Date   
Re: AD9850/51 module pc boards

Steven Dick
 

Looks good. I like the coilcraft transformer on-board as I have used one myself (Coilcraft WB1-6TSLB)
and it allows good wideband performance and adequate output power for me with a lowpass filter and no output power amp which adds harmonics. Some of those amps are marginal, some can oscillate if the layout is not adequate. There is a problem with the xtal oscillator on EBAY AD9851 type II boards. For the ebay modules. the AD9851 was running off 5 volts but its crystal oscillator also ran on 5 volts but was only rated for 3.3v. Make sure the xtal oscillator is really compatible with 5V input voltage.

for AD9850, 125 MHz oscillator labelled MCO 1510A 125 MHz TQG-9h is a 5 volt oscillator made by Shenzhen Great Electronics. Data sheet:

http://www.szgreat.com/UploadFiles/2010322153029.pdf

For AD9851, 30 MHz oscillator Type II AD9851 boards are labelled CETDCJ 30 032. They are made by Taiten Electronics. It is a 3.3V oscillator, much to my dismay. One of them already failed. See part numbering guide:

http://www.taitien.com.tw/db/download/Model%20Numbering%20Guide_XO.pdf

I ordered a CTS CB3-3C-30M0000 which is rated for 5V. Mouser sells them: https://www.mouser.com/ProductDetail/CTS-Electronic-Components/CB3-3C-30M0000?qs=HL%252bYNjdyZ0s2KrBi4NmH2g%3D%3D

See file https://groups.io/g/PHSNA/files/K1RF/Bandpass%20Filter%20measurement%20anomalies%20R2.pdf discussing output amplifier
and file https://groups.io/g/PHSNA/files/K1RF/Expected%20power%20output%20from%20Type%20II%20PHSNA.pdf
-Steve K1RF

-----Original Message-----
From: Terry VK5TM
Sent: Friday, September 28, 2018 10:10 PM
To: PHSNA@groups.io
Subject: Re: [PHSNA] AD9850/51 module pc boards

Here is a preliminary 3 pics of the pcb, combined top and bottom layers, top only and bottom only.
Nothing is set in concrete, so changes can be made. The other style module version could also be done.

Couple of points to note:
Obvious one is there is no on-board filter.
The output is through a 1:1 transformer (Coilcraft), this can be changed, but the transformer gives a better output signal both level wise and harmonic suppression wise.
There are some either/or components depending on whether the square wave output is required, these will be explained fully if this goes any further than just an idea.
Most smd components are 0805 sized rather than the smaller size on the current modules.

Re: AD9850/51 module pc boards

Terry VK5TM
 

I think if the oscillator module has to be replaced, it should be replaced with a TCXO.

It is surprising how much change in frequency there is with just a light draft across the oscillator, even just breathing on them can produce a shift of more than 100Hz.

Re: AD9850/51 module pc boards

Steven Dick
 

I think you'll be hard pressed finding a TCXO that runs on 5V.
-Steve K1RF

-----Original Message-----
From: Terry VK5TM
Sent: Saturday, September 29, 2018 7:29 AM
To: PHSNA@groups.io
Subject: Re: [PHSNA] AD9850/51 module pc boards

I think if the oscillator module has to be replaced, it should be replaced with a TCXO.

It is surprising how much change in frequency there is with just a light draft across the oscillator, even just breathing on them can produce a shift of more than 100Hz.

Re: AD9850/51 module pc boards

Henning Weddig
 

Terry,

Your pcb looks awfully good! Especially that You have included an output transformer whcih is already suggested in an Analog Devices app note.

For the 125 Mhz osc: I think there is still some space left to include a 3V3 LDO for the xtal osc. To my knowledge most of the xtal oscs and especially the SMD types are for 3V3.

I am sure Your board will be a bestseller as the chinese units are getting up and up in price.

73

Henning Weddig

DK5LV

Am 29.09.2018 um 04:10 schrieb Terry VK5TM:

Here is a preliminary 3 pics of the pcb, combined top and bottom layers, top only and bottom only.
Nothing is set in concrete, so changes can be made. The other style module version could also be done.

Couple of points to note:
Obvious one is there is no on-board filter.
The output is through a 1:1 transformer (Coilcraft), this can be changed, but the transformer gives a better output signal both level wise and harmonic suppression wise.
There are some either/or components depending on whether the square wave output is required, these will be explained fully if this goes any further than just an idea.
Most smd components are 0805 sized rather than the smaller size on the current modules.

Re: AD9850/51 module pc boards

Tony Jaques
 

Hi all,
Three things: One is that if the distortion introduced by an AD8008 output stage can not be tolerated the perhaps the AD985x device is not a good enough starting point. (Search the files for "an investigation into output stages" - or something close to it.
The second is that modifying the module to give the oscillator a 3.3 volts supply. Again there is a message somewhere in which I show my solution to that. A search in messages might find it, but i no longer remember what the subject line was.
A third comment is that, if you have a 'scope, do make sure that you allow for Gary A Winblad's sweep modification later, even if you don't add it to start with.

I agree - very nicely done board layout Terry.

Tony Jaques (G3PTD)

Re: AD9850/51 module pc boards

Terry VK5TM
 

I've seen at least three different options for using 3.3v oscillator modules on 5V boards, namely:

a couple of series diodes
a 47 ohm resistor
a proper 3.3v regulator.

The first two options work, but do have their own problems, so I will go with the third option of a 3.3v LDO reg.
I will use the LP2985 which is rated at 150mA so will cope with a TCXO and is readily available at Mouser.

And just in case anybody is wondering why not run the whole thing at 3.3v, the AD9850/51 is not specced to run at it's maximum possible osc speed at 3.3v.
While they do and are being run in that condition on the modules, current consumption and heat generation are increased markedly.

Re: AD9850/51 module pc boards

Terry VK5TM
 

Updated pics with the 3.3v reg underneath the Xtal osc.

I have moved some things around to shorten the length of tracks between IC output, transformer and pin at the edge of the pcb.

Will leave this alone for a day or two, firstly so any comments can be added/updated and secondly, I find getting away from a design sometimes suggests improvements that can be made or something shows that has been missed (like a power track missing in the first set of pics).

Re: AD9850/51 module pc boards

William R Maxwell
 

Nicely done Terry. I'll certainly take a couple once you get to production.

Bill VK7MX

On 30/09/2018 1:34 PM, Terry VK5TM wrote:
Updated pics with the 3.3v reg underneath the Xtal osc.

I have moved some things around to shorten the length of tracks between IC output, transformer and pin at the edge of the pcb.

Will leave this alone for a day or two, firstly so any comments can be added/updated and secondly, I find getting away from a design sometimes suggests improvements that can be made or something shows that has been missed (like a power track missing in the first set of pics).

Re: AD9850/51 module pc boards

Henning Weddig
 

Terry,

I am missing a lof of decoupling caps close to the Vcc pins of the AD9850. Plöease have a look on the AD9850 data sheet, a lot of .1 µF caps are used there, althoug not easily seen where they belong to. My guess as they are on the bottom side they are close to the AD9850 pins too.

Another hint: the Analog Devicers app note "A Technical Tutorial
on Digital Signal Synthesis" from page 45 on gives good reasons to use a transformer on the DAC outputs!

Henning Weddig

DK5LV


Am 30.09.2018 um 05:34 schrieb Terry VK5TM:

Updated pics with the 3.3v reg underneath the Xtal osc.

I have moved some things around to shorten the length of tracks between IC output, transformer and pin at the edge of the pcb.

Will leave this alone for a day or two, firstly so any comments can be added/updated and secondly, I find getting away from a design sometimes suggests improvements that can be made or something shows that has been missed (like a power track missing in the first set of pics).

Re: AD9850/51 module pc boards

Henning Weddig
 

Terry,

anohter idea:  i just read in an Analog Devices app note that the jitter of the reference clock has a large influence of the overlall jitter performance of the DDS. From my own experience a well regulated and low noise supply to the xtal osc and also TCXO will improve the phase nosie performance of the reference clock and therefore minimize the jitter.

The best known low noise regualtor is the LT3045-- unfortunately not cheap and sometimes hard to get due its large demand.  Another idiea ist o use a "super cap" i.e. an emitter follower with a large cap from the the base to ground. This cap will be multipleid wth the hfe of the transistor.

I had the idea for a piggy back pcb on the chinese units, but never tried. See the attached schematic file.

Henning


Am 30.09.2018 um 05:34 schrieb Terry VK5TM:

Updated pics with the 3.3v reg underneath the Xtal osc.

I have moved some things around to shorten the length of tracks between IC output, transformer and pin at the edge of the pcb.

Will leave this alone for a day or two, firstly so any comments can be added/updated and secondly, I find getting away from a design sometimes suggests improvements that can be made or something shows that has been missed (like a power track missing in the first set of pics).

Re: AD9850/51 module pc boards

Steven Dick
 

I would recommend the TI TLV75733PDBVR for a 5V to 3.3V regulator. It uses an SOT23-5 package, is relatively low noise, and inexpensive (.86 at Mouser.) I have used the 5V version with excellent results. The "super cap" is a good circuit but doesn't help high frequency performance. It is best used in low frequency power supply applications. simple regulator will more than do the job in a compact and low cost solution. Here's a link to the Mouser part:
https://www.mouser.com/ProductDetail/Texas-Instruments/TLV75733PDBVR?qs=sGAEpiMZZMsGz1a6aV8DcD1rpA6FsR3JV0%2fewXLxMYU%3d
-Steve K1RF

-----Original Message-----
From: Henning Weddig via Groups.Io
Sent: Sunday, September 30, 2018 7:37 AM
To: PHSNA@groups.io
Subject: Re: [PHSNA] AD9850/51 module pc boards

Terry,

anohter idea: i just read in an Analog Devices app note that the jitter
of the reference clock has a large influence of the overlall jitter
performance of the DDS. From my own experience a well regulated and low
noise supply to the xtal osc and also TCXO will improve the phase nosie
performance of the reference clock and therefore minimize the jitter.

The best known low noise regualtor is the LT3045-- unfortunately not
cheap and sometimes hard to get due its large demand. Another idiea ist
o use a "super cap" i.e. an emitter follower with a large cap from the
the base to ground. This cap will be multipleid wth the hfe of the
transistor.

I had the idea for a piggy back pcb on the chinese units, but never
tried. See the attached schematic file.

Henning


Am 30.09.2018 um 05:34 schrieb Terry VK5TM:
Updated pics with the 3.3v reg underneath the Xtal osc.

I have moved some things around to shorten the length of tracks between IC output, transformer and pin at the edge of the pcb.

Will leave this alone for a day or two, firstly so any comments can be added/updated and secondly, I find getting away from a design sometimes suggests improvements that can be made or something shows that has been missed (like a power track missing in the first set of pics).


Re: AD9850/51 module pc boards

Terry VK5TM
 

Henning,

There are two decoupling caps right at the chip on the underside of the pcb, I can't get them any closer without cutting the top off and mounting them on the die :) <-smiley face
Another one next to the V+ pin on the underside and the 10µF next to the other V+ pin.
Looking at the evaluation board, which has much longer supply lines, it only has one more than this around the DDS chip, the others are associated with the interface chips.

Re the supply to the xtal osc, the LP2985 is listed as a low noise regulator and I believe should be more than adequate.
It is well decoupled and again is on the underside of the pcb.

I did some digging through Analog's forums and engineering notes today and one thing that I noticed was that, in their own words, some of their recommendations for decoupling, seperate ground planes etc was somewhat over the top and simpler schemes made no noticeable difference.

Re oscillator phase noise, well you're already hindered by the limitations of the AD9850 to begin with, so I can't see using ultra pure regulators and massive amounts of decoupling are going to do anything useful. I could be wrong, which wouldn't be anything unusual, but I don't think so in this case. The obvious proviso is of course to use a good quality Xtal Osc module.

The test will be to build a couple and compare them against the standard module and rework it if necessary.

Re: AD9850/51 module pc boards

Terry VK5TM
 

Steve,

Same package and pinout as the LP2985, so a straight drop-in.

Quick look at the datasheet shows not much difference between the two other than 1A verses 150mA

Re: AD9850/51 module pc boards

Steven Dick
 

LP2985 is an excellent part. go for it.
-Steve K1RF

-----Original Message-----
From: Terry VK5TM
Sent: Sunday, September 30, 2018 8:58 AM
To: PHSNA@groups.io
Subject: Re: [PHSNA] AD9850/51 module pc boards

Henning,

There are two decoupling caps right at the chip on the underside of the pcb,
I can't get them any closer without cutting the top off and mounting them on
the die :) <-smiley face
Another one next to the V+ pin on the underside and the 10µF next to the
other V+ pin.
Looking at the evaluation board, which has much longer supply lines, it only
has one more than this around the DDS chip, the others are associated with
the interface chips.

Re the supply to the xtal osc, the LP2985 is listed as a low noise regulator
and I believe should be more than adequate.
It is well decoupled and again is on the underside of the pcb.

I did some digging through Analog's forums and engineering notes today and
one thing that I noticed was that, in their own words, some of their
recommendations for decoupling, seperate ground planes etc was somewhat over
the top and simpler schemes made no noticeable difference.

Re oscillator phase noise, well you're already hindered by the limitations
of the AD9850 to begin with, so I can't see using ultra pure regulators and
massive amounts of decoupling are going to do anything useful. I could be
wrong, which wouldn't be anything unusual, but I don't think so in this
case. The obvious proviso is of course to use a good quality Xtal Osc
module.

The test will be to build a couple and compare them against the standard
module and rework it if necessary.

Re: AD9850/51 module pc boards

K5ESS
 

Does anyone have a reliable Ebay (or otherwise) source of AD9850 and/or AD9851 chips? Ebay has some advertised as low as $9.50 for 9850 chips but I have no experience with any of the suppliers listed. New chips from Mouser are almost expensive as the complete modules from China.
Mike
K5ESS

-----Original Message-----
From: PHSNA@groups.io [mailto:PHSNA@groups.io] On Behalf Of Terry VK5TM
Sent: Friday, September 28, 2018 9:10 PM
To: PHSNA@groups.io
Subject: Re: [PHSNA] AD9850/51 module pc boards

Here is a preliminary 3 pics of the pcb, combined top and bottom layers, top only and bottom only.
Nothing is set in concrete, so changes can be made. The other style module version could also be done.

Couple of points to note:
Obvious one is there is no on-board filter.
The output is through a 1:1 transformer (Coilcraft), this can be changed, but the transformer gives a better output signal both level wise and harmonic suppression wise.
There are some either/or components depending on whether the square wave output is required, these will be explained fully if this goes any further than just an idea.
Most smd components are 0805 sized rather than the smaller size on the current modules.

Re: AD9850/51 module pc boards

Terry VK5TM
 

Depending on how many you need, you may be able to get free samples out of Analog themselves (I haven't had any samples out of Analog for some time, so not sure of their current policy on samples).

No experience of the Ebay sellers, I base whether to buy on their ratings, reviews and comments in forums such as this.

Re: AD9850/51 module pc boards

William R Maxwell
 

I have not found the Analog sample program to be of any practical use in Australia for many years, Terry.

Back in the 1990s, sample requests to Analog were handled on line from Analog itself and they were happy to forward samples to Australian requesters. Some time in the 2000s, that policy changed, with Australian requests being forwarded to the Australian agents, Arrow from memory, who then insisted on project details, likely production numbers, likely monthly purchases etc. My explanation that I was prototyping, essentially for private, hobby use did not seem to cut the ice with them.

I never once succeeded in obtaining samples from the agents, despite Analog's ongoing sample policy. I did once get samples direct from Analog, after contacting them to enquire whether the Analog policy still applied and to ask whether the Australian agent had been made aware of it. Thereafter though, subsequent requests to them were again forwarded to the Australian agent.

I haven't tried again in recent years but I would be delighted to hear that things had changed and samples were again readily available to Australian users.

Bill, VK7MX

On 1/10/2018 10:22 AM, Terry VK5TM wrote:
Depending on how many you need, you may be able to get free samples out of Analog themselves (I haven't had any samples out of Analog for some time, so not sure of their current policy on samples).

No experience of the Ebay sellers, I base whether to buy on their ratings, reviews and comments in forums such as this.


Re: AD9850/51 module pc boards

Terry VK5TM
 

Hi Bill.
Same experience here, I got samples once from the Australian agent after jumping through multiple hoops. That was for a commercial project that ended up crashing anyway.
Would be nice if it had changed, but I doubt it. Hopefully the guys in the US will have better luck.

Ok, the attached pdf is a preliminary set of notes to go with the pcb, so additions, corrections or updates if anybody can think of any please.

You will notice I have been fiddling again and added a couple of pads so that the Rset resistor can be isolated for those that want to experiment with either level setting or AM modulation.

Now is the time to make suggestions to add or change something (providing it can be fitted on the pcb).

Re: AD9850/51 module pc boards

Don Vosper
 

Some years ago  I got a free sample out of AMD but with subsequent requests the reply came via Germany that they did not supply free samples to individuals with my type of email address. I guess that meant that they would only supply trade or perhaps education.

I see on the Farnell site in the UK that one off 9851, 9850 and the useful AD8037 are £25.00, £23.00 and £10.00 each respectively.

All "supplied until stocks are exhausted"

Don m5aky

Re: AD9850/51 module pc boards

Clifford Heath
 

On 1 Oct 2018, at 4:42 pm, Don Vosper via Groups.Io <donald.vosper=btinternet.com@groups.io> wrote:
I see on the Farnell site in the UK that one off 9851, 9850 and the useful AD8037 are £25.00, £23.00 and £10.00 each respectively.
My last 80 or so purchases have been through AliExpress.
Typically you get parts for under the 10,000 unit price.
I think I've twice had a situation where a completely
wrong thing was shipped, and I didn't pay for those after
disputing them.

All other parts have been up to normal spec, and were
shipped free of charge. Most of the world's electronics is
made in China. There are a lot of "partial reels" of perfectly
good parts being sold off through the thousands of vendors
on AliExpress. I'm really struggling to see any downside.

Clifford Heath.