ebay AD9850 Oscillator Board Doesn't Always Start

[bobledoux]

I bought four of these boards for $22. Mine have the blue adjustment trimmer pot at one end.

These things are fickle . One time it will power up a dozen times in a row. The next time it won't output no matter how many times I try.

My test frequency is set to 7.0000 mhz. These AD boards have an on-board low pass filter that kicks in about 30 mhz and provides an output impedance of about 100 ohms.

I've set up three different circuits, each using a 16F690 as controller. I've powered each circuit from external power supply or from the PicKit2. The circuit pulls about 75ma when operating. There is no power supply voltage sag on power up.

The first two circuits employed serial data input for the 40-bit frequency control. In the last circuit I tried using parallel input. This change made no impact on power-up response.

My antenna is a three inch piece of wire. Connecting a high impedance scope or frequency counter doesn't seem to affect power up response.

When it powers up properly, each board is within 100 Hz of test frequency. All the boards show the same propensity towards fickleness.

What am I missing?

 

[JimB]

I vaguely remember that I have been down a road like this one!

I vaguely remember that I had left the RESET line (pin22 of the AD9850) floating.

Just had a look at one of the modules and there is a 2.7k resistor which I have tacked on top of the AD9851 module between RESET and GND.

Why 2.7k ?
Probably because it was close to hand!

JimB

 

[bobledoux]

Thanks for the replay.

N3ZI puts out a VFO using these boards. I took the following comment from his Yahoo group message number 118:

"1) Missing Jumpers: The reset line is active LOW (not high) the manual calls
for a shorting jumper at R1, which connects the reset line to +5v. You should
never leave the reset line floating. In the new version (DDS2) the reset jumper
is hardwired. ..."

I'm pulling up the reset with a 4.7K resistor so this isn't my problem.

The board acts as though it loading up and stopping. The frequency counter will show the start of a signal that then shuts down. Sometimes it only shows 3khz. Other times it shows over one megahertz. But it will then drop back to zero. Other times, it boots up and operates at expected frequency-7mhz.

Neither frequency counter nor scope are the cause of this load. With them disconnected, I can still pick up the presence or absence of a signal on my receiver.

 

[JimB]

Boble

You have a slightly different module from me. Mine does not have the blue trimmer.

There may be extra reset circuitry on you module, but on mine the RESET pin connects directly to one of the terminal pins on the edge of the module and there are no jumpers associated with RESET.

The RESET on the AD9850 pin22 is active high according to the Analog Devices datasheet.

If the 9850 is putting out the wrong frequency, could it be that there is something wrong with the 40bit control word?
Does your PIC repeatedly send the control word, or as I do, only refresh when the frequency changes?

JimB

 

[bobledoux]

My frequency is correct, if the board boots.

I'm confused about the reset status. Figure 7 in the AD9850 datasheet shows the reset pin pulled high to perform reset.

Is the reset pin kept low for normal operation? I've tried it in both states.

There are two clocks working in the AD9850: The first clock is the oscillator (mine is 125mhz) that operates the AD9850. The second clock is the clock pin of the controller which operates the W_CLK, AD pin 7, that passes in the data.

On my program I only pulse the controller clock if I am passing in data.

I've tried my program in refresh and single command mode. I don't see any difference.

I've operated the board in both serial data transfer and parallel mode. The same problem exists in each.

My problems are consistent across all four boards I own. I wonder if I have a bad batch.

Its amazing. I'm 60 Km from the Pacific ocean and you're not far from the Atlantic and the North Sea. Distance knows no bounds.

 

[JimB]

My frequency is correct, if the board boots.

In that case what you are doing cannot be too far wrong.

I'm confused about the reset status. Figure 7 in the AD9850 datasheet shows the reset pin pulled high to perform reset.
Is the reset pin kept low for normal operation? I've tried it in both states.

Yes, pin 22 of the AD9850 is high for reset, and low for normal operation.

There are two clocks working in the AD9850: The first clock is the oscillator (mine is 125mhz) that operates the AD9850. The second clock is the clock pin of the controller which operates the W_CLK, AD pin 7, that passes in the data.

On my program I only pulse the controller clock if I am passing in data.

OK

I've tried my program in refresh and single command mode. I don't see any difference.

I wondered if the program was running away with itself into the realms of fantasy and scribbling silly data to the AD9850.

I've operated the board in both serial data transfer and parallel mode. The same problem exists in each.

I have never tried parallel programming.

My problems are consistent across all four boards I own. I wonder if I have a bad batch.

When I bought my boards I wondered if they would be some cheap knock-off thing, but all three have worked correctly for me.
I started many years ago with a pukka Analog Devices chip and the Chinese modules worked exactly the same with the same software.

The only things I can suggest without seeing the thing is to:
Wait untill the power is stable before initialising the DDS.
Ensure that the supplies are well decoupled.
If you are doing a serial load of the 40 bit control word, ensure that pins 3 and 4 of the AD9850 are pulled up to 5v and that pin 2 is pulled down to 0v.
Pulse W_CLK
Pulse FQ_UD
The AD9850 should now be correctly in serial load mode.

Sorry if I am telling you things which you know already.

Its amazing. I'm 60 Km from the Pacific ocean and you're not far from the Atlantic and the North Sea. Distance knows no bounds.

I am about 16km from the North Sea, not too sure about the Atlantic, probably a good 4 or 5 hours drive and then I would be looking out at the Western Isles rather than the wide open Atlantic.

 

[bobledoux]

Thanks for the help.

I'm going to leave this for a week and approach it again with a clear mindset.

 

[JimB]

I'm going to leave this for a week and approach it again with a clear mindset.

Often a good idea!
Sometimes you look at something and you KNOW that it is right, because it is.
Let your mind relax for a while and that which was so obviously right is now so stupidly wrong.

Been there, done that, more than once.

JimB