Digital (DDS) VFO kit - need 5 watt buffer amp

[K1JJ]

Hola!

I need a 5 watt digital (DDS) VFO to drive the IPA amp in my 4-1000A plate modulated rig. I just ordered the $79 digital VFO kit below. It puts out about 250 mV:

http://www.pongrance.com/super-dds.html


Here is a 2N2222 buffer amp I will build to bring the VFO drive up to 2 volts into a 50 ohm load. (80 mW)
http://www.pongrance.com/dds-buffer.html


* I need another broadband amplifier (160M - 40M) to take 2V @ 50 ohms (80 mW) and bring it up to about 5 watts at 50 ohms.   What is the easiest way to do this?  Anyone know of a SIMPLE circuit or hobby board available?   I wonder if a class A MOSFET using resistors might be easiest - or is a conventional toroidal core circuit is best...

Thanks.

T

[steve_qix]

Hi Tom,

Hey, when is that 4-1000 going to be heard on 75 again?? 

On your driver question - two possibilities:

1) If the 250mV output from the DDS is into a relatively low impedance, you could step it up to TTL levels with a suitable transformer. Then, drive an IXDD614 driver IC (hi Z input), which will then deliver more than 10A of RF current (at 12V).  That's more than enough power.  You could then step the 12V peak RF output up to whatever you need.

2) If the 250mV output doesn't deliver any power, go through the intermediate amplifier and driver the IXDD614 with that..

Simple, simple, simple 

Regards,  Steve

[K1JJ]

Thanks for the suggestions, Steve.

Yes, I forgot I had some of those IXDD614 driver ICs hanging around and will use them.

The IPA amp in the 4-1000A rig uses quad MRF-150s.   Right now, 2 watts is plenty to drive the MRF-150s to 50 watts out for the 4X1s.     I wonder if the IXDD614 has enuff voltage output to drive the MRF-150s?    Or would I need a step up transformer?   I think the bias voltage on the MRF-150s is about 3.5 volts, so maybe we are fine with 5 volts of IXD TTL level drive, I dunno.

The 2N2222A circuit puts out 2-5V @ 50 ohms.  But you say the IXD is high impedance input. I don't know if the VFO output is suitable to drive the IXD directly.

The Digital VFO output info says:

"The output of my DDS is unbuffered. The voltage is rather low (~300mV P-P) but it drives active mixers (e.g. SA612) with no problem. But it is not sufficient to drive a diode ring mixer, or vintage radio. Consequently a buffer amplifier of some sort is needed. "


Hope to be on soon, once I get this VFO system working.

T

[K1JJ]

Hmmm... wait a second...

The VFO output is a pure filtered sinewave.   The MRF-150 amp is a linear amp.   

If I use the IXD chip as an IPA, will that convert the sine wave into a TTL square wave?   I realize it's ultimately driving a class C stage, but do I want a clean sine wave pulse or will a square wave  produce the same results once it is filtered through the 4-1000A L/C circuitry?


T

[KF1Z]

If you are worried about "clean"...
you may be disappointed with your dds purchase.
(or you might have better luck than I did)

I used mine for a day....

have a second new one down at the bottom of the scrap box too never assembled.


As they say... "your mileage may vary".   

[WBear2GCR]

I be no expert, but putting square waves into a linear stage will produce more harmonics than a sine wave.

I'd go for a simple follower to boost the current available from the output of the DDS.

If you are using the DDS for just one band, then you could use some transformers in there, where the transformers form part of a bandpass or low pass filter, this will drop the higher out of band harmonics down to nil. With enough current you can make the transformer a step up. Probably not that much current either.

            _-_-

[W1RKW]

As college electronics project I built a DDS "oscillator" that was driven by a PC that could go up to 15Mhz. When employed as a VFO with my radio stuff there were 2 issues I came across long after. One was impedance matching (i could never seem to get a good match with Johnson V2 with an outboard amp but the other was the noise floor with the DAC. I tried various LPFs but couldn't get the noise down and I attributed it to phase noise. Never proved it. I never took it beyond that or employed it in my radio stuff here so I still use analog all the way for frequency gen. I occasionally use it as an RF generator. I don't know if I was getting noise from the onboard XTAL oscillator or some manifestation in the design. Suffice it to say it was very stable but just not clean.

[K1JJ]

If you are worried about "clean"...
you may be disappointed with your dds purchase.
(or you might have better luck than I did)

Vely interesting comments.   Yes, I am always concerned about cleanliness... :-)

I could cancel the order at this point, so I'm flexible...

Here's the published specs:

•DDS SPECIFICATIONS

◦Total Harmonic Distortion: -66db dBcf
◦Mirror Spur supression -60 db to 30MHz, -45 db to 34MHz.
◦Spurious-Free Dynamic Range (SFDR)
◾Wide Band (0 to Nyquist) -60 dBc
◾Narrow Band (±200 kHz) -74 dBc
◾Performance details

•Main output is a filtered sine wave
◦250mV p-p, when driving 1K load.
◦Secondary unfiltered analog output


I know that cleanliness is critical for a receiving system, but what can I expect to see when transmitting a dead carrier?  Would there be S9 side crud hash noise when I am S9+60dB over?  What does a "mirror spur" sound like?  Maybe the specs should be more like -80dB down.


I have a homebrew FET analog VFO left over from the class E rig which I could use again.  But the DDS seems so slick.

T

[WBear2GCR]

Over on the Viking II VFO thread there is s a link to the WA1FFL DDS, on his site he shows -140dB noise floor... fwiw.

KF1Z, wondering which one you found not up to par?
What were the issues, and how old were these, what chipset?
PM please if you don't want to post negative stuff.

                  _-_-

[K1JJ]

It appears that the WA1FFL DDS is cleaner, like down to -70 to -80dB when clocked at a rate of 400Mhz.

http://www.wa1ffl.com/index.html

Can anyone confirm this?    I may cancel and replace this order.


"A 14-bit D/A converter at the output, and
a much higher rated clock speed than the
98xx family allows a frequency-dependent
SFDR in excess of 70 dB, and approaching
80 dB with a 400 MHz internal clock frequency."


Here's the full article:

http://www.wa1ffl.com/Hagerty_QEX.pdf


** Here is an excellent review write-up:
http://www.eham.net/reviews/review/107788

[w8khk]

I met Jim / FFL in Dayton when he displayed his kit.  He was operating it with HP oscilloscope and spectrum analyzer displays.  I cannot recall the exact numbers, but it was impressive, so I bought the kit. 

If I recall correctly, the display is not a frequency counter; rather it is a computed value to display the synthesized frequency output.

One of the features I liked was the ability to program an IF offset so that it can be used not only as the transmit vfo, but also as the receiver local oscillator for transceive operation with separate transmitter and receiver.

The kit was a breeze to build and operated perfectly when first turned on.  I was going to use it with the viking, as well as the small wonder but have been too busy with other projects to finish these.

If in doubt, contact Jim directly, he is quite a gentleman to deal with.

[K1JJ]

Hi Rick!

Good to hear from you again.

Thanks for the info. Yes it does seem like a nice clean kit. There is also a buffer/driver board available to ramp up the output, but it may not be enough to give me 2 watts out that I need.

I cancelled the other order and may place an order for the FFL DDS.

I'm now toying with the idea of building up a 2 tube VFO. (osc and buffer)  But will probably take the easy way out and go wid the DDS... :-)

T

[w8khk]

Hi Tom,

I think the issue boils down to your earlier question.   If you are just driving a buffer and/or a class C amplifier that is modulated, how critical is the phase noise issue, if at all?  Surely if this was a signal that you were mixing with the output of a balanced modulator for sideband, it would be an issue.  Surely some of the gurus on the forum could shine a light on the subject and help us understand whether it is really an issue for your desired application.

73, Rick

[K1JJ]

Yes, good point Rick.   I don't know the answer myself.  Maybe driving the non-linear class C stage is a different animal with less stringent requirements vs: the mixing and IMD problems a linear amp incurs.  After all, a class E rig uses a square wave for RF drive... :-)

I'll bet someone like Stu would know...

I sent an email to Jim / FFL with my requirements and we'll see what he says.

T

[WQ9E]

I think your biggest issue with phase noise applied to the transmitter is if you have a very strong signal you are raising the noise floor close to your signal, think spread spectrum.   I think the only time this could be an issue in the ham bands would be during a contest with a lot of high power transmitters with significant phase noise and in the typical contest QRM is so intense the increase in noise level would not be significant.

Clean oscillators are far more critical for commercial service.

Of course the other case where minimal phase noise is important is in your receiver synthesizer system.  Here phase noise effectively creates mixing products from stations far removed from the desired signal that will be fed into the IF chain and greatly degrades your effective noise floor.  Some of the early PLL synthesized rigs (like my FT-980) suffer greatly from this issue.   One of the "cheap" techniques for reducing phase noise was to accept the tradeoff of slower lock time thus the "chuffing" tuning of some of the lower cost early receivers.  DDS definitely improved the state and I doubt that phase noise is going to be an issue for you unless you are planning to use a lot of big tubes with big handles in your final

[Radiohead]

here they use an ef184 with el84 tube.

[w4bfs]

I'm going out on a limb here because my info is 20+ years old and I havn't seen the schizmatic of the dds vfo ....

If it is a lookup method of direct synthesis (counter to prom to adc) then there is a phase discontinuity between each entry in the lookup table .... raising the speed of the adc and adding steps per cycle (more samples) lessen this discontinuity and lower undesired products ..... this of course is dependent of the layout and quickness of the parts chosen

even a crystal oscillator has a phase noise pedestal .... its just more like -140 db

[steve_qix]

Despite all of the hi-tech stuff I build, the oscillators (VFOs) for all of my transmitters are simple, free running oscillators because there is little phase noise.  Yes, they drift a bit but so what    After all, it's AM.  On the plus side, in addition to the low phase noise, they're simple and inexpensive to build and are very trouble free.

[DMOD]

Tom, you show me your schematic and I'll show you mine.

Anyway, below is one potential IPA driver implementation using a bipolar circuit.

Q1 will need a heatsink since it would be a class A circuit.


Phil - AC0OB

[K1JJ]

Thanks for the suggestion, Phil.

Jim/FFL has an old 4 watt QRP handbook circuit that I may build up to go from 250mV to 4 watts.   I may go that route wid his DDS VFO board and display.

Then there is the possibility of using my old FT-102, three 6146s in parallel, as a dedicated driver to drive the 4X1 directly.   Still trying to decide.

T

[WBear2GCR]

FFL says on his site that the current product is lower distortion than the QEX article, if I read the site properly... fyi.

                   _-_-

[N2DTS]

I built a vfo and exciter last winter.
6C4 osc using B+W coil stock and a cap, feeds a buffer amp, into a 6146.
Has a digital freq counter as a display, but no digital stuff in it other then that.

I was thinking of building another one for the big rig using an old Collins pto that I have.
Its amazingly stable, I was shocked.

The exciter is very small, but has a remote power supply.

I never liked cheap digital stuff, or cheap modern analog stuff, it all seems to generate a lot of noise.

 

[K1JJ]

I was thinking of building another one for the big rig using an old Collins pto that I have.
Its amazingly stable, I was shocked.

Yeah, I also have a Collins PTO. I pulled out the tube and put in a small FET at 12V.  As you say, it is very stable.   At one point I used it to drive a little 5 watt QRP board for 80M.  That may be another solution.   (And the spur crud wud be down -140dB)

T

[N2DTS]

I was unclear, the homebrew vfo is very stable.
Its the same circuit as used in my home brew receivers as the LO (go with what you know).
From turn on to about 60 seconds it gets stable, and then drifts no more then 100 Hz

My biggest problem was the change of about 100 hz between keyed and unkeyed.
I got it down to about 50 Hz.

Its got the digital display, plus I look on the sdr, and for junk box parts, I think its quite amazing.
The vfo box is too small, as is the coil, but it seems to work....

90 volts on the plate of the 6C4, regulated with VR tubes....

I do not think the pto in my 32v3 is as stable, but its OLD...
 

[DMOD]

Jim/FFL has an old 4 watt QRP handbook circuit that I may build up to go from 250mV to 4 watts.   I may go that route wid his DDS VFO board and display.

If Jim is talking about the one below, it can only put out about 1.5W into 50 ohms according to P = V^2/50.

I think this circuit was originally speced to output the 12V Peak into > 2.2k loads.

I added a slight voltage supply mod to up the output voltage a bit.

Phil -AC0OB