High frequency sine wave on the LM1036 output

[daniel23]

Hello,
I've put together a circuit identical to that in National's Data Sheet. I get a near to 48 MHz sine wave of about 200 mV out of both outputs. This is with no input signals. I get this result with several different chips. I tried first on solderless breadboard and now on a perfboard.
The upper side of the perfboard is a ground plane where all capacitors are tied as short as possible.

I tried first with conventional components (wired) and now with SMT.
But all that without any result...

Any ideas?
Thanks!

[pacificblue]

How about photos?
Are you sure about the 48 MHz? That tone control IC has unity gain and a frequency response limited to 250 kHz.

[sasmit]

Originally Posted by daniel23

Hello,
I've put together a circuit identical to that in National's Data Sheet. I get a near to 48 MHz sine wave of about 200 mV out of both outputs. This is with no input signals. I get this result with several different chips. I tried first on solderless breadboard and now on a perfboard.
The upper side of the perfboard is a ground plane where all capacitors are tied as short as possible.

I tried first with conventional components (wired) and now with SMT.
But all that without any result...

Any ideas?
Thanks!

I have been using LM1036 quite successfully on one of my amps for arnd 6 years now ! It's made on a piece of prototyping board, no noise not even any hiss. If you could post some pictures of your pcb and schematic we could figure out obvious mistakes if any,

[daniel23]

Thanks for your answers,
On the photo the upper and bottom strips are connected to the ground plane on the other side of the PCB.
The photos are less usable...so I've drawn the layout of the SMT parts. The viewing of the drawing is from the top of the PCB. The photograph's was taken from the back.
On the last picture is the waveform on the output of the LM1036.

[sasmit]

Originally Posted by daniel23

Thanks for your answers,
On the photo the upper and bottom strips are connected to the ground plane on the other side of the PCB.
The photos are less usable...so I've drawn the layout of the SMT parts. The viewing of the drawing is from the top of the PCB. The photograph's was taken from the back.
On the last picture is the waveform on the output of the LM1036.

you are using the DIP package of the IC right ? why use smd parts on a normal prototyping board, from the way the soldering appears on the prototyping board it appears you have might have a solder bridge somewhere. If you're going to make it on a prototyping board make it with normal through hole parts it'll much less messy and chances of any shorts, solder bridges reduced.

[daniel23]

No chance that there is any short. All has been checked with ohmmeter and capacimeter with the active components removed.
On the other side, I tried first with conventional components (through hole parts) and the results where the same.
I expected the results would be better with SMD and a ground plane on the other side.
I've use SMD cause there will be a space issue on the board when it will be fully equiped ...

Sorry if it sound a bit confused, but my english is far from perfect...

[Mooly]

Short your scope probes out,the tip to ground lead... should be nothing of course.

Now connect the still shorted probe and ground to the ground on the circuit... does anything show ?

[daniel23]

Thanks for your answer Mooly,

You can't know if I am a beginner or what may be my level of knowledge. I should say first am working in R&D, mainly in EMC (electromagnetic compatibility) since 1982 and I am confident with most of the parasitic effects that can appear.

I've already done all these tests. Power supply is clean, ground is clean, and all other ICs were removed from the PCB thanks to IC-sockets.
At work I could do more precise measurements. The base frequency varies from 44 to 47 MHz and is amplitude modulated at about 3 MHz. This wave is only on both output pins. For this measurement the pots were replaced by 10k/10k divider.

Someone had the same problem in 2003. He had left a message "Help with LM1036? " in sci.electronics.design but without any answer.

[Speedskater]

If it oscillates both at home and at work, that pretty much rules out EMC problems. I would try a different type, value and/or location for capacitor C50.

[Mooly]

Originally Posted by daniel23

Thanks for your answer Mooly,

No problem, I'm just at a bit of a loss what to suggest tbh.

Have you tried "rolling a finger" over the components to get some idea of what/if anything alters. Oscillation is one thing, but AM modded at 3mhz... that's a bit weird. Maybe also try a variable PSU, just to try and get a handle on what's occuring.

Last resort would be to build the circuit using point to point wiring (just to see) and check if anything odd shows then.