What brand, and model oven did you buy?

At 07:31 PM 2/2/2006, you wrote:

Hi Bill,
Although I haven't tried it yet, that is supposed to be how it works,
the surface tension does hold the parts on the bottom side in place.
They say to do the side with the smallest parts first, then do the side
with the heavier parts second, so the lighter parts will be suspended
upside down during the second reflow.
I took the plunge earlier in the week and bought a toaster oven. I
finished my second board last night and I am just amazed at how well it
turned out. I am building a spectrum analyzer project that has almost
all very small smt parts on boards with lots of ground plane. Very
difficult to solder with an iron due to the large thermal mass of the
ground plane. It is hard to get solder to flow properly when soldering
the ground side of the component. These first two boards have turned out
nearly perfect, and I usually am pretty picky about such things. The
only problem I experienced is a little overshoot on the temperature.
Since I am using manual control of the temp I need to shut the oven off
about 40 or 50 degrees short of the target temp and let it coast into
the desired temp range. I suggest a practice run if you are going to put
some expensive parts in the oven. I am really impressed with this
method, not only is it effective, it is loads of fun! This might become
the preffered method of soldering SMT parts for the homebrewer in the
future. I am going to start my third board tonight and it has parts on
both sides, so I will find out how it works, wish me luck!


Bob K6GGO

Bill Dumke wrote:

How do you do double sided PC boards with SMT on both sides. Will
surface tension hold the SMT parts on the upside down side when you reheat?

Bill WB5TCO

Bob Fish K6GGO wrote:

Hi Guys,
I don't know about the skillet, but the toaster oven reflow technique is
getting to be very popular with home
hobbyists. You can buy solder paste in a syringe for easy application.
Apply a very small amount of solder paste to the pads, then stick the
component
in the paste. After filling the board with parts you put it in a toaster
oven and soak it at about 200F for 3 or 4 minutes (to alieviate thermal
stress) then ramp it up to 410F or so. Then turn off the oven and open
the door. You can get exact temperature profiles (temp vs time) from the
manufacturor. It works great. One of the nice parts is the component
will straighten itself on the pads due to the surface tension of the
solder. Do a google search for "toaster oven reflow". Here is the link
that I learned the method from:



Bob K6GGO

Yahoo! Groups Links

Cecil Bayona
KD5NWA
www.qrpradio.com

"I fail to see why doing the same thing over and over and getting the same results every time is insanity: I've almost proved it isn't; only a few more tests now and I'm sure results will differ this time ... "

Des,
yes, that's quite possible.... when designing SDRadio I took into
consideration a rather small imbalance error to correct, based on the
hardware I was experimenting with at that time. So, to make the
adjustment
more smooth, the max amplitude error range that can be corrected is
rather small, and this explains why you find the best rejection with
the slider at one end.
For the next version, I will increase that range. Thanks for pointing
it out.


Perhaps a course as well as a fine adjustment ??

Bob K3DJC

Hi Bill,
Although I haven't tried it yet, that is supposed to be how it works, the surface tension does hold the parts on the bottom side in place. They say to do the side with the smallest parts first, then do the side with the heavier parts second, so the lighter parts will be suspended upside down during the second reflow.
I took the plunge earlier in the week and bought a toaster oven. I finished my second board last night and I am just amazed at how well it turned out. I am building a spectrum analyzer project that has almost all very small smt parts on boards with lots of ground plane. Very difficult to solder with an iron due to the large thermal mass of the ground plane. It is hard to get solder to flow properly when soldering the ground side of the component. These first two boards have turned out nearly perfect, and I usually am pretty picky about such things. The only problem I experienced is a little overshoot on the temperature. Since I am using manual control of the temp I need to shut the oven off about 40 or 50 degrees short of the target temp and let it coast into the desired temp range. I suggest a practice run if you are going to put some expensive parts in the oven. I am really impressed with this method, not only is it effective, it is loads of fun! This might become the preffered method of soldering SMT parts for the homebrewer in the future. I am going to start my third board tonight and it has parts on both sides, so I will find out how it works, wish me luck!


Bob K6GGO

Bill Dumke wrote:

Des wrote:

I have attempted to minimize the unwanted response using the manual
adjustments in V 0.95 with 48 kHz sampling but the best I could get
was about 30 dB of rejection. I attempted to improve on that a number
of times by doing a "Reset" and making the adjustments again. I found
in all cases that the amplitude slider finished up in its upermost
(max?) position before I had reached an optimum adjustment. The sound
card I used at that time was a Soundblaster Live! 5.1 PCI card.

Des,
yes, that's quite possible.... when designing SDRadio I took into
consideration a rather small imbalance error to correct, based on the
hardware I was experimenting with at that time. So, to make the adjustment
more smooth, the max amplitude error range that can be corrected is
rather small, and this explains why you find the best rejection with
the slider at one end.
For the next version, I will increase that range. Thanks for pointing
it out.

73 Alberto I2PHD
P.S. I just received a reprint of an IEEE paper about an algorithm for
automatic I/Q Imbalance Compensation in OFDM Direct-Conversion
Receivers, which maybe can be adapted for our purposes...

Bill Dumke wrote....

So I next swept the input with the low
pass filter output of a DDS-60 fed through a 10 dB/step step attenuator
to set the level at the SoftRock-40 RF input....

End quote:

Hi Bill, thanks for the tip, I did as you described and had some
success but I can still only get rejection of the unwanted signals at
one point in the band. This is within +/- 2 kHz of the band centre. I
cleared the existing cal' data and injected a strong steady signal but
Rocky just seemed to ignore it until it was within a few kHz of the
band centre at which point it "latched" on to it and appeared to give
some rejection of the unwanted signal at that point. But try as I
might I have been unable to get rejection in any other part of the
spectrum. Rocky seems to be picking up a few more cal' points (besides
the one that gives some rejection) but for some reason none of the
additional cal' points seem to do any good. The best rejection I get
is about 15 to 20 dB.

Because of a posting on the KGKSDR mailing list I became aware of a
problem with the Soundblaster Live! series of cards and when a "fix"
was added to KGKSDR I found it solved all my I/Q problems. By setting
the "Delay right by 1 sample" option in KGKSDR I now get pretty good
rejection of unwanted signals using KGK's auto-cal. By performing a
manual calibration I do even better with about 50 dB over most of the
band. I think I can probably improve on that by doing another cal'
with more cal' points.

I do plan to upgrade my soundcard at a later date but the current PC
(Intel P4 1.4 GHZ, 384 meg RAM) is as hot as it will go so I plan to
upgrade the system first and probably update the soundcard at the same
time.

Thanks again for the feedback.

73's to all on the list.

Des (M0AYF)


--- In softrock40@..., Bill Dumke <billd@...> wrote:

Des,

Rocky will self calibrate to some degree if it sees strong signals
(>60dB on the left hand scale) across the band for a long time. Rovky
does not need to be tuned to the frequency. This is all done in the
background. I left mine running for about 3 months , but the image
rejection was only about 18 dB. So I next swept the input with the low
pass filter output of a DDS-60 fed through a 10 dB/step step attenuator
to set the level at the SoftRock-40 RF input. Then I got about 73 or
74 dB image rejection doing it with two different sound cards, an

Audigy

2 and a Delta 44.

Bill WB5TCO

--- In softrock40@..., windy10605@... wrote:

Maybe the group should be working on an all band preselector for the
next generation softrock?

Regards,

Art

YU1LM, Tasa has designed a tunable BPF for his SDR.
See

73,
Marsh, NC7V

Hello Alberto and the group,


Alberto wrote....

SDRadio doesn't have an automatic rejection adjustment, but a manual
one. Have you tried to minimize the unwanted response by adjusting the
sliders in the panel...

End quote:

I have attempted to minimize the unwanted response using the manual
adjustments in V 0.95 with 48 kHz sampling but the best I could get
was about 30 dB of rejection. I attempted to improve on that a number
of times by doing a "Reset" and making the adjustments again. I found
in all cases that the amplitude slider finished up in its upermost
(max?) position before I had reached an optimum adjustment. The sound
card I used at that time was a Soundblaster Live! 5.1 PCI card.

More recently I purchased a Soundblaster Live! 24 bit PCI card which
seems to do exactly the same when I perform the I/Q calibration. I
downloaded V 0.99 a few weeks ago and installed it yesterday evening.
I have found that the audio quality is much better now when sampling
at 96 kHz but the calibration problem remains the same. I performed
the tests with a Softrock V-4 which I had on loan for a few weeks. The
V4 has now been returned and I have a Softrock V5 of my own. Both the
V4 and V5 exhibit the same problem when I attempt I/Q calibration.

I have a second system which I tested the software on with exactly the
same results. The two systems have only two things in common, the
operating system (XP) and they both now have Soundblaster Live cards.
For this reason I began to suspect the soundcards may be the cause of
all the problems.

I should add that I had similar problems with the Rocky software and
KGKSDR. As luck would have it someone else had similar problems to my
own while using a Soundblaster live! 24 bit PCI card and KGKSDR. They
posted a message on the KGKSDR mailing list with the result that
Duncan had a look into the problem with this card. The result was the
addition of the "Delay right (or left) by 1 sample" option added to
the setup screen in KGKSDR. I downloaded the revised software and
found it cured my I/Q problem. Even with no calibration I could see
some reduction in the unwanted signals across the whole band. I later
performed a manual calibration with about 15 points across the band
and now see quite good rejection.

I hope the above is of some help in explaning the problem. SDRadio has
been very successful for me while experimenting with DC receivers. I
just ignored the oposite sideband :-) But when I began experimenting
with I/Q signals and the Softrock I found the problem. Since many
others had good results with SDRadio I put it down to a problem with
the cheap soundcard I am using.

73's to all on the list.

Des. (M0AYF)

About 70 boxes full of parts, boards,etc. I have to go through all the boxes one at a time, then put them back, it's cold in the garage right now.

At 11:04 AM 2/2/2006, you wrote:

?

----- Original Message -----

From: KD5NWA

To: softrock40@...

Sent: Wednesday, February 01, 2006 5:37 PM

Subject: Re: Fw: Re: [softrock40] Re: Bandpass filter smilter

.

I have one in a box in the garage, I've been searching for it, but I have not found it.

?

Joe wa9cgz

That must be one big garage Joe.

?

Bob W2AWS

?

Yahoo! Groups Links

<*> To visit your group on the web, go to:

???

<*> To unsubscribe from this group, send an email to:

??? softrock40-unsubscribe@...

<*> Your use of Yahoo! Groups is subject to:

???

?

Cecil Bayona

KD5NWA

I fail to see why doing the same thing over and over and getting the same results every time is insanity: I've almost proved it isn't; only a few more tests now and I'm sure results will differ this time ...

---

YAHOO! GROUPS LINKS

?Visit your group " " on the web.

?

?To unsubscribe from this group, send an email to:

? softrock40-unsubscribe@...

?

?Your use of Yahoo! Groups is subject to the .

---

Cecil Bayona
KD5NWA
I fail to see why doing the same thing over and over and getting the same results every time is insanity: I've almost proved it isn't; only a few more tests now and I'm sure results will differ this time ...

I think the even harmonics should not be a problem, as indicated.

But, I thought the odd harmonics were supposed to be 6 db down for each step.

I had a spice model once and I think (but cannot be positive) that I had confirmed this in Spice.

But, suffice it to say that harmonics are a problem. In a superhet, the harmonics are converted out to non-existance....not so with the QSD.

For a general coverage softrock, I would think a preselector would be the way to go....even though it would need to be tweaked every 100 to 500 KHz (depending on the frequency of operation). A small price to pay (IMHO) for peace of mind and not having to wonder whether the signal you hear is actually there or is the result of an undesired harmonic response. Your mileage will vary.

I would caution against automatic tracking varactor tuning. They have all the non linearity of crude mixers when exposed to large signals and the effects are well known to those who have tried to operate them in high rf environments.

Maybe the group should be working on an all band preselector for the next generation softrock?

Regards,

Art

开云体育

----- Original Message -----

From: KD5NWA

To: softrock40@...

Sent: Wednesday, February 01, 2006 5:37 PM

Subject: Re: Fw: Re: [softrock40] Re: Bandpass filter smilter

.

I have one in a box in the garage, I've been searching for it, but I have not found it.

?

Joe wa9cgz

That must be one big garage Joe.

?

Bob W2AWS

?
Yahoo! Groups Links

<*> To visit your group on the web, go to:
???

<*> To unsubscribe from this group, send an email to:
??? softrock40-unsubscribe@...

<*> Your use of Yahoo! Groups is subject to:
???
?

Cecil Bayona
KD5NWA
I fail to see why doing the same thing over and over and getting the same results every time is insanity: I've almost proved it isn't; only a few more tests now and I'm sure results will differ this time ...

Hi Daryl,

Please send me your mailing address and I will get one of the 14.06 MHz crystals in the mail to you. No charge when there is a bad part in one of the kits.

Glad to hear things are working properly on 40m for you.

73,
Tony KB9YIG

Folks,

I got my Softrock working fine on 40m, what an amazing piece of hardware !

When I tried to get it running on 14mhz, I found the oscillator dead.
After much circut tracing, I've decided that I've got a bad crystal.

Are there any surplus crystals for the 20m unit kicking around that I could buy ?

Thanks,
Daryl
WD8PTP

Thanks for the reply sir.

ok than, i just ordered it. I hope i can get it to work.

yeah its really not that expensive either.

--- In softrock40@..., KD5NWA <KD5NWA@...> wrote:

I'm not familiar with that chip, most use newer chips, but if that

is

what you can easily get and it it's not horribly expensive, I

would

go ahead and get it.

It will give frequency mobility so you can move around in the
bandpass and find signals to be able to demonstrate that your

project

is functional.

Since you bought as SR-5 you will need a 1X oscillator output,

that

chip should be capable of moving around to all kinds of usable
frequencies, use shielded coax to get the signal to the radio.


At 03:07 AM 2/1/2006, you wrote:

Hello all.

I am still assembling my Softrock these days. But i was thinking

of

purchasing a DDS and an Evaluation board.

The distributors here in Pakistan, only offer to bring the

following

DDS from UK:

AD9832
25 MHz Serial Interface DDS With On-Chip 10-Bit DAC

and its Evaluation Board:

EVAL-AD9832EB

The evaluation board comes with its software and can be controlled
via the computer parallel port.

This will be delivered to me some time in the next 2 weeks, by

than

i should have my softrock working with the crystal that it came
with.

So I was wondering if this DDS setup is any good?? I am going to
make another BPF board for my softrock for 500-1600Khz range. So i
need the DDS for tuning in this range, as well as the 7 Mhz and up
band.

What do you guys think? should i go for it?

I will be using the evaluation board and the software that came

with

it, controlling via the parallel port.

Will this work?

Thanks and sorry for the ultra long mail!

:)






Yahoo! Groups Links

Cecil Bayona
KD5NWA
www.qrpradio.com

"I fail to see why doing the same thing over and over and getting

the

same results every time is insanity: I've almost proved it isn't;
only a few more tests now and I'm sure results will differ this

time ... "

How do you do double sided PC boards with SMT on both sides. Will surface tension hold the SMT parts on the upside down side when you reheat?

Bill WB5TCO

Bob Fish K6GGO wrote:

I'm not familiar with that chip, most use newer chips, but if that is what you can easily get and it it's not horribly expensive, I would go ahead and get it.

It will give frequency mobility so you can move around in the bandpass and find signals to be able to demonstrate that your project is functional.

Since you bought as SR-5 you will need a 1X oscillator output, that chip should be capable of moving around to all kinds of usable frequencies, use shielded coax to get the signal to the radio.


At 03:07 AM 2/1/2006, you wrote:

Hello all.

I am still assembling my Softrock these days. But i was thinking of
purchasing a DDS and an Evaluation board.

The distributors here in Pakistan, only offer to bring the following
DDS from UK:

AD9832
25 MHz Serial Interface DDS With On-Chip 10-Bit DAC

and its Evaluation Board:

EVAL-AD9832EB

The evaluation board comes with its software and can be controlled
via the computer parallel port.

This will be delivered to me some time in the next 2 weeks, by than
i should have my softrock working with the crystal that it came
with.

So I was wondering if this DDS setup is any good?? I am going to
make another BPF board for my softrock for 500-1600Khz range. So i
need the DDS for tuning in this range, as well as the 7 Mhz and up
band.

What do you guys think? should i go for it?

I will be using the evaluation board and the software that came with
it, controlling via the parallel port.

Will this work?

Thanks and sorry for the ultra long mail!

:)






Yahoo! Groups Links

Cecil Bayona
KD5NWA
www.qrpradio.com

"I fail to see why doing the same thing over and over and getting the same results every time is insanity: I've almost proved it isn't; only a few more tests now and I'm sure results will differ this time ... "

--- In softrock40@..., windy10605@... wrote:

You would expect the third harmonic to be the predominant one with

no

filtering. You are using the DDS-60 for the OSC on the SR-40,

right ? set

to receive 1Mhz ? What are you doing to measure the harmonics ?

73 Kees K5BCQ

Kees,

I'm using a DDS30 set to 4mhz connected to base of SR40 OSC transistor
with feedback caps removed. I have a HP654 generator with calibrated
output attenuators. By using one of ROCKY'S DB lines as a reference
the test generator can be adjusted to the same line and the attenuator
read as the difference in db.
Please connect that MiniCiruit transformer and verify my results.

Joe wa9cgz

--- In softrock40@..., windy10605@... wrote:

Exactly ! I would much prefer an old tuneable WJ double tuned frontend
....but I don't have one, was going to adapt a Drake 2B preselector
...haven't done that either, and don't really want to plug and replug

the

BPFs as I scoot around the HF spectrum trying other things.

Looking for the "least undesirable alternative".

73 Kees K5BCQ

Kees,

I just reconnected my test SD40 setup using DDS as 4x vfo although it
will hear the even harmonic (way down) it's the odd that is predominate.

I set DDS to 4mhz for a receive frequency of 1mhz

2mhz down 45 to 50db

3mhz down 10db

5mhz down 12db

7mhz down 15db

Joe

Not just odd and even multiples, it can respond to fractions 5/8....to keep unwanted signals, increased dynamic range, and keep the noise down you need a front end filter as tight as possible. Saying that however, for playing around mobility might be more important issue.

The best of both is if you can find a multi-band front end filter unit (specially the double tuned) from 20 to 50 years ago, you can tune from 100KHz to 30MHz in one small box so you have mobility and better discrimination.

I have one in a box in the garage, I've been searching for it, but I have not found it.

At 03:49 PM 2/1/2006, joe price wrote:

Disregard message below I cheched a 30mtr rx it responds to even? and
odd harmonics need more testing.

> I would expect that the SR40 would respond to even harmonics but mine
> responds ONLY to odd multiples of the rx frequency I don't understand
> why. Can someone else check this? The v5 rx uses a different QSD
scheme
> i'm not sure it's the same I will test and report result.
>
> Joe wa9cgz
>







?
Yahoo! Groups Links

<*> To visit your group on the web, go to:
???

<*> To unsubscribe from this group, send an email to:
??? softrock40-unsubscribe@...

<*> Your use of Yahoo! Groups is subject to:
???
?

Cecil Bayona
KD5NWA
I fail to see why doing the same thing over and over and getting the same results every time is insanity: I've almost proved it isn't; only a few more tests now and I'm sure results will differ this time ...

Disregard message below I cheched a 30mtr rx it responds to even and
odd harmonics need more testing.

I would expect that the SR40 would respond to even harmonics but mine
responds ONLY to odd multiples of the rx frequency I don't understand
why. Can someone else check this? The v5 rx uses a different QSD

scheme

i'm not sure it's the same I will test and report result.

Joe wa9cgz

Disregard message below I cheched a 30mtr rx it responds to even and
odd harmonics need more testing.

I would expect that the SR40 would respond to even harmonics but mine
responds ONLY to odd multiples of the rx frequency I don't understand
why. Can someone else check this? The v5 rx uses a different QSD

scheme

i'm not sure it's the same I will test and report result.

Joe wa9cgz