clemens04) [93.211.72.75] by mail.gmx.net (mp039) with SMTP; 17 Feb 2010 14:16:22 +0100 X-Authenticated: #17214767 X-Provags-ID: V01U2FsdGVkX1+cKbcRvQacVWawGDWms0TSQ4JDl7ceY/Fnz81bHl VKU4qSN0KLH3L8 Message-ID: <008901caafd3$6697b8e0$0201a8c0@Clemens04> From: "Clemens Paul" To: References: <006501caaf3b$53ddb350$0201a8c0@Clemens04> <006601caafbc$28de4c10$0201a8c0@Clemens04> Date: Wed, 17 Feb 2010 14:16:19 +0100 MIME-Version: 1.0 X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 6.00.2900.3598 X-MimeOLE: Produced By Microsoft MimeOLE V6.00.2900.3350 X-Y-GMX-Trusted: 0 X-FuHaFi: 0.45000000000000001,0.45000000000000001 X-Spam-Score: 0.0 (/) X-Spam-Report: autolearn=disabled,HTML_MESSAGE=0.001 Subject: Re: LF: RE: PA matching oddity Content-Type: multipart/alternative; boundary="----=_NextPart_000_0083_01CAAFDB.C66CFB10" X-Spam-Checker-Version: SpamAssassin 2.63 (2004-01-11) on post.thorcom.com X-Spam-Level: X-Spam-Status: No, hits=0.8 required=5.0 tests=HTML_30_40,HTML_MESSAGE autolearn=no version=2.63 X-SA-Exim-Scanned: Yes Sender: owner-rsgb_lf_group@blacksheep.org Precedence: bulk Reply-To: rsgb_lf_group@blacksheep.org X-Listname: rsgb_lf_group X-SA-Exim-Rcpt-To: rs_out_1@blacksheep.org X-SA-Exim-Scanned: No; SAEximRunCond expanded to false x-aol-global-disposition: G x-aol-sid: 3039ac1d40de4b7bec934846 X-AOL-IP: 193.82.116.20 ------=_NextPart_000_0083_01CAAFDB.C66CFB10 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Oh yes,be careful about this precious piece of measuring equipment. But then,with < +/- 0,2dB you come out at an uncertainty within of around +/- 42W at 900W. 73 Clemens DL4RAJ ----- Original Message -----=20 From: Andy Talbot To: rsgb_lf_group@blacksheep.org Sent: Wednesday, February 17, 2010 12:15 PM Subject: Re: LF: RE: PA matching oddity I don't use a Bird wattmeter. Don't trust any directional type of= meter for proper power=20 measurements. I have a 1GHz specified 30dB rated 1kW Bird attenuator (calibrated)= plus HP 432 power meter=20 (reasonably well calibrated) so I reckon power measurement is to with= in 0.2dB However, trying to get more than 1kW would overlaod the attenuator,= which is a very nice=20 little (!) bit of kit to have so probably not a good idea Andy www.g4jnt.com On 17 February 2010 10:29, Clemens Paul wrote: >and I got 950 Watts before the PSU gave up (off 44V). > So it really was giving more power out than it 'ought to' Uncertainty of a bird wattmeter with a 1kW slug would be around +/= -50W, with a 2,5kW element +/-125W. Power master is a bit better with +/- 30W. 73 Clemens DL4RAJ ----- Original Message -----=20 From: Andy Talbot To: rsgb_lf_group@blacksheep.org Sent: Tuesday, February 16, 2010 8:48 PM Subject: Re: LF: RE: PA matching oddity A follow up to this sage After much puzzling over the design, and comment here about how= it must be 3:1 teh only=20 solution was to measure it. So I connected up and measured voltages.= Indeed, it does turn out=20 to be 3:1 - So how could it deliver 1kW plus... Next, fitted the module to a big heatsink, got out the 'telepho= ne exchange' PSU (50V 25=20 A) and wound up the amp to the max it would give. Unfortunately the= PSU current limited when=20 supplying only 800 or so watts, but by winding teh PSU down to 44V, it= current limited at a=20 higher value and I got 950 Watts before the PSU gave up (off 44V). = So it really was giving=20 more power out than it 'ought to' BUT, as Clemens mentioned I'd ignored the saturation element and= in fact, if used as a=20 switching amp can give well over 1kW. I'd used this aspect in the 70= 0W 137kHz Tx,=20 http://www.g4jnt.com/137tx.pdf and forgot it. Enough said. Anyway, tested the PA, and found it would go down to about 5MHz= where the transformers=20 began to give up, and up to about 20MHz where the FETs lost gain / eff= iciency. So as a=20 13.56MHz industrial unit its fine, but of little use as a general HF= amp. BUT, with suitable=20 change of output transformer, could probably be persuaded to give 1.5k= W on 137 or 500kHz - if a=20 suitable PSU were to hand. That's where it sits now. Experimentatio= n done, and any further=20 work can wait for other projects to get pushed of fthe stack. Andy www.g4jnt.com On 16 February 2010 19:07, Clemens Paul wrote: Andy,Rik, after thinking a while over the transformer of Andy's PA and= referring to http://www.g4jnt.com/pamatch.gif I'm sorry to say that I'm still quite (say 99,9%) sure that it is a conventional 1:9= (Z) transfomer. Here are some crucial points which exclude *absolutely* the ab= ove design to be an autotransformer,(1:16 or whatever). 1.) Autotransformers provide no dc isolation, i.e. you have a= dc path from the primary to secondary side. There is no such dc path here. 2.)In a bootstrap design the braid of one coax end is connecte= d to the center conductor at the start of the same coax (1:4) or of the next= coax section (1:9 or=20 higher). This is not the case in this design.The center of the secondar= y is nowhere connected with a braid. BTW there is no reason to worry about the theoretical 900W lim= it in linear mode. If the PA is slightly overdriven you get easily one 1dB more= power on the fundamental frequency which is about 1133W so you have enough margin for= losses of the low pass=20 filter losses etc. It's very unlikely that this PA which was used as industrial= RF heater was designed to be superlinear. 73 Clemens DL4RAJ ----- Original Message -----=20 From: Andy Talbot To: rsgb_lf_group@blacksheep.org Cc: ukmicrowaves@yahoogroups.com Sent: Tuesday, February 09, 2010 10:43 PM Subject: Re: LF: RE: PA matching oddity BINGO - Tnx Rik, thats it, convincing. Sorted So my thing IS 1:4 turns ratio, (1:16 Z ratio) just as the= calculations said it=20 needed to be, and is indeed, effectively, an auto transformer. Otherw= ise known as a=20 bootstrapped coaxial transformer, but just configured with different= reference points to confuse=20 everyone. Your second reference, the Semelab paper, (UKuW see below)= gives the answer. Look at page 10, the 1:9 Ruthroff UnUn . With just two= bits of coax, which are=20 then bootstrapped up on top of the input to give a 3:1 ratio (for 1:9= impedance). Add another=20 turn, so three windings are bootstrapped on top of the primary, rearr= ange the grounding points=20 and there, in all its glory, is my mystery. The reference / grounding points can be shifted as the mosfe= t PA has two separate,=20 non magnetically linked cores. This is is just normal push-pull HF PA= practice, and being so=20 normal in its concept, completely hides hides a big real advantage -= it enables the designer to=20 float just whatever ternainals need to be floated for balanced / unbal= anced operation, and even=20 to provide DC isolation, Any conductor passing through the core allo= ws one end to be=20 completely Rf decoupled from the other end, so that's how balanced ope= ration is permitted with=20 the basic Ruthrof design shown in the paper. On the same page, a single length of coax is shown giving 1:= 2 transformation by the=20 same bootstrapping arrangement, and when the transmision line is shown= as a twisted pair rather=20 than coax, then wrapped on a torroid, it becomes the 1:4 balun beloved= of HF antenna=20 constructors. A natural progression that helps illustrate the concep= t of a loop of coax=20 giving one turn more than it appears to. It all falls out. ... life is worth living again ... ... can retire happy ... Andy www.g4jnt.com ... now, next job, how to sort out the world financial crisi= s ... On 9 February 2010 19:53, Rik Strobbe wrote: Andy, have a look here : http://www.polyfet.com/HFE0503_Leong.pdf http://www.semelab.com/rf/documents/Push-Pull%20Circuits%2= 0and%20Wideband%20Transformers.pdf 73, Rik ON7YD - OR7T -------------------------------------------------------------------- Van: owner-rsgb_lf_group@blacksheep.org [owner-rsgb_lf_gro= up@blacksheep.org] namens=20 Andy Talbot [andy.g4jnt@googlemail.com] Verzonden: dinsdag 9 februari 2010 18:10 Aan: rsgb_lf_group@blacksheep.org; ukmicrowaves@yahoogroup= s.com Onderwerp: LF: PA matching oddity ---------------------------------------------------------------------- Eingehende eMail ist virenfrei. Von AVG =FCberpr=FCft - www.avg.de Version: 9.0.733 / Virendatenbank: 271.1.1/2677 - Ausgabedat= um: 02/09/10 08:35:00 ----------------------------------------------------------------------= ---- Eingehende eMail ist virenfrei. Von AVG =FCberpr=FCft - www.avg.de Version: 9.0.733 / Virendatenbank: 271.1.1/2691 - Ausgabedatum:= 02/16/10 08:35:00 ----------------------------------------------------------------------= -------- Eingehende eMail ist virenfrei. Von AVG =FCberpr=FCft - www.avg.de Version: 9.0.733 / Virendatenbank: 271.1.1/2692 - Ausgabedatum: 02/1= 6/10 20:35:00 ------=_NextPart_000_0083_01CAAFDB.C66CFB10 Content-Type: text/html; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable
Oh yes,be careful about this precious= piece of=20 measuring equipment.
But then,with < +/- 0,2dB you come= out at an=20 uncertainty
within of around +/- 42W at 900W.
 
73
Clemens
DL4RAJ
 
----- Original Message -----
From:=20 Andy Talbot
Sent: Wednesday, February 17,= 2010 12:15=20 PM
Subject: Re: LF: RE: PA match= ing=20 oddity

I don't use a Bird wattmeter.  Don't trust any directional= type of=20 meter for proper power measurements. 
 
I have a  1GHz specified 30dB rated 1kW Bird attenuator (c= alibrated)=20 plus HP 432 power meter (reasonably well calibrated)  so I reck= on power=20 measurement is to within 0.2dB
 
However, trying to get more than 1kW would overlaod the attenua= tor, which=20 is a very nice little (!) bit of kit to have so probably not a good= idea=20
 
On 17 February 2010 10:29, Clemens Paul <cpaul@gmx.net>=20 wrote:
>and I got 950 Watts before the PSU gave up (off 44V)= .
> So it really was giving more power out than it 'oug= ht=20 to'
 
Uncertainty of a bird wattmeter= with a 1kW slug=20 would be around +/-50W,
with a 2,5kW element +/-125W= .
Power master is a bit better with= +/-=20 30W.
 
73
Clemens
DL4RAJ
 
 
----- Original Message -----
Sent: Tuesday, February= 16, 2010 8:48=20 PM
Subject: Re: LF: RE: PA= matching=20 oddity

A follow up to this sage  
 
After much puzzling over the design, and comment here about= how it=20 must be 3:1 teh only solution was to measure it.  So= I connected=20 up and measured voltages. Indeed, it does turn out to be 3:= 1=20 -   So   how could it deliver 1kW plus...
 
Next, fitted the module to a big heatsink,  got out th= e=20 'telephone exchange' PSU (50V 25 A) and wound up the amp to the= max it=20 would give.   Unfortunately the PSU current limited wh= en=20 supplying only 800 or so watts, but by winding teh PSU down to= 44V, it=20 current limited at a higher value and I got 950 Watts befor= e the PSU=20 gave up (off 44V).    So it really was giving mor= e power=20 out than it 'ought to'
 
BUT, as Clemens mentioned I'd ignored the saturation elemen= t and in=20 fact, if used as a switching amp can give well over 1kW. &n= bsp; I'd=20 used this aspect in the 700W 137kHz Tx, http://www.g4jnt.co= m/137tx.pdf  and forgot=20 it.   Enough said.
 
Anyway, tested the PA, and found it would go down to= about 5MHz=20 where the transformers began to give up, and up to about 20MHz= where the=20 FETs lost gain / efficiency.   So as a 13.56MHz indust= rial unit=20 its fine, but of little use as a general HF amp.   BUT= , with=20 suitable change of output transformer, could probably be persuad= ed to give=20 1.5kW on 137 or 500kHz - if a suitable PSU were to hand. = That's=20 where it sits now.   Experimentation done, and any fur= ther work=20 can wait for other projects to get pushed of fthe stack.
On 16 February 2010 19:07, Clemens Paul= <cpaul@gmx.net> wrote:
Andy,Rik,
 
after thinking a while over= the transformer=20 of Andy's PA and referring= =20 to
I'm sorry to say that<= /DIV>
I'm still quite (say 99,9%)&n= bsp;sure that=20 it is a conventional 1:9 (Z) transfomer.
 
Here are some crucial points= =20 which exclude *absolutely* the above design
to be an autotransformer,(1:16 or whatever).
1.) Autotransformers pro= vide no dc=20 isolation, i.e. you have a dc path from
the primary to secondary side= .
There is no such dc= path=20 here.
2.)In a bootstrap design the= =20 braid of one coax end is connect= ed=20 to the center
conductor at= the start of=20 the same  coax (1:4) or of the next  coax section (1= :9 or=20 higher).
This is not the case in this= =20 design.The center of the secondary is nowhere=20 connected
with a braid.
 = ;
BTW there is no reason to wor= ry about the=20 theoretical 900W limit in linear mode.
If the PA is slightly overdri= ven you get=20 easily one 1dB more power on the fundamental
frequency which is about 1133= W so you have=20 enough margin for losses of the low pass filter
losses etc.
It's very unlikely that this= PA which was=20 used as industrial RF= heater=20
was designed to be=20 superlinear.
 
73
Clemens
DL4RAJ
 
 = ;
 
 
----- Original Message -----=
From= : Andy Talbot= =20
Sent: Tuesday, Februa= ry 09, 2010=20 10:43 PM
Subject: Re: LF: RE:= PA matching=20 oddity

BINGO -  Tnx Rik, thats it,=20 convincing.  Sorted
 
So my thing  IS 1:4 turns ratio, (1:16 Z rati= o)=20  just as the calculations said it needed to be,&nb= sp;and is=20 indeed, effectively, an auto transformer.  Otherwise kn= own as a=20 bootstrapped coaxial transformer, but just configured with= different=20 reference points to confuse everyone. Your se= cond=20 reference, the Semelab paper, (UKuW see below)  gives= the=20 answer.  
 
Look at page 10,   the 1:9=20 Ruthroff   UnUn .   With just = two bits=20 of coax, which are then bootstrapped up on top of the input= to give a=20 3:1 ratio (for 1:9 impedance).   Add another turn,= so three=20 windings are  bootstrapped on top of the primary, rearr= ange the=20 grounding points and there, in all its glory, is my=20 mystery.   
 
The reference / grounding points can be shifted as the= mosfet=20 PA  has  two separate, non magnetically=20 linked cores.  This is is just normal push-pu= ll HF PA=20 practice, and being so normal in its concept, completel= y hides=20 hides a big real advantage - it enables the design= er to=20 float just whatever ternainals need to be floated for balanc= ed /=20 unbalanced operation, and even to provide DC isolation, = ; =20 Any conductor passing through the core allows one end to be= completely=20 Rf decoupled from the other end, so that's how balanced= operation=20 is permitted with the basic Ruthrof design shown in the pape= r.
 
On the same page, a single length of coax is shown= giving=20 1:2 transformation by the same bootstrapping arrangement, an= d when the=20 transmision line is shown as a twisted pair rather than coax= , then=20 wrapped on a torroid, it becomes the 1:4 balun beloved= of HF=20 antenna constructors.   A natural progression= that=20 helps illustrate the concept  of  a loop of coax= giving one=20 turn more than it appears to.
 
It all falls out. 
... life is worth living again ...
... can retire happy ...
 
... now, next job, how to sort out the world financial= crisis ...=20

On 9 February 2010 19:53, Rik Strob= be <Rik.Strobbe@fys.kuleuven.be> wrote:


Eingehende eMail ist virenfrei.
Von AVG =FCber= pr=FCft - w= ww.avg.de
Version:=20 9.0.733 / Virendatenbank: 271.1.1/2677 - Ausgabedatum: 02/09= /10=20 08:35:00



Eingehende eMail ist virenfrei.
Von AVG =FCberpr= =FCft - www.avg= .de=20
Version: 9.0.733 / Virendatenbank: 271.1.1/2691= -=20 Ausgabedatum: 02/16/10 08:35:00



Eingehende eMail ist virenfrei.
Von AVG =FCberpr=FCft= - www.avg.de=20
Version: 9.0.733 / Virendatenbank: 271.1.1/2692 - Ausgabedatum:= 02/16/10=20 20:35:00
------=_NextPart_000_0083_01CAAFDB.C66CFB10--